Apart from 3D virtual endoscopy, various other 3D applications are available to aid evaluation. A distinction is made between “additional 3D projections” and functional 3D applications, known as “3D tools.” Additional 3D projections enable one to obtain additional 3D views for further evaluation of detected lesions. 3D tools can be used to optimize 3D evaluation, for instance, by highlighting structural information of and depicting blind spots on 3D images. A particularly interesting point is the development of so-called advanced 3D visualization techniques for time-efficient evaluation of the colon, known as 3D “panoramic” or “virtual dissection” views.
Additional 3D Projections
In addition to virtual endoscopic 3D views, modern computer workstations now also contain a range of 3D imaging features which vary by software manufacturer. These offer, for instance, global views of the colon or combined 2D/3D solutions as well as 3D target views (targeted 3D views of a selected volume of interest). Generally, these views are not intended for primary evaluation of the colon, but they may be helpful for further analysis and interpretation of any detected lesions, for documentation and reporting, and for communication of findings to the referring physician.
Global Views (3D Maps)
Global views of the colon provide an external overview of the entire distended large bowel. They are available on almost all workstations. Global views are either selective, nontransparent “cast” representations (“intraluminal air cast”) or semitransparent double-contrast views (tissue transition projection, TTP) of the entire gas-filled colon. The surrounding structures in the CT dataset are not reproduced. “Cast” views provide an opaque image similar to conventional single-contrast imaging of the colon. Semitransparent double-contrast views are similar to double-contrast barium imaging. These views are not suitable for the primary evaluation of a dataset (Fig. 3.17). While large polyps and stenosing carcinomas are usually well depicted, small endoluminal structures are visualized inadequately, if at all. Global views are especially well suited for the assessment of colonic anatomy and the extent of stenoses. They provide a clearly arranged overview of the length and location of colonic segments. The bookmarks of detected lesions can be overlaid on the image to document their exact location in the colon, and this is why these views are often referred to as “3D maps.” In this way global views also provide a useful “road map” for the endoscopist or surgeon, and should be attached to the radiologist′s report to allow lesions to be more easily found during endoscopy or surgery.
Combined 2D/3D Views
Combined 2D/3D views are hybrid views that, in addition to the endoluminal view, give a 2D representation of the extraluminal surroundings on the same image. This representation is suitable for additional analysis of suspicious findings, but does not replace a standard 2D evaluation. These views are particularly good for demonstration purposes (Fig. 3.18). A variation of the hybrid view is the “undistorted bisegmentation view,” which is to be used in 3D data analysis (see below, “Advanced 3D Visualization Techniques”).
3D Target View
Three-dimensional target views, or coned-down views, are used for detailed analysis of a subvolume of interest in a CT colonography dataset. The target lesion is located in the center of the subvolume. Structures surrounding the subvolume with the lesion, such as internal organs or parts of the bowel wall, are not visualized. This allows one to get a better look at the lesion and, if necessary, to evaluate it from a greater distance than with an intraluminal view (Fig. 3.19).
3D tools are used for optimizing the 3D evaluation. For instance, they will integrate structural information into 3D views or demonstrate blind spots. In general, they allow additional information to be obtained on a 3D visualization that normally would require the radiologist to switch to a 2D view.
When fecal tagging is used, the increased density of labeled stool is not apparent on 3D displays. Electronic labeling assigns colors to different density values. The color information is then displayed on a 3D endoluminal view, improving the differentiation between tagged residual fecal material and polyps (Fig. 3.20).
Digital Subtraction—Electronic Cleansing
An alternative to electronic labeling is digital subtraction of tagged stool and fluid. If residual stool and fluids are adequately (i.e., homogeneously) labeled with an orally administered contrast agent, they can be subtracted from the dataset electronically on the basis of their high density, by means of thresholding. This technique is known as “electronic cleansing” or “digital stool subtraction.” The method can be readily used if fecal tagging is performed in combination with a standard bowel preparation method using laxatives. Limited data are available on the feasibility of digital subtraction in patients in whom fecal tagging has been done without laxatives (“prepless”) as the sole means of bowel cleansing.
Some commercially available workstations include software algorithms that allow digital “removal” of homogeneously tagged stool or fluids. This technique aims to facilitate primary 3D evaluation of fecal tagging datasets, especially if they contain large amounts of residual tagged fluid, which make 3D evaluation without stool subtraction incomplete (Fig. 3.21).
Artifacts. A drawback of any electronic cleansing method is that subtraction of high-density voxels from the dataset may cause artifacts which could lead to misinterpretation. These techniques should therefore be used with caution and in full knowledge of the possible artifacts. It must be assured that true colonic lesions are not removed from the dataset and normal structures are not blurred or visualized as potentially pathological findings. For example, partial volume effects at the air–fluid interface can, after subtraction, give rise to a linear artifact that is easy to recognize on both 2D and 3D views (Fig. 3.22). In addition, partial volume effects at the air–fluid interface, or active movement or motion from intraluminal fluid, may also give rise to polypoid or bizarre artifacts. Incomplete electronic subtraction due to inadequate tagging of fluid and stool, or to the presence of small air bubbles, can generate artifacts that mimic polypoid lesions and carcinomas. The morphology of such pseudolesions should be easy to identify on corresponding unsubtracted 2D prone/supine images (Fig. 3.23). To avoid misinterpretation when using subtraction techniques, the primary, unsubtracted dataset must always be available and interrogated for problem solving.
The potential usefulness of electronic cleansing techniques is limited at present by the multitude of associated artifacts and the extra time needed to interpret them. Hence, these methods are not yet fully practicable for routine everyday practice, and most experts nowadays do not completely rely on electronic cleansing.