Diagnostic potential of three-dimensional ultrasound-based virtual cystoscopy: an experimental study using pig bladders

OBJECTIVE: We sought to investigate the feasibility of 3-dimensional ultrasound (3D US)-based virtual cystoscopy in experimentally designed bladder phantoms using pig bladders. MATERIALS AND METHODS: Ten bladder phantoms with simulated polypoid tumors of different sizes were constructed to provide 3D US-based virtual cystoscopic images. The simulated polypoid tumors were divided by their location into 2 groups: group 1 included the polyps on the anterior and posterior walls, and group 2 included those on the remaining 4 walls. Two genitourinary radiologists independently interpreted the virtual cystoscopic images, with a consensus reached for cases with discrepant results. RESULTS: In a consensus reading, 3D US-based virtual cystoscopy detected 44 (91.7%) of 48 polypoid lesions. All lesions (n = 4) with false-negative results were 5 mm or smaller in diameter, and so the detection rate was 81.8% (18/22) for lesions 5 mm or smaller whereas it was 100% (26/26) for lesions larger than 5 mm. The detection rate for group 1 (93.8%) was not significantly different from that for group 2 (90.6%; P = 1.000). There were 7 false-positive lesions. Of these, 5 lesions were seen in group 1, and reverberation artifacts were the most common source of these errors (n = 4). Blind areas were present in 19 (31.7%) of 60 virtual cystoscopic images and were seen only in group 2. The differences in the false-positive rate and the incidence of blind area were statistically significant between both groups (P = 0.036, P = 0.00008, respectively). CONCLUSION: 3D US-based virtual cystoscopy may have diagnostic potential in the detection of polypoid bladder tumors.