Dating neurological injury - a forensic guide for radiologists, other expert medical witnesses, and attorneys

Dating Neurological Injury

Jeff L. Creasy

Dating Neurological Injury

A Forensic Guide for Radiologists, Other Expert Medical Witnesses, and Attorneys

Jeff L. CreasyAssociate Professor of NeuroradiologyVanderbilt University Medical CenterNashville, [email protected]

ISBN 978-1-60761-249-0 e-ISBN 978-1-60761-250-6DOI 10.1007/978-1-60761-250-6Springer New York Dordrecht Heidelberg London

© Springer Science+Business Media, LLC 2011All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden.The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights.While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibil-ity for any errors or omissions that may be made. The publisher makes no warranty, express or im-plied, with respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

v

Preface

Why do we need a book on the dating of neurological injury? Over the last decade, I have personally reviewed more than 80 medical–legal cases related to neurological disease in which some aspect of the case involved imaging technology. Many cases – in clinical situations, such as alleged birth-related hypoxia or ischemia, surgically-related injuries, or trauma-induced spinal cord or brain substance abnormalities, for example – shared the need both to detect the presence of an injury and to date the time it occurred. While a minority of cases involved the misdiagnosis of an aneurysm, a delayed diagnosis of spinal fracture, or orbital injury during a surgical procedure, the large majority of cases used modern imaging, first to detect if an injury had occurred to the brain substance or the spinal cord (collectively referred to as the central nervous system) and second to determine, if an injury did occur, at what time it occurred.

In a medical–legal setting, the interplay between the radiographic findings and the clinical find-ings has several possible scenarios. On one extreme, the imaging findings may be so unequivocal that no doubt exists as to what occurred and even little doubt about when it occurred. On the oppo-site extreme, the radiologic findings may be either completely noncontributory or may show that an event occurred but offer no insight into when it occurred (and hence its proximate cause). In between is a gray area in which the clinical history can often be very helpful in delineating the imaging findings to more accuracy and specificity; and the reverse may also be true, i.e., that the radiology may help clarify the clinical picture. My hope is that this text will be helpful in all situa-tions – from those in which radiology is clear, to those in which the findings are less certain though still present – by providing guidelines and principles for the application of imaging findings.

The realm of this book is not to discuss specific clinical and radiographic findings at the level of the medical expert radiologist, nor is it intended to be an exhaustive treatise on recognizing the imaging signs of brain abnormality, as that is more appropriately covered in a textbook on medical imaging. Rather, I intend to represent in a systematic fashion the principles involved in the interpre-tation of images of the central nervous system specifically in a medical–legal setting where concern exists about the occurrence and timing of an injury.

What this book uniquely presents is a new way to approach the dating of neurological injury as imaged by modern computed tomography (CT), magnetic resonance (MR), and ultrasound (US). Throughout the text, I describe dating by two distinct but complementary methods. In the first, I explain how knowledge of the dynamic and rapidly changing imaging findings that occur in the first few weeks after an injury permit dating in this acute period. In the second, I illustrate how patterns of injury with specific features can date with some accuracy the time an injury occurs, which may be much earlier than the time when the image was obtained. This tends to be dating that occurs in the chronic period.

Chapters are presented in a logical progression beginning with the general appearance of normal brain and progressing to the way abnormalities manifest themselves on CT, MR, and US images. The emphasis in these discussions is on the appearance of edema and of hemorrhage, as these two findings are the brain’s most common response to injury. I discuss the role of contrast in central

vi Preface

nervous system (CNS) imaging, which will lead to a discussion of how infarction (death of tissue), ischemia (decreased blood flow to tissue that is still potentially alive and recoverable), and hemor-rhage change with time as seen on CT, MR, and US images, and a dialog of what different patterns of injury tells us about the mechanism, severity, and duration of injury. This then permits a state-ment of what I consider the overriding principles of image interpretation as they relate to legal matters and a frank discussion, based on everything mentioned up to this point, of what can and what cannot be said in a medical–legal setting based on the imaging findings. The last chapter is on the root causes for uncertainty in dating neurologic events from imaging studies.

vii

Acknowledgments

The following individuals were instrumental in helping me bring this book to publication. I wish to thank them all.

Dr. James Scatliff and Dr. Robert Whaley - my Neuroradiology mentors at UNC in Chapel Hill•The Vanderbilt radiology fellows and attending staff who assisted with case acquisitions•The members of the Neuroradiology section at Vanderbilt who assisted me •Dr. Jeanette Norton for her assistance in multiple revisions of the first chapter on anatomy.•Janet Staley for her effort as an editorial and grammatical assistant who reviewed the entire •manuscript several times prior to submission.Dominic Doyle for his preparation of the illustrations•Administrative secretarial assistance from Tara Timmons, Debbie Holland and Dewain •PattersonThe Springer editorial staff in both New York and in India•And most importantly, Lynn and our family, Erin, David, John and Carrie. They have continually •supported me in my effort to write this book with their encouragement and understanding. It is to them that I dedicate this book.

ix

Contents

Part I Fundamentals

1 The Structure of the Normal Brain and Its Imaging Appearance ................................. 3Introduction ........................................................................................................................... 3The Layers of the Scalp, Skull, and Meninges ..................................................................... 4

The Scalp .......................................................................................................................... 4The Skull ........................................................................................................................... 4The Epidural Space ........................................................................................................... 5The Dura ........................................................................................................................... 6The Subdural Space .......................................................................................................... 6The Arachnoid Membrane and Subarachnoid Space ........................................................ 7The Pia .............................................................................................................................. 8The Visible Outer Surface of the Brain ............................................................................. 8The Imaging Appearance of the Normal Scalp, Skull, and Meninges .............................. 8

Brief Overview of Brain Anatomy ........................................................................................ 9The Blood Supply to the Brain ............................................................................................. 10

The Arterial Vessels .......................................................................................................... 10The Venous Vessels ........................................................................................................... 14Methods for Imaging the Intracerebral Arteries and Venous Structures ........................... 16

The Cerebral Hemispheres .................................................................................................... 20The Lobes of the Cerebral Hemispheres ........................................................................... 20The Cortical Gray Matter .................................................................................................. 22White Matter of the Cerebral Hemispheres ...................................................................... 22Deep Gray Matter Nuclei of the Cerebral Hemispheres ................................................... 23Imaging Appearance of the Normal Cerebral Hemispheres ............................................. 24

Cerebellum ............................................................................................................................ 27Gray Matter of the Cerebellar Hemispheres ..................................................................... 32Deep Cerebellar White Matter .......................................................................................... 32Imaging Appearance of the Cerebellar Hemispheres ....................................................... 32

Brainstem .............................................................................................................................. 32The Imaging Appearance of the Normal Brain Stem ....................................................... 35

Ventricles ............................................................................................................................... 35The Ventricles – Normal Shape, Size, and Position .......................................................... 35The Nonventricular CSF Spaces (Cisterns) ...................................................................... 37CSF Dynamics .................................................................................................................. 38Imaging Appearance of the Normal Ventricles, CSF Spaces,and CSF Dynamics ........... 40

References ............................................................................................................................. 40

x Contents

2 The General Appearance of Edema and Hemorrhage on CT, MR and US (Including a General Introduction to CT, MR and US Scanning) ................................. 43Introduction ........................................................................................................................... 43CT Scanning: The Absolute Basics ...................................................................................... 43MR Scanning: The Absolute Basics ..................................................................................... 45US Scanning: The Absolute Basics ...................................................................................... 47Edema ................................................................................................................................... 48

Edema on CT Scanning .................................................................................................... 51Edema on MR Scanning ................................................................................................... 52Edema on US Scanning .................................................................................................... 52

The General Appearance of Hemorrhage ............................................................................. 53Hemorrhage on CT Scanning ............................................................................................ 54Hemorrhage on MR Scanning .......................................................................................... 55Hemorrhage on Ultrasound ............................................................................................... 55

Chapter Summary ................................................................................................................. 57References ............................................................................................................................. 58

3 The Basics of Contrast and Its Role in Dating ................................................................. 59Introduction ........................................................................................................................... 59Basic Principles to Understanding the Use of Contrast in the Brain .................................... 59

For CT Contrast ................................................................................................................ 61For MR Contrast ............................................................................................................... 61

MR Contrast Dose and Pulse Sequence Choice ................................................................... 62MR Contrast Effects vs. Flow Void Effects .......................................................................... 62Clinical Importance of Contrast Enhancement ..................................................................... 67References ............................................................................................................................. 67

4 How the Imaging Appearance of Edema and Hemorrhage Change Over Time on CT, MR, and US: Dynamic (Acute) Dating ............................... 69Introduction ........................................................................................................................... 69

Changes of Edema Over Time on CT ............................................................................... 70Changes of Edema Over Time on MR .............................................................................. 71

Locations of Possible Intracerebral Hemorrhage .................................................................. 74Changes of Hemorrhage Over Time on CT ...................................................................... 75Changes of Hemorrhage Over Time on MR ..................................................................... 76

T1 Changes Over Time Within a Hemorrhage on MR Scanning ......................................... 81T2 Changes Over Time Within a Hemorrhage on MR Scanning ......................................... 81FLAIR Changes Over Time Within a Hemorrhage on MR Scanning .................................. 84Gradient Echo (Magnetic Susceptibility-Weighted) Changes Over Time Within a Hemorrhage on MR ................................................................................................ 85

Changes of Edema and Hemorrhage Over Time on US ................................................... 85References ............................................................................................................................. 86

5 Patterns of Parenchymal Injury: Pattern (Chronic) Dating ........................................... 89Introduction ........................................................................................................................... 89Beginning Principles ............................................................................................................. 89Factors Affecting the Outcome of a Region of Ischemia ...................................................... 90

Regional Variations in Perfusion ...................................................................................... 90Variations in Severity of Insult .......................................................................................... 91Variations in Duration of Injury ........................................................................................ 91Variations in Metabolic Activity ....................................................................................... 92

xiContents

Patterns of Parenchymal Injury ............................................................................................. 92Patterns Where There Has Been Total Loss of a Focal Region of Brain Parenchyma ......................................................................................................... 92Injuries That Affect the Brain Diffusely ........................................................................... 94Common Patterns That Show Targeting, with Partial or Total Cell Loss ......................... 95

References ............................................................................................................................. 100

Part II Application to the Medical-Legal Setting

6 Principles of Dynamic Dating in the Medical Legal Setting ........................................... 103Introduction ........................................................................................................................... 103General Comments on Dating an Event by CT and/or MR and/or US in the First 2 Weeks ............................................................................................................... 103Dating Edema by CT Alone .................................................................................................. 104Dating Hemorrhage by CT Alone ......................................................................................... 104Dating Edema by MR Alone ................................................................................................. 105Dating Hemorrhage by MR Alone ........................................................................................ 106Dating Edema by US Alone .................................................................................................. 107Dating Hemorrhage by US Alone ......................................................................................... 107Dating Events by CT in Conjunction with MR, and with US .............................................. 107

7 Principles of Pattern Dating in the Medical Legal Setting .............................................. 111Introduction ........................................................................................................................... 111

Concerning Edema and Infarction .................................................................................... 111Concerning Hemorrhage ................................................................................................... 113

Conclusion ............................................................................................................................ 115

8 Therefore, What Can Be Said Based on the Images, and What Can’t Be Said Based on the Images ................................................................. 117Introduction ........................................................................................................................... 117

9 The Root Causes of Uncertainty in Dating Neurologic Events Based on Imaging Findings ........................................................................................................... 119Introduction ........................................................................................................................... 119The Interpretation of the Available Images Varies from Expert to Expert, with Disagreement as to Whether Certain Findings Are Present or Not .............................. 120Image Findings Are Acknowledged to be Present by All Observers; However, the Interpretation of the Findings Varies from Expert to Expert .......................................... 120Multiple Findings are Present for Which the Neuroradiological Dating Methods We Have Discussed Produce Conflicting Time Periods as to the Probable Occurrence of the Injury .......................................................................... 121Radiographic Findings Which Are at Odds with the Clinical Picture .................................. 121Conclusion ............................................................................................................................ 121

Index ........................................................................................................................................... 123

Part IFundamentals

3J.L. Creasy, Dating Neurological Injury: A Forensic Guide for Radiologists, Other Expert Medical Witnesses, and Attorneys, DOI 10.1007/978-1-60761-250-6_1, © Springer Science+Business Media, LLC 2011

Abstract This chapter is an introduction to the anatomy and terminology necessary for understanding the remainder of the book. It presents the names of the regions and structures of the brain, the various spaces of the intracranial compartment, the significant anatomy of the major structures that surround the brain, and the spaces that contain cerebrospinal fluid both within and around the brain. Because infarctions and hemorrhage are major topics discussed later in the book, this chapter places heavy emphasis on the vessels of the brain, normal vascular anatomy of the brain, normal vascular territories, and the various means of imaging these vascular structures.

Keywords Brain • Brainstem • Cerebellum • Computed tomography (CT) • Magnetic resonance (MR) • Meninges • Neuroanatomy • Scalp • Skull • Spinal cord • Ventricles • Ultrasound (US)

Introduction

This chapter’s anatomic discourse is intended to introduce to those unfamiliar with neuroanatomy the basic concepts necessary to understand the more detailed discussion which will follow. Topics to be covered in this chapter are: (1) layers of the scalp, skull, and meninges (the tissue coverings of the brain and spinal cord); (2) blood supply (arterial and venous); (3) the cerebral hemispheres; (4) the cerebellum; (5) the brainstem; and (6) the ventricles (fluid-filled cavities within the brain) and cerebral spinal fluid (CSF) dynamics. Clinicians, experts, or others with an existing, solid knowledge of neuroanatomy and the normal appearance of the brain on computed tomography (CT), magnetic resonance (MR), and ultrasound (US) may need only to skim this chapter as a review before beginning in earnest in Chap. 2. However, regardless of one’s level of familiarity with these topics, individuals who are unfamiliar with the basic terminology used to describe CT, MR, and US should briefly read the first section of Chap. 2 before tackling this chapter on introductory anatomy and the imaging appearance of the brain.

Finally, before we proceed, we must briefly discuss the standard nomenclature for orientation of images. When referring to different cross-sectional images of the brain, three standard orienta-tions are used. As a reference, if one uses a person standing, facing you, the viewer, then an axial plane is a section parallel to the floor – that is, perpendicular to the long axis that runs from the head to the feet. A coronal plane is at right angles to an axial plane and results in thin sections, as though one were viewing a small slice of the brain from the front. A sagittal plane is perpen-dicular to both of the previous planes and results in a thin section of the patient’s brain viewed from the side (Fig. 1.1). If images of a patient are initially obtained in one plane and then used to produce images in a different plane, the second set of images is said to be reformatted.

Chapter 1The Structure of the Normal Brain and Its Imaging Appearance

4

Fig. 1.1 Anatomic planes. Oblique view of the head showing standard anatomic orientation of a coronal, a sagittal, and an axial (or transverse) plane

1 The Structure of the Normal Brain and Its Imaging Appearance

The Layers of the Scalp, Skull, and Meninges

A discussion of normal brain and skull anatomy is best started with an explanation of the layers of tissue related to the skull, beginning most superficially and extending sequentially into the deeper and deeper tissues. Figure 1.2 is a graphical representation of the layers.

The Scalp

The scalp is the outermost layer of the tissues of the head. Beginning superficially and progressing to successively deeper tissues encountered are the skin (epidermis and underlying dermis), a layer of fat, a layer of dense fibrous tissue (the aponeurosis), another thin layer of fat, and the periosteum, covering the outer surface of the bony skull.

Clinical Note: Blood can collect between any of these layers. If it is located in the skin and underly-ing fat above the aponeurosis, it is termed a caput succedaneum. If located between the aponeurosis and the periosteum, it is a subgaleal hemorrhage. Lastly, if between the periosteum and the outer table, it is termed a subperiosteal hemorrhage – or, alternatively, a cephalohematoma [1].

The Skull

The cranial portion of the skull, composed of bone, has a dense outer layer – the outer table – and a second dense inner layer, the inner table. Between the inner and outer tables is the diploic space, which contains fat and myeloproliferative elements – unlike the thicker, denser bone on either side. The myeloproliferative elements are constituents of the body which produce and are the precursors of white blood cells, red blood cells, and platelets.

5

Fig. 1.2 Layers of the scalp, skull, and meninges. Intracranially, the space between the skull and the dura is the epidural space, the space between the dura and the arachnoid membrane is the subdural space, and the space between the arachnoid membrane and the pia membrane closely applied to the brain surface is the subarachnoid space. Cerebral spinal fluid (CSF) fills the subarachnoid space normally


The major bones comprising the sides and top portion of the skull are the frontal, parietal, temporal, and occipital bones. The underlying lobes of the brain take their names from the bones which overlie them. Sutures are the point of contact between adjacent bones. The paired coronal sutures are the line of connection between the frontal and parietal bones; the single midline sagittal suture is where the two parietal bones meet in the midline at the top of the skull and the posteriorly located, paired lambdoid sutures are where the occipital bone meets the parietal bone on either side. In the neonate the region in the anterior midline at the front of the sagittal suture is not yet completely fused, and this anterior fontanelle permits US examination of the brain up to about 1 year of age (when the fontanelle closes completely) (Fig. 1.3).

The Epidural Space

The epidural space is a potential space; meaning that it has the possibility of existing, but in many normal individuals, does not. It is the space between the inside of the skull (or inner table) and the endosteal layer of the dura (the outermost connective tissue covering the brain) which has produced a periosteum, which is normally closely and firmly applied to the inner surface of the skull. In order for anything to occupy this space, this periosteum must be stripped (separated) from the inner table.

Clinical Note: Blood collections in the epidural space most commonly occur as the result of a skull fracture, which ruptures blood vessels either in the skull or in the dura. As arterial pressure builds behind the blood and strips the periosteum from the inner table, a clot forms within the epidural space. Characteristically, blood in the epidural space has a biconvex shape, meaning both sides of the clot bow outward. This is due to the difficulty the clot has in stripping the dura from the inner table, so that the size of the hematoma is restricted and the margins of the hematoma are sharply defined.

6 1 The Structure of the Normal Brain and Its Imaging Appearance

Fig. 1.3 Bones and sutures of the skull. Side and top view of the skull showing the major flat bones which make up the calvarium – the frontal, parietal, temporal, and occipital bones; and the major connecting sutures – the coronal, sagittal, and lambdoid sutures

Since the dura is very tightly applied to the skull at the coronal, sagittal, and lambdoid sutures, epidural hemorrhages rarely cross cranial suture lines (Figs. 2.12 and 2.13).

The Dura

The dura is the outermost, toughest, thickest layer of the meninges, which collectively are the three connective tissue layers that cover the brain. The meninges consist of the dura, the arach-noid membrane, and the pial membrane. The dura is further subdivided into an endosteal layer (firmly applied to the bone and producing a periosteum as described above) and a meningeal layer. These two portions of the dura are normally closely applied to each other throughout most of the inside of the skull. However, along the top of the skull and the rear of the skull on either side, the two layers separate and produce a venous drainage passage or dural sinus deep in the skull (Fig. 1.2).

The inner portion of the dura also subdivides the major compartments of the brain. The inner meningeal layer of the dura is reflected inward, toward the center of the skull and forms two prin-cipal septae. The first, the tentorium cerebelli, runs horizontally, separating the occipital lobe of the cerebral hemispheres from the cerebellum and brainstem. The second, the falx, is a midline, verti-cally oriented layer of meningeal dura that separates the right from the left hemisphere. This separa-tion is partial, as a side-to-side connection remains, represented by the major white matter tract of the corpus callosum (Fig. 1.4).

The Subdural Space

The subdural space is the space between the meningeal layer of dura and the arachnoid membranes. This space is normally very small, with the potential to enlarge.

7

Fig. 1.4 Dura. The dura consists of an endosteal layer and a meningeal layer. The endosteal layer is applied to the inner surface of the skull and forms the periosteum. The meningeal layer is normally fused to the endosteal layer, but is separated from the latter at locations of dural venous sinuses. It is the meningeal layer that turns inward toward the center of the skull, forming the two major septae – the midline falx cerebri structure which is attached superiorly to the skull but is open inferiorly, and the horizontally oriented tentorium cerebelli which separates the occipital lobe above the tentorium from the cerebellum and brainstem below


Clinical note: Hemorrhage within the subdural space can flow more freely over the cerebral hemi-spheres, but it does not flow down into the sulci (troughs) between the cortical gyri (the rounded, curved linear structures on the surface of the brain). It is not limited by suture lines and has a char-acteristic convex/concave shape – that is, the convex side hugs the inner table and dura and the concave side faces the brain below (Figs. 2.12 and 2.13).

The Arachnoid Membrane and Subarachnoid Space

The arachnoid membrane is normally closely applied to the meningeal layer of dura. The subarach-noid space is located between the arachnoid membrane and the pia (which is closely applied to the brain surface) and is filled with cerebrospinal fluid. This space is important for several reasons. First, the arteries and cortical veins on the surface of the brain lie within this space. Second, it is the space into which aneurysms most commonly rupture. Next, it is also the space into which nonan-eurysmal hemorrhage occurs, most commonly due to trauma. Finally, it is the space in which puru-lent material accumulates in a meningeal infection.

Clinical Note: The most common cause (occurring 80–90% of the time) of subarachnoid hemorrhage in the nontraumatic patient is rupture of an intracerebral aneurysm (a weakened portion of an artery that has abnormally ballooned) [2]. However, overall, the most common cause of subarachnoid hemorrhage is trauma, as trauma is a much more common entity than aneurysmal rupture. Therefore, the CSF can become bloody both from aneurysm rupture and from head trauma. The subarachnoid space can also be filled with inflammatory cells in cases of meningitis (infection or inflammation of the meninges) and in most cases of cerebritis (infection or inflammation of the brain). Other causes of dense CSF include proteinaceous inflammatory fluid or frank pus from a truly fulminate bacterial infection.

The subarachnoid space over the hemispheres freely communicates with the subarachnoid space around the spinal column and with the CSF within the ventricles (fluid-containing cavities


within the brain) via small holes in the fourth ventricle, called the foramen of Magendie and Luschka (see Fig. 1.37).

Finally, the arterial vessels that supply the brain, as well as the cortical veins that drain blood from the brain, run within the subarachnoid space. As the incoming arterial vessels branch progres-sively smaller, they remain on the surface of the brain until they reach a small size, at which point they penetrate the surface of the brain, carrying with them in a circumferential fashion a contiguous extension of pia and the subarachnoid space. This extension of the subarachnoid space around the penetrating arteries and draining veins is called the Virchow–Robin space. The fluid within this space is CSF.

The Pia

This, the last and deepest layer of the meninges, is a thin layer closely applied to the surface of the brain.

The Visible Outer Surface of the Brain

The outermost part of the brain surface is composed of a thin mantle of cortex. The cortex is made up of cells, both neurons and their supporting elements of several different cell types, all collec-tively referred to as glial cells. Because in the fresh brain the neurons on the surface appear gray, the cortex is also referred to as gray matter. Just below the cortex, the tissue that is the axons of neurons – each one surrounded by a fatty sheath – interconnects different portions of the brain. Since these nerve sheaths appear “white” in the fresh brain, they are referred to as white matter. The cellular organization of the brain will be discussed in more detail in the section on the cere-bral hemispheres, below.

The Imaging Appearance of the Normal Scalp, Skull, and Meninges

In general, different elements of the scalp, skull and meninges show up differently on the two major imaging methods: CT and MR. CT demonstrates bony structures better than MR scans. MR is superior in the demonstration of the soft tissues of the scalp superficially and of the meningeal structures that are deep to the cranium. On CT, beginning most superficially in the scalp, a thin, dense (bright) line represents the skin or dermis. Deep to this is a predominantly fat layer of variable thickness that has low density – a CT number of less than zero (for more infor-mation on CT numbers, see the introduction to CT at the beginning of Chap. 2). Immediately deep into the subcutaneous fatty tissue is another dense, fibrous layer that is intimately applied to the outer surface of the skull. Due to its thinness, this structure is usually not discretely imaged. The thick, protective top and sides of the skull consist of both an inner and an outer layer (or table) of dense bone, which appear on CT as an inner and an outer white line, separated by an interposed lower density, more lucent (darker) line that is the diploic or marrow space. The contents of the diploic space are cellular and may represent fatty marrow, and thus the CT numbers can be fairly low. Even if a contrast agent is administered, it is difficult to image the dura on the inner surface of the inner table. Similarly, the pia and arachnoid are not easily visualized on CT scanning. The subarachnoid space, with the black (i.e., low density) CSF filling

9

Fig. 1.5 Appearance of layers of scalp and skull. Axial computed tomography (CT) with (a) brain windows, (b) soft tissue windows, and (c) bone windows. Even with intravenous contrast (not given in this case), the dura is not visible. Axial magnetic resonance (MR) scans – (d) T2-weighted, (e) FLAIR, (f) T1-weighted without contrast, and (g) T1-weighted with contrast. The major layers of the scalp including the fat-containing dermis, the outer table of the skull (ot), the diploic space (ds), the inner table of the skull (it), and the dura on the inside surface of the skull. In G, the enhancing dural structures are visible

Brief Overview of Brain Anatomy

the cortical sulci (clefts between the gyri) as well as the basilar cisterns, is usually well seen (CT of these structures – Fig. 1.5a–c).

MR scanning is unlike CT in that no radiation is used, only radio waves and a high-strength magnet. On MR, the soft tissues of the scalp will be bright on both T1- and T2-weighted imaging. (The terms T1- and T2-weighted refer to specific types of MR sequences. A more complete intro-ductory description of MR will be given at the beginning of Chap. 2.) If fat saturation is employed with either of these pulse sequences, the fatty tissues will appear dark. As for bone, on all pulse sequences bone has no signal, so both the inner and outer tables will show up as black lines. The diploic space contains cellular elements (including fatty tissue) and, therefore, will have some brightness on both T1- and T2-weighting. The dura is best seen only after a contrast agent has been administered. It is important to note that mild degrees of dural enhancement are much better visualized on MR than CT, as in MR one is trying to visualize an enhancing white line next to the black inner table – this is markedly easier to see than the thin, white enhancing line on CT next to another white line of the inner table dense bone (MR of these structures – Fig. 1.5d–g).

Brief Overview of Brain Anatomy

Before beginning a discussion of intracranial arteries and veins, and later CSF spaces – even before a more detailed discussion of the cerebrum, cerebellum, and brainstem – it is appropriate to briefly name the major brain substructures. The accompanying diagram shows the CNS

10

Fig. 1.6 Basic layout of the brain and skull base. (a) View from the side and (b) view from the top, with the top of the skull and the brain removed – looking down at the bony skull base. The cerebrum, brainstem, cerebellum, and spinal cord are labeled and shown in relation to the head and skull. For reference, other key intracranial structures such as the foramen magnum (the opening at the base of the skull where the brainstem meets the spinal cord and enters the spinal column), the pituitary fossa or sella (within which is the pituitary gland), the three (anterior, middle, and posterior) paired intracranial fossae, and the eyes or orbits are indicated. Lastly, directional terms are included, such as anterior (to the front), posterior (to the rear), midline (in the sagittal center of the body), and lateral (extending away from the midline)


structures of cerebrum, brainstem, cerebellum, and spinal cord in relation to the head and skull. For reference, other key intracranial structures, such as the foramen magnum (the opening at the base of the skull where the brainstem meets the spinal cord and enters the spinal column), the pituitary fossa or sella (within which is the pituitary gland), and the eyes, or orbits, are drawn. Lastly, directional terms, such as anterior (to the front), posterior (to the rear), midline (in the sagittal center of the body), and lateral (extending away from the midline) are demonstrated (Fig. 1.6).

The Blood Supply to the Brain

The Arterial Vessels

The vessels that bring blood to the brain consist of a system of larger arteries steadily narrowing to medium vessels, to progressively smaller caliber vessels, all of which lie in the subarachnoid space around the brain. At some point, when arterial vessels are small enough, they penetrate the brain surface, traveling radially from the outside in an inward direction to supply the brain tissue. Upon reaching their target, the arteries continue to branch into smaller and smaller arter-ies, then arterioles and, finally, a capillary system. Nutrients pass from the blood stream to the cells of the brain at the capillary level. These small blood vessels then reaggregate into venules, small veins, and then progressively larger veins. The arterial supply to the brain is divided into the anterior circulation, made up of the right and left internal carotid arteries (or ICAs) and their branches on either side and the posterior circulation, or vertebro-basilar system (Figs. 1.7 and 1.8) [3, 4].

11

Fig. 1.7 Major intracerebral arteries. These four views are diagrammatic representations of the appearance of the intracranial major arteries as viewed from an anterior (from the front) projection (a, c) and from a lateral (from the side) projection (b, d). Major intracranial branches of the right internal carotid artery (R ICA) viewed from the front (in (a)) and from the side (in (b)). Major posterior vessels viewed from the front (in (c)) and from the side (in (d))


The anterior intracranial circulation is made up of the right and left ICAs, which enter the skull base, passing up through the cavernous sinus (a venous structure located on either side of the pitu-itary gland in the very central aspect of the skull base). Just above the cavernous sinus each ICA bifurcates into an anterior cerebral artery (ACA) and a middle cerebral artery (MCA). The MCA passes laterally into the Sylvian fissure and, near its lateral extent, bifurcates into multiple branches which supply the lateral aspect of the brain.

The ACA passes antero-medially forward from the terminus of the ICA and, immediately in front of the pituitary region, a short linking artery, the anterior communicating artery (AComm), connects the right and left ACAs. Beyond this the ACAs continue, first forward and then up over and posteriorly over the corpus callosum (Fig. 1.7).

Posteriorly, the major arteries are the right and left vertebral arteries and the single artery they create when they fuse at or near the base of the brain: the basilar artery. The basilar artery then pro-ceeds superiorly in front of the brainstem until it reaches its terminus, the basilar summit. The blood supply for the cerebellum, brainstem, the rear of the cerebral hemispheres (the occipital and temporal lobes), and a portion of the deep structures within the cerebral hemispheres arises from this vertebro-basilar system. At, roughly, the level of the foramen magnum at the base of the skull, before they join to create the basilar artery, each vertebral artery gives rise to a posterior inferior cerebellar artery (PICA), which supplies the inferior portions of the cerebellar hemispheres and a portion of the medulla. The basilar artery gives rise to the anterior inferior cerebellar artery (AICA), supplying the midportion of the cerebellar hemispheres and, just below the terminus of the basilar artery, to the bilateral superior cerebellar arteries (SCAs) that feed the superior aspects of the cerebellar hemi-spheres. At its terminus, the basilar tip bifurcates into the two posterior cerebral arteries (PCAs).

Between the major arteries of the anterior and posterior circulations, one pair of connecting vessels, the right and left posterior communicating arteries (or PComm), travels from the distal portion of

12

Fig. 1.8 Circle of Willis. The circle of Willis, if complete, is a ring of vessels that connect the internal carotid (anterior) to the vertebro-basilar (posterior) circulation. This ring is often incomplete, as different segments of the ring may be absent in the normal individual. The right half of the circle, beginning posteriorly, starts at the tip of the basilar artery (ba) and proceeds clockwise through the proximal portion of the right posterior cerebral artery (PCA) or P1 segment (R P1), the right posterior communicating artery (PComm), the top end of the R ICA, the proximal portion of the right anterior cerebral artery (ACA) or A1 segment (R A1) and, finally, in the midline anteriorly, the anterior communicating artery (AComm). Similar labels apply to the left half of the circle of Willis. The shaded regions (A1 segment, PComm, P1 segment) denote the most commonly absent segments


the ICA backward to the early portion of the PCA on the same side. In this fashion, potentially, a circle of blood vessels lies at the base of the brain. Known as the circle of Willis, it is composed of the basilar tip itself and then, presuming one were viewing the circle of Willis from above, would consist of the proximal portion of the PCA and the PComm on the left, the distal aspect of the left ICA, the left ACA between the ICA terminus and the AComm (otherwise known as the A1 segment of the ACA). Following the AComm itself are the right A1 segment of the right ACA, the terminus of the right ICA, the right PComm and, finally, the proximal portion of the right PCA, which once again connects to the basilar artery terminus (Fig. 1.8 – Line drawing of circle of Willis).

However, it should be noted that, in the vast majority of patients, the circle of Willis is incom-plete, with the most common absences being lack of a P1 segment of the PCA (from the basilar terminus to the PComm), lack of a PComm, or lack of an A1 segment of the ACA (from the ICA terminus to the AComm) (see shaded vessel portions in Fig. 1.8). In fact, only about 30% of people have an absolutely, completely intact circle of Willis [5]. This fact is relevant to our dis-cussion of ischemia and infarction, as an intact circle of Willis allows shunting of blood to a region of brain that would otherwise progress to infarction. For example, an individual may have a sudden occlusion of the right ICA in the neck, but if, via an intact circle of Willis, blood can flow into the right ACA via the AComm from a normal left ACA, or into the very superior right ICA (and, thence, into the right ACA and right MCA), from a patent right PComm, then an infarc-tion may not occur. On the other hand, a person with an acute right ICA occlusion, who has absence of the right A1 segment of his right ACA and an absent right PComm, is doomed to have a large infarction of the effectively isolated entire MCA territory. Preexisting vascular anatomy

13

Fig. 1.9 Supratentorial arterial vascular territories. The anterior, middle, and PCA territories are indicated by shading. The ACA territory begins inferiorly as a wedge-shaped portion of the frontal lobes and extends superiorly as a band along the midline medial portion of the hemisphere. The middle cerebral artery (MCA) territory is the lateral aspect of the hemisphere beginning at the temporal tip and extending upward and backward until it reaches the anterior circulation territory anteriorly and superiorly and the posterior circulation posteriorly. The posterior cerebral territory is a wedge at the rear of the brain in the occipital lobe extending forward to meet the anterior and MCA territories. It includes the inferior aspect of the temporal and occipital lobes except for the temporal tip


is, therefore, one important factor in determining both if an infarction occurs and, if so, the severity of the injury.

The vascular territories of the supratentorial portion of the brain (that portion of the brain – the cerebrum – that sits above the tentorium cerebelli) are supplied by the anterior, the middle and the PCAs. The ACAs normally supply a wedge-shaped portion of the frontal lobe, beginning at the midline, and extending laterally to involve a wedge-shaped region of the brain. This ACA territory begins inferiorly at the floor of the frontal fossa and then extends up superiorly, narrowing as it gets more posterior and ending in a stripe several centimeters wide that extends posteriorly in a parasagittal location to near the level of the rear of the corpus callosum. The PCA, similarly, mainly supplies a triangular-shaped region of brain occupying the occipital lobe, beginning at the midline and extending laterally, and then extending up over the postero-lateral surface of the brain to meet the rear portion of the ACA distribution. The inferior surface of the temporal lobe, beginning just behind the temporal tip and extending all the way back to the occipital lobe, is also supplied by posterior temporal branches of the PCA. The MCA terri-tory occupies the area laterally in the hemisphere bounded by the anterior circulation anteriorly and superiorly, and the posterior circulation posteriorly and inferiorly. The MCA territory includes the majority of the posterior portions of the frontal lobe, lateral portions of the parietal lobe, and the anterior and superior portions of the temporal lobe. The MCA, therefore, supplies the majority of the lateral portion of the cerebral hemispheres (Fig. 1.9 – Line drawing of supratentorial vascular territories).

In the posterior fossa (that portion of the brain – brainstem and cerebellum – that lies under the tentorium cerebelli), the PICA vessels supply portions of the medulla, as well as the inferior por-tions of the cerebellar hemispheres. The SCA supplies the superior portion of the cerebellar hemi-spheres. The AICA arteries supply the midequatorial portions of the cerebellar hemispheres. The blood supply to the majority of the brainstem comes directly from the basilar artery itself (Fig. 1.10 – Line drawing of infratentorial vascular territories).

Clinical Note: Occlusions of any one of these named vessels or one of their branches above or below the tentorium cerebelli result in a geographic, sharply demarcated infarction (region of cell death), which includes all the tissues (grey and white matter) in the involved vascular supply. Such an

14

Fig. 1.10 Infratentorial arterial vascular territories. The major vascular territories in the posterior fossa are those supplied by the right and left posterior inferior cerebellar artery (PICA), anterior inferior cerebral artery (AICA), and superior cerebellar artery (SCA) and by smaller vessels arising directly from the basilar artery. The PICA territory is the inferior aspect of the cerebral hemispheres and portions of the postero-lateral medulla. The AICA supplies the midportion of the cerebellar hemispheres, while the SCA supplies the superior aspect of the cerebellar hemispheres. The majority of the blood to the brainstem (midbrain, pons, and medulla) itself is from small vessels that originate directly from the basilar artery


infarct is said to be arterial in nature, as it follows well-defined arterial vascular territories. The severity of a total arterial occlusion also depends on the amount of collateral blood flow that reaches the region deprived of blood by occlusion of its normal primary supplying vessel. Additional factors affecting the extent of the infarction will be discussed in Chap. 5.

The Venous Vessels

Venous vascular anatomy is a complement to the arterial system, but does not merely duplicate each vessel. The venous drainage system is different in two important ways. First, there is no venous intracranial counterpart to the major arteries comprising the anterior, middle, and cerebral arteries. In order words, the arterial territories supplied by the arteries are different from the venous territo-ries drained by the veins and dural sinuses. Second, unlike the feeding arteries which all follow the same general pattern, once the blood is collected into medium and large veins, two functionally and anatomically distinct venous drainage systems exist – the deep, central midline venous system and the system that consists of the cortical veins and dural venous sinuses [6, 7].

15

Fig. 1.11 Deep (midline) venous system. The major components, directly in the midline, with blood flowing from anterior to posterior, are the internal cerebral veins, the short vein of Galen, and the straight sinus. Contributing to these midline veins, the paired thalamostriate veins flow forward in the lateral ventricle, going through the foramen of Monro and then joining the septal veins to create the internal cerebral vein. More posteriorly, the basal veins of Rosenthal drain the lateral ventricles and, together with the internal cerebral veins, form the Vein of Galen


In the deep system (Fig. 1.11), as the blood collects from the capillaries into small venules and then into larger veins, the brain drains centrally inward at right angles to the surface of the ventricles in the medullary veins. These veins, in the walls of the body of the lateral ventricles, collect blood that flows forward toward the foramen of Monro and then posteriorly in the roof of the third ven-tricle as the internal cerebral vein. This system is duplicated on both sides. Upon reaching the rear of the third ventricle, these veins are joined by the right and left basal veins of Rosenthal, which drain the temporal lobes, and this set of four veins collectively form the vein of Galen. This midline, normally short, one-centimeter vein swings upward and posteriorly and continues inferiorly and posteriorly as the straight sinus that extends all the way to the rear of the skull, joining the superior sagittal sinus at the confluence of the sinuses.

The second venous drainage pattern involves veins that radiate outward toward the surface of the brain (Fig. 1.12). On the surface of the brain, the cortical veins collect to form increasingly larger veins. These larger cortical veins then drain superiorly into the superior sagittal sinus and laterally and posteriorly into the transverse sinus. For these cortical veins to reach the dural sinuses (channels formed between the endosteal and meningeal layers, or twofolds of meningeal dura), they must cross the intervening subarachnoid space. The veins which traverse these spaces are the bridging veins, which are susceptible to injury from a number of causes and which, if injured, can result in bleeding into the subarachnoid space. Thus, whether the cortical veins are small or one of the larger, named cortical veins, they all, via bridging veins, drain into the supe-rior sagittal sinus, other dural sinuses or smaller venous plexuses around the skull base centrally. The largest dural venous sinus is the superior sagittal sinus, which runs from the frontal region to the confluence of the sinuses in the midline of the skull posteriorly. This superior sagittal sinus has a triangular cross-section, with all three sides formed by layers of dura (Fig. 1.2).

16

Fig. 1.12 Major cortical veins and dural venous sinuses. The diagram demonstrates the major dural venous sinuses: the superior sagittal sinus, the inferior sagittal sinus, the transverse sinuses, the sigmoid sinuses, and the straight sinus. In addition to these major dural venous sinuses are numerous cortical veins, the largest of which are the superiorly located and draining vein of Trolard, which drains in the superior sagittal sinus, and the laterally oriented vein of Labbe, which drains into the transverse sinus


The other significant large dural venous sinus begins where the sagittal sinus joins the straight sinus at the confluence of the sinuses, then splits and runs horizontally on both sides at the junc-tion of the tentorium and the inner table as the left and right transverse sinus. Upon reaching the postero-lateral corner of the skull, this sinus turns downwards in an “S” shaped curve (hence called the sigmoid sinus) and forms the jugular bulb, which is the beginning of the internal jugu-lar vein on each side.

Clinical Note: Occlusions of dural venous sinuses have a propensity to cause venous infarcts due to blockage of venous outflow. Unlike a typical arterial occlusion, a clot within the superior sagittal sinus can cause an infarct on both sides of the midline. Similarly, a clot of a transverse sinus can cause infarcts or ischemia (decreased blood flow that has not yet progressed to infarction) in terri-tories that cross the MCA and PCA territories, or even cause infarctions both above (in the occipital lobe) and below (in the cerebellum) the tentorium. Venous infarcts are also more prone to hemorrhage than arterial infarcts. An interesting imaging correlation of venous occlusion is that symptomatic venous thrombosis (occlusion of a blood vessel caused by an intraluminal blood clot) in the head is almost always accompanied by signal changes on MR [8]. Occlusions and thromboses of the central deep venous system cause abnormalities involving the deeper, more centralized portions of the brain, usually including the basal ganglia.

Methods for Imaging the Intracerebral Arteries and Venous Structures

One has a number of options in imaging the intracerebral arteries and veins, including traditional catheter angiography, routine CT scanning, computed tomography angiography (CTA), routine MR imaging and magnetic resonance angiography (MRA), and magnetic resonance venography (MRV). Each of these modalities has strengths and weaknesses, relative indications and, consequently, each images the intracerebral arteries and veins to different degrees and clarities. Table 1.1 compares these five methods of imaging based on whether contrast is utilized or radiation is employed

17

Table 1.1 Five methods of angiographic imaging

Method Imaging used Contrast Invasive Best used forSignificant artifacts

Catheter angiography X-rays Yes Yes, requires vessel catheterization

Arterial anatomy No

CTA CT Yes Yes, IV only Arteries NoCTV CT Yes Yes, IV only Veins NoMRA MR No No Arteries YesMRV MR No No Veins No


and provides a relative rating of ease of visualization of the major veins, dural venous sinuses, and arterial structure.

Traditional catheter angiography is performed by the sterile and fluoroscopically guided introduc-tion of a small tube into a peripheral artery, usually the femoral artery at the groin. The catheter is directed retrograde up the descending aorta and then selectively into the vessels in the neck that feed blood to the brain, either the ICAs or the vertebral arteries. Following contrast injection, films are made about the head in various projections. This oldest and most traditional method of examining the vascular system in the brain carries with it several risks which fall into three categories: (1) risks associated with the contrast agent, (2) risks associated with injury to the femoral artery at the site of the puncture, and (3) risks associated with embolic or thrombolic processes occurring from the end of the catheter. A true anaphylactic reaction (i.e., a life-threatening, heart stopping, hard-to-breathe reaction to a chemi-cal agent) can conceivably occur with any iodinated contrast (a contrast agent containing iodine – used for traditional catheter angiography, CTA and contrasted CT exams) administration.

Complications that can arise at the site of the puncture include transection of the femoral artery, laceration of the femoral artery resulting in local hematoma, and induced arterio-venous fistulas (an abnormal direct connection between an artery and a vein). Complications occurring from the end of the catheter usually relate to the catheter tip displacing plaque or producing distal emboli (any substance – clot or arterial wall plaque – that forms in one part of the arterial system and travels along the blood vessel, eventually lodging and completely occluding the vessel more distally). Thrombus (arterial occlusion that occurs by a clot forming at a stationary place in the blood vessel without traveling distally) can also occur at the end of the catheter and can distally embolize into the brain, causing strokes or infarctions. Traditional catheter angiography excels at visualizing arteries of the intracranial vasculature. It is relatively less efficient at visualizing the cortical veins and the dural venous sinuses.

CT, when not performed in a way to specifically image the vessels (that would be CTA), can, nevertheless, provide visualization of the intracranial vasculature. Gross lesions of the dural sinuses or arteries often cannot even be seen on routine CT without contrast, but the sensitivity is markedly increased by performing CT with contrast. However, the best manner in which to use CT to visualize the vasculature is to perform a specialized CTA examination. This employs a bolus administration of contrast material injected into a peripheral vein (as opposed to directly injecting it into an intracerebral artery) and using the special capabilities of a modern rapid CT scanner to image multiple thin sections through the region of interest, in this case the brain. This type of examination eliminates complications associated with the arterial puncture and with the catheter, though it retains the possibility of complications due to contrast administration and the ionizing radiation. Because of the relative lack of complications, CTA is becoming the method of choice for visualization of the intracranial vasculature. Numerous reports are now citing the utility of CTA for detecting atherosclerosis, vasospasm, arterial injury, arterial thrombosis, and intracranial aneurysms.

18

Fig. 1.13 Catheter angiography of intracranial arteries; typical catheter angiograms of the intracranial circulation. Angiograms are of the posterior circulation in (a–c) and of the anterior circulation in (d, e). (a) Early arterial phase of an anterior–posterior projection image from an injection of the left vertebral artery. (b) Late arterial phase of an anterior-projection image from an injection of the left vertebral artery and (c) lateral projection from an injection of the left vertebral artery. (d) Anterior–posterior projection image of the head from an injection of the left internal carotid artery in the neck, and (e) lateral projection image of the head from an injection of the left internal carotid artery


Routine MR (without using specialized MRA sequences) is also useful for visualizing gross vascular anatomy. On a routine MR sequence, patent (open) vessels with good flow will show up as areas of signal flow voids (the lumens of patent vessels will appear black). This aspect of MR can be used to demonstrate that a vessel is not occluded. This is velocity sensitive and works best for rapidly flowing blood; hence, flow voids are easier to visualize in intracerebral arteries than they are in intracerebral venous structures. If specific visualization of the vasculature is required, it is helpful to perform a specialized MRA examination, which visualizes the intracranial arteries. The intracranial veins and dural sinuses can be removed by appropriate saturation bands. Similarly, if it is desired to specifically visualize the intracranial venous structures, one can remove the arterial structures by using saturation maps, which produces an MRV. The accompanying figures give demonstrations of these five methods of visualizing the intracranial arteries, veins, and dural venous sinuses (Figs. 1.13–1.17).

19

Fig. 1.15 CTA of the veins and sinuses. Images were produced from CTA source images acquired in the later (more venous) phase. (a) Sagittal reconstruction of the image data at 25 mm-slice thickness. (b) Coronal 10 mm-thick reconstruction at the level of the internal cerebral veins and superior sagittal sinus. (c) Coronal reconstruction more posteriorly through the level of the transverse sinuses

Fig. 1.14 Computed tomography angiography (CTA) of intracranial arteries. (a) Axial source (0.8 mm) single slice from the original data acquisition. (b) Fused axial 3 mm thick slices. (c) Coronal 10 mm thick slices. (d) Anterior 3D surface reconstructions of the entire data set. (e) Superior view of same data set. The major labeled arteries are labeled by the conventions in Figs. 1.6 and 1.8


20

Fig. 1.17 MRV of the intracranial veins and sinuses. This technique is slightly different from the routine intracranial MRA in that the images are usually obtained in the coronal plane. (a) Coronal source images. (b) MIP reconstruction in the anterior–posterior direction. (c) MIP reconstruction in the lateral view

Fig. 1.16 Magnetic resonance angiography (MRA) of the intracranial arteries. (a) Axial source images. (b) MIP reconstruction in the anterior–posterior direction. (c) Lateral MIP reconstruction as viewed from the side (MIP stands for maximum intensity projection – a technique for generating images similar to traditional catheter angiograms from axial source images.)


The Cerebral Hemispheres

The Lobes of the Cerebral Hemispheres

The cerebral hemispheres are organized into lobes (Fig. 1.18). The developmental embryology behind the lobes is far beyond the scope of this book, but, at a descriptive level, the traditional lobes of the hemispheres are the frontal, parietal, temporal, and occipital lobes – each taking its name from the bones that overlie them. However, some texts will also discuss the insular region as a unique portion of the brain, a portion that is not visible from the surface. Certain functions of the central nervous system localize within each lobe of the brain; those gross localizations of specific functions will be discussed in our anatomic survey [9–11].

21

Fig. 1.18 Lobes of the brain and functional localizations. Lateral (in (a)) and medial (in (b)) views of the brain show the frontal, parietal, temporal, and occipital lobes. Superimposed on these regions are the important functional regions of the brain related to motor activity, sensory activity, hearing, speech, vision, and memory


The brain consists of two paired cerebral hemispheres, each a mirror image of the opposing side. Each consists of four lobes and the insular region and each has, with one important exception, fairly rigidly defined margins. The exception occurs on the lateral posterior surface of the hemisphere where the posterior temporal, parietal, and lateral occipital lobes all join. Each frontal lobe occupies the most anterior portion of the intracranial compartment and is the largest lobe of the cerebrum in humans. When viewed from the side or top of the brain, the rear margin of the frontal lobe is a major sulcus called the central sulcus. The central sulcus runs from the midportion of the Sylvian fissure (the large infolding in the lateral surface of the brain which separates the frontal lobe above from the temporal lobe below) up over the lateral aspect of the brain to near the midline. The most posteriorly located gyrus in front of the central sulcus is termed the precentral gyrus, which has the important function of being the motor cortex (the area in which motor movement is initiated). On the medial surface of the brain, the frontal lobe extends all the way back to the counterpart of the central sulcus called the angu-lar sulcus. This defines the rear of the frontal lobe and the front of the parietal lobe.

The temporal lobe, best viewed from the lateral side of the brain, composes that portion of the cortex that is below the Sylvian fissure. The front of the temporal lobe is rounded and lies in the middle cranial fossa (Fig. 1.6). The rear of the temporal lobe, where it merges into the parietal and the occipital regions, is less well defined. Important functions localized to the temporal lobe include speech on the lateral surface; regions important for memory, including a structure called the hippocampus, are located on the medial surface.

The occipital lobe is the most posterior portion of the cortex. It is demarcated anteriorly on the medial surface by the parietal occipital fissure that separates the occipital lobe from the parietal lobe. On the lateral surface of the brain, the point at which the occipital lobe joins the temporal lobe inferiorly and the parietal lobe more superiorly is less well defined. The important function that localizes in the occipital lobe is vision, as the primary visual cortex (also called the calcarine cortex) is located primarily on the medial surface.

The parietal lobe is demarcated on the medial aspect of the cerebral hemispheres by the parietal occipital sulcus posteriorly and by the angular sulcus anteriorly. On the lateral surface of the hemi-spheres the central sulcus forms the anterior margin of the parietal lobe. The more posterior and more inferior edges where the parietal lobe merges into the occipital lobe and temporal lobe, respec-tively, are less well defined. The most anterior gyrus of the parietal lobe, as it sits behind the central


sulcus on the lateral surface of the brain, is termed the postcentral gyrus and is concerned with the sensation of touch, referred to as the somato-sensory cortex.

The last major cortical region is the insula or insular cortex. This portion of cortex forms the upper (from the inferior frontal lobe) and lower (from the upper temporal lobe) boundaries of the Sylvian fissure. At the deepest portion of the Sylvian fissure is additional cortex – contiguous with both the frontal and temporal lobes – that lies just lateral to the basal ganglia. None of this cortex is visible from a lateral view of the brain, but is best seen as a “T” shape on its side on coronal sections – in essence, it is cortex buried within the Sylvian fissure.

In addition to the localization of function mentioned above on each of the lobes, the general concept of eloquent vs. noneloquent brain should be addressed. Eloquent brain is that which is closely tied to a specific, measurable neurologic function, such as speech, hearing, vision or motor, or sensory control of a specific part of the body. Lesions in each of these areas will cause an easily defined clinical deficit. These areas of the brain collectively make up only a small percentage of the cortical surface. Much larger areas of brain are devoted to secondary or higher-order associative functions. For example, the primary visual cortex in the calcarine area allows one to see, but the interpretation of vision and the correlation with other things such as memory occur in secondary and tertiary-associated visual areas. A lesion in one of these areas can be more clinically subtle than a lesion in one of the primary or more eloquent areas of brain. These gross localizations of the brain, as well as indicated areas of important function, are shown in Fig. 1.18.

The Cortical Gray Matter

Within any lobe of the brain, the more detailed (i.e., microscopic) anatomy of the cerebral hemi-spheres shows that they are composed most superficially of a layer of cortex, or gray matter. This gray matter represents a majority of the neuronal cell bodies of the cerebral hemispheres with the exception of those seen in the deeper nuclei, discussed below. A neuron is the major interactive, computational component of the central nervous system. Functionally, a neuron consists of a central cell body which contains the nucleus of the cell, an array of processes (called dendrites) that extend out from the cell body and receive input from other cells, and an axon which transmits electrical impulses to other neurons. The axon is surrounded by a sheath of myelin, which enhances nerve conduction speed and is produced and maintained by nonneuronal cells called oligodendrocytes. These oligodendrocytes are one cell type of many included in the classification of glia – a classification which also includes astrocytes and microglial cells. The astrocytes are felt to produce and maintain the cytochemical supporting architecture of the brain – both normally and postinjury. Microglial cells are thought to participate in the removal of dead cellular material. A diagram of these cell types is shown in Fig. 1.19 [12, 13].

White Matter of the Cerebral Hemispheres

The white matter in the cerebral hemispheres is located deep to the cortex and functionally can be thought of as the wiring that connects cortex to cortex and to other structures. Many white matter tracts originate in one portion of the cortex and extend to other portions of the cortex, or to deep nuclei, the brainstem or all the way down to the spinal cord. At that point, the white matter fiber tracks (axons) join and synapse with additional nerve cell nuclei (a collection of neuron cell bodies) and pass out of the spinal cord and the spinal canal as peripheral nerves to supply both motor and sensory function to the remainder of the body. Usually not just one, but several synapses can occur between a cortical neuron and the final termination of a string of connected neurons that end in bodily tissue.

23

Fig. 1.19 Cellular elements of the brain (neurons, oligodendrocytes, astrocytes, microglia). This diagram demonstrates a simplification of the complex network of neurons and their supporting elements. A single neuron has been high-lighted. Note its extensive dendritic processes connecting with other neurons, as well as its single long axon which can connect locally or distally to additional neurons


Deep Gray Matter Nuclei of the Cerebral Hemispheres

The deep nuclei of the cerebral hemispheres are paired structures, symmetric about the midline, with the most important and radiologically visible ones being the caudate and the lentiform nucleus (com-posed of the globus pallidus and the putamen). Many images will also include the thalamus as part of the “deep gray matter,” although, unlike the aforementioned structures, it is of different embryologic origin. Regardless of their developmental origin, all of these structures are composed predominantly of gray matter [14, 15]. The blood supply to the basal ganglia region does not arise solely from branches of the main supratentorial vessels. Instead, portions of the basal ganglia are fed directly from the M1 segment of the MCA, from vessels arising off of the basilar tip and from the P1 segments of the PCA.

The caudate is a gray matter structure with a curved, comet shape. The largest portion is the head of the caudate, located anteriorly and nestled in the postero-lateral aspect of the anterior end of the lateral ventricle. The body of the caudate tapers considerably as it extends upwards and posteriorly over the body of the lateral ventricle, ending near the rear of the lateral ventricle. The lentiform nucleus consists of a lateral putamen and medial globus pallidus, which, together, are in the shape of a very fat, broad-based lens. The apex of the lens, which points medially, is the globus pallidus,

24

Fig. 1.20 Deep nuclei of cerebral hemispheres. This oblique view shows the relationship of the deep nuclei of the cerebral hemispheres to the midline, the ventricles, and the cortical surface. On either side of midline, the claustrum, caudate, amygdala, putamen, and globus pallidus together are the deep nuclei of the hemispheres. The thalamus is an additional grey matter structure on either side of the third ventricle but, unlike the deep nuclei, it is not of hemispheric origin


the broader lateral aspect is the putamen. The internal capsule sits around the medial apex of the lentiform nucleus and resembles a “V” in axial cross-section, with the apex of the V situated medially. This is a white matter structure containing a large number of long tracts, some of which ascend toward the cortex and some of which descend from the cortex to the more inferior parts of the brain and spinal cord. The thalamus is composed of many nuclei and sits medially to the posterior limb of the internal capsule. Many of the thalamic nuclei are important way-stations for most of the primary sense functions, such as vision and hearing. These structures are demonstrated in Fig. 1.20.

Imaging Appearance of the Normal Cerebral Hemispheres

The imaging appearance of the cerebral hemispheres presented here will follow closely the description of the lobes of the cerebral hemispheres as well as the line drawings in Figs. 1.6 and 1.18. As shown in Figs. 1.21–1.23, the lobes of the brain are well demonstrated on CT in the axial, sagittal and coronal planes. Similarly, these same lobes are well seen on MR in these same three planes, as shown in Figs. 1.24–1.26, and on US in Figs. 1.27 and 1.28.

Normal brain is characterized by the ability to clearly see the cortex as having a distinctly different appearance from the underlying white matter on CT and on most pulse sequences in MR. The normal brain is also symmetric with respect to the midline. Deep grey matter structures of the cere-bral hemispheres, such as the basal ganglia described below, are also easily discernible from their surrounding white matter.

Another attribute of “normal” brain is having an appropriate amount of brain parenchyma for the patient’s age. On a series of CT scans at different ages, Fig. 1.29 shows the normal amount of brain parenchyma that might be expected. As can be seen, as the patient ages the relative amount of brain parenchyma (both white matter and grey matter) decreases. This loss becomes even mildly noticeable only in the third and fourth decades of life, but then becomes increasingly prominent as one gets into the septuagenarian and octogenarian years. If a patient has more loss of brain parenchyma than is age appropriate, then one should consider the possibility of a global process affecting the brain. In addi-tion to the brain having an age appropriate amount of total mass, the amount of gray and white matter also changes with age. Often, the amount of brain parenchyma lost with time is most visible in the

25

Fig. 1.21 Multislice CT of normal cerebrum, source axial images. (a) Near the top of the brain, (b) at the top of the lateral ventricles, (c) at the mid portion of the third ventricle, and (d) at the mid portion of the temporal lobe. The major structures are indicated by the following legends: frontal lobe (fr), occipital lobe (occ), parietal lobe (par), temporal lobe (temp), Sylvian fissure (syl), falx cerebri (falx), lateral ventricle (lv), third ventricle (III), and fourth ventricle (IV)


white matter around the ventricles. Consequently, it is normal for the ventricles to show an increase in size with aging to compensate for this normal, gradual loss of periventricular white matter.

Normal changes of white matter density (on CT) and the signal intensity of white matter (on MR) occur with age. These changes are primarily related to the fact that a newborn brain has very little myelin around axons and the fact that the newborn brain contains comparatively more water within the tissue than the adult brain. Together, these findings show that in the young child the white matter density in absolute CT number is lower relative to gray matter than it is in the adult. On MR scan-ning, aging leads to signal changes consistent with myelination that appears over time in the devel-oping brain. The newborn brain shows very little myelination and this is evidenced by the lack of dark signal in the white matter structures on T2-weighting. The normal structures gradually acquire myelination and an associated decrease in signal intensity on T2, which approaches the adult pattern by approximately 2 years of age (Fig. 1.30).

On US, the normal cerebral hemispheres have symmetry, visible distinction of grey matter from white matter and ventricular size and position as seen on MR and CT. Due to constraints of scanning geometry,

26

Fig. 1.22 Multislice CT of normal cerebrum. The axial source images have been reformatted into the coronal plane: (a) through the frontal lobes, (b) through the anterior horn of the lateral ventricle and the anterior portion of the tempo-ral lobes, (c) at the top of the brainstem, and (d) through the occipital lobes and cerebellum (legends as in Fig. 1.21)

Fig. 1.23 Multislice CT of normal cerebrum. The axial source images have been reformatted into the sagittal plane: (a) through the midline and (b) laterally through the Sylvian fissure and temporal lobe (legends as in Fig. 1.21)


27

Fig. 1.24 MR of the normal cerebrum. Axial T2-weighted images: (a) through the inferior, anterior aspect of the temporal lobes, (b) through the midportion of the temporal lobe and inferior portion of the frontal lobes, (c) through portions of the frontal, parietal, and occipital lobes, and (d) superiorly placed axial image through the frontal and parietal lobes (legends as in Fig. 1.21)

Cerebellum

it is difficult to angle the US transducer through the anterior fontanelle sufficiently to see the very lateral aspect of the skull. Visualizing structures in the posterior fossa using US can also be difficult.

Cerebellum

The cerebellum consists of two small, laterally located hemispheres and a single midline structure (the vermis); all are located in the posterior fossa, which lies below the tentorium cerebelli. The cerebellum also sits behind the brainstem, which runs from top to bottom in the anterior midline of

28

Fig. 1.25 MR of the normal cerebrum. Coronal T2-weighted images: (a) through the midportion of the frontal lobes and the anterior tips of the temporal lobes, (b) slightly posterior to (a), (c) at the level of the brainstem, and (d) through the parietal and temporal lobes and the posterior fossa (legends as in Fig. 1.21)


the posterior fossa and also behind the fourth ventricle (Fig. 1.6). Organizationally, the cerebellar hemispheres are similar to the cerebral hemispheres, with cortex or gray matter on the surface, white matter below the cortex and, most deep and central within the cerebellar hemispheres, deep nuclei composed of gray matter near the fourth ventricle (Fig. 1.31). These three components function in a similar manner to the cerebrum, with a majority of the neurons being present in the cortex, the white matter representing the interconnecting axon bundles and the deep nuclei having additional specific neurologic functions [16–18].

Fig. 1.27 Ultrasound (US) of the normal cerebrum. Coronal images through the anterior fontanelle: (a) frontal lobes, (b) anterior aspect of the temporal lobes and anterior horn of lateral ventricle, (c) brainstem, and (d) oblique axial/coronal image through body of lateral ventricle (legends as in Fig. 1.21)

Fig. 1.26 MR of the normal cerebrum. Sagittal T1-weighted images: (a) through the midline and (b) laterally through the Sylvian fissure and temporal lobe (legends as in Fig. 1.21)

Fig. 1.29 CT of normal brain at different ages. A simple axial image that includes the frontal horns of the lateral ventricles, the temporal lobes, and the occipital lobes of the brain is demonstrated in (a) at 9 day old, (b) at 5 years of age, (c) at 45 years of age, and (d) at 75 years of age. Note the general slow loss of brain substance with increas-ing prominence of the sulci, the Sylvian fissures, and the size of the ventricles

Fig. 1.28 US of the normal cerebrum. Sagittal images through the anterior fontanelle: (a) through the midline and (b) laterally through the Sylvian fissure and temporal lobe (legends as in Fig. 1.21)

Fig. 1.30 MR of normal brain at two different ages. At 1 month of age: (a) A sagittal T1 and (b) an axial T2 image; and at 4 years of age: (c) a sagittal T1, and (d) an axial T2 image. On the T1-weighted images in the 1 month old (in (a)), grey matter is whiter than white matter, while on the T2-weighted images (in (b)) white matter is brighter than grey matter. By 4 years of age (the changes normally occur by 2 years old), white matter is now brighter on T1-weighted images (in (c)), and grey matter is brighter on T2-weighted images (in (d))

Fig. 1.31 Cerebellar hemispheres and vermis. Line drawing demonstrating the major components of the cerebellum


Gray Matter of the Cerebellar Hemispheres

All of the neurons within the cerebellum are located in either the cerebellar cortex or the deep nuclei. As in the cerebral hemispheres, the cerebellar cortex is infolded, though in the cerebellum the gyri are called folia. Also, as in the cerebrum, the prominence of the sulci between the folia increases with advancing age. Whereas the primary functions of the majority of neurons in the cerebral gray matter are concerned with motion, sensation, thought, vision, hearing, etc., the correlation between specific anatomic locations in the cerebellum and a specific action is less precise. Rather, the cerebellum should be thought of as a kinesthetic memory device which coordinates complex actions and serves as the source of muscle or motion memory for complex activities. Damage to the cerebellum most obviously produces incoordination of motor movement.

Extensive edema or any other space-occupying lesion (such as a tumor mass or large clot from hemorrhage) in the inferiorly located cerebellar tonsil – an important structure of the cerebellum – can cause it to herniate downward through the foramen magnum and cause death. This structure is also low-lying in some congenital abnormalities, such as Chiari malformations.

Deep Cerebellar White Matter

Continuing the cerebral hemispheric analogy, the deep white matter in the cerebellum serves to interconnect the cerebellar cortex, the deep nuclei, and the remaining portions of the central nervous system. Diseases involving white matter can involve any of the white matter within the cerebral hemispheres, the cerebellum or the brainstem, and spinal cord.

Imaging Appearance of the Cerebellar Hemispheres

In general, the cerebellum is imaged in the axial plane on CT and in the axial, sagittal, and coronal planes on MR imaging. As in the cerebral hemispheres, on MR most pulse sequences show adequate differentiation of gray and white matter such that the cerebellar cortex is easily discernible from the deeper white matter structures. The normal appearance of the cerebellum on MR is shown in Fig. 1.32; the normal appearance of both the cerebellum and the brainstem on CT is shown in Fig. 1.33 and on US in Fig. 1.34.

Brainstem

The brainstem is divided into three regions – the midbrain, pons, and medulla (Fig. 1.35). The brainstem is anterior to the cerebellum and extends from the base of the cerebral hemispheres infe-riorly, down to the start of the cervical spinal cord at the level of the foramen magnum, the large, central bony opening in the floor of the rear of the skull. The brainstem is a key component of the connectivity of the central nervous system as well as the site of many important clusters of neurons or nuclei. Numerous white matter pathways pass through the brainstem, connecting the brain corti-ces, the basal ganglia, and the cerebellum to one another and to the spinal cord and, thence, to the remainder of the body. The nuclei located with the substance of the brainstem control the function of the majority of the cranial nerves (twelve pairs of nerves that arise symmetrically from the brain and brainstem and are responsible for many functions, including the senses of vision, hearing, and taste; and motor and sensation to portions of the head, neck, and body), as well as participate in the

33

Fig. 1.32 Routine MR images of the cerebellum. In (a) axial T2 image at level of mid medulla through the inferior portion of the cerebellar hemispheres, (b) axial T2 image at the midpons through midportion of the cerebellar hemi-spheres, (c) axial T2 image at the midbrain through the upper portion of the hemispheres, and (d) coronal T2 image through the midportion of the cerebellar hemispheres

Brainstem

coordination of various specific functions of the nervous system. A detailed discussion of these cranial nerves and other brainstem nuclei is not appropriate for this text; however, it is correct to state that lesions within the brainstem can be very discrete in their functional impairment since specific nuclei have clearly defined neurologic functions.

The brainstem components, from superior to inferior, are the midbrain, the pons, and the medulla [19–21]. The midbrain has a short, approximately 1 cm vertical dimension; the pons has a vertical dimension of several centimeters and the medulla likewise. The pons has the largest cross-sectional area and, when viewed from the side, the belly of the pons extends anteriorly farther than the mid-brain and medulla. The belly of the pons consists of large, white matter tracts, extending bilaterally in a postero-lateral direction and sweeping back into the cerebellum on either side. The lowest por-tion of the brainstem, the medulla, connects the pons above to the cervical spinal cord at the level of the skull base (the cervicomedullary junction). With the exception of the first and second cranial nerves, all cranial nerves arise from the brainstem, with the third cranial nerve being the most superior, and progressing inferiorly, successively pass through the fourth, fifth, sixth, and so on, eventually to the twelfth cranial nerve (Fig. 1.35). Thus, the brainstem consists of both white matter tracts connecting the various brain regions and grey matter nuclei.

Fig. 1.33 Axial CT scan at level of cerebral peduncles (a), midbrain and upper parts of cerebellum (b), mid pons (c), and medulla (d)

Fig. 1.34 Routine US mages of the cerebellum and brainstem. Typical coronal image (a), midline sagittal in (b) & (d), oblique sagittal in (c)

35

Fig. 1.35 Brainstem. View of the brainstem in relation to the cerebral hemispheres above and the cerebellum posteriorly. The levels of the brainstem include the upper midbrain, which connects to the cerebral peduncles that connect to the cerebral hemispheres, the middle rounded pons, and the inferiorly located medulla. The brainstem connects to the cerebellum by three pairs of cerebellar peduncles

Ventricles

The Imaging Appearance of the Normal Brain Stem

The brainstem is normally imaged with axial CT images and with sagittal, axial, and coronal MR images. While the internal structure of the brainstem is composed of numerous white matter tracts as well as an additional, fairly large number of gray matter nuclei, this often appears homogenous in signal whether on CT or MR. At the superior aspect of the brainstem, the red nuclei (which is a large, medially located nucleus within the midbrain) is easily discernible on most MR pulse sequences.

The points of connection between the brainstem and the cerebellum, the brainstem and the cere-bral hemispheres, and the brainstem and the cervical spinal cord are areas of smooth transition. No sharp demarcation is evident between these adjacent structures, as the white matter tracts smoothly tie these various components of the central nervous system to one another (Figs. 1.33 [CT], 1.34 [US], and 1.36 [MR]).

Ventricles

The Ventricles – Normal Shape, Size, and Position

The normal shape of the lateral ventricles (the large, fluid-filled cavities within each cerebral hemi-sphere), when viewed from above, is two “C”s positioned back-to-back, with an additional exten-sion (or horn) of the ventricle extending back into the occipital pole. The confluence of this occipital horn with the rear of the “C” on each side is known as the atrium of the lateral ventricle. One can think of the atrium as the point of confluence of the body, occipital horn, and temporal horn. From the atrium, the temporal horn extends laterally and anteriorly from the atrium into the core of the

36

Fig. 1.36 MR images of the brainstem. In (a) sagittal T1 image through the midline, (b) axial T2 image at the level of the cerebral peduncles, (c) axial T2 image at the junction of the cerebral peduncles and midbrain, (d) axial T2 image at the mid pons, (e) axial T2 image at the ponto-medullary junction, (f ) axial T2 image at the medulla, (g) coronal T2 image showing the belly of the pons and the medulla, and (h) coronal T2 image through the cerebral peduncles


37

Fig. 1.37 Ventricles. Oblique view of the ventricles demonstrating the paired “C”-shaped ventricles that have an extension going posteriorly as the occipital horn. These paired lateral ventricles communicate on either side via the foramen of Monro with the midline third ventricle. The midline third ventricle then communicates via the Aqueduct of Sylvius with the fourth ventricle, which then communicates through the single midline foramen of Magendie and the paired lateral foramina of Luschka with the subarachnoid space

Ventricles

temporal lobe. Similarly, the body of the ventricle extends forward and medially with the right- and left-sided structures abutting in the midline. At the anterior extent of the body of the ventricle, the anterior horn diverges laterally and the foramen of Monro opens downward into the anterior aspect of the third ventricle. The third ventricle is a thin, midline structure bounded laterally by the thalami and located immediately below the corpus callosum, the large white matter tract running from side-to-side. The cerebral aqueduct, or Aqueduct of Sylvius, connects the rear of the third ventricle to the midline fourth ventricle, which lies immediately behind the brainstem and in front of the cere-bellum (Fig. 1.37) [22, 23].

The Nonventricular CSF Spaces (Cisterns)

Cerebrospinal fluid is present not only within the ventricular system where it is produced, but also overlies the outside surface of the cerebral hemispheres, brainstem, and cerebellum. In the subarach-noid space, where the arachnoid does not closely follow the contours of the brain, cisterns – focal, enlarged pools of CSF – form [24, 25]. A number of these cisterns have specific names. For example, where the pituitary gland sits at the inferior central base of the brain, the large CSF collection in the midline above the pituitary fossa is called the suprasellar cistern; above the cerebellum in the midline is the superior cerebellar cistern; anterior to the pons is the prepontine cistern; and below the cerebel-lum inferiorly in the midline is the cisterna magna. Additional off-midline CSF spaces, primarily in the posterior fossa, include the ambient wing and quadrigeminal cistern at the level of the midbrain and the cerebellar pontine angle cistern, which is on either side of the medulla near the internal audi-tory canal (Fig. 1.38 extraventricular CSF spaces). These CSF-containing spaces also extend inferi-orly to bathe the spinal cord all the way down to the level of the sacrum, with the large collection of CSF in the lower back or lumbar region referred to as the lumbar cistern.

38

Fig. 1.38 Nonventricular CSF spaces (cisterns). Cerebrospinal fluid is present not only within the ventricular system, but overlying the outside surface of the cerebral hemispheres, brainstem, and cerebellum. Named cisterns that are present in the midline include the suprasellar cistern, the superior cerebellar cistern, the prepontine cistern, and the cisterna magna. Additional off-midline CSF spaces, primarily in the posterior fossa, include the ambient wing and quadrigeminal cistern at the level of the midbrain and the cerebellar pontine angle cistern which is on either side of the medulla near the internal auditory canal


CSF Dynamics

In the normal adult, several hundred milliliters of CSF are continuously produced as an ultrafiltrate of the serum component of blood by the formative organ, the choroid plexus. This specialized structure is located within the lateral, third, and fourth ventricles. There is bulk flow of CSF out of the lateral ventricles on each side, through their respective foramen of Monro and into the anterior third ventricle. CSF then flows from the rear of the third ventricle downward through the cerebral aqueduct into the fourth ventricle. The fourth ventricle has single midline (the foramen of Magendie) and paired lateral (the foramina of Luschka) openings that allow exit of CSF inferiorly into the subarachnoid space at the base of the posterior fossa. Once CSF exits the ventricular system, it bathes both the brain and the spinal cord in the subarachnoid space. After the CSF circulates through the ventricular system and the subarachnoid space, it is eventually returned to the systemic circulation via structures called arachnoid granulations, which protrude into the superior sagittal sinus.

Clinical Note: The size of the ventricles can be increased either because they are being dilated by too much spinal fluid within a normal-sized ventricular system or because the ventricles have dilated in response to a loss of brain tissue [26]. In the first case, too much spinal fluid may be present in the ventricles due to overproduction of spinal fluid, blockage of normal flow of spinal fluid in its attempt to move through the ventricular system, or impaired absorption of spinal fluid once it is outside the ventricular system. In the latter cases, the ventricles expand in response to a loss of brain tissue and, in effect, the ventricles are expanding to fill the “vacuum” created by the loss of brain parenchyma.

39

Fig. 1.39 Imaging of the ventricles and extraventricular CSF spaces by MR, with axial T2-weighted images: (a) at the top of the lateral ventricle, (b) at the midthird ventricle, (c) at the level of the suprasellar cistern, and (d) at the midpons and cerebellar pontine angle cisterns

Ventricles

Overproduction of spinal fluid most commonly occurs in the case of tumors of the choroid plexus. These tumors are usually choroid plexus papillomas (benign tumors of the choroid plexus), but in less than 5% of cases they may be malignant choroid plexus carcinomas. This overproduction of spinal fluid results in the dilatation of the ventricles. Blockage of the ventricular system at any level produces dilatation of the ventricles upstream from the blockage. For example, blockage of a level of the fora-men of Monroe will cause dilatation of the lateral ventricle. A blockage of the lateral ventricle at the atrial region will cause dilatation of the temporal horn. A mass that occludes the rear of the third ven-tricle or the aqueduct will cause enlargement of both lateral ventricles and the third ventricle. A block-age at the level of the normal foramina that allows exit from the fourth ventricle (Magendie and Luschka) will cause dilatation of all of the ventricles. This last could be accomplished, for example, by an adhesive process from a prior infection at the skull base. A similar response to prior infection can block the arachnoid granulations near the superior sagittal sinus and cause impaired resorption of cerebrospinal fluid, leading to diffuse enlargement of all of the ventricles.

Not only can the ventricles enlarge from overproduction of CSF or from blockage of flow, they can also enlarge secondary to marked parenchymal loss. For example, a global anoxic event that does not cause a discrete infarct of brain, but rather a diffuse loss of brain parenchyma, often


presents with abnormally enlarged ventricles for the patient’s age. Alternatively, the ventricles may enlarge as a result of the loss of grey matter and white matter associated with aging.

Imaging Appearance of the Normal Ventricles, CSF Spaces, and CSF Dynamics

The CSF-containing spaces of the brain will appear on CT scans as areas of low (dark) density, with CT numbers close to that of water. This allows CT scanning to very adequately show both the ventricles and the extraaxial CSF collections. Similarly, the CSF spaces are well demonstrated on MR scanning, though the appearance of the spinal fluid will vary significantly with the pulse sequence chosen. On a T1-weighted pulsed sequence, for example, CSF will appear black. On a T2-weighted sequence, the CSF will appear very white within the image. Finally, on a FLAIR sequence, which can be thought of as a T2-weighted sequence that has additional magnetic and radiofrequency pulses, all of the free water in CSF spaces becomes black. Consequently, on both the T1-weighted sequence and on a FLAIR sequence, the spinal fluid will appear black, and on a T2-weighted sequence spinal fluid will be white. The major ventricular structures, as well as the extraaxial CSF spaces, are shown on CT, MR, and US in Figs. 1.21–1.28, 1.32–1.34, 1.36, and 1.39.

References

1. Barkovich AJ. Pediatric Neuroimaging, 2005, Lippincott, Philadelphia, pp 252–253. 2. van Gijn J, Kerr RS, Rinkel GJ. Subarachnoid haemorrhage. Lancet 2007;369(9558):306–318. 3. Kretschmann H-J, Weinrich W. Cranial Neuroimaging and Clinical Neuroanatomy: Atlas of MR Imaging and

Computed Tomography, 2004, 3rd Edition, Thieme, Stuttgart, pp 253–282 (Section 5.4 Arteries of the Brain). 4. Snell RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia – Chapter 17. The Blood

Supply of the Brain and Spinal Cord, pp 473–479. 5. Krabbe-Hartkamp MJ, van der Grond J, de Leeuw F-E, et al. Circle of Willis: morphologic variation on three-

dimensional time-of-flight MR angiograms. Radiology 1998;207:103–111. 6. Kretschmann H-J, Weinrich W. Cranial Neuroimaging and Clinical Neuroanatomy: Atlas of MR Imaging and

Computed Tomography, 2004, 3rd Edition, Thieme, Stuttgart, pp 283–285 (Section 5.5 Veins of the Brain). 7. Snell RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia – Chapter 17. The Blood

Supply of the Brain and Spinal Cord, pp 479–480. 8. Grossman RI, Yousem DM. Neuroradiology: The Requisites, 2003, Mosby, Philadelphia, pp 217–220. 9. Kiernan J. Barr’s The Human Nervous System – An Anatomical Viewpoint, 2009, Lippincott, Philadelphia –

Chapter 13. Topology of the Cerebral Hemispheres pp 211–218 and Chapter 15 Functional Localization in the Cerebral Cortex, pp 227–244.

10. Kretschmann H-J, Weinrich W. Cranial Neuroimaging and Clinical Neuroanatomy: Atlas of MR Imaging and Computed Tomography, 2004, 3rd Edition, Thieme, Stuttgart, pp 307–324 (Section 5.7.5 Telencephalon).

11. Snell RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia – Chapter 7. Cerebrum pp 247–280 and Chapter 8 The Structure and Functional Localization of the Cerebral Cortex, pp 286–301.

12. Kiernan J. Barr’s The Human Nervous System – An Anatomical Viewpoint, 2009, Lippincott, Philadelphia – Chapter 14. Histology of the Cerebral Cortex, pp 219–226.

13. Snell RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia – Chapter 8. The Structure and Functional Localization of the Cerebral Cortex, pp 281–286.

14. Kiernan J. Barr’s The Human Nervous System – An Anatomical Viewpoint, 2009, Lippincott, Philadelphia – Chapter 12. Corpus Striatum, pp 199–210.

15. Snell RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia – Chapter 7 Cerebrum, pp 260–261.

16. Kiernan J. Barr’s The Human Nervous System – An Anatomical Viewpoint, 2009, Lippincott, Philadelphia – Chapter 10 Cerebellum, pp 157–172.

41References

17. Kretschmann H-J, Weinrich W. Cranial Neuroimaging and Clinical Neuroanatomy: Atlas of MR Imaging and Computed Tomography, 2004, 3rd Edition, Thieme, Stuttgart, pp 303–304 (Section 5.7.3 Cerebellum).

18. Snell RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia Chapter 6, pp 225–245 (The cerebellum and its connections).

19. Kiernan J. Barr’s The Human Nervous System – An Anatomical Viewpoint, 2009, Lippincott, Philadelphia – Chapter 6. Brainstem: External Anatomy, pp 79–86 and Chapter 7 Brainstem: Nuclei and Tracts, pp 87–112.

20. Kretschmann H-J, Weinrich W. Cranial Neuroimaging and Clinical Neuroanatomy: Atlas of MR Imaging and Computed Tomography, 2004, 3rd Edition, Thieme, Stuttgart pp 288–303 (Sections 5.7.1 Pons and Medulla and 5.7.2 Midbrain).

21. Snell RS. Clinical Neuroanatomy for Medical Students, 2001 Lippincott, Chapter 5, pp 189–224 (The Brainstem). 22. Kretschmann H-J, Weinrich W. Cranial Neuroimaging and Clinical Neuroanatomy: Atlas of MR Imaging and

Computed Tomography, 2004, 3rd Edition, Thieme, Stuttgart, pp 239–253 (Section 5.3 Cerebrospinal Fluid Containing Spaces).

23. Snell, RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia – Chapter 16: The ven-tricular system, the cerebrospinal fluid and the blood brain and blood cerebrospinal fluid barriers, pp 443–456.

24. Kretschmann H-J, Weinrich W. Cranial Neuroimaging and Clinical Neuroanatomy: Atlas of MR Imaging and Computed Tomography, 2004, 3rd Edition, Thieme, Stuttgart Chapter 16: The ventricular system, the cerebrospi-nal fluid and the blood brain and blood cerebrospinal fluid barriers, pp 456, 457.

25. Snell, RS. Clinical Neuroanatomy for Medical Students, 2001, Lippincott, Philadelphia – Chapter 16: The ven-tricular system, the cerebrospinal fluid and the blood brain and blood cerebrospinal fluid barriers, pp 443–456.

26. Osborn AG, Blaser SI, Salzman KL, et al. Diagnostic Imaging – Brain, 2004, Amerisys, Salt Lake City II:1:12–27.


Abstract This chapter begins with a discussion of the terminology and hardware of computed tomographic (CT), magnetic resonance (MR), and ultrasound (US) imaging. This includes an intro-duction to CT numbers and very basic MR image formation, sequence types, and sequence usages.

The remainder of the chapter is an introduction to edema and hemorrhage. The two basic types of edema are cytotoxic and vasogenic. The appearance of these types of edema is presented on CT, MR, and US. The general imaging of hemorrhage – without following in detail the evolution of these findings with time – and the various spaces in which it occurs in the brain are discussed and demonstrated on CT, MR, and ultrasound.

Keywords Computed tomography (CT) • Magnetic resonance (MR) • Ultrasound (US) • Edema • Cytotoxic edema • Vasogenic edema • Hemorrhage

Introduction

This chapter will build on the anatomy and imaging findings of normal brain detailed in Chap. 1. Knowing the normal appearance of brain is mandatory to make an intelligent assessment of the changes in brain as a result of injury. In this chapter, we begin with a brief introduction on the basics of computed tomography (CT), then magnetic resonance (MR) and finally, ultrasound (US) imaging of the brain. Second, we will discuss the general topic of edema and how it appears on each of the three imaging modalities. Finally, in a broad sense, we will introduce hemorrhage and its imaging appearance. In all cases in this chapter, we will not discuss in detail the manner in which the basic findings change con-siderably over time. That detailed discussion, including the most complicated topic of all – the changing appearance of hemorrhage over time on MR scanning – is reserved for Chap. 4.

CT Scanning: The Absolute Basics

In any discussion of the pathology within the central nervous system (CNS – the brain and spinal cord), one should have an understanding of the basic underlying principles of the modality that is used to image the brain, in addition to an anatomic understanding of the brain. Therefore, at this point, we will provide an overview of the basic principles of CT scanning and, more specifically, those principles that are directly applicable to imaging the CNS [1]. Once one understands CT

Chapter 2The General Appearance of Edema and Hemorrhage on CT, MR and US (Including a General Introduction to CT, MR and US Scanning)

44

Fig. 2.1 CT scanner. A typical CT scanner. The table on which the patient lies, is at the front of the scanner. The table then moves into the central bore of the CT scanner. Within the scanner gantry is a rotating ring with both the X-ray tube and the detector array

2 The General Appearance of Edema and Hemorrhage

technology, we can move beyond that into the specific findings on CT that occur in the setting of injury. Abnormalities are recognized in one of two ways: either anatomy is distorted (i.e., something there that shouldn’t be, or something not there that should be) and/or an alteration occurs in the normal CT numbers of the tissue. CT numbers, as we shall see, are an essential component of the information provided by a CT scan.

A CT scanner consists of the basic components of a patient table (on which the patient to be scanned lies); the scanner gantry (which contains the rotating portion that holds the X-ray tube generator and detector array) and a computer system (for performing the necessary calculations to go from measurements to a viewable image) (Fig. 2.1). The patient lies supine on the movable scanner table as the portion of the body to be scanned passes through the middle of the opening in the gantry. During the scan, the X-ray tube continually generates X-rays, which pass sequentially through the patient and then on to an array of detectors. The computer begins with the detector-made measure-ments of the radiation that is “left over” after passing through the patient and calculates what the tissues in the body must have looked like to have produced the observed measurements.

The term “density” is often used to describe the distribution of matter that must have been present to partially absorb the X-ray beam and produce the measured residual beam at the detec-tors. However, a more technically accurate description is that the scanner computes the amount of radiation absorption that occurs at each scanned point in the body. Fortunately, a close correlation exists between the density of a tissue and its ability to stop X-ray photons. For our purposes we will, therefore, use the shorter term “density,” rather than the technically more accurate (but much more cumbersome) phrase, “relative ability to absorb X-ray photons,” for the remainder of this book.

When the CT scanner computer finishes its calculations for a single image, the result is a cross-sectional image of the brain in which white represents structures that are more dense and dark represents structures that are less dense. The CT image can be thought of as a density map, showing the relative propensity of different portions of the image to absorb the X-rays that the CT scanner beam sends through the patient. Within the scanner, it is the job of the computer to reconstruct and calculate the density distribution which must have existed to produce the measured absorptions, and

45

Table 2.1 CT numbers

Air –1000Fat –100 to –50Water 0Most tissues 20–40Hemorrhage <80Calcium HundredsBone Hundreds to thousands

Typical CT numbers for various tissues. Note that hem-orrhage can have a CT number up to a maximum of 80

MR Scanning: The Absolute Basics

then to assign a CT number to each voxel (very small, square elements which collectively make up the image), from –1,000 to +3,000 (the higher the number the more that voxel absorbs X-rays). Within the range of –1,000 to +3,000, the CT number for air is –1,000; the CT number for water is 0, and the CT number for bone can range from 100 to the maximum of the scale, 3,000. Within the much larger range of CT numbers lays the smaller range of numbers which are normal tissue. Most brain tissue (whether gray or white matter) measures in the 20–40 CT number range, with gray matter being slightly more dense than white matter. Acute hemorrhage can have a CT number from a value between 30 and 40 to a value as high as 80. Because of this, if a suspicious region has a CT number greater than 80, it cannot represent hemorrhage exclusively. Other CT number ranges which can be seen with pathologic entities within the brain include aging hemorrhage, which will have a slowly decreasing CT number from the ranges of acute blood (40–80), down through that of normal brain white matter (20–40) and, eventually, to water density (around 0). Proteinaceous fluid collec-tions can have density numbers ranging as high as 60–80 as well, and entities that have some degree of calcification – whether a large mass of calcification or merely microscopic calcifications – can generally raise the density of tissue into the upper end of the 0–100 range and even into the low range above 100 (Table 2.1 – CT numbers).

MR Scanning: The Absolute Basics

An MR scanner consists of a large magnet, wound in a solenoid fashion on an open central bore. When electrical current is applied to the magnet, a strong magnetic field is generated, running down the bore, and the patient is placed within this bore. A strong radio-transmitter broadcasts radio energy into the patient and a separate receiving system listens to the radio energy as it is emitted back from the patient. This information is then used to construct an image. Other key components of the total scanner system are additional coils to generate spatial information, and a computer system which controls the system (Fig. 2.2)

MR as an imaging modality is intrinsically more complicated than CT scanning [2]. CT scanning represents a single 1:1 correspondence between one physical entity – that is absorption of X-ray photons – and the brightness (or darkness) of a point in the final image. In contrast, MR imaging is due to at least five different imaging factors. Without entering a rather protracted discussion of MR physics, these five factors are: T1 value of tissue, T2 value of tissue, proton density, magnetic susceptibility, and motion. When an MR scan is performed, all five of these factors contribute in varying degrees to the final image. It is impossible to perform an MR scan that does not have some contribution from these five physical factors in the final image. The one reason why the word “density” is not used for MR is that five factors exist, and the terminology used to describe the appearance of a final image is “brightness” or lack thereof.

46

Fig. 2.2 MR scanner. Cut-away drawing showing the internal major components of an MR scanner. During a scan the patient lies within the bore of the main magnet. Smaller gradient coils shape the larger field and allow imaging. Radiofrequency coils send radio signals into the patient and listen to the emitted signals, which are used to create the MR image

Table 2.2 MR pulse weightings

Image factorPulse sequence with this weighting Best used for

T1 T1-weighted Depict anatomyT2 T2-weighted Depict pathologyT2 FLAIR Depict pathologyProton density PD-weighted Seldom usedMagnetic susceptibility Magnetic susceptibility-

weightedDetect calcium and

hemorrhageMotion DWI (microscopic motion) Detect infarctionsMotion MRA (macroscopic motion) Evaluate vasculature

For each of the five components of an MR image (T1, T2, proton density, motion, magnetic susceptibility) the table lists the different sequences that can have that component as their primary weighting, and the main use of those sequences


However, while it is not possible to scan an image that is composed purely of any one of these five, it is possible to scan an image in which one of these five becomes the predominant image contrast in the final image. Such an image is said to be a weighted image. At least one type of imag-ing examination corresponds to each of these five components of an MR image, including T1-weighted images, T2-weighted images, proton density-weighted images, and magnetic suscep-tibility-weighted images. Motion-weighted images have two types of sequences. The first empha-sizes motion on the microscopic level (these are also called diffusion-weighted images); the second weights motion on the macroscopic level and these are MR angiograms, or images of vessels within the brain (Table 2.2 – MR pulse sequence weighting). One of the greatest strengths of MR scanning is that pathology can be detected by an abnormality of signal on a large number of different MR sequences; and the appearance both of normal and abnormal tissues varies from sequence to sequence.

47

Fig. 2.3 Ultrasound (US) scanning. Diagram of hand-held, real-time transcranial ultrasonography of an infant. The transducer is placed at the anterior fontanel and obtains coronal and sagittal images of the brain

US Scanning: The Absolute Basics

US Scanning: The Absolute Basics

An ultrasound machine utilizes high-frequency ultrasound to image patient anatomy and pathology. High-frequency, “ultrasonic” sound waves are produced in a scanner placed on the patient’s skin, directly over the area to be imaged. In order to produce an image, the sound waves must penetrate the patient. As the sound waves pass progressively deeper into the tissues of the patient, some are selectively reflected back and received by the same transducer that generated them only a few mil-liseconds earlier. This produces, in effect, a sonar image of the tissues of the body [3] (Fig. 2.3).

While CT and MR can be more helpful in the older child and adult, in specific clinical situations neonatal head ultrasounds are extremely useful. It is often the case that the only images available of an infant’s head in the first few days of life are ultrasound examinations, and only later are CT and/or MR scans performed. The neonatal head can be scanned with ultrasound because of the presence of an open anterior fontanelle, i.e., a normal developmental defect in the skull, anteriorly. However, by 1 year of age the fontanelle is usually completely fused, and the previously usable acoustic window is no longer available. From that time onward, MR and CT are the only means available for intracranial imaging.

However, in the initial days, weeks and first few months of life, ultrasound may provide an impor-tant adjunct or, in some cases, the only imaging evaluation of the intracranial contents. Consequently, ultrasound needs to be considered in the discussion of the dating of neurologic injury by neuroradio-logical imaging techniques. While, in general, some of the findings on ultrasound are perhaps not as specific as the findings noted on CT and MR, in some clinical situations the ultrasound examination may be a more sensitive test to early changes of injury to the brain parenchyma.

48

Fig. 2.4 CT scan of edema. (a) Extensive low density, representing vasogenic edema, is present throughout the high right centrum semiovale, surrounding a several day old hemorrhage. (b) Extensive cytotoxic edema produces a wedge-shaped region of low density in this patient with an acute left-sided infarction


Edema

Before proceeding further with a discussion of the appearance of edema on MR and CT, it is appro-priate to discuss edema itself. Edema refers to swelling within a tissue due to the accumulation of fluid. Edema occurs as the result of a variety of pathologic conditions. The brain experiences edema as a result of almost any insulting agent – it is seen in and around regions of dead or dying brain, around metastases and abscesses, after traumatic injury, following hypoxic ischemic injury, and around primary brain tumors. Greenfield describes five different types of edema – vasogenic, cyto-toxic, hydrostatic, interstitial, and hypoosmotic [4]. For our purposes, we will concentrate on the vasogenic and cytotoxic types.

In vasogenic edema, the edematous tissue swells due to breakdown of the blood brain barrier. An interesting feature of vasogenic edema is that it can spread to regions that are some distance from the site of the brain abnormality. For example, an abnormal, disrupted blood brain barrier at point A can lead to vasogenic edema in the brain, which can spread to point B, even though point B is several centimeters away, and was otherwise normal brain. This spreading water within the tissues moves more freely through white matter than through gray matter, and as a result it will often halt when it reaches the underside of the cortex. Vasogenic edema occurs around infectious processes like abscesses and both benign (meningiomas) and malignant tumors (gliomas and metastases). Vasogenic edema can also be seen peripherally around a central core of cytotoxic edema in cases of infarction.

Greenfield describes cytotoxic edema as “cellular swelling associated with a reduced extracel-lular space, but with an intact blood brain barrier (at least to macromolecules in the initial stages)” [4]. In cytotoxic edema, tissue swelling occurs because the tissue is severely injured, dying or dead. Such an injury could occur, for example, if an arterial occlusion ceases all blood flow to a demar-cated region of brain. Within that region, all the brain would be equally affected, whether it were gray or white matter. Cytotoxic edema from an arterial infarct involves both white matter and over-lying gray matter. Cytotoxic edema is more commonly associated with ischemic or hypoxic pro-cesses (Figs. 2.4 and 2.5 – Vasogenic and cytotoxic edema on CT and MR).

Three effects of edema are visible on imaging: loss of gray-white matter differentiation, swelling of sulci (shrinking of gyri) and mass effects [5]. The first effect, the loss of gray-white differentiation,

49

Fig. 2.5 MR scan of edema. (a) Extensive vasogenic edema limited to the white matter on this axial FLAIR image, posterior to a large tumor in the right hemisphere. (b) Wedge-shaped area of abnormal diffusion on a diffusion-weighted imaging (DWI) in the right occipital pole, representing cytotoxic edema, consistent with acute infarct. In this early infarction, the FLAIR and T2 scans were normal

Fig. 2.6 CT scan demon-strating loss of gray-white differentiation with fuzzy, indistinct basal ganglia

Edema

is commonly seen in cytotoxic edema. This loss of the ability to discern gray matter from white matter on MR or CT is seen between the gray matter of the cortex and the immediate adjacent underlying white matter, and within the basal ganglia, obscuring the ability to visually isolate the gray matter of the caudate, thalamus and lentiform nucleus from the white matter of the internal capsules (Figs. 2.6 and 2.7 – Loss of cortical G-W differentiation and fuzzy BG on MR and CT).

The second effect of edema on a region of the brain is to cause the involved gyri to expand and the intervening sulci to decrease in size. As the brain continues to swell, not only do the sulci decrease, but all of the CSF spaces of the hemispheres decrease as well. This includes the Sylvian fissure, as well as the basilar cisterns containing CSF around the brainstem and posterior fossa structures (Figs. 2.8–2.10 – Brain edema effects on sulci and ventricles on CT, MR and US).

50

Fig. 2.8 CT scan showing the effect of brain edema on intracranial structures. In (a) a large infarction in the territory of the right middle cerebral artery with marked low density, complete loss of gray-white differentiation within the infarct, mass effect on the right lateral ventricle, and slight midline shift to the left. In (b), a higher axial scan in a different patient with a right-sided infarction shows complete absence (effacement) of the sulci on the effected side

Fig. 2.7 MR scan demonstrating abnormal DWI signal in the left basal ganglia in (a), and abnormal FLAIR signal in the same region in (b)


The third and last effect edema produces is to cause the ventricles to decrease in size. The total volume of the intracranial compartment is fixed, and composed of brain and all of the CSF-containing spaces. As the brain tissues swell, in order for the total intracranial volume to remain constant, the ventricles and extraaxial CSF spaces must decrease in total volume. As the amount of cerebral edema worsens, more and more CSF is pushed out of the lateral, third and fourth ventricles into other CSF spaces of the CNS beyond the cranium – for example into the spinal canal. An excessive amount of edema can cause mass effect which will result in both the displacement of normal ventricles and have secondary effects on the ventricles. In addition to the ventricles being decreased in size, they can also be displaced and moved, as can the normal midline structures of the brain. Severe edema can close

51

Fig. 2.10 Ultrasonic demonstration of brain edema shows abnormally increased echogenicity in this patient bilater-ally, more so on the patient’s right side, with early bilateral basal ganglia infarctions

Fig. 2.9 MR scan of the effect of edema. The cortical sulci are faced in the region indicated by the arrowheads the right lateral ventricle is compressed and there is a slight midline shift to the left

Edema

off the CSF drainage pathways and can cause portions of the ventricular system to increase in size (Fig. 2.11 – Additional effects of edema).

Edema on CT Scanning

As the brain swells (especially from cytotoxic edema), the comparative difference in appearance between gray and white matter decreases on CT. As more and more water enters into the tissues, the relative density of both gray and white matter decreases, as does the difference in density between the two tissues. Consequently, areas affected by cytotoxic edema can show the loss of normal gray-white differentiation. This is manifested by an inability to see the difference between

52

Fig. 2.11 Demonstration of various internal herniations. In (a) axial T2-weighted MR scan in a patient with a large left-sided mass, surrounding vasogenic edema, and marked midline shift from left to right. In a different patient, in (b) axial CT at level of midbrain showing loss of CSF spaces, consistent with tentorial herniation; and (c) axial CT low in the posterior fossa showing very severe posterior fossa edema


gray and white matter in the region of affected brain, whether that involves the cortex and the under-lying white matter, or whether that involves the deeper gray matter structures and their adjacent white matter tracts (see Figs. 2.4, 2.6, and 2.8).

Edema on MR Scanning

On MR, the appearance of edema is similar to that of CT, though often, unlike CT, the underlying anatomy is not completely obscured by the edema. MR can produce many different types of sequences with very different “weightings.” However, of all these different types of weighting, the most sensitive sequence for subtle degrees of edema is diffusion-weighted imaging (DWI) [6]. DWI can detect subtle edema before it is seen on any other type of MR pulse sequence and, certainly, before it is seen on CT.

The ability to detect edema on MR scanning is due to a combination of two imaging findings. The first is abnormal signal, most commonly noted on DWI but also seen on FLAIR and T2-weighted imaging and, to a much lesser extent, on T1-weighted imaging; the second is due to morphologic alteration (i.e., distortion) of the normal appearance of the tissues. While edema typically shows up as bright signal on DWI, T2-weighting and FLAIR imaging, differences in the time, appearance and duration of these signal abnormalities exist and will be discussed in much more detail in Chap. 4 (Figs. 2.5, 2.7, and 2.9 – Edema on MR).

Edema on US Scanning

Edema on ultrasound in the neonate usually occurs around the ventricles in the periventricular white matter. Because this is a watershed territory in newborns, the area around the ventricles and its associated white matter is often involved by ischemic events. The initial examination can be normal. However, in sonography the first detectable abnormality is areas of increased echogenicity around the ventricle. Over time, cysts can develop, normally in 2–4 weeks. At longer time intervals, the cysts can either coalesce or disappear such that they are no longer visible on ultrasound – but the gliotic changes are still visible on MR scanning. An initially normal ultrasound can become abnormal after weeks or months; hence, fol-low-up imaging, even after an initial normal scan, is indicated (see Fig. 2.10).

53

Fig. 2.12 CT scans of subarachnoid hemorrhage in (a), subdural hemorrhage in (b), and epidural hemorrhage in (c)

The General Appearance of Hemorrhage

In addition to the changes seen around the ventricles from paraventricular leukomalacia, more global insults affecting larger areas of brain occur in the setting of global anoxia. This injured anoxic/ischemic brain is diffusely echogenic with poorly defined sulci. The sulci may be lost in the overall increase in echogenicity of the remaining brain parenchyma, the so called “silhouetting of the sulci.”

Another pattern which may be seen on ultrasound is damage to the basal ganglia, which can occur with acute, near-total intrauterine asphyxia. Finally, while focal geographic parenchymal infarction is uncommon in the neonate, it can occur due to underlying causes such as emboli from heart disease or meningitis. In these cases, the region of brain that is affected shows focally increased echogenicity.

In summary, the sonographic signs of cerebral infarction include echogenic parenchyma, mass effect from edema, a pattern of injury which is that of an arterial territory and decreased definition of the sulci [7].


Hemorrhage occurs when blood enters a portion of the brain that is not within the normal vascular system. Hemorrhage can occur in the brain in three different space categories. It can occur within the brain parenchyma (intraparenchymal hemorrhage), within the ventricles (intraventricular hem-orrhage) or in the extraaxial spaces around the brain. Alluding to the earlier discussions of the layers of the scalp, skull, and meninges in Chap. 1, extraaxial hemorrhage can occur in the subarachnoid space, the subdural space or the epidural space (Figs. 2.12 and 2.13 – Subarachnoid, subdural and epidural hemorrhages on CT and MR).

Hemorrhage occurring into the brain parenchyma itself is usually the result of a closed head injury or of bleeding from a vascular lesion such as a malformation, a malignancy, or hypertension. Infarctions can also hemorrhage, as can intrinsic brain gliomas. These intraparenchymal hematomas can be discrete, well-circumscribed collections that consist entirely of blood, or can be areas of brain that are in effect “bruised” and are regions of brain where hemorrhage is interspersed among the normal cellular elements (also called contusions) (Figs. 2.14 and 2.15 – Intraparenchymal hem-orrhage on CT and MR).

Blood in the ventricles can occur from a trauma which ruptures the small vessels that line the ventricular wall. Alternatively, bleeding can occur initially into the brain adjacent to the ventricles, and then rupture into the ventricles secondarily (Fig. 2.16 – IVH on CT and MR).

54

Fig. 2.13 MR scans of subarachnoid hemorrhage in (a) (arrows), subdural hemorrhage on FLAIR in (b), on T2 in (c) and with GE in (d)


Blood in the extraaxial spaces tends to be caused by a short list of etiologies. The most common cause of subarachnoid hemorrhage is trauma. If trauma is eliminated, in 80–90% of cases the most common cause of subarachnoid hemorrhage is rupture of an intracranial aneurysm. Subarachnoid hemorrhage can also occur from the rupture of a vascular malformation or from bleeding from a CNS tumor. Bleeding into the subdural space is almost always traumatic in nature. However, if the jet of leaking high-velocity blood is aimed directly at the dura, a previously-ruptured aneurysm that has developed scar about its dome and ruptured again can rupture all or partially into the subdural space. Epidural hemorrhages are almost always traumatic in nature.

Hemorrhage on CT Scanning

Regardless of the location of the hemorrhage, the appearance of the blood follows a fairly predictable time course such that over days or at most weeks, hemorrhage decreases from its initial density of 40–80, to a density range equal to that of gray or white matter, and finally to that of cerebrospinal fluid [8]. This ordered, sequential progression of density changes over time occurs for any blood

55

Fig. 2.14 CT of a small intraparenchymal hemorrhage in the left frontal brain


collection – but the time each step takes varies with the exact location of the hemorrhage (be it intraventricular, within the brain substance, or outside the brain). As the hemorrhage is resolved or resorbed over time, its size also decreases, as does the amount of surrounding edema. The details of this time course change will be discussed in much more detail in Chap. 4.

Hemorrhage on MR Scanning

The appearance of hemorrhage on MR scanning is very complex. Unlike CT scanning which shows a progressive decrease in density over time, the MR appearance of hemorrhage varies mark-edly over time and is different on each pulse sequence. In fact, on each of several different pulse sequences the image appearance of hemorrhage may change several times over the initial 2–3 week period following its occurrence. It is, therefore, not possible to state that hemorrhage looks a specific way without also addressing and discussing the changes that hemorrhage takes on over time on the different pulse sequences. Therefore, this complex topic will be discussed in much more detail in Chap. 4.

Hemorrhage on Ultrasound

As with the ultrasonography of cerebral edema, the ultrasonography of cerebral hemorrhage is also a discussion which is limited to patients under 1 year of age; ideally, even in the first few weeks and months. After that period of time it becomes increasingly difficult to sonographically view the intracranial contents.

56

Fig. 2.15 MR of bilateral temporal lobe intraparenchymal hemorrhages. (a) T1-weighted axial image, (b) T2-weighted axial image, (c) FLAIR and (d) gradient echo image – all at the same level


The most common site for neonatal intracranial hemorrhage is the subependymal region, a spe-cial portion of the lateral ventricles in neonates which houses the germinal matrix in the thalamo-caudate groove. This is a frequent site of hemorrhage, which can remain subependymal or rupture into the ventricles. In either case, these hemorrhages initially appear as hyperechoic structures, either limited to the immediate subependymal brain or extending into the ventricle. Whereas, the initial appearance is uniformly hyperechoic throughout, with time (1–2 weeks) the central portions of the clot can become more sonolucent. Additional signs of intraventricular hemorrhage include a clot that forms and makes a cast of the ventricle, a thickly echogenic choroid plexus, low-level echoes floating within a ventricle, and CSF-blood fluid levels [9].

In a similar fashion, ultrasound can rather easily locate intraparenchymal hemorrhage. These hemorrhages, again, begin as hyperechoic structures which eventually have more echolucent centers. The final resolution can be complete disappearance of the clot so that no sonographic appearance

57

Fig. 2.16 Intraventricular hemorrhage on CT in (a). MR images of intraventricular hemorrhage – sagittal T1 in (b), axial T2 in (c), axial gradient echo in (d). US of IVH in (e)

Chapter Summary

remains. Alternatively, a small cyst or slit-like hole at the site of the prior hematoma may persist [10]. While hemorrhages can be detected noninvasively on ultrasound, both CT and MR scanning have been found to be more sensitive exams [11, 12].

Chapter Summary

We began this chapter with an introduction to the three imaging modalities of CT, MR, and US. The usefulness of US is limited to the first year of life. In contrast, CT and MR can both be used to image brain injury throughout life. We presented the general imaging findings on CT and MR, but made no attempt to describe how the imaging appearance of hemorrhage or infarction changes with time. Those detailed discussions are reserved for Chaps. 4 and 5.

Regions of both infarction and hemorrhage are shown as areas of brighter signal (increased echogenicity) on US. On CT, regions of hemorrhage begin as high density (whiter) and then decrease in density (become darker) with time, while regions of edema are initially similar in density to brain and then (if infarction occurs) become progressively less dense, eventually approaching the density of water. On MR, regions of edema tend to have abnormally increased (whiter) signal on DWI, FLAIR and T2-weighted sequences. The MR appearance of hemor-rhage is quite complex, and is the reason why a significant portion of Chap. 4 is devoted to that topic alone.

58 2 The General Appearance of Edema and Hemorrhage

References

1. Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The Essential Physics of Medical Imaging, 2nd edition, 2002, Lippincott, Philadelphia. Chapter 13: Computed Tomography, pages 327–372.

2. Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The Essential Physics of Medical Imaging, 2nd edition, 2002, Lippincott, Philadelphia. Chapter 14: Nuclear Magnetic Resonance, pages 373–413.

3. Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The Essential Physics of Medical Imaging, 2nd edition, 2002, Lippincott, Philadelphia. Chapter 16: Ultrasound, pages 469–554.

4. Graham DI, Lantos PL, editors. Greenfield’s Neuropathology, 7th edition, 2002, Arnold, London, pages 203–209.

5. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, 2009, Lippincott, Philadelphia. Acute Cerebral Ischemia-Infarction, pages I:4:108–111.

6. Gonzalez RG, Schaefer PW, Buonanno FS, et al. Diffusion-weighted MR Imaging: Diagnostic Accuracy in Patients Imaged within 6 Hours of Stroke Symptom Onset, Radiology, 1999, 210:155–162.

7. Rumack CM, Wilson SR, Charboneau JW. Diagnostic Ultrasound, 3rd edition, 2005, Elsevier Mosby, St. Louis, pages 1671–1681.

8. Grossman RI, Yousem DM. Neuroradiology: The Requisites, 2nd edition, 2003, Mosby, Philadelphia, pages 200–208.



11. Blankenberg FG, Norbash AM, Lane B, Stevenson DK, Bracci PM, Enzmann DR. Neonatal Intracranial Ischemia and Hemorrhage: Diagnosis with US, CT, and MR Imaging, Radiology, 1996, 199:253–259.

12. Blankenberg FG, Loh N-N, Bracci P, et al. Sonography, CT and MR Imaging: A Prospective Comparison of Neonates with Suspected Intracranial Ischemia and Hemorrhage, AJNR, 2000, 21:213–218.


Abstract This chapter is an introduction to the mechanisms underlying the use of contrast. The concept of the blood-brain barrier (BBB) is presented. Within the brain, contrast is either confined by the BBB and remains within vessels or, in pathologic brain processes, leaks into regions where the BBB has been disrupted. The presence of contrast in abnormal tissue and the time course of the resulting enhancement on magnetic resonance (MR) and computed tomography (CT) aid in dating an injury. These appearances are demonstrated and graphically explained.

Keywords Blood-brain barrier • Iodinated contrast • Paramagnetic contrast • Computed tomography (CT) • Magnetic resonance (MR)

Introduction

It is very important to discuss the role of contrast in imaging of the central nervous system. The use of contrast allows visualization of the disrupted blood-brain barrier (BBB) and can be a sensitive study related to the presence of injury to the brain parenchyma. In general, a region of “contrast enhancement” appears bright on both computed tomography (CT) and magnetic resonance (MR) images, although the physical mechanisms producing the brightness are different for the two modalities. For our purposes of dating the time of a neurological injury, the time interval in which contrast enhancement occurs on CT and MR following an insult is an extremely important compo-nent of the dating process.

On CT scanning, the contrast agent is a small molecule that contains iodine, which is very effi-cient at stopping X-rays. Consequently, areas that have an abnormal increase in contrast concentra-tion stop more photons, i.e., the tissue is “denser” as far as the photons are concerned and, thus, these regions will appear whiter on a CT scan. On MR, the contrast agent has differing effects on T1- and T2-weighted images but, primarily, T1-weighted images are used after contrast has been administered. Similar to CT, the contrast agent, which is a gadolinium-containing compound, causes an increase in brightness in the affected areas on postcontrast T1-weighted images.

Basic Principles to Understanding the Use of Contrast in the Brain

Principle #1: Contrast in normal brain stays intravascular, secondary to the presence of an intact BBB.

Chapter 3The Basics of Contrast and Its Role in Dating

60

Fig. 3.1 The blood-brain barrier (BBB). Unlike the rest of the body, the brain has a much tighter compartmentaliza-tion of the intravascular space. The cells lining the inside of the blood vessels have tighter junctions between them, and the membrane on which the cells rest is also more resistant to the passage of materials than elsewhere in the body

3 The Basics of Contrast and Its Role in Dating

In the case of both CT and MR, the contrast agent is a small molecule which is connected to either an iodine atom (in the case of CT) or a gadolinium atom (in the case of MR). While both types of compounds are chemically very different, the net result is that, when administered to a patient for a contrasted CT or MR scan, both agents stay primarily within the blood vessels that feed and drain the brain. This is qualitatively different from what occurs elsewhere in the body.

Within the brain, the microscopic anatomy of the blood vessels is different than the remainder of the body. Unlike most of the rest of the body, the cells lining the blood vessels in the brain have tight connections between them. These tight junctions prevent the majority of substances in the blood from passing through the vessel wall and into the brain. The result is a relative restriction of the passage of materials into the brain by what is called the blood-brain barrier (BBB) [1–4]. The only substances allowed passage into the brain by the cells in the walls of the blood vessels are those recognized as being normal nutrients, or compounds that the brain requires for its metabolic activity. In the normal state, all of the oxygen, glucose and metabolic substrates that the normal brain needs to survive are allowed passage. However, contrast agents are not allowed passage through these tight junctions of the normal, intact BBB. As a conse-quence, in the normal brain contrast stays intravascular secondary to this intact BBB (Fig. 3.1 – Line drawing of normal BBB).

Principle #2: Following the administration of contrast, the total amount of contrast in a portion of brain increases by a small amount.

Because approximately 2–3% of a given volume of brain is composed of the intravascular compart-ment (the vascular filling percentage) [5], this increase in contrast within the blood vessels is responsible for the generalized mild increase in density seen in the brain on postcontrast CT images, just as a similar generalized increase in signal occurs in the brain following MR contrast administra-tion with T1-weighted images.

61

Fig. 3.2 Normal vascular filling percentage of brain. Of the total volume of a sample of cerebral cortex and white matter, only a small percentage is intravascular, as represented here. Hence, contrast that stays within the vessels in the brain has only a slight effect on computed tomography (CT) number (if CT) or brightness (if MR)

Basic Principles to Understanding the Use of Contrast in the Brain

For CT Contrast

Once contrast is administered into a peripheral vein, the concentration of that contrast agent initially increases within the blood. As the contrast bolus passes into the brain, the CT density of the blood vessels increases and a mild general increase occurs in the CT numbers throughout the brain, both gray and white matter – not due to contrast that has passed through the BBB, but simply due to the fact that 2–3% of the volume of the brain is composed of the space that is within the blood vessels. Consequently, even though one sees a significant increase in density within the blood vessel, the overall increase in density of the brain parenchyma is slight, as the intravascular volume is a small percentage of the overall volume of any given volume of brain (Fig. 3.2 – Line drawing of normal vascular filling percentage in brain).

Therefore, as the bolus is given, the intravascular contrast concentration rises with time and reaches a peak, either during administration or right at the end of the bolus administration. In the normal brain, with an intact BBB, almost all of this contrast remains within the blood vessels. The result is that a slight increase in overall CT number can be seen within the gray and white matter of the brain [6, 7] (Fig. 3.3 – Line drawing/table of bolus, IV, brain CT numbers vs. time).

For MR Contrast

Just as a discussion of hemorrhage on MR is much more complex than the discussion of hemorrhage on CT, the discussion of the effects of contrast on MR scan is more complex. This is true for two reasons:

1. The effects of MR contrast are different on different pulse-sequence images, meaning that the appearance on a T1 image is different than the appearance on a T2 image.

2. The effects of contrast on MR images can be mitigated by other MR image factors. Specifically, for spin-echo pulse sequences, flowing blood tends to produce a signal void that counteracts the tendency of contrast to shorten T1 and, thence, increase the signal within the blood vessel on T1 images.

62

Fig. 3.3 Blood contrast concentration vs. time following bolus administration. After a bolus of intravenous iodinated CT contrast a transient increase is followed by a slower decrease to baseline of contrast concentration (and hence CT number) within the vessels. In a similar way, after a bolus administration of intravenous paramagnetic magnetic resonance (MR) contrast agent, tissues that enhance have an increase and then a decrease in their brightness on T1-weighted images


MR Contrast Dose and Pulse Sequence Choice

A complex relationship exists between the dose of contrast agent given and the resulting image brightness. The impact of contrast on the MR images also is very dependent on the pulse sequence chosen (whether spin-echo, gradient-echo, T1- or T2-weighted).

In almost all physiologic situations in which MR paramagnetic contrast is administered for patient imaging, contrast is given in the dose range that produces a linear response on T1-weighted sequences [8]. That is, as the dose of contrast increases, increased enhancement (and consequential bright signal) occurs. At greater than physiologic doses, a very large amount of contrast produces, even on T1-weighted sequences, a decrease in signal. This biphasic effect is never seen in the nor-mal imaging situation.

T2-weighted sequences are usually employed only in conjunction with contrast for the performance of MR perfusion examinations [9]. In these situations the contrast agent is given rapidly, and rapid scans are performed sequentially throughout the brain. As contrast washes through the brain paren-chyma, a decrease in signal proportional to the amount of blood flow regionally present, is seen in the perfused regions. In the physiologic range of doses, increasing amounts of contrast material produce increasing amounts of lowered signal from within the brain parenchyma. Therefore, no potential exists for the biphasic effect of increasing amounts of contrast on T2-weighted images [10].

MR Contrast Effects vs. Flow Void Effects

The other competing factor to MR contrast enhancement when using a spin-echo T1-weighted pulse sequence is the effect of blood flow. In a typical situation without contrast, blood flowing above a certain threshold velocity will produce a flow void. While present on many different pulse sequences, the effect is often easiest to visualize on a T2-weighted pulse sequence.

Regardless of the sequence used, the administration of contrast shortens the T1 of the blood, such that slower velocity blood may no longer have a flow void. The net effect is that before contrast admin-istration above a certain velocity range, flow voids are seen on any spin-echo pulse sequence – and after contrast, some of the vessels with slower flowing blood (just above the prior threshold) will now lose their flow voids. Therefore, two competing and opposite effects occur on the blood flowing within

63

Fig. 3.4 Concentration of contrast in brain tissue, vessels, and the body vs. time after bolus contrast administration. Note that blood concentration peaks first, followed by concentration in the body; afterward, concentration in the body remains slightly higher than in the vessels, as there is a gradient into the vessels, and then out of the body by excretion. Note also that the concentration is always lower in the normal brain than elsewhere in the body


the vessels. On the one hand, the more rapid the flow, the more the signal tends to “drop out” or disap-pear, producing the “flow void.” On the other hand, giving contrast tends to brighten the signal of blood. Therefore, contrast, given in the doses normally used in spin-echo imaging, can counteract the flow void effect and produce signal in vessels where signal would not otherwise be seen.

On contrast enhanced MR examinations, this ability to see contrast in vessels that contain slow flowing blood is an important finding in the setting of ischemia. For the first few days of an infarc-tion, when there is little or no parenchymal enhancement, enhancement can be seen in the vessels within and adjacent to the infarcted region. This is an early sign of brain injury [11].

Principle #3: The concentration of contrast agent within the vessels of the vascular system throughout the body has an increase and then a decrease after administration of the bolus, due to renal excretion and contrast behavior in all parts of the body except for the central nervous system (i.e., places with no BBB).

After the bolus is given, concentration of contrast within the vessels will transiently increase until the end, or just after the end of administration. After that, the concentration within the vessels will decrease. This decrease in intravascular contrast concentration occurs because as the contrast circu-lates throughout the remainder of the body, it is (1) distributed in other tissues of the body that do not have a BBB and (2) it is excreted by the kidneys [6, 7].

In the brain, one observes a simple decrease in intravascular contrast after the peak is reached, with very little contrast actually entering the brain tissue. A completely different behavior of contrast occurs in the remainder of the body. As no BBB exists, contrast does leak out of the blood vessels into the various tissues and organs. Once enough contrast has leaked into the tissues of the body, the gradient between the higher concentration in the blood and the lower concentration in the tissues disappears and contrast no longer preferentially flows from the blood vessels to the tissues. As the kidneys clear the contrast – lowering the concentration in the blood vessels – in all body tissues except the brain, contrast preferentially moves out of the tissues back into the blood vessels, from which it is again progressively cleared. The net result is that, in the normal brain, contrast concentration rises in the vessels, followed by a fall in contrast concentration in the vessels, which is a reflection of events happening elsewhere in the body (Fig. 3.4 – Time vs. contrast curve in vessels, brain and body).

64

Fig. 3.5 MR of normal enhancing tissues; all images are postcontrast T1-weighted images. (a) Sagittal midline image with pituitary enhancement. (b) Axial image with enhancement of the choroid plexus in the lateral ventricles. (c) Coronal image with enhancement of the dura and visualization of slow-flowing venous blood in the internal cerebral veins

Fig. 3.6 CT of normal enhancing tissues. In (a) enhancing vessels and pituitary stalk. In (b) enhancement of the dura and choroid plexus


Principle #4: Some normal tissues within the cranial vault enhance with contrast.

In interpreting contrast examinations of the brain it is important to recognize that a number of tissues in the brain do enhance following contrast administration. Very simply, these are tissues that do not have a BBB. This is because either they are not part of the central nervous system itself – though they are still intracranial – or they are portions of the brain which simply do not have an intact endothelial barrier. Examples of the former are such structures as the dura and the other two components of the meninges. Examples of the latter are the pituitary stalk and pituitary gland and the choroid plexus. In addition, on any MR or CT scan of the head that includes the brain, addi-tional, extra dural structures will be seen, and these can enhance with contrast as well. The most common, which can be a complicating factor on head scans, is the brisk enhancement of all of the mucosal surfaces within the nasopharynx, oral pharynx, airway, and paranasal sinuses. While this is rarely a distraction on routine MRI images, MRA images of the brain obtained with contrast have these confounding areas of enhancement which make interpretation of the resulting images difficult (Figs. 3.5 and 3.6 – MR and CT of normal enhancing tissues).

65

Fig. 3.7 MR of BBB disruption. (a) Axial T1-weighted postcontrast and (b) axial FLAIR image of an enhancing metas-tasis in the left posterior white matter. (c) Postcontrast axial T1-weighted image of enhancement of a right frontal brain abscess. (d) T1-weighted image with contrast with enhancement in the region of a stroke 6 days after the initial event


Principle #5: Some tissues within the cranial vault do have an intact BBB in the normal situation, but in the presence of disease they will have an abnormal or completely absent BBB.

The list of causative agents which can disrupt the BBB in normal brain includes trauma, infection, inflammation, infarction, ischemia, malignancy, and chemical or toxic agents. This list is both long and very diverse, so that the mere existence of disruption of the BBB is rarely specific enough to allow one to make an exact diagnosis. On the other hand, this uniformity of response and the fact that a large number of conditions will disrupt the BBB makes the presence of enhancement within the brain tissue a sensitive detector of underlying pathology (Figs. 3.7 and 3.8 – MR and CT of BBB disruption).

For the purpose of this book, we are most interested in the disruption of the BBB that occurs with ischemia and infarction. In the setting of ischemia, those portions of the brain that suffer some decreased level, either of blood flow and/or oxygenation, such that the tissue is “injured,” can have varying degrees of blood-brain disruption. Consequently, contrast can be used to detect areas of stroke.

66

Fig. 3.8 CT of BBB disruption. Precontrast image (a) and postcontrast axial image (b) of a metastasis in the posterior left white matter (same patient as in Fig. 3.7a, b). (c) Enhancement around a region of infection


In the reparative phase following a stroke, enhancement can be seen around the margins of the infarct as the brain attempts to repair the damage and/or remove the damaged tissue. A number of different patterns of enhancement can be seen in infarction [12].

Principle #6: Administration of contrast, therefore, reveals those areas with either an absent BBB or with a disrupted BBB.

The administration of contrast, therefore, will result in the enhancement of two different sets of tissues. The first types of tissues are those normal tissues within the intracranial compartment that enhance following contrast (pituitary, pineal gland, choroid plexus, dura and normal vascular structures) and those tissues beyond the cranial vault that normally enhance (mainly mucosal surfaces, salivary glands and the thyroid gland). The second set of tissues that enhance are normal tissues of the brain in which some degree of disruption of their normal BBB has occurred.

Principle #7: Enhancement does not equal pathologic activity, as enhancement persists for some time after infection, inflammation or tumor.

Once the BBB is disrupted by infection, inflammation or tumor, the length of time that the disrup-tion can be seen on routine scanning varies from several weeks postevent on CT [13–16] to up to perhaps 2 months on MR scanning [11, 17, 18]. Consequently, if one sees an enhancing area, it does not necessarily mean that an inciting agent (such as an active infection) is still present or that the infarction was recent, especially on MR.

Therefore, the specific issue of enhancement and, more precisely, persistent enhancement and its place in the evaluation of injury, is extremely important. The key factor here is that visible CT enhancement persists for a shorter length of time than visible MR enhancement. While CT enhance-ment following a period of injury is not normally seen past 2–3 weeks, enhancement following injury on MR can be seen for weeks or months. Thus, enhancement on CT is a good indicator of injury occurring in the 2–3 weeks prior to the scan but the same cannot be said about the enhance-ment noted on MR. Because of this, enhancement alone is insufficient to accurately date the occur-rence of a neurologic event on MR scanning. Other characteristics of the injured brain, such as its morphology, amount of edema, amount of parenchymal loss, and signal change over time (espe-cially if hemorrhage is involved), must be additionally used to date the time of the injury.

However, in the case of MR, the presence of intravascular enhancement is a fleeting phenomenon, seen only for a few days following the infarction. In this case only, the presence of enhancement can be a very reliable indicator that the edema-causing event occurred in the previous few days.

Principle #8: Enhancement on CT and MR is not specific for any one exact type of causative pathology.

67References

As discussed earlier, the brain has a rather systematic way of responding to insult, no matter what the inciting agent. When edema is present, either vasogenic or cytotoxic, disruption occurs to the BBB. Therefore, without depending on the morphologic pattern of the injury or any clinical input, it is often virtually impossible to specify what inciting agent produced the BBB disruption based only on the presence of contrast enhancement.

Clinical Importance of Contrast Enhancement

This chapter intensely details the physics and subtler biological dynamics underlying the use of contrast in MR and CT. The importance of contrast enhancement is to take advantage of the finite time window following injury when an injured brain will enhance on CT. The time window is longer on MR. Therefore, it is very helpful in the dating of a neurologic injury to understand why contrast occurs and to understand the expected boundaries of these time windows.

While the endpoints are not “set in concrete,” in general, CT enhancement of the brain tissue can extend from a few days to several weeks following an ischemic event. On MR, the contrast enhance-ment possibly begins slightly earlier but can extend for weeks or months afterward. However, injuries older than weeks in the case of CT and older than many months in the case of MR will not enhance if the etiologic agent was ischemic in nature. In addition, on MR, intravascular enhancement is usually seen for the first few days following the ischemic event.

As explained in the subsequent chapters, the presence or absence of contrast enhancement on CT or MR becomes yet another data point in the overall matrix of findings used to try to pinpoint the time when a neurologic injury occurred. Contrast enhancement should never be used alone as the means to date an injury but, used together with the changing imaging appearance on US and/or CT and/or MR, it can provide useful additional information.

References

1. Bakay L. The Blood Brain Barrier, 1956, Thomas, Springfield, Illinois. 2. Brierly JB. “The Blood-Brain Barrier: Structural Aspects” in Metabolism of the Nervous System, edited by

Richter D, 1957, Pergamon Press, London, pages 121–135. 3. Dobbing J. The Blood-Brain Barrier, Guy’s Hosp Rep, 1956;105:27–38. 4. Edstrom R. An Explanation of the Blood-Brain Barrier Phenomenon, Acta Physiol Neurol, 1958;33:403–416. 5. Weinmann H-J, Muhler A, Raduchel B. “Gadolinium Chelates: Chemistry, Safety and Behavior” in Methods in

Biomedical Magnetic Resonance Imaging and Spectroscopy, edited by Young IR, 2000, Wiley, New York, Vol 1, page 709.

6. Dawson P. Textbook of Contrast Media, 1999, Isis Medical Media, Oxford, UK, Chapter 6: Pharmacokinetics of Water-Soluble Iodinated X-Ray Contrast Agents, pages 61–74.

7. Dawson P. Textbook of Contrast Media, 1999, Isis Medical Media, Oxford, UK, Chapter 17: Contrast Agents in Computed Tomography, pages 217–228.

8. Lin W, Haacke EM, Smith AS, Clampitt WE. Gadolinium-enhanced High-Resolution MR Angiography with Adaptive Vessel Tracking: Preliminary Results in the Intracranial Circulation, JMRI, 1992;2(3):277–284.

9. Ostergaard L. “Cerebral Perfusion Imaging by Exogenous Contrast Agents”, in Methods in Biomedical Magnetic Resonance Imaging and Spectroscopy, edited by Young IR, 2000, Wiley, New York, Vol 1, pages 537–549.

10. Weinmann H-J, Muhler A, Raduchel B. “Gadolinium Chelates: Chemistry, Safety and Behavior” in Methods in Biomedical Magnetic Resonance Imaging and Spectroscopy, edited by Young IR, 2000, Wiley, New York, Vol 1, pages 705–711.

11. Elster AD, Moody DM. Early Cerebral Infarction: Gadopentetate Dimeglumine Enhancement, Radiology, 1990; 177:627–632.

12. Smirniotopoulus JG, Murphy FM, Rushing EJ, Rees JH, Schroeder JW. Patterns of Contrast Enhancement in the Brain and Meninges, Radiographics, 2007;27:525–551.

68 3 The Basics of Contrast and Its Role in Dating

13. Norman D, Stevens EA, Wing SD, Levin V, Newton TH. Quantitative Aspects of Contrast Enhancement in Cranial Computed Tomography, Radiology, 1978;129:683–688.

14. Caille JM, Guibert F, Bidabe AM, Billerey J, Piton J. Enhancement of Cerebral Infarctions with CT, Comput Tomogr, 1980;4:73–77.

15. Pullicino P, Kendall BE. Contrast Enhancement in Ischaemic Lesions I. Relationship to Prognosis, Neuroradiology, 1980;19:235–239.

16. Hornig CR, Busse O, Buettner T, Dorndof W, Agnoli A, Akengin Z. CT contrast enhancement on brain scans and blood-CSF barrier disturbances in cerebral ischemic infarction, Stroke, 1985;16:268–273.

17. Crain MR, Yuh WT, Greene GM, Loes DJ, Ryals TJ, Sato Y, Hart MN. Cerebral Ischemia: Evaluation with Contrast Enhanced MR Imaging, AJNR, 1991;12:631–639.

18. Karonen JO, Partanen PLK, Vanninen RL, Vainio PA, Aronen HJ. Evolution of MR Contrast Enhancement Patterns during the First Week after Acute Ischemic Stroke, AJNR, 2001;22:103–111.

69J.L. Creasy, Dating Neurological Injury: A Forensic Guide for Radiologists, Other Expert Medical Witnesses,and Attorneys, DOI 10.1007/978-1-60761-250-6_4, © Springer Science+Business Media, LLC 2011

Abstract This chapter focuses on the changes that occur in a region of infarction and edema over the first 2 weeks from the time of occurrence, on MR, CT, and US, presenting the different locations of possible hemorrhage and the changing appearance of recent hemorrhage on MR, CT, and US.

Keywords Magnetic resonance (MR) • Computed tomography (CT) • Ultrasound (US) • Infarction • Edema • Hemorrhage • Dating of lesions

Introduction

Writing a book whose purpose is to date events by their neuroradiology imaging appearance is a daunting task. This chapter will discuss sequentially how edema on CT, edema on MR, hemorrhage on CT, hemorrhage on MR, and edema and hemorrhage on ultrasound (US) mature and change over time. Initially, the changes seen over time with infarction and hemorrhage were described by CT scanning; however, early CT scanning was much less sensitive than modern CT scanning. Descriptions of the CT appearance of both infarction and hemorrhage in the early CT literature used the terms “acute,” “sub-acute,” and “chronic” – with the later addition of “hyper-acute” and “old.” To confuse matters even further, the introduction of MR scanning in the 1980s used the same set of names for the time periods, but these time periods did not always have the same endpoints. For dating, the advent of MR is fortuitous because the multiple changing factors permit estimating the time of occurrence of an event with more precision than CT alone.

For our purposes, we will begin the discussion of dating by using the same terms – but we will precisely define the time limits of each phase and will fix these times for CT and MR, and for both edema/infarction and hemorrhage. In our usage, “hyperacute” will be the time period from the occurrence of the bleed or infarct until 6 h postevent. “Acute” will be from 6 h postevent until 3 days postevent. “Sub-acute” will be from 3 to 14 days, or 2 weeks. “Chronic” will begin at 2 weeks after the event – and at some vague time of many months or even many years, will become “old.”

We will attempt to be as precise as the literature permits, while trying to limit the discussion to human research, experiments, and empirical observations – with mention of animal studies only when they are the sole reasonable manner to document the time changes. Rather than simply say-ing that it is a “sub-acute” finding, we will emphasize the time range, that is, 2–4 days. Otherwise, the literature would offer a large number of possibilities for the precise time period subacute references.

Chapter 4How the Imaging Appearance of Edema and Hemorrhage Change Over Time on CT, MR, and US: Dynamic (Acute) Dating

70 4 How the Imaging Appearance of Edema and Hemorrhage

Changes of Edema Over Time on CT

As already mentioned, the brain has a limited response to an insult and, to most insults, it responds with some degree of edema. This edema follows a very characteristic evolution over time, with the only uncertainty being the variations in duration or length of each individual step as the edema matures and not the sequence of changes from step to step. Similarly, one sees a predictable sequence of changing imaging appearance of hemorrhage on both CT and MR. However, we will first discuss the time evolution of edema on CT scanning.

In the hyperacute phase, edema, whether it is vasogenic or cytotoxic, is usually first visible within, at most, a few hours after the ischemic or anoxic event. Traditional teaching held that some infarcts could not be seen on CT for up to 24 hours; however, with modern scanning this is somewhat unusual. The short version of the evolution of edema findings on CT is that the first signs of the infarct are noted within minutes to hours of the initial event. Such early signs of ischemia include decreased density in the involved gray and white matter, blurring or complete loss of the normal interface between gray and white matter, and generalized brain swelling within the affected area. It is unusual at this early stage to see large degrees of cerebral edema within the affected area, although, in the rare case, significant brain swelling can occur. At this early stage, the infarcted area is characteristically of lower density than normal gray and white matter, but is usually less dense by only a few CT numbers. The margins of the infarct are also poorly visualized [1].

By the end of the acute phase (6 h to 3 days postevent), the brain has entered the period of the maximum amount of swelling. Indeed, maximal swelling traditionally is said to occur between 3 and 5 days following the event [2]. Therefore, it can be difficult, when viewing a brain in which one sees a mild to moderate amount of swelling, to know whether the infarct is being imaged prior to the time of the maximal swelling or following the time of the maximal swelling. However, helpful clues exist. If the brain is imaged before the time of maximal swelling, the infarct tends to be of not very low density and to have poorly defined margins. By the time one is on the far side of the peak of maximal swelling, a similar amount of edema can be seen, but the infarct tends to be much lower in density and to have much more clearly defined margins. During this acute phase of the infarct enhancement can be seen both within the infarct itself and around the periphery of the infarct [3, 4].

Early in the subacute phase (the front end of 3–14 days postevent), the edema on CT scan is near its maximum and by end of this period the brain is beginning to return to its baseline appear-ance, due to the steadily decreasing mass effect. Any geographic infarcts that have resulted are becoming more and more sharply defined, though the core of the infarcted area still may not be at the same low density as cerebral spinal fluid. Enhancement is continuing, though it is begin-ning to level off, and is beginning to become less prominent than it was during the acute period of edema [5].

In the chronic phase of infarction (older than 14 days), the brain has responded by complete lysis and removal of the dead brain tissue. Therefore, the infarcted area is very sharply demarcated and it contains only CSF with a density equal to that of the spinal fluid within the ventricles. At this time, no mass effect is seen on CT and there is no contrast enhancement. With CT, these chronic or old findings are stable and appear the same, whether the CT scan is obtained at 14 days after an infarct, or 2 months, or 6 months, or 5 or 10 years after the acute infarct [6].

Finally, dystrophic calcification can occur in a region of infarction; however, this normally takes at least 7–14 days to develop, even in a child. Calcifications within an area of infarction do not occur sooner than 1–2 weeks from the acute event.

The series of CT images in Fig. 4.1 demonstrate the normal evolution of the edematous changes following an acute ischemic or anoxic event. Note the time course of the appearance of low density, the loss of gray/white differentiation, and the maximal edema.

71

Fig. 4.1 CT exam of acute infarction. In (a), large infarct in the left middle cerebral artery territory with loss of gray–white differentiation and decrease in density. In (b), same patient 12 h later. The margins of the infarct are becoming better demarcated with an increase in edema. In a different patient, in (c) acute infarct in the left MCA territory, with unusually dense cortex due to microscopic hemorrhage into the infarcted cortex. In (d) a CT of the same patient 1 week later; the increase in density in the cortex is now more apparent. In (e) same patient 3 weeks later; the high density in the cortex beginning to abate and with an overall decrease in edema and swelling

Introduction

Changes of Edema Over Time on MR

Edema as a consequence of an inciting ischemic or infarction event is superiorly imaged on MR compared to CT. Not only can changes be seen more rapidly on MR, but, overall, the sensitivity at a given point in time is also better for MR [7]. Because MR employs a variety of pulse sequences, not all sequences are optimized for the detection of edema. The best sequence is the diffusion-weighted sequence, which can detect early changes of edema within minutes of an event.

From a clinical perspective, it is obvious that the injured brain does not differentiate whether MR or CT is the imaging modality, since the changes going on in the parenchyma at a microscopic and macroscopic level are identical, regardless of the modality utilized. But, MR is able to detect the abnormality sooner following the inciting event.

In the hyperacute period, the diffusion-weighted scan is the first sequence that will become posi-tive [8, 9]. This type of MR scan is especially constructed to be sensitive to restrictions on normal diffusion of water on the microscopic level. Since very early edema is manifested on that microscopic


level by swelling of the cells, diffusion is restricted, especially in the extracellular space. This restricted diffusion shows up on a diffusion-weighted image (DWI) as a bright signal. While some artifacts can produce similar bright signal on a diffusion-weighted scan, if an apparent diffusion coefficient (ADC) map is also computed, false positives are effectively eliminated, as a true region of ischemia or infarction will not only be bright on the diffusion-weighted scan, but will also be black on the ADC map. It is likely that the DWI scan is primarily sensitive to cytotoxic edema, rather than to vasogenic edema.

In the following acute phase, a normal fluid-attenuated inversion-recovery (FLAIR) sequence scan will convert to an abnormal FLAIR scan in a matter of hours [3, 10]. Abnormalities visible on a T1- or T2-weighted scan take longer to develop, and it may take hours to a day or two to develop an abnormal-appearing T1- or T2-weighted scan [3, 4, 10]. Usually, one sees no enhancement in this initial period. In the acute period, as seen on CT, the patient’s affected brain begins to show signifi-cant cerebral edema and swelling, which will be maximal at 3–5 days, but may appear prior to and after that peak, so that the entire period of edema can be from several hours to a week or 10 days after the inciting event. It should be noted that some areas that are positive on a diffusion-weighted scan will not necessarily progress to complete infarction. Some diffusion-weighted abnormalities represent reversible ischemic change and, if properly treated, those portions of the brain can recover without infarction.

The subacute phase of edema is a period of rapid changes between the acute and the chronic periods, as seen on MR imaging. Both the morphology of the region of injury and the signal intensi-ties of the affected brain parenchyma are in transition [4, 5, 11]. As to morphology, by the end of the subacute phase, the mass effect has peaked and is subsiding and a region of definite total infarc-tion with a much smaller (if present at all) surrounding or adjacent area of ischemia is now clearly seen. The brightness of the infarction on the DWI is also decreasing and will return to a brightness equal to or less than normal brain by 2 weeks after the event. At the same time, the signal intensity of the injured area continues to increase on FLAIR and T2-weighted sequences. Unlike CT – where maximal enhancement is usually over within 10–14 days – contrasted MR images may show enhance-ment around the injury for weeks or months.

In the chronic phase, edema abates and the amount of residual brain is revealed [6]. In the event of a significant insult, one will see a loss of brain parenchyma compared with the preevent brain. The ischemic areas that have not infarcted will return to completely normal (i.e., a normal back-ground level of gray), while the infarcted areas will be fluid-filled and will appear as dark areas on the diffusion-weighted scan. However, beginning at this time, areas that have suffered ischemic injury without going on to a complete infarction will persist with their FLAIR and T2 signal abnor-malities. This continuation of the bright FLAIR and T2 signal in the injured area will continue for the life of the patient; hence, it is of no help in dating the time of the injury. The brain will have the same appearance whether it is imaged 3 weeks, 6 months, or 5 years after the ischemic or hypoxic event.

Again, as with CT, there is a period of time during which the area of injured brain has a disrupted blood–brain barrier and will enhance if contrast is administered. This maximal period of parenchy-mal enhancement is a time window which is broader than that seen on CT, as MR is a more sensitive examination. Stated differently, relatively more brightness or enhancement will be seen on the MR than on the CT image, presuming identical amounts of blood–brain barrier disruption and the same physiological amounts of iodinated (for CT) and paramagnetic (for MR) contrast are given. Unlike CT, contrasted T1-weighted MR images will have a brief initial period of several days when intra-vascular enhancement is present [4, 12].

In very old infarcts or ischemia, one sees continued stability with loss of brain substance, persis-tent bright signal on FLAIR and T2, and no abnormalities of restricted diffusion. For the changing appearance of edema on MR images, see Fig. 4.2.

73

Fig. 4.2 MRI examination of acute infarctions. In (a), axial diffusion-weighted imaging at the midbody level of the ventricles demonstrates the small area of restricted diffusion in the posterior limb of the internal capsule on the right side. Axial FLAIR image in (b) shows minimal to no findings in this same area, as is also true in the T2-weighted image in (c). In (d), in a different patient, a very large acute left-sided acute infarction is seen at 24 h after the event; and in (e) abnormal FLAIR signal seen at 3 days later (approximately day 4) shows swelling and mass effect with effacement of the left lateral ventricle. In (f) (same patient as in (a–c), now after 2 days, the infarct has increased in size on the diffusion-weighted scan. This new, larger infarct is now visible on the axial FLAIR scan in (g) and the axial T2 scan in (h)

Introduction


Locations of Possible Intracerebral Hemorrhage

Blood occurs in the brain in one or a combination of the following spaces, beginning at the outer surface of the brain and working centrally: (1) epidural hemorrhage, (2) subdural hemorrhage, (3) suba-rachnoid hemorrhage, (4) various types of intraparenchymal hemorrhage, and (5) intraventricular hemorrhage. Each of these five major types of hemorrhage has its own distinguishing imaging characteristics, most common causative factors, and unique evolution and duration of hemorrhage, as imaged by the three primary modalities of CT, MR, and US [13, 14].

Epidural hematomas occur in the space between the bony inner table of the skull or calvarium and the normally closely applied layer of dura. They are almost always caused by trauma in which there has been a fracture of the calvarium, with laceration of vessels lying either within the skull or on the inner surface of the skull. On axial cross-section, epidural hematomas have a characteristic biconvex or lens shape [15] (Fig. 2.12c).

Subdural hematomas occur in the space between the dura and the pia-arachnoid on the surface of the brain. As explained in Chap. 1, this is normally a potential space and it does not exist, but when the membrane on either side is traumatized, blood can enter and expand this potential space so that it becomes a real space. Unlike epidural hematomas that are limited with sharp borders and the characteristic lens shape, subdural hematomas are free to flow more widely over the surface of the brain because they are not bounded circumferentially where the dura is still attached to the inner table. Therefore, subdural hematomas can cover a whole lobe or even a whole hemisphere and can even cross the midline [16] (see Figs. 2.12b and 2.13b–d).

Subarachnoid hemorrhage is blood that occurs in the space between the pia-arachnoid membrane and the surface of the brain [17]. This space is normally filled with cerebrospinal fluid. Trauma is by far the most common cause of subarachnoid hemorrhage. However, in those patients who do not have a trauma history, the most common cause of subarachnoid hemorrhage is rupture of an intracerebral aneurysm. Some confusing crossover may exist between these two different sets of patients. If, for example, a traumatic patient is seen with subarachnoid hemorrhage, it may be that trauma caused the subarachnoid hemorrhage. Alternatively, a ruptured aneurysm could have produced the subarachnoid hemorrhage, which subsequently led to the traumatic event. It should also be noted that the subarachnoid space freely communicates with the cerebral spinal fluid present within the ventricles. Hence, given time, blood present in the subarachnoid space will be seen within the ven-tricles, and vice versa (see Figs. 2.12a and 2.13a).

The category of intraparenchymal hemorrhage is broad, as there are several degrees of hemorrhage and each has unique imaging characteristics. Hemorrhagic injuries also can be thought of as occurring either focally or diffusely. Grading focal intraparenchymal hemorrhage from least to most severe would minimally include the categories of shear injury, contusion, and parenchymal hematoma. The more severe, diffuse form of multiple hemorrhages is also referred to as diffuse axonal injury.

Shear injuries most commonly occur at the gray–white junction where there is a difference in elastic modulus between the gray and white matter. They also occur at regions where the brain slams up against adjacent bone or other rigid structures during rapid linear deceleration. Therefore, these injuries most likely involve the frontal poles, the temporal tips, the occipital poles, and either side of the falx cerebri (Fig. 2.14). If more complex motions produce a closed head injury, especially those with an angular component, deeper shear injuries involving the basal ganglia, corpus callosum, and brainstem can occur. Hemorrhagic contusions can be thought of as slightly more severe than a simple shear injury [18]. The involved brain is “bruised” and, on a microscopic level, has blood oozing out through multiple damaged capillaries into the brain parenchyma. These contusions tend to occur at the same locations and for the same reasons as the shear injuries noted above.

Parenchymal hematomas have the distinguishing characteristic that no normal brain constituents exist within the hematoma cavity. In essence, a hematoma completely displaces all of the surrounding normal

75Locations of Possible Intracerebral Hemorrhage

brain circumferentially, in a radial fashion. Besides trauma, the typical causes of intraparenchymal hemorrhage would include bleeding from an intrinsic tumor of the brain parenchyma (such as a glioma), bleeding from a highly vascular metastasis to the brain, bleeding from the rupture of a portion of a vascular malformation, or a hypertensive hemorrhage.

Diffuse axonal injury affects large regions of the brain parenchyma simultaneously. It is likely due to extreme linear or rotational acceleration or deceleration [19, 20]. Clinically, patients are in a coma, with generally a poor prognosis. Widespread injuries to the white matter are best seen on MR (FLAIR and gradient echo [GE] sequences). The lesions most commonly involve the corpus cal-losum, brainstem, and gray–white junction of the cortex.

Intraventricular hemorrhage is usually seen in conjunction with trauma that tears vessels around the bodies of the ventricles. Vascular tumors that are adjacent to the ventricles or lesions within the ventricles, such as choroid plexus tumors, can also produce blood within the ventricles. Again, remember that this space is contiguous to the subarachnoid space, such that blood in the subarach-noid space over the surface of the brain can eventually be seen within the ventricles. Similarly, intraventricular hemorrhage can move through the ventricles and eventually be seen over the con-vexities in the subarachnoid space (Fig. 2.16).

Changes of Hemorrhage Over Time on CT

Hemorrhage within the brain results in a fluid collection which is higher in density than normal gray or white matter on CT scanning. This occurs because blood of a normal hematocrit is naturally denser than brain parenchyma. Thus, whether the blood is present within the vascular system or outside the vascular system in an abnormal location, it will be of higher density. The three factors which most influence the density of the region of hemorrhage are the patient’s hematocrit, the hem-orrhage maturity (i.e., has clot retraction occurred), and the hematoma age.

The patient’s hematocrit is important, as a very low hematocrit may result in hemorrhage less dense than normal brain, even if it is acute. Regardless of the initial density of the hematoma, some increase in density can be seen immediately following its occurrence as the clot organizes and “retracts,” squeezing some of the fluid out of the clot and leaving the higher density protein structure of the organizing clot behind. As this occurs, the clot density (CT number) increases to the upper CT number limit of hemorrhage, roughly 80–85.

Of all the potential spaces where hemorrhage can occur (epidural, subdural, subarachnoid, intra-parenchymal, and intraventricular), blood clears most rapidly in the subarachnoid space, if only a small amount is present. Larger subarachnoid bleeds, subdural and epidural extraaxial bleeds, and intraventricular and intraparenchymal bleeds take days to weeks to clear. In all cases, however, blood clears by gradually decreasing in density and apparent size on CT. A good example of this is a classic subdural hematoma, which begins as a high density, crescent-shaped collection of blood. Over time, the fluid, blood, and blood-breakdown products within the hematoma decrease in den-sity. Eventually, the hematoma changes from being highly dense to being the same density (or isodense) as brain and then continues decreasing in density until it approaches that of cerebrospinal fluid. However, even older subdural hematomas may retain a CT density slightly higher than spinal fluid due to retained proteinaceous material. The time it takes for a blood collection to evolve and eventually disappear on CT relates in some degree to the amount of blood circulating around the hemorrhage. Therefore, with similar-sized bleeds, an intraparenchymal hematoma will clear more rapidly than a subdural or an epidural hematoma.

Discussing the time periods in hemorrhage, therefore, involves dating similar to that discussed for edema. The hyperacute period is that time when the hemorrhage initially occurs and is at its highest density. This is the case for the first several hours. By the acute phase, clot retraction occurs and the hematoma may slightly increase in density. It is also during this period that the brain often

76

Fig. 4.3 Axial Image of subarachnoid hemorrhage on CT scanning. In (a), acute image demonstrates extensive subarachnoid hemorrhage throughout all of the cisterns (CSF-containing spaces) at the base of the brain. In (b), 10 days later, the subarachnoid blood is completely cleared

4 How the Imaging Appearance of Edema and Hemorrhage

responds to an intraparenchymal hematoma with surrounding edema. Hence, a CT scan obtained at several days will show the hematoma, approximately of its original size, surrounded by a band of edema. With increasing time intervals, the size of the hematoma will steadily decrease, as will its density, from an outward-in direction (like a melting ice cube). Eventually, in the subacute, chronic, and old appearance, an intraparenchymal hematoma may completely disappear except for possibly a small cleft. Over the same period of many days to several weeks, intraventricular hematomas and larger subarachnoid bleeds may completely clear, as well. Subdural and epidural hematomas have a habit of recurring and, while both may completely resorb without therapy, both are often treated with surgical evacuation if a midline shift occurs. Untreated, many subdural and epidural hemato-mas will eventually resolve, though chronic epidural hematomas may produce calcification of the adjacent underlying dura.

The different locations of hemorrhages, therefore, have different time evolutions. A small suba-rachnoid hemorrhage may no longer be visible on CT in as little as 1–3 days after the event. Larger subarachnoid hemorrhages, subdural and epidural hemorrhages, and large parenchymal and intra-ventricular hemorrhages may take as much as 2 weeks or longer to clear completely. A recent intraparenchymal hemorrhage may not appear to change size at all for the first several days, but at the same time a ring of edema will usually develop around the blood collection.

The accompanying CT images (Figs. 4.3–4.7) show the imaging appearance from the acute event over the first 10–14 days for each of the above-mentioned types of hematomas.

Changes of Hemorrhage Over Time on MR

In the first 2 weeks following hemorrhage, the appearance on MR undergoes the most dynamic and far-ranging changes of all the processes discussed in this book. The physiologic basis for this change is complex and beyond the scope of this text, as it involves what is occurring with the blood-breakdown products on the microscopic level. Simply put, the blood-breakdown products contain an iron molecule in the hemoglobin and this undergoes varying degrees of reduction or oxygen-ation; this, then, changes the number of unpaired electrons present, which has an effect on the signal characteristics.

77

Fig. 4.4 CT examination of time course of subdural hemorrhage. In (a), acute large right-sided subdural hemorrhage surrounds the hemisphere and is also present in the midline between the hemispheres, with swelling of the right brain. In (b), isodense left-sided subdural hematoma of approximately 7–10 days age. The density of this subdural hema-toma is identical to that of gray matter. In (c), a more superior placed image of an aging, almost chronic subdural hematoma that is equal in density to CSF. In (d), there is calcification of a very long-standing chronic subdural hema-toma, seen on bone windows


In effect, what happens in an evolving hemorrhage, as seen on MR, is an evolution and chemical change of the hemoglobin present within the blood. Hemoglobin in a fresh hemorrhage is primarily oxyhemoglobin, which is located intracellularly, as the red blood cell membrane is still intact. With decreasing oxygen tension, this hemoglobin progresses from oxyhemoglobin to deoxyhemoglobin then to methemoglobin and, finally, to large iron storage complexes, such as ferritin or hemosiderin. At the same time, the red cell membranes are losing their integrity and, subsequently, the blood-breakdown products become increasingly extracellular as the bleed ages. The ability of each of these different blood products or their breakdown components to function as influencers of T1 or T2 relaxation varies; consequently, the signal intensity on both T1 and T2 images of the hematoma dramatically varies over time. The hematoma appearance is affected not only by the presence and amount of oxyhemoglobin vs. deoxyhemoglobin vs. methemoglobin, but also by whether these moieties are distributed symmetrically throughout the clot or are local-ized within intact red blood cells (before they lyse later in the evolution of the hematoma). Lastly, MR images of the hemorrhage are affected by the hydration state of the clot, whether it is fresh

78

Fig. 4.5 Time course of epidural hematoma on CT scanning. In (a), acute right-sided small epidural hematoma seen in the anterior aspect of the right middle cranial fossa. Sequential images demonstrate the resolution of this hemor-rhage and its decrease in density. In (b), on day 2, slightly increased edema is visible around the epidural; in (c) on day 6, a slight decrease in density of the hematoma is seen; in (d), on day 10, there is a further decrease in density; and in (e), on day 14, the hemorrhage is now isodense to gray matter with only the interface between the hemorrhage and the brain being seen as a thin white line


and somewhat juicy, more contracted and solid, or undergoing lysis peripherally as the brain attempts to repair itself. Suffice it to say that these factors all produce a dynamically changing MR appear-ance on T1-weighted, T2-weighted and GE images that, for our purposes, will be described only empirically [21–26].

Having detailed this much of the underlying physical basis that accounts for imaging changes over time, we will repeat that it is not the purpose of this book to specify the actual physical basis for changes in signal intensity to any more depth than was given in the preceding paragraph. Suffice it to say that some of the signal changes are accounted for by concepts such as the number of unpaired electrons in the hemoglobin breakdown products, and by interactions between protons and electron dipoles (the wordy version is – proton–electron dipole–dipole proton relaxation enhance-ment). We will leave that detailed discussion to an appropriate chapter on hemorrhage within neu-roradiology textbooks.

Most descriptions of the time evolution of hemorrhage appearance on MR break the time intervals into hyperacute, acute, subacute, chronic, and old; similar to the phases we have used since the start of this chapter. In each of these five phases, the hematoma has a distinct appear-ance, varying on each different pulse sequence. One approach to addressing this complicated evolution of imaging appearance for each type of sequence is to look at a number of factors at each time period – such as the signal intensity of the central part of the hematoma, the signal

79

Fig. 4.7 Time course of intraventricular hemorrhage on CT examination. In (a), initial bleed is present within the right lateral ventricle and a smaller bleed within the right thalamus. In (b), 5 days later, the amount of the intraven-tricular hemorrhage is decreasing, and parenchymal hemorrhage is still present. In (c) at 3 weeks after presentation, the intraventricular component is now completely cleared. Low density persists in the basal ganglia

Fig. 4.6 Time course of intraparenchymal hemorrhage on CT scanning. In (a), acute scan showing hemorrhage into the cerebellum on the day of presentation. In (b), on day 2, (c) day 5, (d) day 7 and, eventually, (e) on day 16, the hemorrhage slowly decreases in density until it is the same density as brain


80

Fig. 4.8 Table graphically demonstrates the time course of signal changes within and around the hematoma on T1- and T2-weighted MR imaging. The horizontal line represents material that is isointense to brain parenchyma. Above the line is hyperintense clot; below the line is hypointense clot


intensity of the periphery of the hematoma, the appearance of the edge of the hematoma, and the reaction of the adjacent brain. Figure 4.8 gives a graphic depiction of the evolution of these various factors on T1- and T2-weighted images over the life of an intraparenchymal hematoma.

81T2 Changes Over Time Within a Hemorrhage on MR Scanning

With the understanding that we will attempt to limit the detailed explanation of evolving biochemical and physical changes, what follows is an empirical description of the signal changes on the most commonly performed sequences, i.e., T1- and T2-weighted. However, because many centers currently image hemorrhage additionally with FLAIR and GE (heavily magnetic suscepti-bility-weighted images), in the following sections we will discuss, in order, the time-varying appearance of hemorrhage on T1-weighted, T2-weighted, FLAIR, and finally, GE sequences [21–26].

T1 Changes Over Time Within a Hemorrhage on MR Scanning

On T1-weighted scans, an intraparenchymal hematoma has an initial hyperacute phase in which the signal intensity is isointense to slightly hyperintense (only oxyhemoglobin is present). The duration of this hyperacute phase is measured in minutes to hours.

In the acute phase, the signal intensity is slightly hypointense to isointense on T1. Oxyhemoglobin and, increasingly, deoxyhemoglobin are present.

In the subacute phase, the signal intensity rapidly increases on T1-weighted imaging, going from slightly hypointense to isointense, to become hyperintense before the end of the subacute period. These changes are a consequence of the production of methemoglobin and the lysis of the red cell membrane. This changing of the hematoma from isointense to hyperintense occurs first from the periphery of the hematoma and then moves to the center of the hematoma.

By the chronic phase, the intensity remains uniformly bright on T1 throughout the hematoma, primarily due to a large collection of dilute, extracellular methemoglobin.

One can recognize a fifth or “old” phase, in which all the blood products have been removed. At this point, the hemorrhage acts as a collection of fluid with very low protein content and, conse-quently, appears like spinal fluid with a resultant low signal on T1-weighted images.

The changes described to this point are those occurring within the hematoma itself as seen on T1-weighted images. Additional changes are occurring in the brain immediately adjacent to the hemorrhage. While hemosiderin accumulates in time, it affects T2-weighted images much more than T1-weighted images. Consequently, even a chronic or old hematoma which has a discernible black surrounding rim on T2-weighted images (see the below discussion) will have, at most, a very faint dark rim on T1-weighted images. The adjacent brain will also show changes consistent with edema, which peak during the acute and subacute phase and are rela-tively less prominent in the hyperacute phase and not visible in the chronic and old phases (Figs. 4.8–4.11).


Switching our attention to the changes on T2-weighted images of the hematoma itself, in the hyperacute phase, the signal intensity is increased on T2-weighting. This hyperacute phase lasts for the same minutes to hours as described under T1-weighted imaging appearances. The signal is hyperintense due to the high fluid content in fresh blood and is not offset by any of the paramagnetic effects, due to the lack of blood-breakdown products or proton–electron dipole interactions.

In the acute phase there is low signal within the hematoma on T2-weighting, due on the micro-scopic level to the inhomogeneous distribution of deoxyhemoglobin within red cells that have

82

Fig. 4.9 Initial presentation of an acute hemorrhage on different pulse weightings. This acute hemorrhage into a region of infarction is several hours old and is shown in (a) with T1-weighting, in (b) with T2-weighting, in (c) on a FLAIR image, and in (d) on a gradient recalled image. Note that the acute bleed is low signal on T1 and very dark on the remaining pulse sequences. This hemorrhagic infarct in the left cortex is the same patient as in Fig. 4.1 images (c–e)


not yet lysed. This, in fact, overrides the propensity of the fluid collection by itself to be bright on T2-weighting.

Early in the subacute phase, the signal intensity remains low on T2-weighting, again, primarily due to the inhomogeneous distribution of the paramagnetic deoxyhemoglobin modules. However, as this phase progresses, red cell lysis occurs, as does the production of an increasing amount of methemoglobin. This results in extracellular methemoglobin, which is initially concentrated and then becomes more dilute with time. As a consequence, on T2-weighted images the signal is mildly decreased when the methemoglobin is concentrated and the T2 effects are dominant; and the signal is increased when the methemoglobin is dilute and the T1 effects are predominant. (In essence, these T1 effects override the shortened T2 in the hematoma.) The net effect is that, by the end of the subacute period, the signal intensity within the hematoma has begun to increase and become significantly hyperintense. Note that this increase in signal intensity from iso- or hypointense to

83

Fig. 4.10 MR examination of a late acute, early subacute hemorrhage 3–4 days old. Right occipital bleed is shown on T1-weighting in (a), T2-weighting in (b), FLAIR in (c), and gradient recalled imaging on (d). Note that the hem-orrhage is bright on T1, but remains dark on T2, FLAIR, and GRE. In (e), a CT examination on the same day as the MR shows that the hemorrhage is old enough to have a rim of edema


hyperintense occurs at the end of the subacute period on T2-weighting, which occurs later than the same change on T1-weighting, which occurs at the beginning or middle of the subacute period. Therefore, the change to hyperintensity appears earlier by several days on T1-weighting than on T2-weighting.

In the chronic stage, the signal intensity remains bright on T2-weighting. In an “old” hematoma the signal remains bright in the area due to the high water content, even if there has been complete removal of all of the blood-breakdown products.

Changes that occur around the hematoma in the adjacent brain are more prominent on T2-weighted images than on T1. As the hemosiderin accumulates with time around the periphery of the hema-toma, a black, thin rim of decreased signal becomes visible. This is usually seen by the time of the late subacute stage and, certainly, by the chronic stage. Microscopically, this is due to the presence of hemosiderin and/or ferritin. Edema, which follows the same time course as on T1-weighting, is also present in the adjacent brain, but has the opposite signal characteristics. That is to say, the edema shows up as bright on T2 signal, but as the edema wanes, this increased signal usually returns to normal by 1–2 weeks.

The accompanying figures demonstrate changes of the hematoma center and periphery, the adja-cent brain, and the hematoma rim on T1- and T2-weighting with time (Figs. 4.8–4.11).

84

Fig. 4.11 Chronic hemorrhage on MR weighting. This hemorrhage is definitely 20 days out from the initial bleed that is documented by CT. The hemorrhage, on sagittal T1-weighting in (a), axial T2-weighting in (b), axial FLAIR imaging in (c), and GRE in (d), shows at this chronic stage that the center of the hemorrhage is white on all pulse sequences. A rim of very black signal is most apparent on the T2 and the gradient recall images. Increased signal persists around the hemorrhage on the T2, FLAIR, and GRE images. In a very old hematoma, the central portion will often eventually become dark again on T1-weighting


FLAIR Changes Over Time Within a Hemorrhage on MR Scanning

FLAIR imaging is an MR technique which has been available in most centers since the mid-1990s. This technique uses special pulses to attenuate the normal bright signal that arises from free fluid on a T2-weighted sequence. Consequently, lesions (which usually also tend to be bright on routine T2-weighted images) that are not pure fluid stand out better against this fluid-attenuated background.

Numerous papers cite the beneficial and increased imaging capability of MR over CT to find small amounts of hemorrhage, both in the subarachnoid space and within the ventricles [27, 28, 29]. However, all of these studies also note that the FLAIR sequence has limitations, namely that signifi-cant pulsation artifacts appear in the third and fourth ventricles, which may be problematic for visualizing blood in those ventricles as opposed to the lateral ventricles. Second, it has been shown that, in the case of subarachnoid hemorrhage, MR is only slightly more sensitive than CT for the detection of small amounts of blood. However, in some situations, both the CT and the FLAIR MR sequences are negative, yet a lumbar puncture will be positive for blood within the cerebral spinal

85Gradient Echo (Magnetic Susceptibility-Weighted) Changes Over Time

fluid. Having presented these shortcomings, let us describe the typical time course of hemorrhage, both into the ventricles and into the subarachnoid space.

In the case of both of these hemorrhages, it is well documented that hyperacute and acute hemor-rhages are almost uniformly bright on FLAIR within the first 48 h. This occurs whether the hemor-rhage is in the subarachnoid space overlying the convexities and extending down into the sulci, or is within the lateral ventricles. At and after 48 h, the appearance is much more inconsistent. In the case of the ventricular hemorrhage, after 48 h, due to varying rates of breakdown, the hemorrhage may still continue to be hyperintense, but could also become isointense or hypointense. Especially in the case of hypointense hemorrhages in the ventricles, it becomes almost impossible to see on FLAIR because it matches the signal intensity of the CSF.

Similarly, blood in the subarachnoid space is usually well seen in the first 48 h (hyperacute and early acute phases). However, as this blood ages and goes into the subacute and chronic stages, it becomes iso- and then hypointense and is lost against the background of the normal sulci and the normal black fluid within the CSF space. Most subarachnoid hemorrhages are not acutely visible for as long a period of time as the typical intraventricular hemorrhage is seen on FLAIR imaging (Figs. 4.9–4.11).

Gradient Echo (Magnetic Susceptibility-Weighted) Changes Over Time Within a Hemorrhage on MR

GE imaging is performed differently from routine T1, T2, DWI, FLAIR, or MRA sequences. The nature of the MR appearance of hemorrhage on GE (T2*) images is also significantly different than all of the aforementioned sequences. While T1, T2, and DWI images of hemorrhage change sequentially and dynamically over the initial 2 weeks, after a hemorrhage GE-T2* images are initially normal for a “short time.” The duration of this short period of normal appearance is up to some debate, but, at most, it is several hours. By the end of the hyperacute phase (6 h after the bleed), the area of hemorrhage definitely appears dark on this sequence. It continues to be very dark or black – often the remainder of the patient’s life – through the acute, subacute, and chronic phases [30, 31, 32].

Therefore, from a dating perspective, hemorrhage on GE looks one way (normal) for at most sev-eral hours. After that short initial period, all or portions of the region of hemorrhage remain dark for a very long time. Consequently, GE is relatively poor as an aid in dating, with one exception. In a patient who is known by multiple serial exams to have had a hemorrhage at a specific site, if a GE scan appears fairly normal at some previous time and subsequent GE scans convert to abnormal dark signal, then the hemorrhage can be dated to, at most, the 3–6 h prior to the “normal” appearing scan.

Despite the inability of this sequence to be routinely useful in dating time of occurrence of a hemor-rhage, it does have a very important attribute; it is very sensitive to the presence of calcium and blood-breakdown products. In fact, it is more sensitive to small amounts of bleeding than any other routinely utilized pulse sequence. In the clinical setting of hemorrhagic contusions or diffuse axonal injury (characterized by many posttraumatic hemorrhages distributed widely throughout the brain), it will often detect hemorrhages that are missed by CT or other MR pulse sequences (Figs. 4.9–4.11).

Changes of Edema and Hemorrhage Over Time on US

Little needs to be added to the comments on the appearance of edema and hemorrhage on US that were given in Chap. 2. However, some additional discussion and summary of imaging findings are appropriate at this point.

Fig. 4.12 Large, acute hemorrhage on ultrasound (US) into the left frontal lobe seen on day zero in (a), day 5 in (b), and day 16 in (c). Note that over the course of time, the hemorrhage becomes contracted within a larger cystic cavity

Fig. 4.13 US examination of periventricular infarction. In (a), initial scan shows periventricular increased echogenicity on the left side on the first day. In (b), on day 6, the area of abnormal echogenicity lateral to the anterior horn is more defined. In (c), on day 21, the infarction area has undergone cystic change (cavitation)

87References

Hemorrhage presents initially as a region of high echogenicity which usually does not have shadowing. This hemorrhage will slowly decrease in size over time. Associated edema surrounding the bleed will grow maximally over several days and then decrease to normal by 1 1/2 to 2 weeks. The center of the lesion often becomes relatively hypoechogenic before the periphery of the lesion. Other associated findings in the setting of hemorrhage include fluid-fluid levels in the ventricles, particulate material within the ventricles, and the delayed development of hydrocephalus (Fig. 4.12).

Early edema may not be visible on ultrasonography. Whether the earliest appearance of swollen or injured brain is best seen on very carefully performed US evaluation or on MR scanning with diffusion weighting is controversial. Regardless, at some early time after the initial injury, edema may be visible on ultrasonography as a region of increased echogenicity which acutely has poorly defined margins. Shortly thereafter, the margins of the affected region become more sharply defined. With time, this region of abnormal echogenicity will resolve in one of two ways – either the area of injury will completely resolve, leaving no residual whatsoever, or the edematous area will go on to complete infarction and an area of encephalomalacia will result (Fig. 4.13).

References

1. Kucinski T. Unenhanced CT and acute stroke physiology. Neuroimaging Clinics of North America. 2005; 15(2):397–407.

2. Grossman RI, Yousem DM. Neuroradiology: The Requisites, 2nd edition, Mosby, Philadelphia, 2003 184. 3. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Acute Cerebral Ischemia-Infarction I:4:108–111. 4. Beauchamp NJ, Barker PB, Wang PY, van Zijl PCM. Imaging of acute cerebral ischemia. Radiology. 1999;

212:307–324. 5. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Subacute Cerebral Infarction I:4:112–115. 6. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Chronic Cerebral Infarction I:4:116–119. 7. Baird AE, Warach S. Magnetic resonance imaging of acute stroke. Journal of Cerebral Blood Flow and Metabolism.

1998;18:583–609. 8. Gonzalez RG, Schaefer PW, Buonanno FS, et al. Diffusion-weighted MR imaging: diagnostic accuracy in

patients imaged within 6 hours of stroke symptom onset. Radiology. 1999;210:155–162. 9. Provenzale JM, Sorensen AG. Diffusion-weighted MR imaging in acute stroke: theoretic considerations and

clinical applications. AJR. 1999;173:1459–1467. 10. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Stroke Overview I:4:2–5. 11. Ricci PE, Burdett JH, Elster AD, Reboussin DM. A comparison of fast spin-echo, fluid-attenuated inversion-recovery,

and diffusion-weighted MR imaging in the first 10 days after cerebral infarction. AJNR. 1999;20:1535–1542. 12. Elster AD, Moody DM. Early cerebral infarction: gadopentetate dimeglumine enhancement. Radiology. 1990;

177:627–632. 13. Kidwell CS, Wintermark M. Imaging of intracranial hemorrhage. Lancet Neuroradiology. 2008;7:256–267. 14. Huisman TAGM. Intracranial hemorrhage: ultrasound, CT and MRI findings. European Radiology. 2005;

15:434–440. 15. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Epidural Hematoma I:2:10–14. 16. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Subdural Hematomas I:2:14–27. 17. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Traumatic Subarachnoid Hemorrhage I:2:28–31. 18. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Cerebral Contusion I:2:32–35. 19. Liu AY, Maldjian JA, Bagley LJ, Sinson GP, Grossman RI. Traumatic brain injury: diffusion-weighted MR imag-

ing findings. AJNR. 1999;20:1636–1641. 20. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd edition, Lippincott Williams & Wilkins,

Philadelphia, 2009, Diffuse Axonal Injury I:2:36–39.

88

21. Parizel PM, Makkat S, Van Miert E, Van Goethem JW, van den Hauwe L, De Schepper AM. Intracranial hemor-rhage: principles of CT and MRI interpretation. European Radiology. 2001;11:1770–1783.

22. Atlas SW, Thulborn KR. MR detection of hyperacute parenchymal hemorrhage of the brain. AJNR. 1998; 19:1471–1507.

23. Clark RA, Watanabe AT, Bradley WG, Roberts JD. Acute hematomas: effects of deoxygenation, hematocrit, and fibrin-clot formation and retraction on T2 shortening. Radiology. 1990;175:201–206.

24. Gimori JM, Grossman RI, Hackney DB, Goldberg HI, Zimmerman RA, Bilaniuk LT. Variable appearances of subacute intracranial hematomas on high-field spin-echo MR. AJR. 1988;150:171–178.

25. Gimori JM, Grossman RI. Mechanisms responsible for the MR appearance and evolution of intracranial hemor-rhage. Radiographics. 1988;8:427–440.

26. Grossman RI, Yousem DM. Neuroradiology: The Requisites, 2nd edition, Mosby, Philadelphia, 2003, 208–215. 27. Noguchi K, Ogawa T, et al. Subacute and chronic subarachnoid hemorrhage: diagnosis with fluid-attenuated

inversion-recovery MR imaging. Radiology. 1997;203:257–262. 28. Noguchi K, Seto H, Kamisaki Y, Tmizawa G, Toyoshima S, Watanabe N. Comparison of fluid-attenuated

inversion-recovery MR imaging with CT in a simulated model of acute subarachnoid hemorrhage. AJNR. 2000; 21:923–927.

29. Bakshi R, Kamran S, et al. Fluid-attenuated inversion-recovery MR imaging in acute and subacute cerebral intra-ventricular hemorrhage. AJNR. 1999;20:629–636.

30. Ripoll MA, Stenborg A, Sonninen P, Terent A, Raininko R. Detection and appearance of intraparenchymal hae-matomas of the brain at 1.5T with spin-echo, FLAIR and GE sequences; poor relationship to the age of the haematoma. Neuroradiology. 2004;46:435–443.

31. Lin DD, Filippi CG, Steever AB, Immerman RD. Detection of intracranial hemorrhage: comparison between gradient-echo images and b(0) images obtained from diffusion-weighted echo-planar sequences. AJNR. 2001; 22:1275–1281.

32. Liang L, Korogi Y, Sugahara T, Shigematsu Y, Okuda T, Ikushima I, Takahashi M. Detection of intracranial hemorrhage with susceptibility-weighted MR sequences. AJNR. 1999;20:1527–1534.


J.L. Creasy, Dating Neurological Injury: A Forensic Guide for Radiologists, Other Expert Medical Witnesses, and Attorneys, DOI 10.1007/978-1-60761-250-6_5, © Springer Science+Business Media, LLC 2011

Abstract This chapter focuses on patterns of change within the brain parenchyma that occur as the result of a previous injury – but not one occurring in the immediately previous 2 weeks. These chronic changes persist for the life of the patient.

Keywords Magnetic resonance (MR) • Computed tomography (CT) • Ultrasound (US) • Infarction, Hemorrhage • Watershed regions • Duration of injury

Introduction

While the first major method of dating an injury is by computed tomography (CT) and/or magnetic resonance (MR) scans that ideally occur within the first 1–2 weeks after a neurological event, there is a second very valuable method of dating injury as well. This method relates to the existence of patterns of neurologic injury that are so characteristic in their appearance that the likely cause of an injury, and in some cases the probable timing of the injury can be identified. This method is not as powerful, or as accurate as the dynamic dating method but it still is useful in a setting where the timing of events is important. This chapter will discuss the appearance of several of the more common classic patterns – and where possible, the usual age groups of occurrence, the usual causative agents, and any information that can be deduced about timing. The majority of these more chronic patterns are most useful in the setting of edema and infarction, however, the principles also apply to older (i.e., late sub-acute and chronic) hemorrhages. The appropriate application of this information in the pattern dating setting will be discussed by using three illustrative examples in Chap. 7.

As mentioned, several patterns are age specific, that is to say that the patterns that are “classic” for in utero injury are different from those that occur in the neonatal period (around the time of birth), which are different from those occurring in the pediatric period, which are different from adult patterns. Whereas in all cases, regardless of the age of the patient, normal scans are significant for side-to-side symmetry, the presence of all normal structures, and normal density (on CT) or signal (on MR) in every tissue – in the pathologic patterns we will discuss, the abnormalities are either of structure, or of density (on CT) or signal (on MR), or of both.

Beginning Principles

In describing the patterns of parenchymal injury due to ischemia and infarction, one must pay atten-tion to the normal sequence of events that occurs with ischemic brain. Initially, as a portion of brain becomes ischemic, one of three things can happen. In the first situation, the area of ischemic brain

Chapter 5Patterns of Parenchymal Injury: Pattern (Chronic) Dating

89

5 Patterns of Parenchymal Injury: Pattern (Chronic) Dating

completely recovers with no imaging sequelae. In the second instance, the area of ischemic brain goes on to have subtotal or partial or selective loss of substance. This can be thought of as a situation in which instead of 100% cell death within the affected area, a subpopulation of the cells within the affected area die [1]. Depending upon the severity of the injury and the percentage of population that is injured vs. the percentage that is spared, the imaging appearance can vary from normal to minimally abnormal (e.g., increased FLAIR or T2 signal), to severely abnormal with cystic change and partial parenchymal loss [2] on MR examinations. In the third case, all of the ischemic tissue goes on to infarction – complete death of all of the cells within the affected region. On imaging this outcome is detectable as a well-demarcated area of deficient brain. The age of the patient affects the appearance of the brain after injury in the second and the third cases, as it is more common to see some minimal residual structures (strands of tissue, multiple cysts, and septae) within the infarcted neonatal brain than in the older brain [3].

On the concept of ischemia and infarction, a region of ischemia most commonly evolves to a region of infarction beginning centrally. The periphery of the region of ischemia may then experience complete recovery, partial injury, or complete infarction. It will not remain in a state of ischemia for a protracted period of time.

Factors Affecting the Outcome of a Region of Ischemia

Why do some regions of the brain that undergo ischemia go on to complete recovery? Why do some areas of brain appear to be more sensitive to ischemia than other areas? Or stated another way; in general, what factors determine the differences between the relative sensitivities and eventual out-comes of different regions of the brain? The answer to this question can be broken down into four different collections of factors, which all interact and contribute to the eventual final outcome of an insult to the brain. These factors are as follows:

1. Different perfusions to different regions of the brain.2. Different severity of injury-partial vs. complete.3. Different durations of injury-short vs. prolonged.4. Different metabolic rates of the at risk tissue.

Regional Variations in Perfusion

In considering different perfusions, the place to begin is with any variations from normal vascular anatomy. An example of a preexisting anatomic variation would be a patient who does not have the complete circle of Willis as is discussed in Chap. 1. This decreases the patient’s ability to shunt blood into a vessel that is occluded more proximally because there is a lack of distal collateral con-nections from normal vessels to the abnormal vessels beyond their point of narrowing or occlusion. A second cause of different perfusion for different regions of the brain is superimposed acquired vascular pathology on top of the patient’s preexisting anatomy. An example is atherosclerotic disease that has significantly narrowed a vessel or vessels or branches of several vessels, decreasing the ability of those vessels to perfuse the distal brain.

It is also important to discuss the concept of watershed territories. While the basic brain perfusion territories were discussed in Chap. 1, the junction of these different territories creates “ watershed regions.” Within these watershed regions, which can be thought of being at the end of the vascular territory on both sides, the brain tissues at that spot are relatively the first impacted by a global decrease in perfusion. The anatomic location of the watershed territories changes as the patient ages. Thus, the “water-shed territories” in a neonatal or young child are different from the watershed territories in an adult.

90

Factors Affecting the Outcome of a Region of Ischemia

In the child, the watershed region occurs at the interface in the hemispheric white matter between the vessels that supply the region immediately around the ventricles, and the second set of vessels that supply the more superficial white matter more immediately below the cortical gray matter [4, 5, 6]. As a result, watershed infarcts in neonates tend to occur high in the parasagittal region and deep in the white matter around the ventricles. In the adult brain, the watershed regions are at the boundary between the middle and the anterior cerebral artery territories, and between the middle and posterior artery territories. Adult watershed infarcts have a characteristic “string-of-pearls” appearance occurring in an anterior–posterior line in the white matter of the centrum semiovale. They also occur anteriorly in the lateral frontal area between the ACA and MCA territories, or posteriorly in the lateral occipital region between the MCA and PCA territories.

Variations in Severity of Insult

Due to the different survivability of different tissues, which is secondary to the different metabolic needs of different tissues (see below), different levels of severity of insult produce different amounts of injury. While the effects of anoxia are dependent both on its duration (see below) and the age of the patient; they also depend on the severity of the insult. For example, it is not surprising that at any given age, the outcome for the same length of time of complete anoxia is different from the outcome from the same duration of partial anoxia. Expressed in another way, two identical tissues, with equivalent duration of injury, will be unequally affected if one insult is complete, and one insult is partial.

These processes are especially important in the neonate, where severity of injury produces a dif-ferent outcome in both the neonate and the infant. An example of the difference in severity of out-come occurs in both of the neonate and in the older child. In the term neonate, mild to moderate hypertension produces classically a parasagittal watershed injury. In contrast, severe hypotension produces injuries that are more extensive and involve the brainstem, cerebellum, thalamus, basal ganglia, and the region of the cortex around the central sulcus (the perirolandic cortex) [7]. Similarly, in the older child mild to moderate hypotension also produces a parasagittal watershed injury; but profound hypotension produces basal ganglia and diffuse cortical injury [8, 9].

Variations in Duration of Injury

Similar to the prior concept, two identical tissues with identical severity of insult will have different degrees of injury if the duration of the insult is different. Past a certain threshold, for the same degree of decrease of oxygenation or perfusion, the longer the duration of this deficit, the more severe the injury (presuming the tissues being compared have similar metabolic needs and sensitivities). Tissues which have the same metabolic sensitivity will be affected differently by different lengths of time of insult – the longer the insult, the more damage will occur. For example, in the setting of complete anoxia, no effect is seen in neonates until the duration of complete anoxia is 10–15 min, while in the adult 4–6 min of complete anoxia produces irreversible brain damage. In both age groups, there is a sudden abrupt change in outcome when that time threshold is crossed [10].

On the other hand, with partial anoxia in the neonate there is little injury for an insult of short duration, but partial anoxia of long duration produces chronic changes both within the brain and central nervous system and within the rest of the body [11]. In partial anoxia of a mild or moderate degree, depending on the length of the insult, the body can attempt to increase cerebral flow by shunting blood to the brain from the remainder of the body. If this is done for a short period of time, no deficits may be found in the brain or elsewhere in the body. Conversely, prolonged periods of partial anoxia, especially in the neonate, can produce damage in both regions.

91


Variations in Metabolic Activity

This final factor is in reality the common pathway that determines whether the variations in perfusion, degree of insult, or duration of insult produce an injury to the brain. Each portion of the brain has unique metabolic needs, metabolic rates, and sensitivities to injury – and these attributes change as the patient ages [12]. Whether a particular tissue is injured or not depends on whether the insult from whatever cause, exceeds the tissues’ ability to survive the insult. The reason why one portion of the brain is affected, and a second different region of the brain is spared depends on the differing metabolic needs and toxic sensitivities of the two tissues. A clinical example is the setting in which a global decrease in oxygenation or perfusion selectively targets specific portions of the brain, such as the basal ganglia, which are very metabolically active [8, 9].

Patterns of Parenchymal Injury

By training, an interpreter of radiologic images approaches clinical situations from the mindset that a large amount of information is present on the images. Frequently, little or no clinical information is provided. In a medical-legal situation, we believe it is always appropriate to first review the images, with the aid of as little, or no clinical information as possible before attempting to correlate and syn-chronize the radiographic findings and their implications with the clinical picture. If one approaches the images in this way it becomes very important to look for specific morphological patterns, as the identification of a pattern then has direct implications on causation, and often on timing.

Therefore, in this section we will classify conditions based on the morphology and site of the injury. First we will consider regional, total injuries; then global anoxia and its common aftermath global atrophy. However, global anoxia can also produce focal geographic infarcts, and targeted injury to specific brain tissues such as the basal ganglia, periventricular white matter, regions around the central sulcus, and watershed regions. We will discuss these varying patterns below.

Patterns Where There Has Been Total Loss of a Focal Region of Brain Parenchyma

Patterns in which total parenchymal loss of a geographic portion of the brain occurs tend to be those patterns in which there has been an occlusion of arterial inflow, venous outflow, or a general-ized decrease in perfusion which tends to affect the watershed regions. In each of the situations, the area of affected brain tends to have the classic core of complete parenchymal loss, with a surrounding area of ischemia which often goes on to a subtotal or partial neuronal loss.

Focal or Geographic Infarct

Perhaps the most straightforward injury to discuss is the injury that results in focal infarction of brain parenchyma. This is most commonly due to occlusion of the feeding artery or its branches (for example a branch of the middle cerebral artery as opposed to the entire middle cerebral artery). If the infarction involves an arterial territory or part of an arterial territory then an arterial occlusion is most likely; whereas, if the area of infarction is geographic but crosses vascular territory boundaries, then either one must presuppose a lesion involving two feeding arteries, or a draining venous struc-ture in which a drainage catchment area crosses vascular territories.

92


Focal or geographic infarcts can occur in all ages (in utero, neonatal, pediatric, and adult). The appearance of the edema from the infarction follows that as discussed in Chap. 2, with the time course as described in Chap. 4. Chronic geographic infarcts look very similar regardless of the age of the patient. In the chronic setting geographic infarcts on CT appear very low in density (identical to CSF) with sharp margins, no enhancement, and no edema or mass effect. Surrounding the infarct the immediately adjacent surviving but injured brain will also be decreased in density (but not as low as that of CSF). On MR a chronic geographic infarction may have bright signal on a DWI scan (but the ADC confirms there is no real restricted diffusion), and no edema or mass effect will be present. The margins of the region of total infarction will be sharply demarcated, and the region of the infarct will be filled with fluid that is identical in appearance to CSF on all pulse sequences. The adjacent, injured but surviving brain will continue to have abnormal T2 and FLAIR signal for the life of the patient. Several examples of chronic focal or geographic infarcts on CT and MR are shown in Fig. 5.1.

93

Fig. 5.1 Old infarctions on magnetic resonance (MR) and computed tomography (CT). (a) CT scan of old left puta-men infarction. (b) CT of old left parietal infarction. (c, d) MR scans of an old left frontal infarction on FLAIR in (c), and a more inferiorly located FLAIR image in the same patient demonstrating atrophy of the left cerebral peduncle from this old event in (d)


Injuries That Affect the Brain Diffusely

Let us now explore the situations in which there is an anoxic insult to the entire brain. We will first consider global anoxia – which is presented here under patterns – even though it is an acute process. It is appropriate to place it here, however, as it is often the antecedent condition to global atrophy. It can also lead to regional, geographic insults; or to selective injury to specific brain tissues [8, 9].

Global Anoxia

Global anoxia is an event due to decreased oxygen to the brain, seen in all ages of patients. Global anoxia has a similar appearance in the neonatal, pediatric or adult patient. Any etiology that causes either a global decrease in the brain perfusion and/or a global decrease in the amount of oxygen reaching the brain can cause global anoxia. For example, normally oxygenated blood may be pres-ent, but due to a blockage at the level of the heart, or just simple bradycardia, not enough blood gets to the brain. Alternatively the brain may be perfused by a normal amount of blood, but due to some interference with normal respiration as in asphyxia or drowning, there is no oxygen present within the blood. Either one produces the same clinical appearance of global anoxia.

In this setting, the entire brain shows abnormalities on both CT and MR scanning [13]. On CT, one sees diffuse loss of density and an appearance of uniformity to all of the brain substance with no discernable difference in gray and white matter. The pediatric patient may have some preservation of normal density involving the basal ganglia and the posterior fossa which is referred to as “the reversal sign” [14]. On MR scanning, during the acute phase, large areas of abnormal signal are present on the diffusion weighted scan. This may involve such large areas that the entire scan is bright, and no regional involvement is evident, or there may be large confluent patchy areas of cortical and white matter signal abnormality. After the very early stages, in addition to the diffusion weighted abnormal-ity, abnormal signal is also evident on FLAIR and T2-weighting (Figs. 5.2 and 5.3).

Global Atrophy

Global atrophy is a delayed finding which occurs in the brain at some period after a global insult. One way to think of global atrophy is as an injury that has diffusely damaged and/or killed a subset of the

94

Fig. 5.2 CT of global anoxia. CT scans of three different patients each demonstrating large areas of low density, within which there is no discernable difference in density between gray and white matter. In (a), scan obtained at mid third ventricle; in (b) axially above the ventricles; and in (c) at level of thalami


neurons and glial cells in the brain but has not produced a focal geographic region in which there is total cell death. On a microscopic level, a view of the involved brain would show a patchy loss of neurons and glial cells, possibly with scarring but with a preservation of some percentage of the nor-mal cell architecture. However, this degree of microscopic cellular specificity is not achievable with imaging modalities which only look at gross macroscopic collections of cells. Therefore, the imaging correlation of a microscopic patchy loss of cell parenchyma is a diffuse decreased amount of normal brain parenchyma that often will have signal abnormalities on FLAIR or T2-weighted MR scanning (Fig. 5.4).

Another imaging finding that is indicative of global parenchymal loss is an otherwise normal appearing, but age inappropriate amount of brain. Presuming that initially the brain was normal in appearance in all respects, if a global anoxic event occurs, a subsequent exam months or years later may show a markedly decreased amount of brain parenchyma which is now no lon-ger age appropriate. This can all occur in the setting of the absence of any well-demarcated geometric or geographic focus in which there has been complete loss of brain parenchyma, and in the absence of any significant abnormal density (CT) or signal (MR) in the remaining brain.

The finding of global atrophy is a chronic finding, and consequently its presence is a poor predic-tor of the time at which the injury to the brain occurred. What can be said is only that the incident occurred at least several weeks, but possibly months or years prior to the time of the scan that shows the actual atrophy.

Common Patterns That Show Targeting, with Partial or Total Cell Loss

The major categories of neonatal hypoxic-ischemic encephalopathy (HIE) [14, 15] are described by Volpe [16]. He breaks the injuries into four major groups: focal and multifocal ischemic brain necrosis, periventricular leukomalacia, selective neuronal necrosis, and parasagittal injury. We have discussed the first group, focal and multifocal ischemic brain necrosis, in the preceding sections on geographic infarcts and global anoxia. Periventricular leukomalacia, which is a condition that occurs in children, but its effects can be seen in later scans in adulthood, will be discussed below. Two of the more commonly seen examples of selective neuronal necrosis – those cases with involve-ment of the basal ganglia and the perirolandic region – will then be covered. Parasagittal injury,

95

Fig. 5.3 MR of global anoxia. Global anoxic insult has produced bilateral basal ganglia and focal right cortical signal abnormalities in (a–c)


which is a watershed injury both in children and in adults, will be covered last. Except for PVL, all of the patterns described do occur in both children and adults.

PVL/White Matter Loss

Classic periventricular leukomalacia involves white matter near the lateral ventricles. The term PVL does not describe a specific disease, but rather the response and appearance of the brain after a large number of possible etiologic agents [17, 18, 19]. It is seen almost exclusively in infants in the neonatal period. PVL is usually first imaged on ultrasound (US). It first is seen as increased echogenicity around the later ventricles, which often progresses to cyst formation. The cysts may be small at first, but later coalesce and eventually merge with the ventricle, producing an enlarged ventricle with decreased adjacent white matter between the ventricular surface and the underside of cortex. MR scanning demonstrates the cystic areas and loss of white matter, especially posteri-orly. We find the best way to assess for possible PVL is to use an image acquired in the coronal plane, obtained posteriorly through the rear of the lateral ventricles. This permits one to very care-fully assess the thickness of normal white matter between the surface of the ventricle and the underside of the cortex (Fig. 5.5).

96

Fig. 5.4 MR and CT of global atrophy. Initial CT in (a) with loss of gray-white differentiation diffusely. In (b), CT 1 day later increasing low density is present in the left frontal region and posteriorly. In (c), DWI MR at 6 days postevent, restricted diffusion is present posteriorly and in the left insular cortex. In (d), axial T2-weighted MR at the same time as “C,” a normal amount of brain tissue is present. In (e), T2-weighted MR exam 10 weeks after “C” there is marked diffuse loss of brain tissue, most prominently in the frontal regions and the right occipital region


The next two discussions of specific patterns fall within the broader heading of what Volpe refers to as diffuse neuronal injury. While selective injury can occur to a number of structures including the cortex and perirolandic region, the brainstem, the basal ganglia, the cerebellum, and the spinal cord; we will emphasize two structures which are commonly involved and are easily recognized on imaging studies. These structures are very frequently injured in conjunction with other regions of the brain.

Basal Ganglia

One sees injuries to the basal ganglia in both premature and newborn children as one component of usually much more extensive injury, in patients with an episode of anoxia. Injury to the basal gan-glia with simultaneous injury to the thalamus is typical of neonatal HIE [20, 21]. The basal ganglia involvement can be selective (i.e., on imaging only the basal ganglia appear involved) or part of a more diffuse involvement of brain, in which they are the most severely affected tissues. The accom-panying Fig. 5.6 illustrates both patterns of involvement.

97

Fig. 5.5 Periventricular white matter loss in two patients. In the first patient, extensive loss of white matter is seen bilaterally posteriorly on Axial T1 MR in (a) and coronal T1 MR in (b). In (c) (same patient) axial FLAIR shows abnormal signal in the remaining white matter in the occipital region. In the second young patient, ultrasound (US) examination initially shows increased echogenicity around the frontal horns of the lateral ventricles in (d). In (e), US obtained 2½ weeks later, there has been extensive cystic change in the involved periventricular white matter


Perirolandic Region

The second example of selective neuronal injury is the cerebral cortex, in the region around the central sulcus [20]. In pediatric patients with this component of diffuse neuronal injury, one can see evidence of restricted diffusion in the tissues about the central sulcus on diffusion weighted imaging, with FLAIR and T2 signal abnormalities. In the chronic situation, parenchymal injury is evident by cyst formation and by persistent MR signal abnormality in this region. Perirolandic involvement is a good indicator of anoxia in the neonatal period (Fig. 5.7).

98

Fig. 5.6 CT and MR of severe anoxic injury with more severe involvement of the basal ganglia. Recent anoxic insult producing restricted diffusion (bright regions) in the basal ganglia on DWI MR in (a), and low density on CT in the same regions in (b). Second patient, a 7 month old child in (c) and (d) shows large homogenous low density regions, especially posteriorly, with superimposed even lower density of the basal ganglia bilaterally


Parasagittal Injury

Parasagittal injury in the neonate is the expression of a watershed lesion in that age group [22]. The same string-of-pearls appearance in the same location is also seen in watershed events in the adult patient. In the child, this parasagittal region represents the junction between the deep and superficial vessels supplying the deep white matter. However, in the adult this region is the junction between the distal portions of the ACA territory medially, and the upper portions of the MCA territory more laterally (Fig. 5.8).

99

Fig. 5.7 MR of injury to the central sulcus region. (a) FLAIR image of bilateral cysts immediately adjacent to the central sulcus. (b) T2 image of extensive signal abnormality in the same region

Fig. 5.8 MR of parasagittal or watershed infarction in the adult. (a, b) Axial DWI images with characteristic “string-of-pearls” appearance in the left centrum semiovale


References

1. Volpe JJ. Neurology of the Newborn. Saunders, Philadelphia, 2001, page 299. 2. Frigieri G, Guildi B, Costa Zaccarelli S, Rossi C, Muratori G, Ferrari F, Cavazzuti GB. Multicystic

Encephalomalacia in Term Infants. Child’s Nervous System. 1996;12:759–764. 3. Barkovich AJ. Pediatric Neuroimaging, 4th Ed. Lippincott Williams & Wilkins, Philadelphia, 2005, page 191. 4. Barkovich AJ. Pediatric Neuroimaging, 4th Ed. Lippincott Williams & Wilkins, Philadelphia, 2005, page 206. 5. Volpe JJ. Neurology of the Newborn. Saunders, Philadelphia, 2001, pages 303–306. 6. Grossman RI, Yousem DM. Neuroradiology: The Requisites, 2nd Ed. Mosby, Philadelphia, 2003, page 176. 7. Barkovich AJ. Pediatric Neuroimaging, 4th Ed. Lippincott Williams & Wilkins, Philadelphia, 2005, page 205. 8. Pasternak JF, Predey TA, Mikhael MA. Neonatal Asphyxia: Vulnerability of Basal Ganglia, Thalamus, and

Brainstem. Pediatric Neurology. 1991;7:147–149. 9. Pasternak JF, Gorey MT. The Syndrome of Acute Near-Total Intrauterine Asphyxia in the Term Infant. Pediatric

Neurology. 1998;18:391–398. 10. Barkovich AJ. Pediatric Neuroimaging, 4th Ed. Lippincott Williams & Wilkins, Philadelphia, 2005, page 205. 11. Barkovich AJ. Pediatric Neuroimaging, 4th Ed. Lippincott Williams & Wilkins, Philadelphia, 2005, page 207. 12. Johnston MV, Hoon AH. Possible Mechanisms in Infants for Selective Basal Ganglia Damage from Asphyxia,

Kernicterus or Mitochondrial Encephalopathies. Journal of Child Neurology. 2000;15:588–591. 13. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd Ed. Lippincott Williams & Wilkins,

Philadelphia, 2009. Term and Adult Hypoxic Ischemic Injury, pages I:4:90–97. 14. Han BK, Towbin RB, De Courten-Myers G, McLAurin RL, Ball WS. Reversal Sign on CT: Effect of Anoxic/

Ischemic Cerebral Injury on Children. American Journal of Neuroradiology. 1990;10:1191–1198. 15. Barkovich AJ. Pediatric Neuroimaging, 4th Ed. Lippincott Williams & Wilkins, Philadelphia, 2005, pages

195–244. 16. Volpe JJ. Neurology of the Newborn. Saunders, Philadelphia, 2001, page 297. 17. Volpe JJ. Neurology of the Newborn. Saunders, Philadelphia, 2001, page 307–315. 18. Barkovich AJ. Pediatric Neuroimaging, 4th Ed. Lippincott Williams & Wilkins, Philadelphia, 2005, pages

209–220. 19. Osborn AG, Salzman L, Barkovich AJ. Diagnostic Imaging: Brain, 2nd Ed. Lippincott Williams & Wilkins,

Philadelphia, 2009. White Matter Injury of Prematurity, pages I:4:86–89. 20. Volpe JJ. Neurology of the Newborn. Saunders, Philadelphia, 2001, page 299. 21. Garg BP, DeMyer E. Ischemic Thalamic Infarction in Children: Clinical Presentation, Etiology and Outcome.

Pediatric Neurology. 1995;13:46–49. 22. Volpe JJ. Neurology of the Newborn. Saunders, Philadelphia, 2001, pages 303–306.

100

Part IIApplication to the Medical-Legal Setting


Abstract This chapter utilizes the information in the first four chapters to focus on dating events that have occurred in the previous 2 weeks. In the first 2 weeks after an infarction or hemorrhage, the imaging appearance of the involved brain undergoes dynamic changes that allow a more delim-ited estimation of the probable time of occurrence.

Keywords Magnetic resonance (MR) • Computed tomography (CT) • Ultrasound (US) • Hematoma • Infarction

Introduction

This is the first chapter in the second part of the book, in which the principles of imaging findings presented in the first five chapters will now be applied specifically in the medical legal setting. Based on the facts of medical neuroradiological image interpretation, there are some guiding prin-ciples. This does not change the fact that medical image interpretation is still an art as well as a science and, while some findings are very straightforward, many findings are open to interpretation of the individual expert. A more thorough discussion of the possible root causes of disagreements between experts is given in detail in Chap. 9, but this chapter will concentrate on the principles of neuroradiological image interpretation in the medical legal setting. This chapter will also limit its discussion to imaging findings and dating that occur in close proximity – less than 2 weeks – after the acute event.

General Comments on Dating an Event by CT and/or MR and/or US in the First 2 Weeks

Several important features emerge from Chaps. 2 to 4, on the appearance of edema, hemorrhage and contrast enhancement behavior on CT, MR, and US, and the time evolution of changes. Together, these allow some accuracy in dating a neurologic event if the appearance occurs during the most rapid and dynamic changes, either on CT or MR. In general, the total time period for the most rapid changes on CT occurs within 7–10 days from the inciting event, whereas, on MR, dynamic changes can continue to be seen in edema up to 2 weeks, and even longer than that if one attempts to date a hemorrhage. We find it helpful to explain the rationales behind dating the occurrence of an event by presenting situations where only CTs are available, where only MRs

Chapter 6Principles of Dynamic Dating in the Medical Legal Setting

104 6 Principles of Dynamic Dating in the Medical

are available, and where both CT and MR are available for dating events in the first 2 weeks. It is also helpful to attribute dating to both edema and its time changes and to hemorrhage and its time changes.

Dating Edema by CT Alone

First, consider the use of CT in the first 7–10 days after the event, at a time when the predominant or only imaging characteristic is edema. Early on, no findings may be present; imaging findings of a severe acute infarct in the first hour or so are rare. Certainly, by the first or second day a low density, poorly demarcated area may appear. At this point cytotoxic edema is visible on CT, as well as some enhancement around the area of injury. By 2–4 days, the edema becomes more profound, more sharply demarcated, of lower density, and with more mass effect and more enhancement. By 4–6 days after the event, the injury becomes very sharply demarcated, even lower in density, with edema that is beginning to subside, and with subsiding contrast enhance-ment. By 6–10 days following the event, the infarct typically is very well-demarcated and very low in density, with very little mass effect. After 10–14 days, one sees a well-demarcated infarct with sharp borders, CSF density, no edema and minimal or no enhancement. Thus, whether a scan is done at 14 days after the event, 3 months after the event, or 5 years after the event, it will have a similar appearance. Consequently, if a scan shows old or chronic changes, based on the imaging alone with no correlation of the clinical history, there can be no dating of the inciting event. However, during the first 10 days, an assessment of the status within the evolving continuum of the injury as to demarcation of the border, internal density, amount of mass effect, and enhance-ment does allow one to place the lesion with more specificity than just “sometime in the prior ten days” (see Fig. 4.1).

Dating Hemorrhage by CT Alone

In the CT dating of hemorrhage, discerning a range of dates for the occurrence of the bleed is heav-ily dependent upon the space into which the hemorrhage occurs and the amount of blood present in the hemorrhage. On the shorter end of the time spectrum, small subarachnoid hemorrhages or small contusions within the brain resolve most rapidly, on the order of several days. By 4 or 5 days fol-lowing the event, a patient who had a subarachnoid hemorrhage or multiple contusions of the brain at the time of an acute event can have a CT scan that appears completely normal. At the opposite extreme, a large intraparenchymal hemorrhage or a large subdural or epidural hematoma may con-tinue to show dynamically changing appearance on CT for 3–4 weeks following its initial occur-rence. The disparate time course is in part due to the different mechanisms involved in “clearing” the blood from the given space, which, to some extent, can be related to the amount of circulation and the ready availability of cell types which are capable of digesting and removing the blood and the blood breakdown products over time.

In the case of a small amount of subarachnoid hemorrhage, the blood is rapidly cleared from the cerebral spinal fluid over a matter of days. What one sees on an imaging study is a gradually decreasing density in the subarachnoid space until the time when the subarachnoid space returns to the same density as normal cerebral spinal fluid. In the case of large intraparenchymal hematomas, one can see the hemorrhage shrink over time, and this is a process that occurs from the outside

105Dating Edema by MR Alone

inward. That is to say, the overall size of the apparent area of increased density decreases in time. What also is occurring in both a developing ring of edema, and blood products that are being degraded so that they go through a density change from being hyperdense to the brain to isodense or even hypodense to brain. Therefore, an acute, large intraparenchymal hemorrhage is very high density with a very sharp rim and is surrounded by brain of normal density. By 2–4 days, the size of the hemorrhage may be stable to slightly decreased with a band of lower density now seen around the hematoma. This band is due both to the edema in the adjacent brain and to the early degradation of the clot periphery (see Fig. 4.6).

Subdural hematomas undergo a similar course of decreasing density with time. Traditionally, a hyperdense subdural hematoma is thought to occur in the first 4–6 days. By 6–14 days, the hema-toma density slowly decreases until it is isodense to brain. Past that time, the hematoma actually becomes hypodense compared to brain. This may persist for weeks or months, or at this point it may resorb, leaving no trace at all on a CT scan (see Fig. 4.4).

Therefore, several possibilities exist for dating a hemorrhage based on the time-evolving appear-ance on CT: (1) A small amount of subarachnoid hemorrhage must have occurred in the last few days. (2) A well-demarcated, large parenchymal hematoma without any surrounding low density must have occurred in the last 1–2 days. (3) A large, intraparenchymal hematoma with a surround-ing rim of low density probably occurred in the prior 2–4 days. (4) A hyperdense, subdural hema-toma is less than 4–6 days old. (5) An isodense subdural is 6–12 days old. (6) A hypodense subdural hematoma is 10–14 days or older. Finally, dynamic changes no longer occur after the 3- to 5-day period for small subarachnoid hemorrhages or the several week periods for intraparenchymal, sub-dural or epidural hematoma.

Dating Edema by MR Alone

As with CT, one has two opportunities for dating neurologic events based on MR images acquired in the first week or two following the incident. For an infarct, these dynamic changes stretch out over a slightly longer time period than normally seen with CT. Hence, MR scanning may be more helpful than CT in dating an event, at least in the sense that a longer time can have elapsed between the inciting event and the acquisition of the scan, though still permitting some narrowing of the time window when the event could have occurred.

Let us begin, then, with the discussion of the ability to date edema from ischemia or infarction, if scans are acquired in the first 2 weeks. Whereas, CT scanning is dependent solely on looking at a density change and the associated morphologic rearrangement of the brain that happens from edema, the MR scan has several different aspects of imaging that can be used for dating.

In dating edema by MR, an area of ischemia that does not go on to complete infarction (with subsequent removal and filling of the defect with CSF) has a persistent abnormal appearance on MR. An area of ischemia or injury will have persistent abnormalities on both T2 and FLAIR imaging, often for the remainder of the patient’s life. In the absence of outright infarction, the T1 changes are often more subtle, but also will persist for the life of the patient. Diffusion weighted abnormalities, which become abnormal at a few hours, reach their peak by several days and completely disappear by 2 weeks. Intravascular enhancement is present for the first several days after the acute event. Finally, contrast enhancement of the parenchyma of the brain begins at several days and variably persists for weeks or months following the event. Therefore, based on the persistence of findings alone one has ample capability to date the time of an ischemic event by MR (see Fig. 4.2).

106 6 Principles of Dynamic Dating in the Medical

For example, any event with diffusion abnormalities must have occurred in the previous 2 weeks. FLAIR abnormalities become apparent in the first day (after the DWI becomes abnormal). T2 changes are then visible, followed by T1 changes. While these signal changes occur within the lesion on various pulse sequences, the morphology of the infarction and edema evolves in a predictable manner. Specifically, the same sorts of changes occurring dynamically as described above in the CT section also occur on MR. So, similar to CT, the involved area becomes better demarcated, and takes on increasingly bright signal on FLAIR and T2 within the 2-week window. Edema occurs, reaching its peak at 3–5 days and abating over the course of the 2-week window. Parenchymal contrast enhancement is not very helpful at narrowing within that window, as it occurs early on and can continue past the 2-week period. However, intravascular enhancement, if present, is a good marker for the first few days after the acute event. Based on imaging criteria alone with no additional clini-cal information, MR images obtained later than 2 weeks are poor at delineating when the event occurred.

Dating Hemorrhage by MR Alone

When obtained in the first 2 weeks, even a single MR examination can be a powerful tool in fixing, within a range, the time a hemorrhage occurs. This dating of a hemorrhage is perhaps most accurate if the hemorrhage is intraparenchymal, as the time relationships are best understood for an intraparen-chymal hemorrhage vs. one within the ventricles or in any of the extraaxial spaces (subarachnoid, subdural, or epidural). The ability to date a hemorrhage on the basis of the single MR comes from the imaging appearances of a hemorrhage discussed in Chap. 2, as well as an understanding of the time changes of hemorrhage, which is discussed in Chap. 4.

The easiest way to understand the time changes in hemorrhage on MR is to parallel the different imaging appearances on T1- and T2-weighting – of the hematoma, the rim of the hematoma and the adjacent brain – at each time interval, i.e., hyperacute, acute, subacute, chronic, and old. A shorthand means of remembering the T1 and T2 signal intensities within the clot at the five different ages, beginning with hyperacute (T1 mentioned first, then T2) is iso-bright, iso-dark, bright-dark, bright-bright, dark-bright. Of the five time periods, the hyperacute is the most difficult to observe because, to fall in that time range, a bleed must have occurred, at most, within hours prior to the time when the image was obtained. This therefore, usually is not seen unless the patient bleeds immediately before, or when the patient is in the MR scanner.

Other than the rarely seen hyperacute bleed, there is no situation other than the very old hema-toma in which the brightness of the hematoma on T2 is brighter than the brightness on T1. Indeed, in the acute period the hematoma is isointense on T1 and dark on T2, and the increase in signal in the subacute period occurs earlier on T1 than on T2, until by the chronic period the signal intensity is bright on both T1 and T2. The circumstances in which the signal is isointense or dark on T1 and bright on T2 is, again, seen only in the uncommonly-captured hyperacute hemorrhage.

Therefore, if, by history, a patient has an acute event and is scanned within the first few hours, it is possible to have a hematoma that is isointense on T1 and bright on T2. After that period, hematomas can be separated into acute, subacute, and chronic based on the relative T1 and T2 appearance. For example, a hematoma that is isointense on T1 but dark on T2 is an acute hematoma. A hematoma that is isointense to somewhat bright on T1 and is still dark on T2 is a subacute hematoma. A hematoma that is bright on both T1 and T2 is a chronic hematoma. In addition, at some point in the mid- to late-subacute phase and extending into the chronic phase, the hematoma will usually begin to show a black hemosiderin ring on T2-weighting. Edema surrounding the hematoma, whether imaged as low signal on T1 or bright signal on T2,

107Dating Events by CT in Conjunction with MR, and with US

reaches its maximum extent during the acute and early subacute phases, fading off by the late subacute, chronic, and old phases (see Figs. 4.9–4.11).

Dating Edema by US Alone

Even rough-dating the age of the region of the ischemia by ultrasound is difficult. To review, initially, ischemia may not be visible for several hours. When it first becomes visible, it is an area of faint echogenicity which is poorly marginated. Over time, this region either disappears and reverts to a normal appearance, goes on to cystic change which is visible on ultrasound, becomes a band of increased echogenicity (as in the case of periventricular leukomalacia), results in a geographic infarct with a well-demarcated region of fluid, or goes on to global atrophy or parenchymal loss without a focal deficit. Because of this broad range of possible outcomes, seeing the poorly margin-ated area of increased oxygenicity only suggests a recent infarct, at best. In a clinical setting, cor-relation of history, physical exam or other findings may be strongly indicative of an infarct, but US by itself is still very poor at localizing the time of occurrence.

Dating Hemorrhage by US Alone

Ultrasound is perhaps best at demonstrating hemorrhage in the germinal matrix, within the ventri-cles and within the brain parenchyma. While large subdural and epidural hematomas can be seen, smaller and/or more laterally located supratentorial extraaxial bleeds and posterior fossa lesions are harder to detect. Ultrasound is the poorest choice for demonstrating subarachnoid hemorrhage.

Again, by way of review, acute hemorrhages are best demonstrated as a regional increase in echogenicity within the brain parenchyma. Intraventricular hemorrhage occurs as a cast within the ventricles or as a fluid-fluid level. However, other than the fact that the presence of a hemorrhage means that a hemorrhage has occurred, the echogenic appearance is poor at time-dating the age.

Dating Events by CT in Conjunction with MR, and with US

Adding MR to CT in the dynamic period allows additional narrowing of the possible window for the time the event occurred. Either modality by itself can narrow the time windows as described if the scan is obtained in the first week or two, but the two modalities used together provide addi-tional cross checks and correlations dating the time of an injury. For example, if only a single CT is obtained when there is a moderate degree of edema from an ictus of unknown time of occur-rence, it may be difficult to determine whether the scan occurred at the front end or the back end of the peak of maximal edema. However, if, in addition to the CT, an MR scan is available, the diffusion-weighted MR scan will be much brighter on the front end of the period of edema, than on the tail end of the edematous period – and, hence, will aid in discriminating between the two, further narrowing the time window over CT alone. As another example, a CT may have returned to a “chronic appearing” appearance by day 8–10, but persistent diffusion abnormalities are seen on MR out to day 14. Therefore, if the CT alone was obtained at day 10 it might be difficult to date the lesion, but if the MR scan is obtained also on day 10 or 9 or 11 or 12 and shows present but mild diffusion abnormalities, this can be dated as an event that occurred in the previous 1–2 weeks (Figs. 6.1 and 6.2).

108

Fig. 6.1 Combined dating with CT and MR. (a) Hemorrhage on CT in anterior aspect of left temporal lobe with surrounding edema. The hemorrhage is less dense than very acute blood, implying the bleed is at least several days old – but beyond that further dating is indeterminate. MR scans acquired the same day – (b) sagittal T1, (c) axial FLAIR, (d) axial T2 and (e) axial GRE show signal intensity within the central aspect of the clot as iso-, iso-, iso- and dark respectively. The periphery of the hematoma is becoming bright on T1, and bright signal signifying edema is present around the hematoma on T2 and FLAIR. Taken together, the MR findings are of acute to early subacute hemorrhage, implying a 2–5 day range, which is nevertheless more specific than could have been deduced from the CT scan alone

Fig. 6.2 (continued) Combined dating with CT and MR. Initial CT scans with (a) a hyperdense middle cerebral artery (HMCA – represents clot within the vessel) and (b) signs of very early edema with loss of normal gray-white differentiation laterally and fuzzy basal ganglia on the right. MR scan obtained 2 h later shows abnormal diffusion in DWI image in (c), and very minimal changes on FLAIR image in the affected region in (d). Lastly, CT scan obtained 4 days later demonstrates the HMCA in (e) and more mature infarction, with well defined borders and low density in (e, f). Without using images (a) through (d), images (e) and (f) would imply an infarc-tion of several days age. But, having an abnormal DWI scan with an almost normal FLAIR scan definitely means the event occurred just before (c) and (d), a more precise dating of the event

6 Principles of Dynamic Dating in the Medical

In conclusion, images obtained after the initial 2-week period are poor at dating the time of the injury but can confirm the presence of injury. The most definitive way that these images can confirm the presence of an injury is by the existence of a preevent scan that shows the lack of the chronic changes. If no such scan exists, then one or two assumptions are usually made: (1) that the amount of brain parenchyma for a patient that age should be within the norm for that age person (considering any deviation as the result of a prior insult), and (2) that clinical data may exist which more accurately positions occurrence time. Serial images confirm the chronic nature of the injury and may show continuing additional global atrophic changes from a global insult.

109Dating Events by CT in Conjunction with MR, and with US


Abstract This chapter utilizes the information in Chaps. 1–3 and 5 to focus on dating events that have occurred more than 2 weeks prior to the time of imaging. It emphasizes different patterns of injury from edema and infarction and discusses the underlying causes of a particular imaging appearance.

Keywords Magnetic resonance (MR) • Computed tomography (CT) • Ultrasound (US) • Hematoma • Infarction • Chronic injury patterns

Introduction

This second chapter of Part II emphasizes the primary application of the different patterns of injury from edema and subsequent brain parenchyma infarction, as described in Chap. 5. While Chap. 6 emphasizes the use of the dynamically changing appearance of these insults as a means to detect the occurrence of an event and the likely time of that occurrence, this chapter concentrates on the correct method to apply chronic patterns of injury in order to time the occurrence of an event involving the brain.

Principle #1: The best pattern dating (occurring more than 2 weeks after the acute event) for edema/infarction occurs when a clearly recognizable set of findings is characteristic for a known mechanism of injury. Older hemorrhage is a special instance of this principle.

Concerning Edema and Infarction

To review, all the dynamic changes seen on computed tomography (CT) following ischemia or infarction return to a normal appearance by, at most, 2–3 weeks after the event. If at that time, or any time later, a CT were to be obtained and compared with a pre-event scan, any permanent injuries – geographic infarcts, white matter loss or abnormally decreased density, or even global loss of parenchyma – would be visible as new findings. That is to say, by 3 weeks almost all changes would be fully manifested, with the single exception that a diffuse ischemic event may continue to produce general atrophic changes for the ensuing weeks or months.

Chapter 7Principles of Pattern Dating in the Medical Legal Setting

112

Fig. 7.1 Example of pattern dating in a pediatric patient with hypoxic-ischemic encephalopathy or HIE. (a, b) CT scans of a 3-year-old child. (c) US image of same child as an infant 5 days old, (d) MR T1-weighted scan at 2½ weeks of age. (e) Second US exam at 3 weeks of age. For full discussion see accompanying text

7 Principles of Pattern Dating in the Medical Legal Setting

Similarly, a magnetic resonance (MR) examination should have reached a stable appearance for the remaining uninfarcted brain by 2–3 weeks after the acute event. As on CT, areas of brain that have died will appear as regions of encephalomalacia. Unlike CT, regions that underwent significant ischemic injury, though not complete infarction, will continue to show FLAIR and T2 signal abnor-malities for the life of the patient. Areas with restricted diffusion during the acute phase that did not go on to infarction will have returned to a normal appearance.

If a characteristic pattern, such as those described in Chap. 5 is present, then, at the very least, the mechanism of injury can be identified – though not always the time of the injury. To illustrate this point, an example will be helpful.

Example 1: The first two images (a and b) are from a brain CT scan of a 3-year-old child. One sees extensive loss of brain parenchyma (Fig. 7.1). Only the posterior fossa structures and the basal ganglia are normal in appearance. Above the tentorium, a thin rim of tissue at the cortical surface and around the margins of the dilated lateral ventricles is all that remains of the cerebral hemi-spheres. The following three figures (c–e) document the brain injury that occurred during the first 3 weeks of life. (c) Is a head ultrasound (US) examination at 5 days of age, (d) a T1-weighted image from an MR examination at 2½ weeks of age, and (e) a second head US at 3 weeks of age. Collectively, they demonstrate increasing echogenicity of the brain, increasing ventricular size and progressive loss of total brain parenchyma.

113

Fig. 7.2 Example of pattern dating in a middle-aged patient with two different subdural hematomas. On the initial CT scan in (a) a chronic right-sided subdural hematoma is present, with a small acute component. In (b) at 7 weeks after the first scan, the subdural has completely cleared, without surgery. In (c) obtained 4 weeks after the second scan, one sees a new left-sided, iso-dense subdural hematoma. See text

Introduction

This pattern, seen in a 3-year-old child, is highly suggestive of a severe anoxic injury that occurred previously in the neonatal period. Based on the images, a more specific time frame is not possible – but the appearance certainly is not that of any kind of recent injury.

Concerning Hemorrhage

Patterns of hemorrhage and the aging of hemorrhage were described in Chap. 4. Hyperacute, acute, and subacute bleeding most appropriately fall into the dynamic dating category, as the hematoma continues to change in appearance in those phases. Very late subacute and chronic hemorrhage is, however, a condition that may be stable in appearance for very long periods of time. Hence, it is not unreasonable to consider an older bleed as a “pattern.” Therefore, the logic surrounding a patient with two different subdural hematomas is presented next.

Example 2: The patient is a middle-aged man, imaged by CT at three different times (Fig. 7.2). On the initial CT scan in (a), a chronic right-sided subdural hematoma is present, with a small acute component. In (b), at 7 weeks after the first scan, the subdural has completely cleared, without surgery. In (c), obtained 4 weeks after the second scan, one sees a new left-sided isodense subdural hematoma. Given these three different CT scans, what can we learn about the time the subdural hemorrhages occurred? One approach is to analyze the information from each scan.

The first scan shows a chronic, subdural hematoma with a small acute component. Therefore, the major portion of the subdural is at least 2 weeks old, and possibly much older than that. The small region of high density within the fluid collection means that, on top of the chronic bleed, one sees a newer component that is, at most, several days old. The second scan documents that the original subdural has completely disappeared. In addition, it establishes a point in time when the scan was normal, prior to the abnormal third scan. The third scan shows an isodense subdural hematoma, which must, therefore, be between roughly 6–12 days old, consistent with a previous normal exam obtained 28 days earlier.

If the third scan had shown a chronic subdural hematoma, then the hemorrhage that caused that hematoma could have occurred anytime after the normal second scan, but no later than 2 weeks prior to the third scan. In effect, it would have generated a 2-week window during which the hemor-rhage would have occurred. Or, said another way, a chronic SDH does not fix the time of occurrence

114

Fig. 7.3 Example of pattern dating in an older infarction. Single CT scan in an older man shows old infarctions bilaterally. No evidence of hemorrhage or edema is seen. See text for more complete discussion

7 Principles of Pattern Dating in the Medical Legal Setting

of the bleed except to say it did not occur in the previous 2–3 weeks before the scan was obtained. However, other factors (in this case a previous scan) can affect the start point of the time window.

Principle #2: Unless a recognizable pattern of injury is present with known causes, time of occurrence and method of injury, a single study after the early dynamic changes have occurred is not very useful at dating the time of an injury.

A single scan obtained 2 weeks after the event will have a relatively stable appearance on CT or MR, whether the inciting ischemic/hypoxic event occurred 2 weeks, 6 weeks, 3 months, 2 years, or 10 years prior to the time of imaging. Therefore, in the absence of clinical correlation, a single image which lacks a classic morphologic pattern of injury cannot date the time of the neurologic injury – except to say that if no dynamic changes are present, the injury did not occur in the preced-ing 2 weeks.

If the patient’s clinical history is devoid of any possible inciting events except for a single epi-sode in the past, then a deductive link is often assumed between the imaging findings and that one clinical event. However, at that single point in time, nothing on the images themselves can conclu-sively link the changes to the known previous event.

On the other hand, a single scan obtained many weeks to months to years remote from the inciting event may confirm that the brain has suffered an injury. Even better confirmation is a pre-event scan showing a normal amount and overall appearance of brain parenchyma. In the event no baseline, premorbid scan is available, then the most reasonable estimate of how the brain should appear is to use the appearance of a normal brain in a person of the same age as a standard. The following example demonstrates an infarct that is poorly datable as to its time of occurrence.

Example 3: This is an older man whose CT scan demonstrates an old, well-demarcated infarction in both hemispheres (Fig. 7.3). Similar to the chronic subdural hematoma, thess areas of infarction clearly shows that an event did occur. However, the brain does not assume this appearance, i.e., with very demarcated edges, no mass effect, and the central portion of the infarct being equivalent to CSF in density, until the infarct is at least several weeks old. As in the previous example, this means that,

115Conclusion

clearly, there has been an event, and it did not occur in the previous 2–3 weeks. But, one could see this same appearance in an infarct that was a month, 2 or 10 years old.

Principle 3: In general, after the first several weeks, the appearance of the remaining brain is stable on CT and MR scanning.

The heart of Principle 3 is that patterns of injury will not evolve over time after the initial few weeks after the event. This is the primary reason why these patterns are of minimal use, except for putting a very broad range of possible dates on the time of event occurrence.

While this is, in general, a very reliable rule by which to interpret images, there is one important exception. In younger patients, especially after an anoxic event, a continued evolution of imaging findings can stretch over months or years. Changes following a severe anoxic event can continue to accrue for weeks or months after a single inciting event.

Conclusion

Pattern dating, while not as intrinsically accurate as dynamic dating, still can be very useful. In some situations, it can accurately localize a very specific injury to a discrete period in time. However, even if that is not possible, the nature of chronic injury – and the time it takes dynamic changes to cease – still is useful in excluding or including time periods. Finally, stable or evolving chronic changes can also confirm a definite previous injury.


Abstract This chapter addresses the specific application of material in the preceding seven chapters in the medical–legal setting. Based on the information that is both present, and not present on the available images, and understanding the underlying pathological processes, it is possible to make definitive statements on timing. On the other hand, there are some statements that cannot be supported on the basis of the available images.

Keywords Chronic pattern dating • Dynamic dating • Medical–legal

Introduction

This chapter follows closely on the thoughts developed in the preceding two chapters, namely, the principles of image interpretation. While similar, there is a difference between the imaging findings themselves (whether dynamic or chronic in nature), and the deductions that can be made from those same findings. With that in mind, what follows are several rules that concern not so much the obser-vations themselves, but rather the conclusions that can be drawn based upon the observations, the known changes of imaging findings with time and the known patterns of disease. We would con-sider, therefore, the following rules as a kind of “expert opinion box,” outside which one should not go as an expert on image interpretation.

Rule #1: One “can” say that, if edema/infarction or hemorrhage is imaged in the dynamic period, the injury occurred between two definable points in time within the 2-week period.

When attempting to date the time a neurological injury occurred, by far the most optimal scans to have available for review are those obtained immediately after the alleged injury. For CT, this optimal time inter-val is the first 7–10 days, whether for infarction, ischemia, or hemorrhage. For MR, whether the incident is ischemia, infarction, or hemorrhage, the optimal time window is perhaps somewhat longer, stretching out to 2 to 2½ weeks. To say that this time range is optimal does not necessarily mean that the observations made from the images and the consequential conclusions provide firm and rigid guidelines for the time of occurrence. For example, a computed tomography (CT) scan that shows maximal edema with moderately well-defined borders, low density, and contrast enhancement is consistent, from an observation standpoint, with the maximal edema one sees 3–5 days after an infarction. The time window of 3–5 days must be understood to represent an average range and, if pushed, one could say that the event could have occurred 2½ days previously or 5½ or 6 days previously. The converse is also true that such an appearance of maxi-mal edema, mass effect, and contrast enhancement probably did not occur from an event that was an hour ago or that was 2 weeks ago. So dynamic pattern dating usually can not only limit an event to the first 2 weeks after its occurrence, but can also limit it to a smaller time window.

Chapter 8Therefore, What Can Be Said Based on the Images, and What Can’t Be Said Based on the Images

118 8 What Can Be Said Based on the Images

Rule #2: One “can” say that, if a classic chronic pattern is present, the injury occurred at a certain time, often from a certain inciting cause.

From Chap. 5, when classic findings are present that allow reliable dating, the most common sce-narios occur in the pediatric population, associated with anoxic events. To a lesser extent, classic patterns for older hemorrhages (late subacute, chronic and old) also permit dating to confined time windows. In both cases, the major pitfall is to be too precise about the time of event occurrence. As stated previously, pattern dating is much better at eliciting the cause of a neurologic event than at precisely determining the time of its occurrence. In addition, uncertainties can creep into a specific clinical situation – and these uncertainties are described in more detail in Chap. 9. Having men-tioned the cautions which must be considered, the patterns of global anoxia, parasagittal/watershed, and rolandic injury are best for placing the time of occurrence.

Rule #3: One “can” say that, if the pattern is not classic, an event did occur – but the most that can be said is that it didn’t occur recently (in the past 2 weeks).

Often, the radiological image expert witness is provided with a set of examinations that were obtained many weeks, months, or years following the time of the alleged injury and in which, while an injury is present, there is no characteristic pattern of a specific cause or time of occurrence. In that case, based on imaging criteria alone, the examinations do not assist in the ability to localize the time of the injury.

On CT, for example, after a geographic arterial blockage, infarction, or intraparenchymal hemor-rhage, the brain will undergo dynamic imaging changes for the first 1–2 weeks. However, after that time, the changes remain fixed. With a single, nonclassical-appearing examination long after the inciting event and in the lack of clinical history, it is not possible to link the damage to a specific time period, based only on the images. What is possible to say unequivocally is that damage did occur. Further, lack of any dynamic changes – i.e., edema, abnormal DWI scan, hyperacute/acute/or subacute blood products – means that the inciting event was at least 2–3 weeks prior to the date of the imaging examinations.

Rule #4: One “cannot” say, based on imaging criteria alone – either on dynamic or pattern dating – that the injury unequivocally occurred at this precise point in time.

On the basis of the first three rules, it is often possible to determine from the provided imaging studies that an injury likely occurred within a certain time window. Establishing that window will usually remove some individuals from possible consideration as contributors and will place other individuals at that point in time when the event probably took place. However, it must be remem-bered that imaging studies alone cannot prove or disprove that any one person or persons did or did not do a certain action at a certain point in time.

The question often posed to expert witnesses at this junction is, “Could the candidate clinical event have produced these findings?” In many instances, the appropriate ethical response is in the affirmative. The question is then often stated differently as, “Did the clinical event produce the radiographic appearance?” In that case, the answer is, “It cannot be said, based on the images alone.” However, there are times, based on training and experience that the expert can say (in most jurisdictions the key phrase is “beyond a reasonable degree of medical certainty” – i.e., >50%), “It is likely that this event happened at that time and caused this appearance.”

Finally, if a pattern of injury is present, one should note whether it is descriptive of an acute, subacute, or a chronic appearance. This is important, as the injury pattern does not fix the time of injury if it is a chronic injury pattern (past the period when dynamic dating is of use); it only con-firms that an injury did occur. It does say that the injury could have happened at a specific previous time, but it does not necessarily say that it did happen at a specific previous time.


Abstract The final chapter discusses the possible causes of uncertainty in the dating of a neurologic event based on imaging findings – variations in interpretation from expert to expert and variations in the manner that consensus findings are interpreted by different experts, multiple different imaging findings that localize to different points in time, and imaging findings that do not appear to correlate with the clinical picture.

Keywords Uncertainty • Chronic dating • Pattern dating

Introduction

We have now discussed in some great detail the normal brain and the changes in the brain of edema or hemorrhage in response to injury as they appear on CT, MR, and US. We have discussed the evolution of these findings on computed tomography (CT) and magnetic resonance (MR) with time ranges for the occurrence of the changes. We have further discussed what this progression of find-ings allows one to say on the basis of imaging studies acquired at discrete points in time following the inciting event. What we have not discussed up to this point is why there is any uncertainty in dating neurologic events or, put another way, why two different medical experts could look at the same set of images and come up with different interpretations either on the presence or absence of injury at all, and on the time when the injury occurred.

Reasons for this uncertainty in dating form a short list of possible causes. These are the following:

1. The interpretation of the available images varies from expert to expert, with disagreement as to whether certain findings are present or not.

2. Image findings are acknowledged to be present by all observers; however, the interpretation of the findings varies from expert to expert.

3. Multiple findings are present for which the neuroradiological dating methods we have discussed produce conflicting time periods as to the probable occurrence of the injury.

4. The radiographic findings and time periods deduced from the images are at odds with the clinical picture.

Each of these four main causes of uncertainty in dating neurologic events is now discussed.

Chapter 9The Root Causes of Uncertainty in Dating Neurologic Events Based on Imaging Findings

120 9 The Root Causes of Uncertainty in Dating Neurologic Events

The Interpretation of the Available Images Varies from Expert to Expert, with Disagreement as to Whether Certain Findings Are Present or Not

The first root cause of uncertainty in dating neurologic events is the uncertainty that arises from the presence or absence of subtle findings. For example, the patient may have undergone a clinical event, which possibly could have produced an anoxic or ischemic injury. Subsequent CT or MR examinations may be interpreted by one observer as being normal and by a second observer as showing subtle signs of early ischemia. The assumption, therefore, made by the two different inter-preters based on the images would be, in the first case that the images support no evidence of isch-emic or anoxic injury and, in the second case that the images are consistent with an early and/or a mild degree of anoxic or ischemic injury.

This example presumes that the controversial findings are subtle, such that there can be a valid difference of opinion between two equally qualified experts as to whether the findings are present or not. This situation is distinctly different from the following situation in which the area of dis-agreement is not on the findings (often in these cases, the presence of an abnormality and/or injury is obvious to all observers) but in the meaning and implication of the findings.

Image Findings Are Acknowledged to be Present by All Observers; However, the Interpretation of the Findings Varies from Expert to Expert

This cause of uncertainty in dating neurologic events based on radiological findings is explained by the fact that the interpretation of scans is an art and not a science. The first reason mentioned for disagreement between experts concerns the situation in which the imaging findings are subtle, and their mere presence is open to debate. However, in this case, the issue is not what the findings are, but rather what they mean.

For example, presume that a patient has had a definite hypoxic-ischemic event and the region of affected brain undergoes a period of swelling which is maximal at 3–5 days. Consider a single scan, obtained 3 days after the acute event (for example, Figs. 4.1d or 4.2e from Chap. 4), reviewed by two different experts, each unaware of when the scan was performed after the acute event. The first expert looks at the scan, notes some edema is present, but not as much as would represent maximal swelling. This expert assesses the sharpness of the edge of the infarction and the amount of decreased density within the infarct compared to adjacent uninfarcted brain and concludes that the scan must have been obtained before the period of maximal swelling – i.e., on day 2 or 3.

The second observer views the same exact set of images and also assesses the sharpness of the margin and density of the infarct and also concludes that there is submaximal swelling. However, this radiologist feels that the images were obtained after the period of maximal swelling – i.e., on day 6 or 8. That is the manner by which two experts look at the same images, note the same findings, but come to two different conclusions as to what the images say about the age of the infarction.

Similarly, different impressions of the images as they relate to the size of the hemorrhage, the den-sity of the hemorrhage on CT, and the maturation of the hemorrhage in the time sequence as assessed by MR may contribute to a discrepancy between two observers when assessing the same images.

Two observers may also be in complete agreement as to the presence of the findings and describe to a high degree of concordance the identical injury, loss of brain, abnormal signal, presence of blood of a certain density (on CT) or signal (on MR) and still disagree on the points that those find-ings fix along a timeline of events. In other words, no matter how elegant and convincing one’s guidelines and time points are, other experts’ time points for identical findings may differ by degree from one’s own.

121Conclusion

Multiple Findings are Present for Which the Neuroradiological Dating Methods We Have Discussed Produce Conflicting Time Periods as to the Probable Occurrence of the Injury

This cause of disparity relates to the fact that some patients will present with more than one potentially “datable” lesion. When the principles of dating are applied to multiple lesions, not all of the lesions will “date” to the same time in some patients. For example, the patient may have multiple infarcts which, by the dating characteristics, seem to have occurred at different times; or the patient may have an infarct and a hemorrhage which, by dating methodology, do not appear to have occurred at the same time. Any of these cases presents a dilemma which must be solved for the expert’s own opinions to be consistent. The expert is presented, really, with only one of two options.

On the one hand, the expert may decide that the dating of the two separate processes is correct. This means that the two (or more) events did not occur at the same time. Alternatively, after rere-viewing the available examinations, the expert may decide that the initial dating of one or both of the events may have been incorrect. The initial assessment may have used time windows that were too narrow. Could the beginning and end points for the possible injury in each case be widened to the extent that the two events were possibly concurrent? That is an assessment that an expert must make. If, based on all the available imaging material available, that determination cannot be made, then the first option is the most likely explanation for the findings.

Radiographic Findings Which Are at Odds with the Clinical Picture

This is, perhaps, the most difficult and perplexing situation that can confront an expert. In this set-ting, the imaging studies point to a certain time window, but the time window when the injury should have – even must have – occurred is not in accordance with the clinical setting. This circum-stance is an expansion of the preceding situation. As in that setting, the expert has a limited number of options.

One’s first option is to consider that the radiologic dating and the clinical dating point to and describe two separate events. This would be the case when another potentially injurious event occurred, separate from the one thought to have occurred by the dating method. It may be that either the radiologic event or the alleged clinical event was “silent” and only one of the two events pro-duced real clinical symptoms.

Alternatively, as in the previous case, it may be that the dating was in error. So again, a reassess-ment must be made of the imaging findings and the time windows they imply to see if it might be possible that a revision of the radiologic time window endpoints could include the date implicated by the clinical history.

Conclusion

This chapter has highlighted the most common causes for discrepancies between experts in their evaluation of radiological images. An understanding of these dynamics is important, as the members of the expert community seek to use the principles described within this book to skillfully, expertly, and ethically serve the patients and families who are affected by these neu-rologic situations.

123

AAnterior cerebral artery (ACA), 11Anterior inferior cerebellar artery

(AICA), 11Apparent diffusion coefficient (ADC)

map, 72

BBasal ganglia infraction, 97–98Blood-brain barrier (BBB), 60Brain

anatomical planes, 3–4anatomical view, 9–10blood supply

arterial vessels, 10–14imaging method, 16–20venous vessels, 14–16

brainstemcomponents, 32CT images, 34MR images, 36nuclei function, 32–33regions, 32–34US images, 34

cerebellumcomponents function, 28cortex, gray matter, 32CT images, 34deep cerebellar white matter, 32MR images, 33structure, 27–28, 32US images, 32, 34

cerebral hemispheresCT, different ages, 22, 24–25, 28, 30CT imaging, 24–26deep gray matter nuclei, 23–24gray matter, 22, 23lobes, 20–22MR imaging, 24, 27–29US imaging, 24, 29–30white matter, 22

image orientation, 3tissue layers

arachnoid membrane, 6–8

cortex, outersurface, 8dura, 5, 6, 7epidural space, 5–6, 53graphical representation of, 5imaging, 8–9pia, 8scalp, 4skull, 4–5, 6subarachnoid space, 7–8, 37subdural space, 6–7, 53

ventriclebrainstem images, 36–39, 40cerebellum, 33–34, 40cerebrum images, 25–30, 39, 40cisterns, CSF spaces, 37–38CSF dynamics, 38–40shape, size and position, 35, 37

Brainstemcomponents, 32CT images, 34MR images, 36nuclei function, 32–33regions, 32–34US images, 34

CCerebellum

components function, 28cortex, gray matter, 32CT images, 32, 34deep cerebellar white matter, 32MR images, 32, 33structure, 27–28, 32US images, 32, 34

Cerebral hemispheresCT and MR imaging, 24–29deep gray matter nuclei, 23–24gray matter, 22, 23lobes, 20–22white matter, 22

Chronic injury pattern datingedema and infarction (see also Parenchymal injury,

chronic dating)computed tomography, 111

Index

124 Index

Chronic injury pattern dating (cont.)magnetic resonance, 112pediatric patient, HIE, 112

hemorrhage (see also Hemorrhage)middle-aged patient, 113–114older infarction, 114–115subdural hematoma, 113

Chronic pattern dating, 118Chronic subdural hematoma, 113Computed tomography (CT), 43–45

brainstem, 34, 35cerebellum, 32, 34cerebral hemispheres, 24–26chronic injury pattern dating, 111chronic subdural hematoma, 113components, 44contrast, role

BBB disruption, 65–66bolus, 61, 62for causative pathology, 65–67enhancing tissues, 64vascular filling percentage, brain, 60

central nervous imaging system, 59dating process, 67enhancement

clinical importance, 67cranial vault, 64CT contrast, 60dura and choroid plexus, 64MR contrast, 62

MR scanningdose and pulse sequence choice, 62effects vs. flow void effects, 62–63T1-and T2-weighted images, 62

principles, 60administration of, 60for CT, 59, 60intact blood-brain barrier, 60for MR, 60

density, 44dynamic dating

edema, 70, 104hemorrhage, 75–76, 77, 78, 79,

104–105dynamic dating, medical-legal setting,

103–105edema (see also Edema)

acute and subacute phase, 70acute infarction, 70–71chronic infraction, 70general appearance, 51–52hyperacute phase, 70

geographic infarction, 92–93global anoxia, 94global atrophy, 94–95hemorrhage

axial image, 76–79patients hematocrit, 75time evolution, 76time period, dating, 69, 70

numbers, 45x-ray beam absorbtion, 44

Computed tomography angiography (CTA), 16, 17, 19

Cytotoxic edema, 48. See also Edema

DDiffuse axonal injury, 75Diffusion-weighted imaging (DWI), 46, 50, 71–72Dynamic dating, medical-legal setting

computed tomographyedema, 70, 104hemorrhage, 75–76, 77, 78, 79, 104–105

CT with MR, 107–109image interpretation, 118magnetic resonance

edema, 71–72, 73, 105–106hemorrhage, 76–85, 106–107

EEdema

change over time, CTacute and subacute phase, 70acute infarction, 70–71chronic infraction, 70hyperacute phase, 70

change over time, MRacute infarctions, 72–73acute phase, 72diffusion-weighted sequence, 71–72hyperacute, 71–72parenchymal enhancement, 72subacute and chronic phase, 72

change over time, US, 85–86CT scanning, 51–52, 70–71cytotoxic, 48definition, 48dynamic changes, 70–73dynamic dating, CT and MR, 103–104,

105–106effect on

bilateral basal ganglia infarctions, 49, 50, 51

intracranial structure, 49–50ventricles, 50–51

loss of G-W matter differentiation, 48–50mass effect, 50–51MR scanning, 49–51, 52US scanning

abnormallity, 51, 52silhouetting of the sulci, 53

vasogenic, 48ventricle size reduction, 50

FFluid-attenuated inversion-recovery (FLAIR)

sequence, 40, 46, 71–72

125Index

GGradient echo imaging, 48, 85

HHemorrhage

change over time, CTaxial image, 76–79patients hematocrit, 75time evolution, 76time period, dating, 69–70

change over time, MRblood-breakdown product, 76–77FLAIR imaging, 82–85gradient echo imaging, 82–85hemoglobin, 76hydration state, clot, 77–78signal changes, 78time evolution, 78, 80T1-weighted images, 80–84T2-weighted images, 80–84

change over time, US, 85–86chronic injury pattern dating, 113–115CT scanning, 54–55dynamic dating, CT, 103, 104–105intraparenchymal hemorrhages

CT scan, 55MR scan, 56

intraventricular hemorrhages, images, 57location

epidural hematoma, 53, 74intraparenchymal hemorrhage, 53, 74–75intraventricular hemorrhage, 53, 75subarachnoid hemorrhage, 53, 54, 74subdural hematoma, 53, 54, 74

magnetic resonance, 106–107MR scanning, 55occurrence, 53subarachnoid hemorrhage

CT scan, 53MR scan, 54

US scanning, 55–56Hypoxic-ischemic encephalopathy (HIE), 95–96, 112

IImage interpretation principle

chronic pattern, certain timings, 118dynamic period, 2-week, 117imaging criteria, precise timings, 118non classic pattern, past 2 weeks, 118uncertainty causes

acknowledged, all observer, 120varies, expert to expert, 120

Internal carotid arteries (ICA), 10–11Ischemia region, variations

duration, 91metabolic activity, 92perfusion, 90–91severity of insult, 91

MMagnetic resonance (MR)

brainstem, 35, 36cerebellum, 32, 33cerebral hemispheres, 24, 27–29chronic geographic infarction, 93chronic injury pattern dating, 112components, 45–46contrast, role

dose and pulse sequence choice, 62effects vs. flow void effects, 62–63T1-and T2-weighted images, 62

dynamic datingedema, 105–106hemorrhage, 106–107

dynamic dating, medical-legal setting, 105–107edema (see also Edema)

acute infarctions, 72–73acute phase, 72diffusion-weighted sequence, 71–72general appearance, 52hyperacute, 71–72parenchymal enhancement, 70, 72subacute and chronic phase, 72

factors, 45–46global anoxia, 94, 95global atrophy, 94–95, 96hemorrhage

blood-breakdown product, 76–77chemical changes, hemoglobin, 76–78FLAIR imaging, 82–85gradient echo imaging, 82–85hydration state, clot, 77–78signal changes, 78time evolution, 78, 80T1-weighted images, 80–84T2-weighted images, 80–84

injury, central sulcus region, 95–96, 98pulse weightings, 46weighted image, 46

Magnetic resonance angiography (MRA), 16, 18Magnetic resonance venography (MRV), 16, 18Middle cerebral artery (MCA), 11

PParasagittal infarction, 99Parenchymal injury, chronic dating

characteristic, 89–90factors affecting, ischemia

duration, 91metabolic activity, 92regional variations, perfusion, 90–91severity of insult, 91

patternbasal ganglia, 97–98geographic infarct, 92–93global anoxia, 94global atrophy, 94–95neonatal HIE, 95–96, 112

126 Index

Parenchymal injury, chronic dating (cont.)parasagittal injury, 99perirolandic region, 98, 99PVL/white matter loss, 96–97

principle, 89–90Posterior cerebral artery (PCA), 11–12Posterior inferior cerebellar artery (PICA), 11

SShear injury, 74Superior cerebellar artery (SCA), 11

TT1-weighted images, 45–46, 81T2-weighted images, 45–46, 81–83

UUltrasound (US) scanning

cerebellum, 32, 34cerebral hemispheres, 24, 29–30change over time, edema and hemorrhage,

85–86

edema, 51, 52 (see also Edema)hemorrhage, 55–57PVL/white matter loss, 96–97sonar image, tissues, 47

Uncertainty causes, image findingsinterpretation

acknowledged, all observer, 120varies, expert to expert, 120

multiple findings, 121radiologic dating, 121

VVasogenic edema, 48. See also EdemaVentricle

brainstem images, 36–39, 40cerebellum, 33–34, 40cerebrum images, 25–30, 39, 40cisterns, CSF spaces, 37–38CSF dynamics, 38–40shape, size and position, 35, 37

WWhite matter loss, 96–97

Dating neurological injury - a forensic guide for radiologists, other expert medical witnesses, and attorneys

Health & Medicine

springer sciencebusiness

human nervous

central nervous

axial t2 image

pulse sequence

flow void

cervical spinal

traditional