Performance of computed tomographic urography for the detection of bladder tumors in patients with microscopic hematuria.

OBJECTIVE Our objective was to evaluate the sensitivity, specificity, predictive value, and accuracy of computed tomographic urography for the detection of bladder tumors in patients with microscopic hematuria. MATERIAL AND METHODS Patients with microscopic hematuria initially evaluated with computed tomography and cystoscopy from January 2006 to December 2009 were evaluated. Computed tomography detecting a bladder lesion suspicious of malignancy was considered positive. Cystoscopy was classified as positive when a lesion requiring biopsy or resection was found. Performance characteristics of computed tomography were determined by comparing with cystoscopic and pathological findings. RESULTS A total of 112 patients were eligible for analysis. Seven tumors were found on cystoscopy; of these, 2 were correctly diagnosed by computed tomography and 5 were missed. An additional case was considered erroneously positive. The results are a sensitivity of 29%, specificity of 99%, positive predictive value of 67%, negative predictive value of 95%, and accuracy of 95%. CONCLUSIONS Although computed tomography has a high specificity its sensitivity is limited. For this reason conventional cystoscopy should be considered the standard for bladder evaluation of patients with microscopic hematuria.