Objective To evaluate the urine cytology silver staining combined with ultrasonography（USG）in the detection of bladder transitional cell carcinoma （TCC） recurrence after transurethral resection of bladder tumor（TURBT）in terms of sensitivity and specificity. Methods Cystoscopy was used as “gold standard”. Urine cytology combined with USG or cystoscopy was measured separately and blindly. AgNORs protein stained by silver were used in cytology with Kappa of inter-observers 0.81. For the USG, the patients were scanned with trans-rectal probe with Kappa of inter-observers 0.76. The results of urine cytology combined with USG （Positive when urine cytology and/or USG positive. Negative when both urine cytology and USG negative） were compared with “gold standard”. Results The 148 consecutive superficial TCC patients with TURBT one year previously were included in this study. Fifty seven recurrenced cases were detected. Recurrence rate was 38.51%. The sensitivity and specificity of urine cytology silver stain were 89.47% （95% CI 0.82 to 0.98） and 87.91% （95% CI 0.81 to 0.95）. Area under ROC curve was 82.22%. The sensitivity and specificity of USG were 57.90% （95% CI 0.45 to 0.71 ） and 90. 11% （ 95% CI 0.84 to 0.96）. Area under ROC curve was 73.13% . The sensitivity was improved to 94. 74% （95% CI 0.89 to 1.00） when cytology combined with USG. But specificity decreased to 84. 62% （95% CI 0.77 to 0.92 ）. Area under ROC curve was improved to 98.28%. Conclusions Urine cytology silver stain combined with USG improves the high sensitivity for follow-up TCC patients after TURBT. The non-invasive protocol is suggested.
B-type ultrasound images have important applications in medical diagnosis. However, the widely spread intensity inhomogeneity, low-scale contrast, constructed defect, noise and blurred edges all make it difficult to implement automatic segmentation of lesion in the images. Based on region level set method, a subordinate degree region level set model was proposed, in which subordinate degree probability of each pixel was defined to reflect the pixel subjection grade to target and background respectively. Pixels were classified to either target or background by calculation of their subordinate degree probabilities, and edge contour was obtained by region level set iterations. In this paper, lesion segmentation is regarded as local segmentation of specific area, and the calculation is restrained to the local sphere abide by the contour, which greatly reduce the calculation complexity. Experiments on B-type ultrasound images showed improved results of the proposed method compared to those of some popular level set methods.
