|
|
The main advantages of 3D virtual colonoscopy over barium enema include patient comfort and speed. Much less patient discomfort results using the 3D virtual colonoscopy since there is no need to reposition the patient. In addition, the entire process for obtaining the helical CT images (including the preparation phase) should take only 15 to 20 minutes of patient time, as compared to 45 minutes or more for barium enema. Also, the radiation dose used in a 3D virtual colonoscopy study is approximately one half to one quarter of the average exposure in barium enema. Furthermore, an unlimited number of viewing angles can be exploited with the visualization techniques. Compared with conventional colonoscopy, this technique is expected to be less expensive (about half of the cost), involves less discomfort and risks to the patient. There is also no need for sedation, and the patient can therefore resume normal activities immediately after the CT scanning procedure.
A major contribution of this paper is the incorporation of a set of volume visualization techniques for the exploration of the CT-scanned colon data sets. In particular, the idea of extending the "peel onion" technique from 2D image processing to the 3D volumetric domain to automatically produce the flythrough animations along the inside of the colon is, to our best knowledge, an original contribution to biomedical visualization. Furthermore, unlike other approaches to 3D virtual colonoscopy [9,4] which extract the colon surface and represent it as a list of triangles, our approach is volume-based, thus not only allowing the user to visualize the mucosal surface of the colon, but also providing the flexibility of exploring the data beyond the surface. For example, a virtual surgery cut can be applied to a polyp or the outer layers of the polyp can be rendered translucently to reveal the hidden information inside the polyp.
Our testing results on the two pipe data sets and the Visible Human data set are very encouraging. The images, interactive flythroughs, and automatically-generated animations that we achieved with the VolVis system clearly demonstrate the feasibility of diagnosing colonic polyps by our non-invasive, safe, and relatively comfortable procedure. We are currently working on obtaining real colon data sets from patients at the University Medical Center of SUNY Stony Brook and evaluating the application of this method on clinical cases.