Report CT scanners for coronary CT angiography (CCTA) in challenging patient groups King's Technology Evaluation Centre (KiTEC) Department of Medical Engineering and Physics King's College Hospital NHS Foundation Trust Denmark Hill London, SE5 9RS phone: +44 (0) 203 299 1626 fax: +44 (0) 203 299 3314 Certificate No. FS36209
129
Embed
CT scanners for coronary CT angiography (CCTA) in ...kitec.co.uk/perch/resources/cardiac-report-for-kitec-website.pdf · CT SCANNERS FOR CARDIAC IMAGING .....45 6.1 CT scanner models
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Report
CT scanners for coronary CT angiography
(CCTA) in challenging patient groups
King's Technology Evaluation Centre (KiTEC) Department of Medical Engineering and Physics King's College Hospital NHS Foundation Trust Denmark Hill London, SE5 9RS phone: +44 (0) 203 299 1626
fax: +44 (0) 203 299 3314
Certificate No. FS36209
1
Table of Contents Abbreviations ............................................................................................................5
3.1 Patients groups posing additional challenges .................................................... 19
3.2 Key technical requirements in cardiac CCTA imaging ........................................ 20
4. PRINCIPLES OF CCTA SCANNING ...................................................................... 22 4.1 Essential components of a CT scanner and key scanner parameters ................ 22
4.2 Scan modes used in CCTA ................................................................................ 24
5. RADIATION DOSE IN CCTA .................................................................................. 31 5.1 Justification of the CCTA examination ............................................................... 33
5.9 Z-axis overbeaming and overscanning .............................................................. 43
6. CT SCANNERS FOR CARDIAC IMAGING ............................................................ 45 6.1 CT scanner models included in the report .......................................................... 45
6.2 Brief description of scanner models ................................................................... 46
6.2.1 GE Healthcare ................................................................................................ 49
8. ADVANTAGES, UNCERTAINTIES AND RISKS OF COMPARING SCANNERS USING TECHNICAL SPECIFICATIONS ....................................................................... 95
ImPACT Former UK centre for CT scanner technical evaluations, based at St. George Hospital, London, (disbanded 2011)
6
KHP King’s Health Partners
KiTEC Kings Technology Evaluation Centre
kV Kilovolt
kW Kilowatt
LAD Left anterior descending artery
LCA Left coronary artery
LCX Left circumflex artery
mA milliampere
MI Myocardial infarction
MSCT Multi-slice computed tomography
MSR Multi-segment reconstruction
MTF Modulation transfer function
NA Not applicable
NFE Non-fatal event
NFMI Non-fatal myocardial infarction
NHS National Health Service
NICE National Institute for Health and Care Excellence
NIHR National Institute for Health Research
NR Not reported
NGCCT New generation cardiac computed tomography
OR Odds ratio
PCI Percutaneous coronary intervention
PTA Prospectively ECG-triggered axial
PTH Prospectively ECG-triggered helical
QALY Quality-adjusted life year
RCA Right coronary artery
RGH Retrospectively ECG-gated helical
ROC Receiver operating characteristic
Se Sensitivity
SNOMED-CT Systematized nomenclature of medicine clinical terms
Sp Specificity
SROC Summary receiver operating characteristic
UK United Kingdom
7
Glossary
Arrhythmia Any disturbance of the normal rhythmic beating of the heart or
myocardial contraction.
Artefact Error in the representation of any visual information introduced by
the technique(s) or equipment.
Blooming artefact High-attenuation structures, namely calcified plaques or stents,
with enlarged appearance due to partial-volume effects,
obscuring the adjacent lumen.
Calcium scoring A technique by which the extent of calcification in the coronary
arteries is measured and scored.
Contrast resolution Ability to differentiate between different tissue types in an image.
Contrast-to-noise
ratio
A measure related to image quality. Defined as the difference
between the signal in the region of interest and the surrounding
background divided by the average variation in the background.
Coronary
angiography
A diagnostic imaging procedure which provides anatomical
information about the degree of stenosis (narrowing) in a
coronary artery.
Coronary artery An artery that supplies the myocardium.
Coronary artery
disease
A condition in which atheromatous plaque builds up inside the
coronary artery leading to narrowing of the arteries which may be
sufficient to restrict blood flow and cause myocardial ischemia.
CTDIw A radiation dose index used in CT to represent the average
absorbed dose in the scan plane of a standard perspex phantom.
CTDIvol As CTWIw but taking into account the exposure variation in the z
axis for non-contiguous scans.
Double sampling Sampling technique that utilises two measurements per detector
in order to prevent deterioration of resolution and the generation
of aliasing artefacts. It can be performed in x-y plane and also z-
axis.
Dual-source multi-
slice computed
tomography
A dual source scanner has two pairs of x-ray sources and multi-
slice detectors mounted at approximately 95 degrees to each
other.
8
Dynamic focal spot
(“flying focal spot”)
An electromagnetic shifting of the location of the focal spot on the
x-ray tube anode relative to the detectors. This technique is used
to increase the data sampling density (see ‘double sampling’).
ECG gating Technique used to improve temporal resolution and reduce
artefacts due to cardiac motion. The patient’s ECG provides real-
time information of the cardiac cycle and is used to identify the
point to trigger image acquisition or to reconstruct images.
Focal spot The area on the anode of an x-ray tube that is struck by electrons
and from which the resulting x-rays are emitted.
Gantry Part of the CT scanner that accommodates the x-ray source,
detectors and rotating mechanism to allow cross-sectional views
of patient anatomy.
Graft Surgical procedure to move tissue from one site to another on
the body, or from another person, without bringing its own blood
supply with it.
Invasive Coronary
Angiography
Invasive procedure used to image the blood vessels of the heart.
It involves manipulation of cardiac catheters from an artery in the
arm or top of the leg. A contrast medium is injected into the
coronary arteries, and the flow of contrast in the artery is
monitored with a rapid series of x-rays. It is considered the “gold
standard” for providing anatomical information and defining the
site and severity of coronary artery lesions.
Inter-observer
variability
The differences occurring between individuals performing the
same and especially a visual task.
Isotropic resolution Equal resolution in all three dimensions of the voxel, i.e. in the
scan plane (x-y), and in the z-axis.
Iterative
reconstruction
Algorithms used to reconstruct 2D and 3D images by performing
repeated reconstructions to improve image fidelity.
Modulation
transfer function
Illustrates the fraction (or percentage) of an object’s contrast that
is recorded by the imaging system, as a function of size (i.e.
spatial frequency) of the object. The MTF is a used as a
complete description of the resolution properties of an imaging
system.
Multi –segment
reconstruction
An approach used in cardiac gated scans whereby data from two
or more heart beats is used to reconstruct an image .
Multi-slice CT
coronary
angiography
A non-invasive investigation that provides anatomical information
about the degree of stenosis (narrowing) in the coronary arteries.
The scanner requires a high rotation time and as the technology
9
has advanced the number of slices in each rotation has
increased.
Negative predictive
value
Indicates the likelihood that a patient does not have a disease
when the test result is negative.
Padding Padding – additional x-ray beam on time; the x-ray exposure
window may be extended to cover a longer phase of the cardiac
cycle to allow for greater a range of cardiac phases available for
image reconstruction.
Pitch The ratio of table feed per rotation to slice collimation. A pitch
less than 1 represents overlapping scans.
Positive predictive
value
Indicates the likelihood of disease in a patient when the test
result is positive
Revascularisation The restoration of blood supply to the affected tissues.
Sampling rate In CT, the frequency at which data are collected as the gantry
rotates around the patient.
Sensitivity Proportion of people with the target disorder who have a positive
test result.
Spatial resolution Ability of an imaging system to accurately depict anatomical
features in an image.
Specificity Proportion of people without the target disorder who have a
negative test result.
Stenosis A narrowing of the arteries leading to a reduction in blood flow.
May be due to the build-up of atherosclerotic deposits of fibrous
and fatty tissue or may be a congenital defect.
Stent An implantable device designed to be inserted into a vessel or
passageway to keep it open.
Temporal
resolution
In CT, it is the period of time over which medical images are
captured or recorded. The ‘intrinsic spatial resolution’ is that
which can be achieved by the scanner without using special
modes such as multi-segment reconstruction or motion
correction techniques.
Volume coverage Refers to the length of the detector array along the z-axis and
can be calculated from the sum of the dimensions of each
detector row along this axis.
z-axis The direction that the scanning table travels in (i.e. head to toe or
vice versa).
10
1. Report summary
This report discusses the technical approaches used to address the challenges of CCTA
imaging and compares the relevant technical specifications of scanners available on the
market suitable for cardiac imaging of challenging patient groups.
A full comparison of eleven CT scanner models is included in this report from the four
main CT scanner manufacturers (GE, Philips, Siemens and Toshiba) contacted. All these
CT scanners are available on the UK market. An additional two scanners are only
described briefly due to lack of data at the time this report was compiled. The vast
majority of the data included in the report were provided by the manufacturers via
responses to a questionnaire1 (Appendix 2).
The principal performance parameters compared in this report are: temporal
resolution, spatial resolution, anatomical volume coverage, and x-ray flux, because
they are considered key in CCTA imaging in the difficult to image subgroups. This
approach was adopted because it helps to clarify the impact of individual technical
features on the quality of the images produced. However, it should be noted that
technical specifications do not provide a complete description of the performance of a
CT scanner and therefore should not be used to draw final conclusions on the
comparative clinical performance of the different scanners. Other technical and non-
technical factors such as post-processing tools, patient characteristics and disease
state also have an impact on the overall clinical performance of the scanner.
1.1 Comparative technical specifications
The key technical specification data for each scanner are tabulated in chapter 6. All
models are third generation CT scanners and have between 64 and 320 detector rows
with dimensions of between 38.4 mm and 160mm in the craniocaudal (z-axis) direction.
Minimum gantry rotation times range from 250 ms to 420 ms. Two models are dual
source systems2. Other specifications tabulated and discussed are: gantry bore diameter,
maximum scan field of view, weight allowed on couch, x-ray generator power, tube
voltage range and tube current range.
1 At the time of the exercise we were informed by some manufacturers that not all data requested
was available, particularly on their newest systems. Additionally, for some of the claims made,
published peer-reviewed evidence is still not available.
2 Although the manufacturer of dual source systems now offers a second dual source model, at
the time this report was compiled sufficient technical data was not available to include this
model in the full comparative specifications.
11
1.2 Approaches used by the manufacturers to
address the challenges in CCTA of difficult to
image patients
CT manufacturers have taken different approaches to meeting the challenges posed by
the various difficult to image patient groups and these are discussed in full in chapter 7.
They are summarised in the paragraphs below.
Whilst each group is discussed separately it is certainly possible that patients will present
with more than one challenge and when scanning such patients this will need to be
considered.
1.2.1 Patients with high heart rates
High temporal resolution and selection of the most static cardiac phase for image
reconstruction are essential for obtaining CCTA images that are free from cardiac motion
artefacts.
The scan mode used is often dependent on the patient’s heart rate. Current
recommendations are to scan patients with lower heart rates with prospectively ECG-
triggered axial (PTA) mode, or prospectively ECG-triggered helical (PTH) mode where
this mode is offered, as these tend to be lower dose modes. However, these modes
generally allow less flexibility in the cardiac phases available for image reconstruction.
Dual source systems, such as the Siemens Somatom Definition Flash Stellar and
Siemens Somatom Force, have a high intrinsic temporal resolution, of particular value for
imaging patients with high heart rates, and on the these systems CCTA scans can be
performed using PTA scan mode even at heart rates above 85 bpm.
Scanning using the same scan mode regardless of a patient’s heart rate is possible on
scanners with a large z-axis coverage, for example the GE Revolution CT and Toshiba
Aquilion ONE models, where scanning in PTA Volume mode is recommended even at
heart rates greater than 75 bpm.
On the Philips Brilliance iCT it has been shown that diagnostic CCTA scans can reliably
be achieved in PTA mode at heart rates up to 75 bpm but in practice diagnostic images
have been obtained at higher heart rates.
A software solution for dealing with motion artefacts is offered on GE systems. The
SnapShot Freeze (SSF) motion correction algorithm is claimed to provide an effective
temporal resolution of 29 ms although peer-reviewed evidence is not yet available.
Toshiba have also recently implemented Adaptive Motion Correction, an algorithm that
compensates for motion in the vessels, myocardium and valves.
12
1.2.2 Arrhythmia
The issue of whether CCTA is a suitable examination for excluding CAD in patients with
atrial fibrillation remains contentious (Vorre et al, 2013; Schuetz et al, 2013). Limited
evidence is available, largely due to the relatively recent technological developments that
have allowed improvements in this area.
GE claims that its adaptive gating algorithm can predict the heart rate of the next cardiac
cycle allowing the scanning window to be adapted accordingly. In addition GE
recommends the use of the SSF motion correction algorithm and multi-segment
reconstruction techniques for improving the effective TR. The Revolution CT scanner
provides a ‘Smart arrhythmia management’ tool that avoids scanning during an irregular
beat.
The Philips iCT Elite can cover the cardiac volume in two to three heart beats in PTA
scan mode and studies (on the Brilliance iCT) have shown that patients with atrial
fibrillation can be successfully imaged (Muenzel et al, 2011, Chao et al, 2010). All Philips
CT scanners supporting card CCTA have automated arrhythmia handling tools that
enable diagnostic quality scans through detection or rejection of ectopic beats.
A meta-analysis has shown that the high intrinsic temporal resolution of the Siemens
dual source scanner makes it suitable for ruling out CAD in patients with atrial fibrillation
(Sun G et al, 2013). On all Siemens cardiac scanner models the ‘adaptive sequence’
axial mode will omit or repeat a scan when an ectopic beat is detected. For patients with
known arrhythmia Siemens recommends the use of retrospectively ECG-gated helical
(RGH) mode with automatic temporary suspension of ECG-dose modulation if arrhythmia
is detected.
Toshiba’s solution to scanning patients with arrhythmia on the Aquilion PRIME scanner is
to switch from PTH to RGH mode if arrhythmia is detected (and vice versa if the heart
rate returns to normal). On the Aquilion ONE scanners, exposure is delayed if an
arrhythmia occurs and will not take place until the heart returns to its normal rhythm. A
recent study which included a small number of patients with heart rhythm irregularities
(Uehara, 2013) showed that equivalent results could be obtained in these patients as on
those with regular heart rhythm.
1.2.3 High calcium scores
A high spatial resolution can reduce the amount of calcium ‘blooming’ and will be
particularly beneficial in the presence of high calcium levels. Manufacturers therefore
generally recommend the use of sharper reconstruction filters in these cases, although
this will result in increased image noise.
13
Other than a high intrinsic spatial resolution, a good temporal resolution is important to
reduce blurring due to motion, and other features such as scanning at higher tube
potentials (kVs) are recommended. Additionally, the use of iterative reconstruction
algorithms, calcium subtraction techniques and dual energy scans, where available, are
proposed.
GE recommends the use of the High Definition (HD) mode available on the Discovery
CT750 HD3 and the Revolution CT. On the CT750 HD use of the dual energy cardiac
mode (GSI Cardiac) to obtain monochromatic (keV images) is also recommended.
Philips claims that the use of iterative algorithms, iDose4 or IMR (Iterative Model-based
Reconstruction) results in improved spatial resolution and therefore reduced blooming
(Philips, 2011).
Siemens recommends using SAFIRE or ADMIRE iterative reconstruction for reduced
blooming in patients with high calcium scores.
On Toshiba scanners, the iterative reconstruction algorithm AIDR 3D which includes
artefact reduction software is recommended. Toshiba also recommends using SURESubtraction Coronary for removing calcium, an approach claimed to have fewer
drawbacks than thresholding techniques.
1.2.4 Coronary artery stents
In common with high calcium scores, the presence of stents can also lead to beam
hardening artefacts and blooming that can make the diagnosis of in-stent restenosis
difficult. Different types of stent materials result in varying levels of blooming but
generally image interpretation becomes problematic for stent diameters less than 3 mm.
Manufacturers generally use the same approaches for dealing with stent imaging as for
high calcium levels. Toshiba’s SURESubtraction Coronary can also be used to remove
stents and check patency, an approach claimed not to be possible with thresholding
techniques alone.
1.2.5 Patients with coronary artery bypass grafts (CABG)
The challenges in the assessment of native coronary arteries in patients after CABG (as
discussed in section 7.5) are related to poor run-off, more extensive calcification and
3 The GE Discovery CT750 HD has been replaced by the Revolution GSI and Revolution HD
CT scanner models. The GSI has dual energy capability whereas the HD does not.
14
diffusely narrowed arteries with small dimensions (Sun et al, 2012). Additionally scanning
of longer lengths to include the whole thorax may be required.
On the Revolution CT scanner GE suggests the use of two table positions with ‘Smart
collimation’ to enable one beat acquisition of the heart and avoid a volume boundary over
the heart.
On the Siemens Somatom Definition Flash Stellar the use of high pitch helical ‘Flash’
mode is claimed to offer diagnostic quality when extended coverage is required.
The results of a small study with the Philips Brilliance iCT scanner showed that CCTA
with a low dose protocol achieved a high diagnostic accuracy in patients with CABG
(Aunt Minnie, 2014).
On the Toshiba Aquilion ONE, upon setting the desired range, Wide Volume mode
automatically calculates the two acquisitions volumes required to cover the heart and
graft. In addition Toshiba claims that AIDR 3D iterative algorithm and a large range of
reconstruction filters available will aid in providing good image quality in CCTA on
patients with CABG.
1.2.6 Obese patients
Manufacturers generally suggest the use of a higher tube potential if an increased photon
flux is required at the detectors. However, the introduction of iterative reconstruction
algorithms has somewhat reduced the problem of photon flux limitation even in obese
patients as, with the same exposure settings, lower noise values are achieved than with
traditional filtered back projection (FBP) methods. The approaches suggested by
manufacturers to scan obese patients are summarised in the following paragraphs.
GE promotes the use of the ASIR iterative reconstruction algorithm, or the more
advanced ASIR-V on the Revolution CT. Another suggested approach for achieving an
increased photon flux in obese patients is to decrease rotation time, which will result in a
higher tube-current-time product (mAs). However, the reduced noise will be achieved at
the expense of temporal resolution. On the Revolution CT automatic exposure control
(AEC) is available for CCTA scans and so the tube potential (kV Assist) and tube current
(Auto mA/Smart mA) will automatically be optimised for patient size.
Philips proposes the use of the iterative algorithm iDose4 and claims that it facilitates
In CCTA these two scan modes can be further subdivided, and the various modes are
summarised in
Figure 4 and Table 2 .The advantages and disadvantages of these scan modes are given
in Table 3.
C. Retrospectively ECG gated helical (PGH) scan mode
(i) Without ECG-gated tube current modulation
(ii) With ECG-gated tube current modulation
D. Prospectively ECG-triggered helical (PTH) scan mode
(i) Low pitch
(ii) High pitch
27
C. Retrospectively ECG gated helical (PGH) scan mode
(i) Without ECG-gated tube current modulation
(ii) With ECG-gated tube current modulation
D. Prospectively ECG-triggered helical (PTH) scan mode
(i) Low pitch
(ii) High pitch
A. Prospectively ECG-triggered axial (PTA) scan mode
(i) No padding
(ii) With padding
B. Prospectively ECG-triggered axial ‘Volume’ (PTAV) scan mode
(i) No padding
(ii) With padding
28
Figure 4 Scan modes employed in coronary CT angiography A & B show axial (sequential) scan modes
and C & D show helical scan modes
29
Table 2 - Description of scan modes used in CCTA.
Scan mode Description
A Prospectively ECG-triggered axial
(PTA) (Fig. 4A)
The scan mode is based on prospective sampling of the patient’s ECG to predict the
optimal phase of the cardiac cycle for data acquisition, determined by the heart rate.
The x-rays beam is triggered at the pre-defined phase of the cardiac cycle and x-ray
exposure only occurs during the period required for data acquisition. The x-ray
exposure window may be extended to cover a longer phase of the cardiac cycle
(referred to as padding) to allow for a greater range of cardiac phases available for
image reconstruction.
Following the exposure of a slab of the cardiac volume, the couch position is
incremented during the subsequent heartbeat, and then the adjacent slab is exposed
during the same cardiac phase as previously. This process is repeated over various
beats until the whole cardiac volume is imaged.
Prospectively ECG-triggered axial
Volume (PTA V) (Fig. 4B)
This mode is similar to the PTA mode described above but no couch incrementation
is required. It is a mode available on CT scanners with an x-ray beam of sufficient
width to acquire the whole cardiac volume in a single rotation within one heartbeat.
Similarly to PTA this mode can operate with or without padding. This scan mode has
a lot of flexibility; the whole cardac cycle can be exposed with or with ECG-gated tube
current modulation. Alternatively data from multiple heartbeats can be acquired and
combined for improved temporal resolution.
Helical
Low pitch (~0.15 – 0.3)
Retrospectively ECG-gated helical
Without or with ECG gated tube
current modulation (RGH) (Fig. 4C)
In this mode the cardiac volume is continuously irradiated while the couch moves at a
low pitch through the gantry bore. Data are acquired throughout the whole cardiac
cycle for image reconstruction in any cardiac phase.
More commonly, this mode is employed with ECG-gated tube current modulation.
The ECG signal is used to select the cardiac phase for viewing the coronary arteries
and outside this phase the mA is reduced The width of the maximum mA window can
be increased for more flexibility of cardiac phases available at full image quality.
There is also flexibility in the % reduction of mA outside the maximum mA window.
Low pitch (~0.15 – 0.3)
Prospectively ECG-triggered helical
(PTH – low pitch) (Fig. 4D(i)
This mode is similar to prospectively ECG-triggered step-and-shoot mode in that
irradiation occurs only during a predetermined phase of the cardiac cycle, but with the
couch continuously moving through the gantry at a low pitch. Padding can be used for
greater flexibility in cardiac phases available for image reconstruction.
High pitch ( >3)
Prospectively ECG-triggered helical
(PTH – high pitch) (Fig. 4D(ii)
This mode is only available on dual source CT scanners. The couch moves
continuously at a high pitch through the gantry and the x-rays are triggered at a pre-
defined phase of the cardiac cycle. The whole cardiac volume is acquired within a
single heartbeat.
30
Table 3 - Advantages and disadvantages of various CCTA scan modes. Qualitative comparisons of dose are given. They provide guidance of relative dose in different scan modes for a given manufacturer, assuming the same patient characteristics
Scan mode Advantages Disadvantages
Axial
Prospectively ECG-triggered axial
(PTA)
Significantly lower radiation dose than
retrospectively gated helical.
Elimination of helical artefacts where
relevant.
Not recommended for high heart rates
Functional cardiac data not usually
available.
Slightly longer scan time than
equivalent helical scan.
Prospectively ECG-triggered axial
Volume
(PTA V)
Single heart beat acquisition eliminates
misregistration and step artefacts
particularly for irregular heart rhythm.
Short overall scan time.
Elimination of helical artefacts where
relevant.
Use of multisegment reconstruction can
improve temporal resolution
Low dose mode
Cone beam artefacts may be an issue.
Helical
Low pitch (~0.15 – 0.3)
Retrospectively ECG-gated helical
Without ECG gated tube current
modulation
(RGH)
Functional data available.
Use of multisegment reconstruction can
improve temporal resolution
Highest radiation dose.
Low pitch (~0.15 – 0.3)
Retrospectively ECG-gated helical
with ECG-gated tube current
modulation
(RGH + ECG mA modulation)
Functional data available.
Where a low minimum mA is available
in ECG-gated tube current modulation,
doses can approach PTA scans.
Increased radiation dose.
Low pitch (~0.15 – 0.3)
Prospectively ECG-triggered helical
(PTH – low pitch)
Shorter overall scan time than PTA
mode scan with equivalent radiation
dose.
Functional cardiac data not available.
High pitch ( >3)
Prospectively ECG-triggered helical
(PTH – high pitch)
Single heart beat acquisition eliminates
misregistration and step artefacts
Short overall scan time.
Low dose mode.
Not suitable for high heart rates due to
cardiac phase difference in images
acquired within one heartbeat.
Functional cardiac data not available.
31
5. Radiation dose in CCTA
MDCT has been one of the most significant areas of progress in medical imaging over
the past 10 years. Manufacturers of CT scanners have invested effort in research and
development of CT technology and each generation of scanners has offered improved
hardware and software features aimed at increasing the accuracy of cardiac imaging.
As discussed in the following chapters the major advantages of modern CT scanner
systems are high temporal resolution and high spatial resolution combined with fast
volume coverage. These features meet the challenges of cardiac imaging and provide
fast acquisition of non-invasive images of the heart and coronary arteries with
exceptionally good quality and accuracy. However, the radiation dose to patients
undergoing CCTA examinations is an additional challenge that needs to be considered.
Dose in CCTA depends on various factors such as the type of scanner and dose
reduction features available, the protocol used, including patient preparation, and the
patient body habitus and stage of disease.
Major contributions to dose reduction have been provided by manufacturers of CT
scanners and also by users and clinical research groups who have tested the available
tools and further developed optimised protocols for CCTA. Important progress has been
achieved in reducing dose in CCTA with manufacturers claiming dose reductions in
CCTA down to sub milliSievert levels with no detrimental effects on image quality. Figure
5 shows example CCTA images with diagnostic quality obtained at standard and ultra-
low dose. Despite this, doses to patients in the ‘challenging to image’ subgroups can be
higher than average as the lowest dose protocols cannot always be used.
In this report a comparison of patient doses for the different scanner models has not
been carried out. It was considered that it would not be possible to perform a fair inter-
system comparison using evidence from the literature or from information provided by the
manufacturers. For this reason the remainder of this section focuses on describing the
effect on dose from various scan protocols and on the technical features available for
dose reduction, but makes no comparisons between doses achieved on different
systems.
32
Figure 5 Three-dimensional volume rendered images from standard (A-C) and ultra-low dose (D-
F) coronary computed angiography of the same patient acquired with 1.2mSv and 0.2mSv. CCTA
was performed on a 64 slice CT scanner (Discovery CT750 HD, GE Healthcare) using prospective
ECG-gating. Adaptive statistical iterative reconstruction (ASIR; GE Healthcare) was used to
reconstruct standard dose images and ultra-low-dose images were reconstructed using a new IR
algorithm (MBIR5; GE Healthcare). The effective radiation dose from CCTA was calculated as the
product of the dose-length product (DLP) and a conversion coefficient for chest [k=0.014
mSv/(mGy.cm)] (Fuchs et al, 2014).
There is a growing body of evidence that the appropriate use of dose reduction strategies
in CCTA can result in substantial dose reduction whilst maintaining diagnostic image
quality. Some of these strategies are available to users on all manufacturers’ CT scanner
systems, whereas a subset is manufacturer specific.
5 It should be noted that MBIR (known as Veo on commercial GE systems) is not currently
available in cardiac scanning on commercial GE CT scanners.
33
5.1 Justification of the CCTA examination
The first step to limit the radiation burden of CCTA is to adhere to appropriate use criteria
and guidelines for the procedure (Taylor AJ, 2010). The CCTA examination is generally
recommended for diagnosis and risk assessment in patients with low or intermediate risk
or pre-test probability of coronary artery disease. It is also appropriate for structural and
functional evaluation.
NICE recommends CCTA (using a 64-slice scanner or above) to assess arteries and
identify significant stenosis in people with an estimated likelihood of coronary artery
disease (CAD) of 10–29% and a calcium score of 1–400 (NICE, 2010). CCTA is
recommended as an option for first-line evaluation of disease progression, to establish
the need for revascularisation, in people with known CAD.
Various other indications have been investigated and shown to have great promise such
as the follow-up of bypass grafts and the evaluation of ventricular function and cardiac
valves. New indications are currently being investigated to extend cardiac CT to the
assessment of myocardial perfusion and viability. CCTA is not recommended as a first
line option in situations in which immediate revascularisation is being considered and
should not be used as a screening tool.
5.2 Limiting Z-axis scan length
A simple and effective way to reduce radiation dose associated with cardiac CT is to limit
the coverage of the anatomy along the craniocaudal axis (z-axis) to the minimum
required. The extent of the region of interest of the coronary arteries needing to be
imaged may be selected (and reduced) using, for example, calcium scoring images
acquired before the contrast-enhanced CCTA examination.
5.3 Use of small scan field of view
In CCTA protocols a small scan field of view is frequently specified as this utilises the
‘small’ beam shaping filter if one is available. These beam shaping filters have a greater
thickness of attenuating material in the periphery of the full field of view, thereby reducing
dose outside the central region of interest.
5.4 Tube current modulation
The main principle of tube current modulation (TCM) techniques is to vary the tube
current (mA) depending on patient attenuation characteristics. The mA modulation can
be done across the patient anatomy or in specific intervals of the cardiac cycle (temporal
34
modulation), or both. Currently, manufacturers provide a range of tube current
modulation techniques and an understanding of their operation principles and pitfalls is
essential for their proper use.
5.4.1 Automatic (spatial) tube current modulation (ATCM)
In this approach, rather than delivering a constant mA throughout the entire scan, the mA
is varied according to body size, shape and attenuation properties of the anatomy.
Typically, mA is increased for thicker (more attenuating) regions of the anatomy
compared to thinner (less attenuating) regions. The tube current can be modulated in the
x-y plane (angular modulation), the z-axis (longitudinal modulation), or a combination of
the two.
To produce images with more uniform noise levels, and consequently adequate
diagnostic quality at reduced dose some form of image quality index has to be defined in
the protocol by the user. This image quality index is termed “noise index” by GE,
“mAs/rotation” or “dose index” by Philips, “quality reference mAs” by Siemens and
“standard deviation” by Toshiba.
All manufacturers enable ATCM on their scanners (termed AutomA/SmartmA by GE,
DoseRight by Philips, CAREDose 4D by Siemens and SureExposure by Toshiba). GE
currently does not allow automatic tube current modulation for CCTA scans on the
Optima 660 or Discovery CT 750HD scanners models, but currently only on the
Revolution CT. On the former two scanners, mA and kV are set from a user prescribed
mA/kV table through CT localiser radiograph scout-based BMI or through user input BMI.
Most ATCM techniques are based on calculations obtained from the CT localiser
radiograph and when using this technique it is very important to ensure that the patient’s
anatomy is well centred (at the isocentre) in the gantry.
The appropriate image quality may vary according to patient size, anatomic region, tube
potential, and diagnostic task and the image quality index should be set accordingly.
Recommended indices are usually provided by the manufacturer, and optimisation
studies are important to investigate their adequacy for non-standard patients (e.g. those
with a high BMI). Considering that user input is required, this technique is in fact not truly
“automatic” and its outcomes are significantly dependent on the user’s understanding
and experience of the technique.
35
5.4.2 Temporal (ECG-gated) tube current modulation
The general principle of retrospectively ECG-gated helical (RGH) scan acquisitions with
ECG-gated tube current modulation consists of decreasing the tube current (mA) outside
the cardiac phase in which data for reconstructing the coronary arteries are acquired.
Accurate evaluation of coronary arteries is typically dependent upon diagnostic
information obtained in late diastole (usually at 70% or 75% of the R-R interval) however,
at higher heart rates, the end systole is often the most diagnostic and motion-free phase
for evaluation of the coronary arteries (Hausleiter et al, 2006). Using this knowledge,
temporal tube current modulation techniques based on the patient’s ECG, have been
developed with the aim of reducing dose to the patient.
The user can define the percentage level of mA reduction and the width of the maximum
mA window within the cardiac cycle (Figure 6). The mA can generally be reduced to 20%
of the maximum mA value set, or on some scanners, down to 5%. Individual patient
characteristics (e.g. heart rate, arrhythmia) influence the adequacy and effectiveness of
this technique to reduce patient dose. It is more efficient in patients with lower and stable
heart rates as lower levels of mA and narrower window widths can be used.
For patients with a high heart rate (e.g. > 80 bpm) and unstable heart rhythm RGH with
ECG-gated tube current modulation is usually not recommended because CCTA images
may need to be reconstructed outside this phase.
Studies using ECG-gated tube current modulation reported overall potential dose savings
of approximately 20% to 50%, depending on the exact protocol used (Entrinkin et al,
2011).
Figure 6 - Diagram illustrating ECG-based modulation of the X-ray tube current. The width of the
temporal window with maximum tube current (Max mA) can be selected by the user, while the
temporal width of the image reconstruction window is fixed. The level of the minimum mA can also
be varied down to 5% 0r 20% depending on the scanner model.
Max mA
Time (s)
Min mA
X-rays ‘on’
Image reconstruction phase
36
5.5 Optimisation of tube potential
It has been routine to scan adult patients using a tube potential of 120 kV. The principle
behind the benefit of lower kV in some clinical applications is that the superior
enhancement of iodine achieved at lower energies potentially results in an improved
contrast to noise ratio. At the same time, if all other parameters are held constant, a
reduction in tube potential from 120 kV to 100 kV results in a reduction in dose of
approximately 30%. This has promoted the investigation of scanning at lower tube
potentials.
Reducing tube potential from 120kV to 100kV has been recommended in CCTA primarily
in non-obese patients (body weight ≤85 kg or BMI ≤30) due to the increase in image
noise above acceptable levels in larger patients. (Hausleiter et al, 2006) (Kalendar et al,
2009) (Blankstein et al, 2011).
A non-randomised study of CCTA with DSCT on 294 patients (both prospectively ECG-
triggering and retrospectively ECG-gating protocols were included) also showed that
lowering the kV from 120kV to 100kV resulted in a substantial reduction in radiation dose
of approximately 45% (Figure 7) with no compromise in diagnostic performance. Despite
higher image noise being observed in the images produced at 100 kV compared to 120
kV, the contrast-to-noise ratio (CNR) and the signal-to-noise ratio (SNR) (Table 4) were
considerably higher at the lower potential.
Figure 7 - Radiation dose of CCTA using 100 kV versus 120 kV. The observed dose reduction was
statistically significant for all scans acquired with a retrospectively gating and prospectively
triggering acquisition protocols (Blankstein et al, 2011).
37
Table 4 - Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) was calculated using the
mean density of the contrast-filled left ventricular chamber and mean density of the left ventricular
wall (adapted from Blankstein et al, 2011).
Parameter 100 kV 120 kV p-value
CNR* 6.9 6.0 0.0176
SNR 9.4 8.3 0.0244
At present there is growing interest in lowering the kV even further, to 80 kV in paediatric
patients and in selected groups of smaller adult patients. Promising results have been
reported. Jun et al, 2012 showed a 70% and 88% dose reduction with an 80 kV protocol
and with 80 kV and ECG-based tube current modulation, respectively, compared to the
120 kV protocol in adults with normal BMIs. Importantly, the benefits of lowering the kV in
CCTA from 100 kV to 80 kV have been demonstrated across multiple CT systems
(Labounty et al 2011). Additional benefits of using 80 kV protocols include the potential to
significantly reduce contrast agent doses in CCTA (Cao et al, 2014).
The methods reported for selection of tube potential for individual patients vary but
usually include a metric of body habitus (BMI, physical inspection by a clinician and/or
visualization of scout and axial bolus test images prior to the scan). This is a subjective
process and it is important to develop more objective recommendations for selection of
tube potential as this seems to be a key aspect to promote the effectiveness of the
technique.
All Siemens cardiac scanners models included in this report have automatic kV selection,
CAREkV, available. More recently GE and Toshiba have also introduced automatic kV
selection on their scanners. These are named kV Assist and SUREkV respectively.
iPatient is available on all Philips scanners which have 64 slice capability or greater
utilizing Exam Cards claimed to facilitate patient-specific dose management.
Further evidence is required to better understand the utility and limitations of using lower
kV techniques in CCTA and the potential additional benefits of combining this with other
dose reduction techniques.
38
5.6 Prospectively ECG-triggered scan modes
As described in section 4.2 various scan modes are available for performing CCTA scans
and the advantages and disadvantages of the various modes were summarised in Table
3.
In general, prospectively ECG-triggering, whether in axial or helical scan mode (Figure 4
A, B and C) will result in a significant reduction in radiation dose compared to
retrospectively gated scan mode (Figure 4,C) as only the cardiac phase to be used for
image reconstruction is irradiated. In the past these scan modes were limited to patients
with low heart rates, but with increases in temporal resolution and z-axis detector
coverage the limiting heart rate for prospectively ECG- triggered scan modes is being
raised.
An important consequence of the prospectively ECG-triggered techniques is that the
radiologist may be entirely dependent on images acquired from a single cardiac phase to
make the diagnosis as no data are acquired in the other phases. However, users can
define an additional amount of exposure time (padding) just before and just after the
selected phase that allows for greater flexibility in phase reconstruction. Padding results
in additional exposure to the patient (Figure 4 A ii).
In prospectively ECG-triggered axial mode (PTA) all cardiac CT scanners provide a user-
definable padding parameter. For example if data for image reconstruction can be
obtained in 220 ms and the user defines 50 ms of padding the total exposure time will be
320 ms (padding is added before and after the selected cardiac phase). If motion
artefacts are present in the primary phase selected for reconstruction, the other phases
may have no such limitations and may therefore provide diagnostic information. The
amount of padding should be selected taking into consideration the patient
characteristics and diagnostic task. Low heart rates are essential if using little or no
padding. A systematic review showed that use of PTA leads to a significant reduction in
radiation dose compared to RGH while offering comparable image quality and diagnostic
value (Sun and Ng, 2012).
Scanners with a detector extent of 160 mm in the z-axis generally employ the so-called
Volume scan mode. This is a fundamentally similar approach to conventional PTA
scanning but the whole cardiac anatomy is generally acquired within one heart beat with
no requirement for couch incremantation to cover successive portions of the heart. This
mode results in a very short scan time and does not suffer from banding artefacts due to
spatial misregistration of data from successive heartbeats.
The high pitch, prospectively ECG-triggered helical (PTH) mode involves scanning at a
pitch of approximately 3.4 which is a factor of more than 10 times higher than pitch
values typically used in retrospectively ECG-gated (RGH) mode (ranging from ~ 0.15 to
0.3). This allows the whole cardiac volume to be acquired within a single heartbeat
(Figure 4 D ii) and Figure 8). The implementation of this technique requires a dual source
39
CT scanner with a table technology that allows extremely fast movement and specific
reconstruction algorithms that combine data acquired by the two sets of detectors. To
achieve the required image quality this scan mode is generally limited to a low, regular
heart rate (stable HR ≤60 bpm) typically requiring the administration of beta blockers
(Entrikin et al, 2011).
.
Figure 8 - Prospectively ECG-triggered helical acquisition at high pitch minimises dose through a
combination of extremely short exposure time (280 ms in the example), dynamic pre-patient
collimation (eliminating overscanning) and near complete elimination of overlap in the cone beam
thereby minimising extraneous patient exposure (adapted from Entrinkin et al).
One of the most recently introduced scan modes is prospectively ECG-triggered helical
mode at a low-pitch (ranging from 0.15 to 0.3) with radiation exposure only occurring
during the phase required for imaging the coronary arteries. Similarly to padding in PTA
mode, the width of the exposure window can be increased to allow for more flexibility in
the cardiac phase available for image reconstruction. Toshiba is currently the only
manufacturer that enables this CCTA scan mode. However, no evidence was identified in
the literature search undertaken on clinical use of this mode.
5.7 Development of standardized protocols
The techniques briefly described above offer opportunities for dose reduction in CCTA.
However it is through the combination of various strategies that the greatest potential for
dose reduction can be achieved.
40
LaBounty and colleagues compared the use of standardized and non-standardised
CCTA protocols at three sites equipped with single-source 64-slice CT scanners (GE
750HD) (LaBounty et al, 2010 - a). Standardized protocols considered patient heart rate,
BMI and clinical indications to systematically define the use of prospective (versus
retrospective) triggering, tube potential, tube current, padding duration and ECG-based
tube current modulation. The scanners offered iterative reconstruction (IR) but this was
not used. Therefore, the results are more generalizable to others scanners.
The standardized protocols resulted in an overall reduction in estimated effective dose of
65% when compared with the non-standardized approach. Additionally, no differences in
the sensitivity, specificity and accuracy were found between the two protocols.
The authors recommended that tube current should be specified based on chest wall
diameter, BMI or estimated noise measures and should vary on a case by case basis.
Tube potential should also be based on a patient’s BMI, and z-axis coverage should be
determined by the coverage clinically required. The authors proposed standardization of
scan parameters (Table 5) for CCTA on patients with controlled heart rates (≤65 bmp) for
use on most 64-MDCT scanners with cardiac capabilities. It should be noted that whilst
the methodology reported in this study may be applicable to the optimisation of protocols
on other CT scanner models the specific optimised protocol parameters must not be
transferred to other models without validation. In particular the tube current (mA) settings
will vary between scanner models.
The above approach is recommended on scanners that do not have the capability for
automatic tube current modulation and/or automatic tube potential selection. The
automated approach is considered to be superior to manual selection.
Table 5 – Example of a proposed standardised CCTA protocol. The authors of the study
recommend that prospectively ECG-triggered axial (step and shoot) should be utilised in all
patients with regular heart rates (≤65 bpm). If heart rate is consistently ≤ 55 bpm and regular, use
of padding is not recommended. *80kV could be considered in patients with a small thoracic
diameter and body mass index <25 kg/m2. PSS – prospective step and shoot (or PTA, using the
terminology adopted in this report) (adapted from LaBounty et al, 2010 a).
41
Figure 9 - Examples above are curved multiplanar reformats through the right coronary artery
generated from PSS (i.e. PTA) scans obtained on a 64-slice scanner without the use of IR
techniques in three different patients. Both A and B were scanned with identical imaging
parameters. However, in B the entire desired z-axis coverage was obtained in three axial steps
resulting in nearly 25% lower dose compared with A. In C, the sub milisievert dose estimate was
achieved through a combination of PSS technique, decreased mA, no padding and decreased
scan length (shows the improvement in image quality following the application of a standard dose
reduction protocol) (Entrinkin et al, 2011).
As discussed, a good understanding of the impact of scan parameters and scan modes
on image quality and patient dose is important to optimise CT protocols.
Centres that wish to optimise their protocols may start by establishing standard protocols
for groups of patients with similar characteristics. A further step in the optimisation
process is tailoring the CT scanner parameters to the individual characteristics of each
patient. To help users in this task, manufacturers provide features on their scanners to
perform automatic selection of various key scanning parameters and scanning modes
based on the characteristics of individual patients. Examples of these features are:
kV selection according to patient size and clinical application;
mA selection and mA modulation according to patient attenuation;
ECG-triggering phase according to heart rate;
Width of maximum mA window in RGH mode according to heart rate;
Amount of padding according to heart rate.
Automatic selection of scanning parameters in CCTA may be particularly useful for
patients presenting with non-standard characteristics (e.g. high BMI, HR pattern)
providing images with adequate quality consistently, whilst taking patient dose into
consideration.
42
5.8 Iterative reconstruction (IR)
Traditionally filtered back projection techniques have been employed to transform the
raw image data obtained on a CT scanner into a final reconstructed image. However,
these techniques employ various assumptions regarding the underlying physics
processes and CT geometry that typically result in increased noise and less accuracy in
the reconstructed images (Entrinkin et al, 2011). In contrast IR techniques use a
statistical model of photon noise and may also include modelling of scanner
characteristics to allow a more accurate integration of the physics and scanner geometry
(Moscariello et al, 2011).
A drawback of IR techniques is the computer power and processing time required which
have the penalty of significantly longer reconstruction times and have delayed
introduction of IR in clinical routine.
Despite these challenges all manufacturers have succeeded in developing simplified IR
techniques feasible for clinical use and they are currently available on every CT scanner
platform. The original algorithms used have now been further developed and provide
improved performance in terms of dose and artefact reduction.
IR itself does not reduce patient dose. However, the noise reduction it provides allows for
the reduction of scan parameters (e.g. mAs, kV) during image acquisition allowing patient
dose savings.
Each manufacturer has developed specific IR algorithms that operate in different ways
and are implemented at different stages of the reconstruction process. Additionally
different ‘generations’ of IR algorithms are available from each manufacturer. It is
important that the radiologist and CT scanner operator have an understanding of how IR
works on the scanner in question, and that they are satisfied with the resulting image
appearance that may differ from that with which they are familiar.
Manufacturers often use blending of iterative reconstruction and FBP to provide images
that are more clinically acceptable. For example, a high percentage of IR may be used
when scanning obese patients and the mAs is kept constant so that image noise is
reduced without an increase in dose. Alternatively, when scanning patients with normal
or low BMI a certain percentage of IR is used and the mAs reduced so that image quality
is maintained at a decreased radiation dose.
43
5.9 Z-axis overbeaming and overscanning
Overbeaming is the additional dose beyond the edge of the detector rows of a MDCT,
which is required because the x-ray beam penumbra must lie outside the active detector
length (Figure 10). With wider z-axis x-ray beam coverage, the percentage dose
contribution from this penumbra region decreases therefore the contribution of this effect
to unnecessary patient dose has reduced on modern scanners with wide z-axis beam
The Optima 660 is the lowest specification model of the GE scanners included in this
report. It has a minimum gantry rotation time of 350 ms. The detector array consists of 64
x 0.625 mm detector rows providing a z-axis coverage of 40 mm.
Revolution GSI & Revolution HD (replace Discovery CT750 HD)
The Revolution GSI launched in 2014 is a direct replacement for the GE Discovery
CT750 HD. The Revolution HD, also launched in 2014, has the same specifications as
the GSI but without dual energy capability. Both scanner models have new features
including, a touch screen interface in the scanner gantry for improved patient workflow,
cardiac motion correction and software for automatic selection of x-ray tube potential.
Both scanner models have the same minimum gantry rotation time and detector array
configuration as the Optima 660, but a different detector material, Gemstone™, with a
faster response time (Jiang et al, 2008). The scanners can operate in high definition
mode (HD) that provides 2.5 times the number of views per rotation as in its standard
mode. On the Revolution GSI, the faster detector material also enables dual energy
scanning with ‘fast kV switching’, acquiring data at the high and low energies in alternate
views during a gantry rotation.
Revolution CT
The Revolution CT is the highest specification scanner offered by GE. It was launched at
the Radiological Society of North America (RSNA) annual meeting in December 2013.
The system has 256 x 0.625 mm detector rows giving a z-axis coverage of 160 mm
which provides full organ coverage in a single gantry rotation. It uses the same detector
material, Gemstone™, as the CT750 HD and so can achieve the same spatial resolution
through the use of HD mode. However, it is not yet have dual energy capability using
rapid kV switching. The scanner allows a minimum x-ray tube potential of 70 kV
compared to the minimum 80 kV available on other GE systems.
6.2.2 Philips Healthcare
Ingenuity
The Ingenuity CT scanner has a minimum gantry rotation time of 420 ms. The detector
array consists of 64 x 0.625 mm detector rows providing a z-axis coverage of 40 mm.
iCT Elite (Referred to as Brilliance iCT in following figures)
The iCT Elite has a minimum gantry rotation time of 270 ms enabled by the AirGlide
bearing. The detector array consists of 128 x 0.625 mm rows providing a z-axis coverage
of 80 mm. The scanner has z-axis flying focal spot technology enabling the acquisition of
50
256 slices per rotation. The imaging chain on the iCT Elite is supported by technologies
like the 2nd generation NanoPanel Elite integrated and modular detector, Eclipse
collimation, 2D Antiscatter Grid and iterative algorithms such as iDose4 and model-based
IMR.
IQon Spectral CT
Introduced in 2013, the IQon Spectral CT has a novel dual-layer detector. Dual energy
data can be obtained retrospectively from all scans without the need for special modes.
Due to this detector-based approach there are no special limitations in temporal
registration, field of view, or the use of dose modulation.
IQon Spectral CT is based on the NanoPanel Prism detector which uses an Yttrium-
based top scintillator with 25% higher light output and Elite electronics adapted from the
NanoPanel Elite detector. Philips claim that one of the benefits of a detector-based
approach to Dual Energy CT is that simultaneous detection allows projection-domain
processing without temporal misregistration for high-motion procedures such as cardiac
CT.
The IQon Spectral CT has a minimum gantry rotation time of 270 ms and the dual-layer
detector array provides a z-axis coverage of 40 mm. The IQon includes Philips’ latest
innovations in iterative reconstruction, including IMR (Iterative Modelled Reconstruction).
Due to the incomplete technical data this scanner is not included in further discussions or
comparisons within this report.
6.2.3 Siemens Healthcare
SOMATOM Definition AS+
The Siemens Somatom Definition AS+ is the lowest specification Siemens scanner
included in this report. It has a minimum gantry rotation time of 300 ms. The detector
array consists of 64 x 0.6 mm detector rows giving a z-axis coverage of 38.4 mm. The
scanner utilizes a flying-focal spot in the z-axis (z-sharp) to double-sample the detector
rows, so 128 slices are acquired per gantry rotation. As for all other Siemens scanners
included in this report, this model has a minimum tube potential of 70 kV.
SOMATOM Definition Edge Stellar
The Siemens Somatom Definition Edge Stellar has many design aspects in common with
the Definition AS+ but with some upgraded features. These include a faster minimum
gantry rotation time of 280 ms and Stellar detectors. The Stellar detector combines the
photodiode with the analogue to digital converter (ADC) in one integrated circuit to
reduce electronic noise and provide sharper slice profiles through reduced cross-talk
between neighbouring detector rows (Ulzheimer et al, 2012).
51
At RSNA 2014 Siemens introduced an upgrade to the Definition Edge Stellar which
enables Dual Energy scanning by splitting the x-ray beam (in the z-axis direction) into a
high and low energy component
SOMATOM Definition Flash Stellar
The Siemens Somatom Definition Flash Stellar is, like its predecessor the Definition
Flash, a dual source CT scanner with two detector-tube assemblies each with 64 x 0.6
mm detector rows, enabling the acquisition of 128 slices per rotation using z-sharp (flying
focal spot) technology, but with the new stellar detector. This scanner also has a
minimum gantry rotation time of 280 ms It has dual energy capability, through operating
the two x-ray tubes at different tube potentials, and allows high helical pitch scanning
(Flash mode) for fast volume coverage. Dual source technology halves the intrinsic
temporal resolution compared to a single source scanner with the same rotation time.
SOMATOM Force
The SOMATOM Force was launched at RSNA December 2013 and is a dual source CT
scanner with two detector-tube assemblies. The system has a reduced gantry rotation
time of 250 ms and therefore an intrinsic temporal resolution of 66 ms. Other upgraded
features include 2 x 96 x 0.6 mm detector rows acquiring 192 slices per rotation using z-
sharp technology, a larger field of view FOV (350 mm) on the second assembly, an x-ray
tube potential range from 70 kV to 150 kV, and a new design of tin filter (Selective
Photon Shield II) which, it is claimed, improves energy separation by 30% for dual energy
scanning.
The scanner also includes a new iterative reconstruction algorithm (ADMIRE) also
available as an upgrade on the Somatom Definition Flash Stellar and Somatom Definition
Edge Stellar models. The system is designed to be a high-end general and cardiac
scanner with a focus on research activities and future applications (e.g. cardiac, dual
energy and preventive medicine).
Due to the incomplete technical data this scanner is not included in further discussions or
comparisons within this report.
6.2.4 Toshiba Medical Systems
Aquilion PRIME
The Aquilion PRIME is the lowest specification Toshiba scanner presented in this report.
It has a minimum gantry rotation time of 350 ms. The detector array consists of 80 x 0.5
mm detector rows providing a z-axis coverage of 40 mm.
52
Aquilion ONE
The Aquilion ONE has the same minimum gantry rotation time as the PRIME, but its
detector array consists of 320 x 0.5 mm detector rows providing a z-axis coverage of
160 mm. This wide detector coverage enables whole organ coverage in a single gantry
rotation.
Aquilion ONE VISION Edition
The Aquilion ONE Vision edition has many specifications in common with the original
Aquilion ONE but with some upgraded features principally a minimum rotation time of
275 ms and a larger, 100 kW, x-ray generator.
53
6.3 Comparison of technical specifications
The principal performance parameters considered key in investigations of coronary artery
disease (CAD) in the challenging patient groups are:
Temporal resolution
Spatial resolution
Volume coverage
X-ray flux
This sub-section includes updated comparisons of these performance parameters. The
data presented was obtained from manufacturers’ responses to the KiTEC specification
questionnaire (Appendix 2). The responses were checked for accuracy and consistency
and additional questions were forwarded to the manufacturers where clarification was
required. Where a manufacturer could not provide a full explanation to a given question
within the timeframe of this project this has been highlighted.
6.3.1 Temporal resolution
A high temporal resolution in conjunction with ECG-gating to reconstruct images in the
optimal phase of the cardiac cycle is essential for sharp depiction of the coronary arteries
with minimum blurring from cardiac motion. This requirement is particularly critical in
patients with high heart rates. In recent years manufacturers have placed strong
emphasis on increasing the temporal resolution of CT scanners through both hardware
and software approaches.
The temporal resolution (TR) of a CT scanner can be considered in terms of the intrinsic
TR and the effective TR. In this report the intrinsic TR is defined as the time taken to
acquire 180° of data, the minimum usually necessary for image reconstruction. The
effective TR is that achieved when various methods for improving the intrinsic TR (e.g.
motion correction algorithms, multi-segment reconstruction) are applied. The intrinsic and
effective TR of the scanner models are tabulated in Table 8 and presented graphically in
Figure 12 and Figure 13.
54
Table 8 - Temporal resolution (TR) for CCTA scans as per information provided by the
manufacturers.
Scanner model
Min.
gantry
rotation
time
(ms)
Intrinsic
TR
(ms)
Maximum
number of
segments -
axial mode
Effective TR -
axial mode (1)
(ms)
Maximum
number of
segments -
helical
mode
Effective TR
- helical mode (1)
(ms)
GE
Optima 660 350 175 1 29 with SSF(2) 4 44 – 175
(29 with SSF(2))
Discovery CT750 HD 350 175 1 29 with SSF(2) 4 44 – 175
(29 with SSF(2))
Revolution CT 280 140 1 24 with SSF(2) N/A N/A
Philips
Ingenuity 420 210 1 210 4 53 - 210
iCT Elite 270 135 1 135 4 34 - 135
Siemens
Somatom Definition
AS+ 300 150 1 150 2 75 - 150
Somatom Definition
Edge Stellar 280 142 1 142 2 71 - 142
Somatom Definition
Flash Stellar (4) 280 75 1 75 2(3)
38– 75 (low pitch)
75 (high pitch)
Toshiba
Aquilion PRIME 350 175 N/A N/A 5 35 - 175
Aquilion ONE 350 175 5 35 - 175 N/A N/A
Aquilion ONE Vision 275 137 5 27 - 137 N/A N/A
1 Range of TRs: minimum is optimal value achieved with maximum available number of segments in multi-
segment reconstruction; maximum value given is for single segment reconstruction 2 SSF – SnapShot Freeze (manufacturer specific term) motion correction software is claimed to remove
blurring due to motion and achieve an effective TR of 29 ms 3 Multi-segment reconstruction is not available in high pitch helical mode 4 Dual source system
N/A Not applicable - scanner not used in this mode for CCTA scans
55
Figure 12 - Intrinsic temporal resolution (TR) of CT scanners (lower values represent a better
temporal resolution).
^ On GE scanners axial mode TR includes use of SnapShot Freeze algorithm. On Toshiba scanners axial mode TR is the optimal value with multi-segment reconstruction using 5 segments.
* In helical mode the optimal TR with maximum number of segments available is given. On GE scanners helical mode TR is with SnapShot Freeze algorithm.
Figure 13 - Effective temporal resolution (TR) of CT scanners (lower values represent a better
temporal resolution).
6.3.2 Spatial resolution
Spatial resolution is another key parameter for successful CCTA imaging. Intrinsic spatial
resolution values quoted are usually measured in stationary phantoms. However, the
effective spatial resolution will also be dependent on the scanner’s temporal resolution as
this affects the degree of additional blurring due to cardiac motion.
0
20
40
60
80
100
120
140
160
180
200
220
GE Optima
660
GE
Discovery
CT 750HD
GE
Revolution
CT
Philips
Brilliance
Ingenuity
Philips
Brilliance
iCT
Siemens
Definition
AS+
Siemens
Defintion
Stellar Edge
Siemens
Definition
Flash Stellar
Toshiba
Aquilion
PRIME
Toshiba
Aquilion
ONE
Toshiba
Aquilion
ONE Vision
Intr
insi
c TR
(m
s)
0
20
40
60
80
100
120
140
160
180
200
220
GE Optima
660
GE Discovery
CT 750HD
GE
Revolution
CT
Philips
Brilliance
Ingenuity
Philips
Brilliance
iCT
Siemens
Definition
AS+
Siemens
Defintion
Stellar Edge
Siemens
Definition
Flash Stellar
Toshiba
Aquilion
PRIME
Toshiba
Aquilion
ONE
Toshiba
Aquilion
ONE Vision
Effe
ctiv
e T
R (
ms)
Temporal resoution: Axial scan mode^
Temporal resoution: Helical scan mode*
56
Spatial resolution must be considered in three dimensions, therefore in this section the
axial (x-y plane) and the z-axis direction resolution are discussed separately to highlight
key developments and facilitate comparison of the various systems. Broadly speaking,
intrinsic x-y plane spatial resolution on CT scanners has not increased significantly in
recent years, as it has always been relatively high compared to that along the z-axis.
The emphasis has been on matching the resolution in the z-axis to that in the x-y plane
to achieve isotropic (equal in all dimensions) resolution. Multi-slice scanners, particularly
those with 64 and more detector rows, are largely capable of meeting this aim.
Axial (x-y plane) spatial resolution
Table 9 compares the principal scanner features that affect the limiting x-y plane spatial
resolution. It also presents data for the spatial resolution achieved with two different
reconstruction kernels recommended by each manufacturer: a standard resolution kernel
for general CCTA scans and a higher resolution (sharp) kernel for CCTA in patients with
stents and/or high calcium scores.
The spatial resolution is defined as the minimum distance (mm) between two small, high
contrast objects that allows the two objects to be resolved in the image. They have been
calculated from the 2% modulation transfer frequency (MTF) values provided by the
manufacturer for the two kernels (standard and high resolution). Where the 2% MTF
value was not given, data was interpolated between the 10% and 0% MTF values
provided. This was a practical approach considered adequate to provide a reasonable
estimate and allow comparison of the data. The spatial resolution values are presented
graphically in
Figure 14.
Some manufacturers may recommend a slightly sharper kernel for improved spatial
resolution at the expense of increased noise.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
GEOp ma660GEDiscoveryCT750HD
GERevolu onCT
PhilipsIngenuity
PhilipsBrillianceiCT
SiemensDefini onAS+
SiemensDefin onStellarEdge
SiemensDefini onFlashStellar
ToshibaAquilionPRIME
ToshibaAquilionONE
ToshibaAquilionONE
Vision
Axialplanespa
alresolu
on(mm)
Standardkernel
Sharpkernel
57
Table 9 - Spatial resolution in x-y (scan) plane as per information provided by the manufacturer.
Scanner models Focal spot
sizes(1) (mm)
Number detectors per row
Sampling density (No. of views per
360, per slice for min. gantry rotation time without FFS(2))
1 Focal spot sizes quoted according to IEC 336/93. Dimensions for small (S) and large (L) focal spot sizes are given 2 Flying focal spot (FFS) in x-y plane, available in cardiac mode on some scanners, oversamples detector elements for increased spatial resolution. On GE scanners this mode is referred to as High Definition (HD) mode. 3 Calculated from 2% MTF values provided by manufacturers for kernel quoted. High resolution kernels may be recommended for patients with stents or high calcium scores to reduce blooming artefacts. 4 Dual source detector system.
N/P data not provided by the manufacturer; N/A – not applicable
58
Figure 14 - Axial (x-y) plane spatial resolution of CT scanners (smaller values represent a better
spatial resolution).
z- axis spatial resolution
The z-axis dimensions and z-spatial resolution of the CT scanners are shown in Table 10.
The z-axis spatial resolution is defined primarily by the z detector dimension but can be
enhanced in a number of ways described in Table 10.
As for the x-y plane, z-axis spatial resolution values are quoted in terms of the minimum
distance (mm) between two small high contrast objects that allows the two objects to be
resolved in the image. These values were derived from the 2% MTF values provided by
the manufacturers. Where the 2% value was not given, data was interpolated from the
10% and 0% MTF values provided. This was a practical approach considered adequate
to provide a reasonable estimate and to allow comparison of the data. The specifications
are compared graphically in Figure 15.
Figure 15 - z-axis spatial resolution of CT scanners (lower values represent a better spatial
resolution).
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
GEOp ma660GEDiscoveryCT750HD
GERevolu onCT
PhilipsIngenuity
PhilipsBrillianceiCT
SiemensDefini onAS+
SiemensDefin onStellarEdge
SiemensDefini onFlashStellar
ToshibaAquilionPRIME
ToshibaAquilionONE
ToshibaAquilionONE
Vision
Axialplanespa
alresolu
on(mm)
Standardkernel
Sharpkernel
0
0.1
0.2
0.3
0.4
0.5
0.6
GE Optima
660
GE Discovery
CT 750HD
GE
Revolution
CT
Philips
Ingenuity
Philips
Brilliance iCT
Siemens
Definition
AS+
Siemens
Defintion
Stellar Edge
Siemens
Definition
Flash Stellar
Toshiba
Aquilion
PRIME
Toshiba
Aquilion ONE
Toshiba
Aquilion ONE
Vision
z-ax
is s
pat
ial r
eso
luti
on
(m
m)
Axial mode
Helical mode
59
Table 10 - Spatial resolution in z-axis direction as per information provided by the manufacturer
1 Calculated from 2% MTF values provided by manufacturer
N/P. data not provided by manufacturer
N/A Not applicable - scanner not used in this mode for CCTA scans
Aquilion ONE Vision 0.5 0.43 N/A ConeXact 3D algorithm
Overlapping image recons every 0.25 mm in axial
60
6.3.3 Volume coverage
Scanners with a wide z-axis detector can acquire the cardiac volume in a smaller number
of heartbeats (ideally a single heartbeat) and CCTA images therefore do not suffer from
misregistration artefacts that may result from irregular heart rates. Dual source scanners
are capable of acquiring the cardiac volume in a single heartbeat by using a high pitch
(>3) (Flash mode) and therefore also have the advantage of eliminating misregistration
artefacts, although the use of this high pitch mode is generally not recommended for high
heart rates.
The z-axis detector length of the various scanner models is presented in Table 11. It is
pertinent to note that quoting the ‘number of slices’, the approach often used to define
the level of a CT scanner, does not necessarily provide an indication of the length of z-
axis coverage.
The number of heartbeats required to scan a cardiac length of 140 mm with various scan
modes is tabulated in Table 12 and shown graphically in Figure 16. In general the
number of heartbeats required the cover a given volume is related to the z-axis
dimension of the detector, with scanners with a z-axis detector dimension of 160 mm
capable of covering the volume within a single heartbeat. The exception to this is if multi-
segment reconstruction is used for high heart rates. The Siemens Dual source scanners
can also cover the cardiac volume within a single heartbeat in their high pitch Flash
mode that is limited to lower heart rates.
61
Table 11 - Length of z- axis detector coverage, number of detector rows, number of slices
acquired per rotation and number of reconstructed slices per rotation as per information provided
by the manufacturer
Scanner models No. of detector rows
No. of slices acquired per rotation
No. of slices reconstructed per rotation (axial mode)
Length of coverage in z-axis (mm)
GE
Optima 660 64 64 64 40
Discovery CT 750HD 64 64 64 40
Revolution CT 256 256 512(1) 160
Philips
Ingenuity 64 64 (2) 64 40
iCT Elite 128 256 (3) 256 80
Siemens
Somatom Definition AS+
64 128 (3) 384 38.4
Somatom Definition Edge Stellar
64 128 (3) 384 (4) 38.4
Somatom Definition Flash Stellar
64 per detector -
tube assembly
128 (3) per detector - tube assembly
384 (4) per detector - tube assembly
38.4
Toshiba
Aquilion PRIME 80 80 160 (1) 40
Aquilion ONE 320 320 640 (1) 160
Aquilion ONE Vision 320 320 640 (1) 160
1 Two overlapping slices per detector row reconstructed in axial mode from raw data
2 Raw data sampling is increased using a high-order interpolator to provide twice the
number of rows of detector data for reduction of ‘windmill’ artefact in helical mode
3 z-axis flying focal spot provides increased sampling by acquisition of two slices per
detector row
4 Three overlapping slices per detector row reconstructed in axial mode from raw data
62
Table 12 - Volume coverage: number of heartbeats required to cover cardiac volume at heart
rates of 60 bpm and 80 bpm as per information provided by the manufacturer.
1 Calculated for pitch recommended in manufacturer’s protocols. Additional rotations required for image reconstruction from helical overscan have not been included in calculations.
N/A. - Not applicable - scanner not used in this mode for CCTA scans: n.a - this mode not available.
Scanner models
No of heartbeats within time taken to scan 140 mm cardiac length
Prospectively ECG-
triggering axial (PTA)
Retrospectively ECG-gated
helical low pitch
(RGH low pitch) (1)
Prospectively ECG-triggered
helical (High or low pitch)(1)
(PTH)
60 bpm 80 bpm 60 bpm 80 bpm 60 bpm 80 bpm
GE
Optima 660 7 N/A N/A 7
(pitch 0.2) n.a. n.a.
Discovery CT750
HD 7 N/A
7
(pitch 0.2) n.a. n.a.
Revolution CT 1 1 N/A N/A n.a. n.a.
Philips
Ingenuity 7 10 10
(pitch 0.16)
13
(pitch 0.16) n.a. n.a.
iCT Elite 3 4 3 (pitch 0.16) 4
(pitch 0.16) n.a. n.a.
Siemens
Somatom
Definition AS+ 7 7 N/A
7
(pitch 0.18) n.a. n.a.
Somatom
Definition Edge
Stellar
7 7 N/A 7
(pitch 0.18) n.a. n.a.
Somatom
Definition Flash
Stellar
N/A 7 N/A N/A 1 (pitch 3.4) n.a.
Toshiba
Aquilion PRIME N/A N/A 5
(pitch 0.22)
7
(pitch 0.2)
5
(pitch 0.22) N/A
Aquilion ONE 1 1 N/A N/A N/A N/A
Aquilion ONE
Vision 1 1 N/A N/A N/A N/A
63
Figure 16 – Number of heartbeats required to scan cardiac 140 mm length in various scan modes.
6.3.4 X-ray flux
The available x-ray flux can be an important consideration in CCTA scanning as the short
gantry rotation times needed for good temporal resolution require a high tube current
(mA) to deliver an adequate signal at the detectors for sufficiently low image noise level.
Achieving a high enough x-ray flux may be a particular problem when scanning obese
patients, although the recent implementation of iterative reconstruction algorithms in CT
(see section 5.8) reduces this problem.
X-ray generator power is often used as a measure of a scanner’s x-ray flux capabilities
as this determines the maximum mA available at a particular tube potential. However, a
better measure of x-ray flux is the dose in a standard phantom (CTDIW) obtained with the
maximum mA and minimum gantry rotation time as presented in Table 13. The caveat to
the approach used here is the assumption that the scanner models all have a similar
dose efficiency. If the scanners vary significantly in terms of dose efficiency then this
measure is not appropriate as a given level of dose will not result in the same image
noise.
Calculations of CTDI were made for two x-ray tube potentials, 100 kV, commonly
recommended in standard patient size CCTA protocols, and 120 kV, the tube potential
typically used in obese patient protocols.
0
2
4
6
8
10
12
GEOp ma660
GEDiscoveryCT
750HD
GERevolu on
CT
PhilipsIngenuity
PhilipsBrillianceiCT
SiemensDefini on
AS+
SiemensDefin onStellarEdge
SiemensDefini onFlashStellar
ToshibaAquilionPRIME
ToshibaAquilionONE
ToshibaAquilion
ONEVision
No.o
fheratbeatsfor140m
mcardiaclength
axial(60bpm)
Axial(80bpm)
lowpitchhelical(60bpm)
lowpitchhelical(80bpm)
highpitchhelical(60bpm)
64
Table 13 - CTDIw for minimum rotation time and maximum tube current available at tube
potentials of 100 kV and 120 kV.
(CTDIw is the absorbed radiation dose in a 32 cm diameter cylindrical Perspex phantom which is approximately
equivalent, in terms of x-ray attenuation, to a large patient. Assuming all the scanner models have a similar dose
efficiency, scanners with a higher CTDIw value will have lower image noise. The CTDIw figures do not rank the scanners
in terms of dose efficiency, but indicate the maximum dose they are able to achieve if used at maximum mA settings in
cases where this might be required e.g. for very large patients.)
Scanner models
X-ray
generator
power
(kW)
x-ray tube potential
100 kV
x-ray tube potential
120 kV
Max tube
current
(mA)
CTDIW (1)
(mGy)
Max tube
current
(mA)
CTDIW (1)
(mGy)
GE
Optima 660 72 480 8.1 600 16.0
Discovery CT750 HD 107 800 12.4 835 20.5
Revolution CT 103 720 N/P 740 17.8.
Philips Ingenuity 80 667 10.6 600 16.6
iCT Elite 120 1000 9.2 1000 15.9
Siemens
Somatom Definition
AS+ 80 650 6.1 666 10.7
Somatom Definition
Edge Stellar 100 650 5.5 800 11.6
Somatom Definition
Flash Stellar
100
(per tube) 600 5.8 800 11.9
Toshiba
Aquilion PRIME 72 600 13.9 600 22.5
Aquilion ONE 70 600 13.2 600 21.4
Aquilion ONE Vision 100 900 15.5 750 21.0
1 CTDIw/100 mAs values were provided by the CT manufacturers
N/P. data not provided by manufacturer
65
7. Technical approaches adopted in current
systems to optimise image quality in ‘difficult
to image’ patient groups
This chapter describes the approaches used by each manufacturer on their cardiac CT
scanner models for imaging the various groups of challenging patients. Although each
patient group is discussed separately it must be noted that some patients will present
with two or more challenges e.g. high heart rate and high calcium levels.
To achieve a diagnostic CCTA scan at the lowest possible dose for each patient requires
not only suitable equipment but also an optimised protocol. A CCTA protocol is
comprised of numerous factors, including patient preparation and iodine contrast
administration, as well as the scan protocol. In CCTA scans the key element of the scan
protocol is the scan mode used. The various scan modes available for CCTA were
described in section 4.2 along with a summary of their comparative advantages and
disadvantages. Choice of scan mode is determined by the scanner model and also by
patient characteristics, primarily the heart rate. Other factors that must be considered in
the scan protocol include the scan range, scan field of view, x-ray beam collimation,
gantry rotation time, x-ray tube potential and tube current, as well as reconstruction
parameters such as slice width, reconstruction algorithm and reconstruction kernel. The
following sections present each manufacturer’s approach for meeting the challenges of
the various patient groups and these are summarised in Table 14 a - f.
7.1 Patients with high heart rates (> 65 bpm)
Obtaining CCTA images that are free from cardiac motion artefacts, particularly on
patients with high heart rates, requires a scanner with a high temporal resolution and
also the selection of the most static cardiac phase for image reconstruction. At lower
heart rates this is generally in diastole (~70% R-R interval), whereas for higher heart
rates end–systole may be the optimal phase (~45% R-R interval) (see Figure 17).
The scan mode used largely determines the flexibility in selection of the cardiac phase
for image reconstruction. Although the scan mode of choice is principally based on the
patient’s heart rate, the cut-off heart rate for a given mode varies with scanner model and
is mainly dependent on the scanner’s temporal resolution. The scan modes
recommended by the manufacturers for different heart rates are given in Table 14.
Generally, the current recommendations are to scan patients with lower heart rates with
prospectively ECG-triggered axial (PTA) mode, or prospectively ECG-triggered helical
(PTH) mode where available, as these are lower dose scan modes. However, these
66
modes generally allow less flexibility in the cardiac phase available for image
reconstruction.
Figure 17 - Coronary CTA on dual source CT scanner with retrospectively ECG-gated helical
mode in a patient with a heart rate of 82 bpm (a) Curved MPR of the RCA at end-systole (45 % of
the RR interval) with excellent image quality (b) Curved MPR of the RCA during diastole (75 % of
the RR interval) displaying poor image quality (Kim HY et al, 2012).
Scanning using the same scan mode regardless of a patient’s heart rate is possible with
certain scanners, and this reduces the possibility of selecting a non-optimal scan mode.
Scanners that have this flexibility are those with a large z-axis coverage, namely the GE
Revolution CT and Toshiba Aquilion ONE models, where scanning in PTA mode is
recommended even for heart rates greater than 75 bpm.
On the GE Revolution CT, at higher heart rates data can be acquired in two cardiac
phases in a single heartbeat through ‘dual-peak phase’ triggering of x-rays in systole and
diastole, with the tube current being reduced to 20% of the maximum value outside these
cardiac phases. The cardiac phase in which the x-rays are triggered can be automatically
selected according to heart rate or else the user can over ride this if preferred. GE also
recommends the use of the SnapShot Freeze motion correction algorithm and multi-
segment reconstruction, techniques for improving the effective TR.
On the Philips Brilliance iCT it has been shown that diagnostic CCTA scans can be
reliably achieved in PTA mode at heart rates up to 75 bpm (Figure 18), but diagnostic
images have been obtained in this mode at higher heart rates.
67
Figure 18 - Philips Brilliance iCT: Receiver-operator characteristic (ROC) curve to establish a cut
off mean heart rate value up to which diagnostic image quality can be consistently achieved using
prospectively triggered axial mode. The curve suggests a cut off of 75 bpm beyond which
coronary image quality was affected (Area under curve [AUC] = 0.92; 95% CI: 0.87, 0.98; P <
0.05) (Hou et al, 2012).
Dual source scanners such as the Siemens Somatom Definition Flash Stellar have a
very high intrinsic temporal resolution. On this system the temporal resolution for all heart
rates and in all scan modes is 75 ms. This scanner can acquire the 180° of data required
for image reconstruction in a quarter of a rotation instead of the usual half of a rotation.
This allows scanning in PTA mode up to heart rates in excess of 85 bpm without the
need to revert to retrospectively ECG-gated helical (RGH) mode (Figure 19) (Xu, 2010;
Sun, 2011).
On Siemens single source systems, the Somatom Definition AS+ and the Somatom
Definition Edge Stellar, diagnostic image quality can be obtained in PTA modes on
patients with heart rates up to 70 bpm and also a slight heart rate variability.
68
Figure 19 - CCTA images acquired on a dual source CT scanner in adaptive sequential mode.
Mean HR = 96 bpm (range 86 – 105 bpm). Image quality was rated as score 1 (absence of motion
artefacts or noise-related blurring, excellent vessel opacification, and no structural discontinuity).
(a) Volume rendered image of coronary arteries (b) Curved multiplanar reformations show right
coronary (c) Curved multiplanar reformations show left anterior descending artery (Xu et al, 2010).
On Siemens scanners the amount of padding used in PTA scans is automatically
selected according to the patient’s heart rate (adaptive sequential mode). When using
RGH scan mode they utilise ‘adaptive ECG pulsing’ with ECG-gated tube current
modulation. This automatically adjusts the width of the maximum mA window according
to heart rate.
Toshiba is currently the only manufacturer with a low pitch prospectively ECG –triggered
helical (PTH) scan mode available, and this is recommended on the Aquilion PRIME up
to heart rates in excess of 65 bpm. As in prospectively triggered axial scanning,
increased padding is recommended with increasing heart rate. Toshiba have also
recently implemented Adaptive Motion Correction, an algorithm that compensates for
motion in the vessels, myocardium and valves.
On Toshiba’s Aquilion scanners, the cardiac phase in which the x-rays are triggered in
PTA Volume scan mode is automatically selected according to heart rate and the amount
of padding is increased for added flexibility. On the Aquilion ONE, it has been shown that
although the image quality decreases with increasing heart rate, good image quality is
still achieved (Figure 20), with the image quality score maintained at between 1
69
(excellent) and 2 (good) for heart rates between 60 bpm and 100 bpm (Sun G et al,
2012).
(a) (b)
(c) (d)
Figure 20 - Images of a 63-year-old patient with an HR of 77.1 bpm and an HRv of 17.1. Excellent
and good images could be reconstructed. (a, b) Curve planar reconstruction showed stenosis in
right coronary artery (RCA) and left anterior descending artery (LAD) (arrows). (c, d) Invasive
coronary angiography confirmed the diagnosis (adapted from Sun G et al, 2012).
As mentioned in section 6.3.1 various methods are used for boosting a scanner’s intrinsic
temporal resolution. One of these is the use of multi-segment reconstruction (MSR),
whereby data from a number of consecutive heartbeats is taken to reconstruct images at
a given z-axis position. In this way the time over which data is acquired within a singe
heartbeat is reduced and so the temporal resolution is improved. On the majority of
systems it is possible to use MSR in low pitch RGH scan mode. On some scanners,
namely the Toshiba Aquilion ONE and ONE Vision models, MSR is also available in PTA
scan mode. MSR is particularly suited to scanners with wide z-axis coverage as fewer
cardiac cycles are needed to acquire the cardiac volume so there is a lower likelihood of
misregistration artefacts. However, the MSR approach does suffer from various
drawbacks (Tomizawa, 2012). On the Revolution CT, GE does not recommend multi-
segment reconstruction, but instead, for heart rates above 65 bpm advises ‘dual peak
phase’ acquisition and/or the use of SnapShot Freeze (SSF) motion correction software.
Because in the ‘dual phase peak’ method, x-rays are triggered in two different phases of
the cardiac cycle, the optimal phase for image reconstruction can be selected for each
70
coronary artery. The SSF algorithm uses information from adjacent cardiac phases within
a single cardiac cycle to characterise vessel motion and determine the actual vessel
position at the prescribed target phase (Leipsic et al, 2012). SSF can be applied in both
PTA and RGH scan modes and is available on all GE’s cardiac CT scanner models. GE
claims that its implementation results in images with an effective temporal resolution of
29 ms (Figure 21). Early results with the SSF algorithm are promising (Carrascosa et al,
2015) and a prospective, international, multicentre trial (ViCTORY) is currently in
progress. Its primary aim is to determine the diagnostic accuracy of SSF for the
diagnosis of CAD, compared to ICA (Min et al, 2013).
Figure 21 - Case study of SnapShot Freeze motion corrected CCTA. 54 year old male patient,
shows mild coronary artery disease in the LAD with no stenosis (GE Healthcare, 2012).
71
Table 14 - Scan modes used at different heart rates as per recommendations provided by the manufacturer. Scanner models Low heart rate (bpm) Medium heart rate (bpm) High heart rate (bpm)
GE
Optima 660 <65
Prospective axial - Step and shoot (PTA (~15% padding)
65 -75 Retrospective helical single segment (RGH)
(Snapshot Plus)
>75 Retrospective helical Multi-segment (RGH)
(Snapshot Burst)
Discovery CT750 HD <65
Prospective axial - Step and shoot (PTA) (~15% padding)
SnapShot Freeze; Multi-sector reconstruction (SnapShot Burst/Burst Plus mode).
Revolution CT Single beat modulated two-peak; Snapshot Freeze; Acquisition phases determined by the system based on patient’s heart rate variability
Philips Ingenuity Rate responsive technology and adaptive multi-cycle reconstruction.
iCT Elite Rate responsive technology and adaptive multi-cycle reconstruction.
Siemens
Somatom Definition AS+
Dual-segment acquisition protocol in axial and helical at high HR; Flex padding in axial; Adaptive ECG pulsing to widen window at high HR
Somatom Definition Edge Stellar
Dual-segment acquisition protocol in axial and helical; Flex padding in axial; Adaptive ECG pulsing to widen window at high HR.
Somatom Definition Flash Stellar
Dual source acquisition equivalent to gantry rotation time of 0.14 s; Dual-segment acquisition protocol in axial and helical; Flex padding in axial;
Adaptive ECG pulsing to widen window at high HR.
Somatom Force Dual source acquisition equivalent to gantry rotation time of 0.125s; Dual-segment acquisition protocol in axial and helical; Flex padding in axial; Adaptive ECG pulsing to widen window at high HR.
Toshiba Aquilion PRIME
The system automatically selects the appropriate factors including mA, kV, pitch and exposure phase of the cardiac cycle according to the patient heart rate. If heart rate is high this may reduce pitch or increase rotation time.
Aquilion ONE / ONE Vision
The system will automatically select the appropriate phase of the cardiac cycle for data acquisition according to the patient heart rate. If high this may suggest two rotations rather than one, but in clinical practice it is possible to narrow the exposure window to ensure one beat acquisition.
1Further clarifications on the operation of some features were not provided by the manufacturer within the timescale of the project.
90
(b) Clinical challenge: Arrhythmia
Scanner model Approaches to meeting challenges of patients with arrhythmia (1)
Revolution CT Adaptive gating; Smart arrhythmia management in case of irregular heart rate.
Philips
Ingenuity Automatic arrhythmia detection and management.
iCT Elite Automatic arrhythmia detection and management
Siemens
Somatom Definition AS+
Adaptive sequence axial mode will omit/repeat scan when ectopic beat detected; Retro helical recommended for known arrhythmia – ECG-modulation temporarily suspended if arrhythmia detected; ECG editing; Phase reconstruction performed with relative (%) or absolute (ms) parameters.
Somatom Definition Edge Stellar
Adaptive sequence axial mode will omit/repeat scan when ectopic beat detected; Retro helical recommended for known arrhythmia – ECG-modulation temporarily suspended if arrhythmia detected; ECG editing; Phase reconstruction performed with relative (%) or absolute (ms) parameters.
Somatom Definition Flash Stellar
High temporal resolution of dual source allows arrhythmia imaging with single segment; Adaptive sequence axial mode will omit/repeat scan when ectopic beat detected; Retro helical recommended for known arrhythmia – ECG-modulation will temporarily suspend if arrhythmia detected;
ECG editing; Phase reconstruction performed with relative (%) or absolute (ms) parameters.
Somatom Force High temporal resolution of dual source allows arrhythmia imaging with single segment; Adaptive sequence axial mode will omit/repeat scan when ectopic beat detected; Retro helical recommended for known arrhythmia – ECG-modulation will temporarily suspend if arrhythmia detected; ECG editing; Phase reconstruction performed with relative (%) or absolute (ms) parameters
Toshiba
Aquilion PRIME Heart rate is monitored in real time and if arrhythmia is detected during prospective acquisition the system will switch into retrospective mode to ensure a diagnostic outcome.
Aquilion ONE/ONE Vision
With no need to move the table, the Aquilion ONE scanners can monitor heart rate in real time and only exposes when there is a suitable interval in the cardiac cycle. If an arrhythmia occurs, exposure will not take place until the next suitable interval. The user can set a number of arrhythmias to reject.
1Further clarifications on the operation of some features were not provided by the manufacturer within the timescale of the project.
91
(c) Clinical challenge: High calcium score
Scanner model Approaches to meeting challenges of patients with high calcium scores (1)
GE
Optima 660 High-frequency kernel.
Discovery CT750 HD High Definition (HD) mode; Dual Energy cardiac (GSI Cardiac) mode.
Revolution CT High Definition (HD) mode.
Philips
Ingenuity Iterative algorithms such as IMR & iDose, Sharper edges to reduce blooming; Magic Glass visualization.
iCT Elite Iterative algoritms such as IMR & iDose, Sharper edges to reduce blooming; Magic Glass visualization.
Siemens
Somatom Definition AS+
Use of Iterative reconstruction (SAFIRE) will reduce the calcium blooming artefact through higher resolution..
Somatom Definition Edge Stellar
Use of Iterative reconstruction (SAFIRE) will reduce the calcium blooming artefact through higher resolution.
Somatom Definition Flash Stellar
Use of Iterative reconstruction (SAFIRE) will reduce the calcium blooming artefact through higher resolution.
Somatom Force Use of Iterative reconstruction (SAFIRE) will reduce the calcium blooming artefact through higher resolution.
Toshiba
Aquilion PRIME Small detector elements and high sampling rate reduce blooming artefacts; AIDR 3D includes artefact reduction software;
Aquilion ONE/ONE Vision
Small detector elements and high sampling rate reduce blooming artefacts; AIDR 3D includes artefact reduction software; SURESubtraction accurately eliminates calcium without drawbacks of thresholding;.
1Further clarifications on the operation of some features were not provided by the manufacturer within the timescale of the project.
92
(d) Clinical challenge: Stents
Scanner model Approaches to meeting challenges of patients with stents (1)
GE Optima 660 High-frequency kernel.
Discovery CT750 HD High Definition (HD) mode; Dual Energy cardiac (GSI Cardiac) mode.
Acquisition using z-sharp technology will optimise spatial resolution to allow small vessel and in-stent visualization.
Somatom Definition Edge Stellar
Stellar detector with Edge technology:
- Improves in-plane spatial resolution for more accurate in-stent visualisation.
- Reduces electronic cross-talk between neighbouring detector rows delivering a spatial cross-plane resolution of 0.30 mm.
Acquisition using z-sharp technology optimises spatial resolution to allow small vessel and in-stent visualization
Somatom Definition Flash Stellar
Stellar detector with Edge technology:
- Improves in-plane spatial resolution for more accurate in-stent visualization;
- Reduces electronic cross-talk between neighbouring detector rows delivering a spatial cross-plane resolution of 0.30 mm;
Acquisition using z-sharp technology will optimise spatial resolution to allow small vessel and in-stent visualization.
Somatom Force
Stellar detector with Edge technology:
- Improves in-plane spatial resolution for more accurate in-stent visualization;
- Reduces electronic cross-talk between neighboring detector rows delivering a spatial cross-plane resolution of 0.30 mm;
Acquisition using z-sharp technology will optimize spatial resolution to allow small vessel and in-stent visualization.
Toshiba Aquilion PRIME
Small detector elements and high sampling rate give best x. y, z resolution to see fine detail in small structures. AIDR 3D and large range of reconstruction algorithms gives good image quality.
Aquilion ONE/ONE Vision
Small detector elements and high sampling rate give best x. y, z resolution to see fine detail in small structures. AIDR 3D and large range of reconstruction algorithms gives good image quality. SURESubtraction accurately eliminates calcium without drawbacks of thresholding;.
1Further clarifications on the operation of some features were not provided by the manufacturer within the timescale of the project.
Revolution CT Two table positions with smart collimation to enable 1 beat acquisition of the heart and avoid volume boundary over the heart.
Philips Ingenuity Prospective Step and Shoot Complete
iCT Elite Prospective Step and Shoot Complete
Siemens
Somatom Definition AS+ The high speed scan acquisition in 128 slice acquisition and flexibility of the scan volume allows whole thorax coverage to identify graft origins to base of heart as required.
Somatom Definition Edge Stellar
The high speed scan acquisition in 128 slice acquisition and flexibility of the scan volume allows whole thorax coverage to identify graft origins to base of heart as required.
Somatom Definition Flash Stellar
The high speed scan acquisition in 128 slice acquisition and flexibility of the scan volume allows whole thorax coverage to identify graft origins to base of heart as required.
The Definition Flash can also use its unique Flash cardio spiral to acquire extended scan ranges with acquisition speeds up to 458 mm/s. This enables a standard thorax (300mm) to be imaged in 0.6 s.
Somatom Force The high speed scan acquisition in 192 slice acquisition and flexibility of the scan volume allows whole thorax coverage to identify graft origins to base of heart as required. The system can also use its Flash cardio spiral to acquire extended scan ranges with acquisition speeds up to 737 mm/s in Turbo Flash Mode. This enables a standard thorax (300 mm) to be imaged in 0.4s.
Toshiba Aquilion PRIME
Small detector elements and high sampling rate give best x. y, z resolution to see fine detail in small structures. AIDR 3D and large range of reconstruction algorithms gives good image quality.
Aquilion ONE/ ONE Vision Small detector elements and high sampling rate give best x. y, z resolution to see fine detail in small structures. AIDR 3D and large range of reconstruction algorithms gives good image quality. Wide Volume mode automatically calculates the two acquisition volumes required for coverage.
1Further clarifications on the operation of some features were not provided by the manufacturer within the timescale of the project.
94
(f) Clinical challenge: Obesity
Scanner model Approaches to meeting challenges of obese patients (1)
Revolution CT AEC: Auto kV and mA adjustment based on measured patient attenuation; ASIR-V iterative reconstruction.
Philips Ingenuity
IMR, iDose4, iPatient dose and image quality management. NanoPanel Elite detector with better Signal/Noise.
iCT Elite IMR, iDose4, iPatient dose and image quality management. 120 kW power Elite detector with better Signal/Noise.
Siemens
Somatom Definition AS+
CAREDose 4D auto mA modulation and CAREkV auto kV selection; 80kW generator (option 100kW) provides the power reserves required in obese imaging to obtain the image quality required to maintaining fast rotation time.
Somatom Definition Edge Stellar
Stellar Detector with TrueSignal Technology: detector is specially optimized for low-signal imaging.
CAREDose 4D automated mA modulation and CAREkV automatic kV selection; 100kW generator provides power reserves required in obese imaging.
Somatom Definition Flash Stellar
Stellar Detector with TrueSignal Technology: detector is specially optimized for low-signal imaging.
CAREDose 4D auto mA modulation and CAREkV auto kV selection provides patient specific parameters; 100kW generator provides the power required in obese imaging; Use of two tubes simultaneously in a cardio obese mode that uses flexible temp.res.(75 to 140ms) and additional projections (90 to 180 degrees of rotation) to improve signal-to-noise ratio. This is coupled with the power of 2 x 100kW generators to provide up to 1600 mA for bariatric cardiac imaging.
Toshiba Aquilion PRIME
AIDR 3D applicable to cardiac. High capacity patient couch. Lateral couch movement aids accurate positioning. Large gantry bore. SureExposure takes account of fat in patient when calculating dose required.
Aquilion ONE/ ONE Vision
AIDR 3D applicable to cardiac. High capacity patient couch. Lateral couch movement aids accurate positioning. Large gantry bore. SureExposure takes account of fat in patient when calculating dose required.
1Further clarifications on the operation of some features were not provided by the manufacturer within the timescale of the project.
95
8. Advantages, uncertainties and risks of
comparing scanners using technical
specifications
8.1 Bias
Technical specifications are objective indicators and therefore reduce the chance of bias
introduced by subjective metrics that are affected by human preference and decision
criteria. Comparing technical specifications allows experienced users of CT technology to
identify equipment that may offer the greatest potential for superior performance when
used in clinical practice.
8.2 Impact on clinical performance
The value of comparing specifications of medical technologies has been recognised and
used as the basis for selection and purchasing of medical equipment in the NHS for many
years. Specialised centres (e.g. KCARE, ImPACT, MagNET) developed relevant expertise
and fruitful cooperation with industry over many years to undertake comparative
specification studies and technology performance assessments widely used by NHS
healthcare staff in the selection of medical devices suitable for their needs.
8.3 Access to technical specifications
Detailed technical specifications may not be readily available from manufacturers as they
are sometimes reluctant to disclose details to protect their intellectual property.
Nevertheless it is reasonable to assume that it is likely to be easier to obtain technical
specifications in a shorter time than to produce diagnostic performance indicators
describing the clinical utility of the equipment.
8.4 Quality of the technical data
Technical specification data are strongly dependent on measurement methodology that
may vary between manufacturers. This may potentially lead to distorted comparisons.
However, it is relevant to point out that any one manufacturer is likely to be consistent in
the specifications provided for their different CT scanners models, allowing a direct
comparison between old and new models and between different ‘levels’ of CT scanner
model (low end and high end) from the same manufacturer.
96
8.5 Expertise required
Modern CT scanners are complex items of technology and adequate expertise is required
to understand the differences between technical data for different scanners and infer how
these differences may impact on diagnostic performance for specific patient groups.
8.6 Manufacturer-specific technical features
Although the fundamental principles of operation of various CT scanners may be
identical, some features are manufacturer-specific and therefore cannot be compared
across systems based on technical specification. Two examples follow:
To achieve a good temporal resolution, one manufacturer has developed a hardware
approach using dual sources whereas another uses a software approach (SnapShot
Freeze algorithm) to correct for motion artefacts and achieve a good effective temporal
resolution.
To achieve fast anatomical coverage and image the cardiac volume within one
heartbeat, some manufacturers have developed detector arrays with long z-axis
dimensions, whereas another manufacturer achieves single beat coverage using a
narrower detector but scanning at a high helical pitch.
8.7 Software upgrades
The discussions in this report mainly address the hardware features of CT equipment.
However there are many software features available for CCTA scanning that can enhance
image quality at the stages of data acquisition and image reconstruction, as well as at the
reporting stage. Moreover, software developments are rapid and upgrades are frequently
implemented (on new and old scanner models) to improve performance.
8.8 Multi-factorial effects
Another area of uncertainty relates to the matching of a given clinical ‘difficulty’ (e.g. high
heart rate, arrhythmia, stents) to a specific key technical feature (e.g. rotation time, z-axis
x-ray beam width, spatial resolution). Although this approach is based on sound physical
considerations, a given technical improvement should result in improved image quality and
is likely to benefit all patient groups.
For example, for obese patients with normal heart rates, a better temporal resolution will,
in theory, provide no improvement in clinical performance, as the key technical
requirement is an increase in CNR. However, clinical studies show that obese patients are
97
successfully imaged on systems with a good temporal resolution (Brodoefel, et al., 2008).
Two different conclusions can be drawn from this:
a) the particular patients in the study could also have been imaged successfully on a
system with a lower temporal resolution; or
b) the clearer depiction of structures due to decreased blurring from cardiac motion,
on the system with improved temporal resolution, compensates for the poorer CNR
in obese patients.
Another example relates to the fact that the spatial resolution of a CT image is not only
dependent on the technical features of the equipment but is very much influenced by the
scan protocol adopted. Variation in the acquisition or reconstruction protocols between
centres is common and might lead to noticeable differences in the spatial resolution
achieved on a given CT scanner. Furthermore, motion artefacts, whose severity will vary
with the temporal resolution of the scanner, can significantly influence the ‘effective spatial
resolution’.
8.9 Technological advances
Step changes result from rapid technological developments in CT, and may make one or
more specification features selected for comparison inadequate or obsolete. For example,
until recently, a high x-ray flux was considered an important requirement for the successful
imaging of obese patients. However, with the introduction of iterative reconstruction
algorithms it may be possible to achieve a given image quality at approximately half the
radiation dose, so the need for a high specification in this respect becomes less relevant.
8.10 Patients with multiple conditions
Finally, it should be acknowledged that the definition of subgroups of difficult to image
patients overlooks the reality that patients may present with multiple ‘difficult to image’
characteristics (e.g. obese patients with high heart rate).
In summary, comparing specifications of CT scanners has value but also has limitations. If
these are taken into consideration technical specification data may provide preliminary
useful information to help with the selection of a suitable scanner for CCTA.
The conclusions drawn from this exercise should be validated with further studies that
provide metrics more directly linked with the clinical diagnostic performance of the
technology.
98
9. Further work
Perform a systematic literature review on the current clinical performance of CT
scanners suitable for CCTA focusing on studies reporting on the relevant patient
subgroups;
As good, bias-free evidence from the literature is limited, the feasibility of
performing a retrospective study of CCTA examinations in the challenging patient
groups should be explored. The study would aim to quantify the limits of the
‘difficult factors’ on different scanner models. For example, what is the heart rate at
which clinical image quality is affected on a particular scanner? This analysis would
require detailed information on the scan protocols used and of patient
characteristics;
To gain a better understanding of the fundamental performance of CT systems in
CCTA examinations and obtain a complete set of unbiased data under controlled
conditions a study could be designed using appropriate dynamic cardiac phantoms
and a standard methodology. It is acknowledged that phantoms may not
adequately simulate the complex motion and structure of the coronary arteries.
However, these studies would be feasible within a relatively short period of time
(compared to clinical studies) and their findings could be valuable to formulate
hypotheses on which robust future clinical studies could be based; and
Develop and maintain a register of CT technical specifications to accompany and
maintain the currency of this report and inform future potential users/purchasers.
The undertaking of these further activities would allow a better understanding of the level
of difficulty posed by the subgroups of challenging patients. Additionally they would
provide important evidence on which to base conclusions as to the comparative diagnostic
performance of scanners used in cardiac CT.
99
Appendix 1: Clinical evidence
A literature search was performed to identify relevant studies reporting on the clinical
performance of CCTA and the impact of technology developments particularly in imaging
difficult patient groups. The tables below summarise details of some of the studies.
A – Diagnostic performance of CCTA - Impact of iterative reconstruction techniques
Record # 1
Date: 2011
Authors: Moscariello et al
Publication: European Radiology 21:2130-2138
Title: Coronary CT angiography: image quality, diagnostic accuracy, and potential for
radiation dose reduction using a novel iterative image reconstruction technique –
comparison with traditional filtered back projection (FBP).
Details of the study:
Comparison of image noise, IQ and diagnostic accuracy of CCTA for CCT using IR software
(SAFIRE) compared to traditional filtered back projection (FBP). Estimate the potential for radiation
dose savings. 65 patients underwent CCTA and ICA. Full radiation dose data was reconstructed with
FBP and 50% of the projections (to simulate the dose) were used to reconstruct the data with the
alternative IR method. CCTA was performed with a CT scanner Definition Flash (Siemens) at 120,
100 and 80 kV for patients with BMI > 25kg/m2, <25kg/m2 and BMI<20 kg/m2, respectively.
Summary of outcomes:
IR significantly reduced image noise without loss of diagnostic information and holds potential for
substantial radiation dose reduction from CCTA. The per-patient accuracy, Se, Sp, PPV and NPV
results are summarised below.
Half dose SAFIRE Full dose FBP p-value
Diagnostic. accuracy 96.9 93.8
Se (%) 100 100
Sp (%) 94.6 89.2 0.001-0.025
PPV (%) 93.3 87.5
NPV (%) 100 100
100
Record # 2
Date: 2011
Authors: Moscariello et al
Publication: European Radiology 21:2130-2138
Title: Coronary CT angiography: image quality, diagnostic accuracy, and potential for
radiation dose reduction using a novel iterative image reconstruction technique –
comparison with traditional filtered back projection (FBP).
Details of the study:
Comparison of image noise, IQ and diagnostic accuracy of CCTA for CCT using an IR software
(SAFIRE) compared to traditional filtered back projection (FBP). Estimate the potential for radiation
dose savings.
Sample of 65 patients underwent CCTA and ICA.
Full radiation dose data was reconstructed with FBP and 50% of the projections (simulation of 50% of
the dose) were used to reconstruct the data with the alternative IR method.
CCTA acquired with a CT scanner Definition Flash (Siemens) at 120, 100 and 80 kV for patients with
BMI > 25kg/m2, <25kg/m2 and BMI<20 kg/m2, respectively.
Summary of outcomes:
IR significantly reduced image noise without loss of diagnostic information and holds potential for
substantial radiation dose reduction from CCTA. The per-patient accuracy, Se, Sp, PPV and NPV
results are summarised below.
Half dose SAFIRE Full dose FBP p-value
Diagnostic. accuracy 96.9 93.8
Se (%) 100 100
Sp (%) 94.6 89.2 0.001-0.025
PPV (%) 93.3 87.5
NPV (%) 100 100
101
B - Diagnostic performance of CCTA - Impact of scan mode
Record # 3
Date: 2013
Authors: Neefjes LA et al
Publication: European Radiology 23 (3): 614-2
Title: Diagnostic accuracy of 128-slice dual-source CT coronary angiography: a
randomized comparison of different acquisition protocols
Details of the study:
Comparison of diagnostic performance and radiation exposure of 128-slice dual source CCTA
protocols to detect coronary stenosis (> 50% obstruction).
459 symptomatic patients randomised between 2 groups
Group A: high pitch spiral vs. narrow window sequential (HR < 65 bpm)
Group B wide window sequential vs. retrospective spiral (HR > 65 bpm).
Diagnostic performance of CCTA was compared with coronary interventional angiography in 267
patients. The mean effective dose for each protocol is compared.
high pitch spiral vs. narrow window sequential (HR < 65 bpm)
All images acquired with Somatom Definition Flash, Siemens Healthcare.
Summary of outcomes:
Sequential CCTA should be used in patients with regular heart rates using 128-slice DSCT.
Diagnostic quality was found comparable in both groups. Pool estimates (for 95% CI) of per segment
Se, Sp, PPV and NPV of prospectively ECG-gating CCTA for diagnosis of coronary stenosis with
more than 50% obstruction are summarised below.
Low heart rate
(mean:58±7bpm)
High heart rate
(mean:75±11bpm)
high pitch
spiral
narrow window
sequential
wide
window
sequential
retrospective
spiral
Se (%) 89* 97* (p=0.01) 94 92
Sp (%) 95 96 95 95
PPV 62 73 67 66
NPV 99 100 99 99
Radiation dose
(mSv)
1.16 ± 0.60
3.82 ± 1.65
(p<0.001)
*6.12 ± 2.58
8.13 ± 4.52
(p<0.001)
102
Record # 4
Date: 2012
Authors: Sun Z and Ng KH
Publication: International Journal of Cardiovascular 28 (8):2109-19
Title: Diagnostic value of coronary angiography with prospectively ECG-gating in the
diagnosis of coronary artery disease: a systematic review and meta-analysis.
Details of the study:
Meta-analysis of the diagnostic value or prospectively ECG-gating CCTA in the diagnosis of
significant stenosis.
14 studies included (910 patients, 3531 coronary arteries and 12056 segments)
Studies included CCTA performed with 64-slice single source CT, 64-slice dual source CT (first and
second generation) and 320-slice CT scanners were included.
Pooled Se, Sp, PPV and NPV were are summarised below.
Summary of outcomes:
CCTA with prospective ECG-gating allowed for reduced radiation exposure without sacrifice in IQ.