Image composing represents a postprocessing application that allows for combination of coherent MRI data sets acquired at different times and displaying different parts of an anatomic system. Because of a limited field of view (FOV), it is not always feasible to display an entire anatomic system, for example the whole spine, with one sequence acquisition. However, technological advances have made it possible today to perform whole body MR imaging within a reasonable scan time. This opens the door to new diagnostic fields, such as screening for metastases with whole body short tau inver- sion recovery (STIR) and whole body contrast-enhanced magnetic resonance angiog- raphy (MRA). In addition, imaging of the whole spine is feasible, with high-resolution cervical, thoracic, and lumbar scans all acquired in a single setting. These imaging approaches share in common the consecutive acquisition of mul- tiple exams or segments to achieve coverage for the extended anatomic region to be evaluated. A requirement for efficient workflow is the presence of coils or a matrix of coils covering the entire region to be evaluated, in essence in most cases the entire body. In a second step, with modern software, the acquired data can be composed to a single data set to enhance image evaluation/diagnostic interpretation. However, there are certain restrictions regarding sequence parameters that one has to consider to perform image composing. Different MR images cannot be com- posed arbitrarily, with the requirement being that some basic sequence parameters must conform to each other. This means that one must know before the sequence acquisition whether the image data are to be composed or not. Image composing is in its infancy, and some of the restrictions currently in place will likely with time be lifted. For example, differing slice thickness and pixel size may prevent image com- position, depending upon the software available. Limitations may exist regarding the relative tilt of the two image sets to be composed. Furthermore, if correction is made for geometric distortion, then corrected and noncorrected images cannot be mixed. Figure 50.1 presents a contrast-enhanced MR angiogram of the abdominal, pelvic, and lower-extremity vessels in a 76-year-old man with advanced atheroscle- rosis. After intravenous gadolinium chelate administration using a power injector, three coronal 3D FLASH scans have been acquired consecutively at different anatomic levels, starting with the lower abdomen and pelvis, then covering the upper legs, and finally the lower legs and feet. With this approach, the MR scans follow the con- trast bolus from the distal aorta through the vessels of the lower legs. Figure 50.1A–C presents maximum intensity projection reconstructions of the corresponding three subtracted raw data sets in a coronal view. Note that at the upper and lower borders of the respective fields of view, anatomic structures are displayed twice, once in the upper image and once in the adjacent lower image (arrows). This is due to an overlap in the acquisition of the corresponding MRA sequences, performed to ensure com- plete coverage of the anatomy in question. However, in this setting, the radiologist interpreting the exam has to be well aware of the normal anatomy in order not to mistake a single lesion for two, due to it appearing twice, and to be able to integrate the scans. Figure 50.1D demonstrates the same consecutively acquired and subtracted data sets after image composing. Because of slight differences between the three respective data sets, one can still detect that the MRA images were obtained with 102 50 Image Composing
Welcome message from author
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
Image composing represents a postprocessing application that allows for combinationof coherent MRI data sets acquired at different times and displaying different parts ofan anatomic system. Because of a limited field of view (FOV), it is not always feasibleto display an entire anatomic system, for example the whole spine, with one sequenceacquisition. However, technological advances have made it possible today to performwhole body MR imaging within a reasonable scan time. This opens the door to newdiagnostic fields, such as screening for metastases with whole body short tau inver-sion recovery (STIR) and whole body contrast-enhanced magnetic resonance angiog-raphy (MRA). In addition, imaging of the whole spine is feasible, with high-resolutioncervical, thoracic, and lumbar scans all acquired in a single setting.
These imaging approaches share in common the consecutive acquisition of mul-tiple exams or segments to achieve coverage for the extended anatomic region to beevaluated. A requirement for efficient workflow is the presence of coils or a matrix ofcoils covering the entire region to be evaluated, in essence in most cases the entirebody. In a second step, with modern software, the acquired data can be composed toa single data set to enhance image evaluation/diagnostic interpretation.
However, there are certain restrictions regarding sequence parameters that onehas to consider to perform image composing. Different MR images cannot be com-posed arbitrarily, with the requirement being that some basic sequence parametersmust conform to each other. This means that one must know before the sequenceacquisition whether the image data are to be composed or not. Image composing isin its infancy, and some of the restrictions currently in place will likely with time belifted. For example, differing slice thickness and pixel size may prevent image com-position, depending upon the software available. Limitations may exist regardingthe relative tilt of the two image sets to be composed. Furthermore, if correction ismade for geometric distortion, then corrected and noncorrected images cannot bemixed.
Figure 50.1 presents a contrast-enhanced MR angiogram of the abdominal,pelvic, and lower-extremity vessels in a 76-year-old man with advanced atheroscle-rosis. After intravenous gadolinium chelate administration using a power injector,three coronal 3D FLASH scans have been acquired consecutively at different anatomiclevels, starting with the lower abdomen and pelvis, then covering the upper legs,and finally the lower legs and feet. With this approach, the MR scans follow the con-trast bolus from the distal aorta through the vessels of the lower legs. Figure 50.1A–Cpresents maximum intensity projection reconstructions of the corresponding threesubtracted raw data sets in a coronal view. Note that at the upper and lower bordersof the respective fields of view, anatomic structures are displayed twice, once in theupper image and once in the adjacent lower image (arrows). This is due to an overlapin the acquisition of the corresponding MRA sequences, performed to ensure com-plete coverage of the anatomy in question. However, in this setting, the radiologistinterpreting the exam has to be well aware of the normal anatomy in order not tomistake a single lesion for two, due to it appearing twice, and to be able to integratethe scans. Figure 50.1D demonstrates the same consecutively acquired and subtracteddata sets after image composing. Because of slight differences between the threerespective data sets, one can still detect that the MRA images were obtained with102
50 Image Composing
c50 10/31/08 3:22 PM Page 102
103
three consecutive acquisitions. However the gaps (arrowheads, D) at the borders ofthe fields of view are no longer seen and the double display (arrows, A–C) of adjacentanatomic structures at the interfaces between the three scans, due to the overlap dur-ing sequence acquisition, is no longer present.
Fig. 50.1
c50 10/31/08 3:22 PM Page 103
Related Documents
