General: Potential Pitfalls with Multidetector/Multislice CT (MDCT/MSCT) Scanning: Protocol Design

Elliot K. Fishman, M.D.

Although MDCT/MSCT offers significant advantages over single detector CT scanning, there are several potential pitfalls that need to be considered in a busy clinical practice. Although some of these pitfalls may not be "unique" to MDCT/MSCT they are potentially a more common problem and can result in errors in these situations.

One can look at MDCT/MSCT as a natural progression of CT from the era of axial imaging to spiral CT scanning to MDCT/MSCT. The radiologist’s traditional needs for study optimization; faster data acquisition, faster reconstruction times, better image resolution and more image processing capability are the direction CT scanning has gone the last two decades. Today, we have reached a new plateau with subsecond, true volume acquisitions in what we believe will be but a step before scanners are designed with multiple x-ray tubes to speed up data acquisition times to the 50-100 millisecond range. This will help with development in such areas as coronary artery angiography, evaluation of mesenteric ischemia and evaluation of cardiac function.

With faster data acquisition the need for timing of the contrast bolus and data acquisition becomes even more critical. Not surprisingly with second scan times the injection bolus delivery is important but even more so with a very fast subsecond acquisition. If a MDCT/MSCT can be done in 10-25 seconds, then the timing of the injection is especially critical and the window of success is narrowed. When looking at the charts in Fig. 1-2 it is obvious that one potential error is to scan the patient when too little contrast is on board. Even a 3 cc/sec injection with a 20 sec delay means that at the initiation of scanning only 60 cc of contrast has been injected. Assuming the study acquisition time is 15 seconds only 105 cc of contrast will have been delivered at the conclusion of the study, but perhaps more importantly, only 60 cc at the start of the study. For examinations where arterial phase imaging (i.e. R/O hepatoma, R/O aneurysm, R/O vascular liver metastases) is critical, one will need to be careful that enough contrast is delivered at the time of data acquisition in order to make the correct diagnosis. Contrast injection rates may need to be routinely increased while scan delay times may need to be extended.

Similarly, the delay time in dual phase imaging must now be carefully addressed. The original spiral scanners could not do dual phase acquisitions without a long interscan delay. The newer single detector CT scanners can produce dual acquisitions with a minimum of a 5 sec interscan delay. The challenge with these systems was to obtain two phases of acquisition that provided both arterial and venous phase imaging of the liver.

With MDCT, the ability to obtain the two phases is no longer time limited. Even with a 25 sec delay one can typically be ready for the second acquisition at 50-55 sec without any difficulty. Yet, this may in fact prove to be too little a difference between the two studies. If the liver is examined with this protocol one would be obtaining essentially a true arterial phase (25-40 sec) followed by an early (45-60 sec) arterial-portal phase. Similarly, in evaluation of the kidney, both phases would provide nearly similar images with sharp cortico-medullary differentiation but well before a nephrographic phase occurred.

In the design of our protocols we have taken these factors into account when choosing our contrast volumes, injection rates and data acquisition times. We are able to lower the contrast volumes from 120-150 cc to 100-120 cc. We are also able to limit injection rates to 3 cc although faster injection rates may be valuable in select situations. We have also commonly extended the delay before acquiring the venous phase of acquisition. With these words of caution we have been successful in doing our CT angiograms without the need for using a computer-assisted program or for a test bolus. This obviously takes a bit of practice in order to be successful.

Another set of potential pitfalls that is especially prominent with MDCT relates to preset image filming window width/center levels. Because of the high level of enhancement routinely achieved, the classic CT soft tissue (WW-410, WC-7), and liver (WW-108, WC-43) settings often no longer apply. Lesions in the kidney and liver will routinely be obscured with these settings. Care must be taken to modify the WW/WC to more appropriate level. We have found that a window width equal to 550 and a window center equal to 50 is a good starting point.

Other applications where filming can produce false negative diagnosis include evaluation of suspected pulmonary embolism and aortic dissection. In these cases an embolism or a flap could easily be overlooked without the correct filming protocol. In time this will be less of an issue as review of data on workstations will make it easier to change the WW/WL at the radiologist's discretion.

Another potential problem relates to the absolute measurement of cysts or cyst-like lesions in the kidney. Maki et al. showed how problems with pseudoenhancement could result in lesion appearance denser (15-20 HU) than they truly were. This may be a bigger problem with MDCT where renal enhancement is unusually ‘rich’ due to the speed of data acquisition.

Finally, an issue that has also been addressed in the literature is the detection of an increased number of lesions in an organ like the liver which are 2-4 mm in size. These lesions are too small to adequately characterize, too small to biopsy and often were missed on pre-MDCT scans (so it may appear to be a new lesion on a follow-up study). Obviously we all recognize that a true malignant lesion is small before it becomes large, it is a potential problem to pick up multiple ‘indeterminate lesions’. Our current practice is to be more descriptive and use caution before we suggest tumor. Obviously liver metastases affect the staging of nearly all patients as well as treatment options. We would for example not want to overstage a patient with a resectable rectal cancer where the treatment options would significantly change. More work will need to be done to develop strategies in this area.

Another vascular pitfall that has been addressed with SDCT but is more common with MDCT is the detection of venous thrombosis. Although we are all familiar with pseudo-thrombus in the IVC and femoral veins we are seeing many cases of pseudo-thrombus in the SMV. This is especially problematic in studies such as those staging a pancreatic or liver mass. Once again the use of dual phase imaging (with a sufficient interscan delay) rather than arterial phase imaging only will help avoid this problem. Finally, when in doubt a couple of delayed scans will avoid this false positive diagnosis.

Despite these pitfalls (and other to be defined in future updates of this article), MDCT is the state of the art in CT scanning and represent the very bright future of CT scanning.

Table 1: Contrast Injection Rates and Volumes. Injected at Start of Study

Table 2: Contrast Injection Rates and Volumes. Injected During the Study

Table 3: Single Acquisition

Table 4 Common False Positive Diagnosis on CT

venous thrombosis
• liver metastasis
• venous invasion/thrombosis by tumor
• inadequate venous phase imaging to detect tumor