Virtual CT cystoscopy appears to have the same potential for identifying cancers invasive to the bladder wall that virtual colonoscopy is recognized for having in a similar manner in the colon, according to study data presented on Sunday at the 2008 meeting of the Radiological Society of North America (RSNA). According to the literature, thinner slice reconstructions have been reported to improve tumor detection rates without impairment of image quality for virtual CT colonoscopy.

Bladder cancer is one of the most common tumors of the urinary tract, causing 4.5% of all newly diagnosed tumors and 1.9% of cancer deaths. Initial evaluation usually consists of cystoscopy with cold cup biopsy, where a cup-shaped device is used to obtain urothelium without electric current. Computed tomography (CT) is used mainly to demonstrate extravesical extension of the tumor and distant metastasis. CT imaging follow-up studies usually are reserved for patients with invasive tumors. Recently, three-dimensional (3-D) computer methods have led to virtual imaging. Virtual imaging allows intraluminal navigation through any hollow shape thus mimicking endoscopy. This technique of virtual endoscopy has been applied to many organs, including the colon, bronchus, stomach, and now the bladder. However, in terms of the optimal variables for virtual CT cystoscopy, this remains to be defined. "The trend in virtual colonoscopy is to look at thinner slice thickness, so in order to look at slice thickness in virtual CT cystoscopy we selected 10 CT scans from a larger study to evaluate the role of the technique in bladder cancer," reported Isabel Castellano PhD, Consultant Physicist at the Royal Marsden Hospital who presented the study. (Read the abstract here.)

In total, 10 delayed contrast enhanced CT examinations (GE Lightspeed 16; 120kV; 16x0.625mm collimation; rotation 0.8sec; pitch 0.938; auto mA, noise index 12) were performed in patients with bladder tumors and reconstructed at different slice thicknesses and intervals. For each of these CT scans, a number of reconstructions were run, starting with slice thickness at 5mm down to 0.6 mm at a selection of slice thickness intervals and the results were rendered into 3-D images. The volume rendered images were created with endoluminal visualization software. Manual thresh holding was used to gain optimal images. The CT number and standard deviation were recorded for bladder contents and bladder wall at three levels through the bladder. The mean standard deviation was used as a measure of noise, and the contrast to noise ratio between the bladder contents and the bladder wall. A formula was used to calculate the mean bladder contents and the bladder wall across the three levels was used for analysis. "Each image quality measure was compared to noise and bladder contents and the bladder wall measurements."

The 3-D images are not perfect, as they suffer from artifacts that can be classified as ridging, dimpling, holes and floaters. Artifacts were noted and scored on each reconstruction sets. Also, tumor definition was scored and an overall virtual image score was provided from 1 (best) to 7 (worst). The study also looked at how the image quality would compare to the quantitative image quality parameters of the images, such as the noise in the bladder wall, the bladder contents and the contrast to noise ratio of the bladder wall against the contents. "There was a very good correlation between the CT number threshold used to produce the virtual images related to the contrast enlargement ratio of the bladder wall which demonstrates that well trained radiologists can do this task very reliably," Dr. Castellano said in an interview with MedicExchange. (Watch the video of the entire interview here.) "It also provides a means for the manufacturer to automate this step in the generation of the virtual images."

The presence of holes and floaters did not correlate with any of the variables assessed. The ridging artifacts were most pronounced in the thickest slices, which were the 5mm slice thickness. The dimpling artifacts were most pronounced in the thinnest slices, in the 0.6mm slice thickness. "The best way to avoid these two artifacts is the middle range of the slice thickness, such as the 1mm and the 1.2mm slice thicknesses." There was a degree of correlation between the dimpling and image noise which suggested that the dimpling was a consequence of the noise in the image. "When we looked at tumor definition, it was confirmed that we had the best definition score in the intervening 1.2 mm and 2.5 mm slice thicknesses. When we looked at the overall score, it became clear the best overall score was the 1.2 mm slice thickness."

Dr. Castellano concluded that the thinnest slices were not necessarily the best slices to choose for virtual CT cystoscopy images, but that 1.2 mm slices would provide the best compromise in terms of minimizing artifacts and visualizing the tumor and bladder wall. "This procedure has been implemented into our standard clinical technique so we are ready to move forward with a reliable protocol in which we are confident. The next step is to use this protocol by moving ahead into the clinic," Dr. Castellano told MedicExchange.

TrackBack URI for this entry

Name

Website

Title

Comment smaller | bigger

Subscribe via email (registered users only)

I have read and agree to the Terms of Usage.

Add Comment