CT Examination Technique
In the years since CT colonography was first introduced, significant advances have been made in scanner technology and in postprocessing software. The first attempts at colonic examination in the mid-1990s used single-slice spiral CT scanners. Although the preliminary results of CT-supported colon studies produced very promising early results, the spatial resolution was inadequate for the detection of smaller lesions or even for computing 3D images of sufficient quality.
Multidetector CT. The introduction of four-slice multidetector CT scanners (MDCT) in 1998 provided adequate spatial resolution for detection of polyps measuring less than 1 cm. Now 16-slice (from 2001 onward) and 64-slice scanners (from 2004 onward) can obtain slice thicknesses of less than 1 mm, offering high imaging quality. For CT colonography, the high acquisition rate offered by the most recent generation of scanners is particularly advantageous. With a single-slice scanner, an abdominal scan with 5-mm-thick slices took more than a minute. Using a 64-slice multidetector scanner with 0.6 mm collimation, the same examination can be done in just 7–8 seconds.
It is important for the examiner to be familiar with the technical specifications of the scanner being used, and to be able to design the examination protocols with these specifications in mind.
Multidetector CT Protocols for CT Colonography
Ideally, the goal of any examination protocol is to achieve the highest possible spatial resolution with the lowest possible radiation dose in the shortest length of time. A fundamental requirement for CT colonography is to achieve a CT scan of the whole abdomen with narrow collimation within one breath-hold. The thinner the reconstructed slice thickness, the closer one comes to a true isotropic volume CT dataset that will allow multiplanar reformatting in any plane and high-resolution 3D images. Recent multidetector scanners can scan the abdomen and pelvis much more quickly, which shortens the breath-hold. This in turn reduces the number of respiration and movement artifacts, thereby significantly improving the image quality. In one study, movement artifacts were reported in 61% of examinations conducted with a single-slice CT, while respiration artifacts occurred in only 16% when a 4-slice scanner was used. These rates may be further decreased by using 16-slice or 64-slice scanners, and in one of our own studies, out of 100 examinations performed with a 64-slice scanner, no movement artifacts occurred. Multidetector scanners with at least 4 detector rows are recommended for CT colonography, and scanners with 16 or more detector rows are preferred. Useful scanning parameters for CT colonography are listed in Tables 2.2 and 2.3.
Collimation and Slice Thickness
Basically, the smallest usable slice thickness defines the smallest lesion that can be detected. This has a direct influence on the detection of colonic lesions. The thinner the individual reconstructed slice of the CT dataset, the smaller the partial volume effects (partial volume averaging) (Fig. 2.24). Thicker slices may also reduce specificity; that is, the number of false-positive diagnoses increases, since typical morphological features of stool such as tiny pockets of air may be less conspicuous.
Collimation. In CT colonography, detection of endoluminal lesions relies heavily on precise imaging of the surface morphology of the colonic mucosa. Partial volume effects can cause blurring between the contours of polyps and normal intestinal anatomy. In addition, stair-step artifacts can considerably compromise image quality. This is especially true for single-slice scanners, in which collimation for a colon examination typically is between 3 and 5 mm. This is why CT colonography requires a high-resolution CT dataset of the wohle abdomen acquired with narrow collimation within one breath-hold. For a 4-slice or 8-slice scanner, collimation should be 2.5 or 1.25 mm, depending on the equipment manufacturer. In 16-slice scanners the narrowest collimation is 0.75 mm; in a 64-slice scanner it is 0.5–0.625 mm. The maximum collimation should not exceed 2.5 mm. Ideally, for CT colonography the narrowest collimation to which the scanner can be set should be selected. The exception to this is 4-slice scanners, as using the narrowest collimation may result in a scanning time that is too long for breath-hold. Here the imaging quality will be better (breathing artifacts reduced) if a compromise is made between an acceptable scanning time and a somewhat wider collimation of 2.5 mm. It may be preferable to use a scanner with at least 16 slices instead.
Scan direction. To minimize breathing artifacts, CT colonographic scans should be performed in the craniocaudal direction.
Slice thickness and data volume. For high-quality images, individual slices must be overlapped and reconstructed. For 16-slice and 64-slice scanners, for instance, 1-mm-thick slices are reconstructed with a reconstruction increment of 0.7 (= 30% overlap) (Table 2.2). This results in a large number of single images, typically as many as 1200, which can only be properly interpreted with the aid of a special off-line workstation. It is important that the internal data transmission system or PACS (picture archiving and communication system) is suitable for supporting the volume of data required for CT colonography. Some manufacturers also offer installation of the interpretation software on the CT scanner console using “shared memory,” eliminating the need for image transmission.
In addition to the thin slices needed for 3D reconstructions, it is also advisable to reconstruct thicker slices (3–5 mm). Due to the reduced image noise, these slices are better suited to the evaluation of extracolonic structures.