Pancreas: Tumors: Multidetector CT: Basic Principles for Optimizing Clinical Studies
Elliot K. Fishman, M.D.
MDCT: AN OVERVIEW
As with any advancement in technology it is usually built upon those achievements that proceeded it. Multidetector or multislice CT scanning (MDCT) is built upon many of the advances of spiral CT over conventional CT scanning. Yet, the technology of MDCT is far more than an upgrade of single detector CT (SDCT), and in many instances provides technology and capabilities that were never even considered possible a few short years ago. In this review, we will look at some of the core concepts and principles of MDCT with a special focus on the differences between MDCT and SDCT.
Although most of the major scanner manufacturers offer at least one version of MDCT there are definite differences in technology and capability between the various vendors. For example, the detector array used, one of the key features on any scanner differs between GE and Siemens, the two largest scan vendors. GE (and Toshiba) offer a fixed array design while Siemens (and Marconi) offer an adaptive array design. Although each manufacturer claims superiority, we will try to define the specific advantages of one or the other techniques. Please note that at Hopkins we use a Siemens VolumeZoom scanner and our knowledge of the system and our hands-on experience with it are far more extensive that any information we could gather on any other scanner.
With a fixed array detector systems, the detector size is constant at 1.25 mm and there are 16 elements (20mm total). Elements can be added together in groups of 2, 3 or 4 prior to sampling resulting in four 2.5 mm channels, four 3.75 mm channels or four 5 mm channels. One potential problem of the fixed detector array are that the multiple joints between detector rows reduce the efficiency of the detectors which can result in poor signal. Additionally, the outer edges of the detector produce a shadow effect further deteriorating image quality. Another limitation of the fixed array system such as the GE LightSpeed scanner is that typically only 2 different pitches are used; pitch 3 (HQ or high quality mode) and pitch 6 (HS or high speed mode).
With scanners that use "Adaptive Array Detectors" such as the Siemens VolumeZoom the detector array consists of a series of 8 detectors set in pairs from the middle out of 1 mm, 1.5 mm, 2.5 mm and 5 mm for a total of 20 mm across (Figure 2). These detectors are then used to create either sets of 4 detectors of 1 mm, 2.5 mm or 5 mm thickness. The pitch can be chosen with equal quality anywhere from 1 to 8. In addition 0.5 mm thick sections can also be generated but only a pitch of 4 can be used in these cases. Please note however that 0.5 mm thick sections result in isotrophic datasets rather than the classic anistrophic datasets.
The number of available parameters provides for maximum flexibility when designing study protocols. One way of looking at MDCT is as the 4 "Cs:
• continuously rotating tube/detector system
• continuous radiation
• continuous data acquisition
• continuous table feed
Although this list can in great part be true in a single detector system, the differences between MDCT and SDCT are crucial and can be broken up into several categories. They include:
• pitch
• scan slice profile
• noise
• radiation dose
• collimation
Let us now look at each of these individually in detail.
Pitch
With the newest versions of SDCT a pitch of up to 3 can be obtained. The advantage of a higher pitch is that a specific volume can be scanned in a shorter period of time which is critical in such applications as pediatric CT, CT angiography and virtual imaging. The radiation dose for that volume will be decreased accordingly as long as the mA is not changed. However, these advantages come at a price as the slice width profile widens with increasing pitch. The slice is widened by up to 27% with a pitch of 2. With MDCT the pitch can be selected anywhere between 1 and 8. Despite the increased pitch the slice width is kept constant with MDCT because of the reconstruction algorithm is in fact independent of pitch. This algorithm is the Adaptive Axial Algorithm.
One of the confusing differences among the various manufacturers is the definition of pitch depending on whether a SDCT or MDCT is being discussed. In single slice CT the definition is straightforward:
Pitch = table movement per rotation/slice collimation.
So for a typical 1 second rotation scanner a pitch of 2 means the table traveled 10 mm with a 5 mm slice width or collimation. With multislice CT it is more complicated as collimation and width may be different because a single collimator can produce several different scan widths. For multislice CT the definition is:
Pitch = table movement per rotation/single slice collimation. With a 0.5 sec scanner there are 2 rotations per second. So if the table travels 12 mm in a second and a 1 mm collimator is used then the pitch would be 6 (6 mm/ 1mm). Please note that this definition is the one used by Siemens Medical Systems and is not universally accepted. Because of the many definitions of pitch there has been a recent attempt to develop a standardized definiton. This would look at the nominal slice thickness and number of detectors (i.e. 2,5 mm collimation and 4 detectors equals 10 mm) as well as the distance traveled per rotation (i.e. 15 mm). This would then give a ratio nominally called pitch. In the above example the pitch would be 1.5 while with the other definiton would be a pitch of 6. Personally, I am in favor of any system that provides uniformity and minimizes confusion. Radiology needs to avoid situations where a study can be done poorly because of a lack of understanding of the available parameter selections. After all we are doing MDCT and not trying to land spaceships on Mars (sorry NASA ).
Slice Width and Collimation
With SDCT the true width or the "effective slice thickness" of the reconstructed image is influenced by pitch and the reconstruction algorithm used (wide vs. slim). With a pitch of 2 and a slim algorithm the "effective slice thickness" may increase by 27%. With MDCT and the use of the "Adaptive Axial Interpolation" both the scan width and collimation are correct without any blooming.
With SDCT the slice collimation and the slice width are the same. This parameter is selected prior to the study and can not be modified in the reconstruction sequence or in post processing. With MDCT there is increased flexibility. The user selects in advance the detector collimation of the study. On our scanner the typical selections are for 4 detectors and are 1 mm, 2.5 mm and 5mm. What is important however is that each detector can provide several true slice widths. They are listed in table 1 but a closer look at the 1 mm collimation shows that scans can be reconstructed anywhere from 1.0 mm to 5 mm to 10 mm. This allows the user in the same study to have images with high resolution but increased noise (1 mm slice thickness) and images with standard resolution but less noise (5 mm slice thickness). In practice this provides unparalleled capabilities including obtaining true volume datasets for CT Angiography and standard 5mm scan slices for looking at the abdominal organs.
Radiation Dose
With a pitch of 8 to 1 one would at first glance think that MDCT should significantly decrease radiation dose for patients. This is based on the realization that with SCDT that the applied dose decreases linearly with increasing pitch. In practice MDCT can provide low dose studies (i.e lung cancer screening) but as set up routinely the applied dose is independent of pitch. There is no decrease for increasing the pitch as the user selects the mAs for the desired slice width and the tube current will adapt to maintain image quality, independent of pitch. Although this feature seems like a good idea it is important to monitor exactly what dose an individual patient receives: The key to lowering dose is using a lower mAs and kVp.
Image Noise
In both SDCT and MDCT, the image noise is dependent on a number of factors which may or may be controlled by the assigned study protocol. They are:
- mAs
- kV
- kernel or reconstruction algorithm
- slice thickness
- patient size
- collimation
- image display
MDCT can help limit noise by modifying the mAs and slice thickness as needed for each individual case. Because of the number of slice widths that can be selected from a single collimation one is able to for example get the necessary volume data for a 3D CT angiogram while at the same time get noise-free images for looking at the liver. The advanced processing algorithms and faster computers allow between 2 and 3 images to be reconstructed per second depending on the scanner available.
Detector Collimation
The collimation used by the scanner has a number of fixed selection choices which can then be expanded as needed. Although there are many choices that can be made retrospectively it is important to be aware that the final choices are limited by the collimation selected.
MDCT and its Impact on Clinical CT Scanning
The evolution of CT technology has ensured the durability of CT as a powerful clinical tool. In particular, Kalender's description of spiral CT in 1990 revolutionized CT and opened up whole new applications including CT angiography. Spiral CT also proved to be a quantum advance for the acquisition of volumetric datasets for 3-D medical imaging. In contrast to conventional CT which images a slice of the body, incrementally moves the patient, obtains another slice, etc., spiral CT acquires data continuously as the patient travels through the CT gantry. This approach markedly improves acquisition speed: an entire chest and abdomen could be imaged in less than 1 minute, compared to 8 to 12 minutes with prior techniques. Despite initial skepticism about the spatial resolution of spiral CT compared to conventional incremental CT, spiral CT rapidly supplanted conventional CT for virtually all body imaging applications by providing three fundamental advantages over incremental CT: (1) reduced or eliminated artifacts caused by respiration or other sources of patient movement, (2) imaging during the optimal window of contrast enhancement for maximal image contrast, and (3) the ability to reconstruct axial images at arbitrary intervals in order to better center the sections over focal lesions or improve the quality of multiplanar and 3-D reconstruction. These three basic advantages have made it possible to use CT for high quality vascular imaging and have provided significantly improved performance for a wide range of traditional imaging applications including lung nodule detection, liver lesion detection, and musculoskeletal imaging.
The newest advance in CT technology, multidetector (MDCT) or multislice CT provides new capabilities which are especially valuable for angiographic applications or any application where volume datasets are needed. Although there is variability between the different types of MDCT available several concepts are similar. These include:
• larger volumes can be scanned
• scan times decrease as pitch is typically 6-8 instead of 2 with single detector scans
• narrower collimation is available with better resolution
• tube heating is no longer a problem
The large volume of data generated by modern spiral scanners and multidetector scanners challenge traditional methods for viewing radiological studies. Where a conventional CT study might have provided 4 sheets of images (12 images per sheet) which could easily be reviewed by a radiologist sitting in front of a light box, today's scanners can generate hundreds or even a thousand images which require many sheets of film to display. This problem has fueled the development of computer workstations which allow the radiologist and clinician to interactively explore spiral CT datasets using a variety of display formats including standard axial slices, reconstructed slices in any plane, or high quality 3-D images. In the case of vascular CT imaging, 3D imaging is no longer a luxury but a necessity. The entire paradigm of how we look at image data is beginning to change and much of this change will focus on 3D volume rendering.
Vascular Imaging: CT Angiography
The area in which spiral scanning has made the greatest impact is in vascular imaging. CT angiography has been shown to be effective in a number of vascular applications including the renal arteries, the aorta and its mesenteric branches, the cerebral circulation, the portal venous system, and the pulmonary arteries. Maximum intensity projection (MIP) and surface rendering were the 3-D techniques which had been primarily applied to CT angiographic applications. Both of these techniques have important shortcomings which ultimately limit the clinical usefulness of the resulting images. We have applied volume rendering to the display of CT angiographic data with promising results. In vitro phantom studies have shown that volume rendered images of simulated vascular stenoses accurately depict the degree of narrowing, the single most important piece of information when deciding whether to treat an arteriosclerotic lesion. Several recently published clinical studies comparing volume rendering with surface rendering and/or MIP confirm that the theoretical advantages of volume rendering do in fact hold up in practice. Volume rendering is becoming the preferred 3-D rendering technique for CT angiography. Volume rendering is the preferred technique in our laboratory and is used for nearly all applications.
Review: 3-D Rendering Techniques Used for Medical Imaging
Surface Rendering
Surface Rendering was one of the earliest methods for 3-D display, and is now available in most commercially available 3-D medical imaging packages. In this method, each voxel within the data set is determined to be a part of or not a part of the object of interest, usually by comparing the voxel intensity to some threshold value, thereby defining the "surface" of the object. With the surface determined, the remainder of the data is discarded. Surface contours are typically modeled as a collection of polygons and displayed with surface shading. The resulting image is a simplified, possibly misleading representation of a structure, particularly if the surface is difficult to determine precisely as is often the case in medical imaging. By converting the data from a volume to a surface, a large portion of the data available is forfeited in exchange for faster, easier computation. While this can be an advantage by allowing real-time rendering and thereby enhancing user interactivity, the usefulness of surface rendered medical images is generally limited by their inconsistent image fidelity.
Maximum Intensity Projection (MIP)
Like surface rendering, MIP is also commonly available in commercial 3-D software packages and so has been extensively clinically evaluated, particularly with respect to its usefulness in creating angiographic images from CT and MRI data. The MIP algorithm evaluates each voxel along a line from the viewer's eye through the image and selects the maximum voxel value as the value of the corresponding display pixel. The resulting images are typically not displayed with surface shading or other devices to help the user appreciated the "depth" of the rendering, making three-dimensional relationships difficult to assess. If there is another high intensity material along the ray through a vessel (such as calcification) the displayed pixel intensity will only represent the calcification and will contain no information from the intravascular contrast. Selection of the highest pixel value also increases the background mean of the image, particularly in enhancing structures such as the kidney and liver, thereby decreasing the visibility of vessels in these structures. What is worse, volume averaging coupled with the MIP algorithm commonly leads to MIP artifacts: a string of beads appearance in MIP images of normal vessels passing obliquely through a volume. While MIP has a number of important artifacts and shortcomings, it has been studied extensively and usually does provide superior accuracy to surface rendering for CT angiography. MDCT datasets are especially useful for MIP imaging.
Volume Rendering
As the name implies, this technique renders the entire volume of data rather than just surfaces, and so potentially conveys more information than a surface model. Volume rendering techniques sum the contributions of each voxel along a line from the viewer's eye through the data set. This is done repeatedly to determine each pixel value in the displayed image. Because the information from the entire data set is incorporated into the resulting image, much more powerful computers are necessary to do volume rendering at a reasonable speed. We view volume techniques as the most advanced form of 3-D rendering currently available for creating accurate, clinically useful medical images. Volume rendering is just now being incorporated into commercially available software packages - with general availability and continued increases in computer power, volume rendering will likely become the most important rendering technique for 3-D medical imaging.
Bibliography
(see www.ctisus.com for an expanded bibliography on all aspects of spiral CT, MDCT and 3D imaging)
1. Ney DR, Drebin RA, Fishman EK, Magid D. Volumetric rendering of computed tomographic data: principles and techniques. IEEE Comput Graphic Appl 1990;10:24-32.
2. Fishman EK, Magid D, Ney DR, et al. Three-dimensional imaging: State of
the Art. Radiology 1991;181:321-337.
3. Kuszyk BS, Ney DR, Fishman EK. The current state of the art in 3D oncologic imaging: an overview. Int J Radiation Oncology Biol Phys 1995; 33:1029-1039.
4. Fishman EK, Kuszyk BS, Heath DG, Gao L, Cabral B. Surgical planning for liver resection. IEEE Computer 1996;29(1):64-72.
5. Kuszyk BS, Heath DG, Bliss DF, Fishman EK. Skeletal 3-D CT: Advantages of volume rendering over surface rendering. Skeletal Radiol 1996;25:207-214.
6. Drebin RA, Carpenter L, Hanrahan P. Volume rendering. Computer Graphics 1988;22:65-74.
7. Kalender WA, Seissler W, Klotz E, Vock P. Spiral volumetric CT with single-breath hold technique, continuous transport, and continuous scanner rotation. Radiology 1990;176:181-183.
8. Cabral B, Cam N, Foran J. Accelerated Volume Rendering and Tomographic Reconstruction Using Texture Mapping Hardware. In: ACM/IEEESymposium on Volume Visualization. 1994. Washington, DC.
9. Kuszyk BS, Heath DG, Ney DR, Bluemke DA, Urban BA, Chambers TP, Fishman EK. CT angiography with volume rendering: imaging findings. AJR 1995;165:445-448.
10. Johnson PT, Heath DG, Kuszyk BS, Fishman EK. CT angiography with volume rendering: advantages and applications in splanchnic vascular imaging. Radiology 1996;200:564-568.
11. Smith PA, Marshall FF, Urban BA, Heath DG, Fishman EK. Three-dimensional CT stereoscopic visualization of renal masses: impact on diagnosis and patient management. AJR 1997;169:1331-1334.
12. Uchida M et al. Three-Dimensional Imaging of Liver Tumors Using Helical CT During Intravenous Injection of Contrast Medium. J Comput Assist Tomogr 1999;23(3): 435-440.
13. Coll DM et al. 3-Dimensional Volume Rendered Computerized Tomography for Preoperative Evaluation and Intraoperative Treatment of Patients undergoing Nephron Sparing Surgery. J Urology 1999; 161:1097-1102.
14. Smith PA, Marshall FF, Corl FM, Fishman EK. Planning Nephron-Sparing Renal Surgery Using 3D Helical CT Angiography. J Comput Assist Tomogr 1999;23(5):649-654.
15. Johnson PT et al. Renal Artery Stenosis: CT Angiography - Comparison of Real-Time Volume Rendering and Maximum Intensity Projection Algorithms. Radiology 1999;211:337-343.
16. Hofmann LV et al Three-Dimensional Helical CT Angiography in Renal Transplant Recipients: A New Problem-Solving Tool AJR 1999;173:1085-1089.
17. Lerner LB et al. Interactive 3-dimensional computerized tomography reconstruction in evaluation of living renal donors. J Urology 1999;161:403-407.