Résumé
PURPOSE: We investigated the factors that predicted later transitional cell carcinoma (TCC) in a subgroup of patients with abnormal cytology and negative initial evaluations. MATERIALS AND METHODS: From January 2002 to June 2007, we retrospectively identified 58 patients. Cases were considered discordant if a work-up of urine cytology was abnormal although initial cystoscopy, upper tract evaluation, and biopsies resulted in a negative or benign diagnosis. Patients who could complete a urine cytology test after 6 to 8 weeks and who were followed up for at least 1 year were included in this study. According to later TCC demonstration, we compared risk factors for TCC between the later TCC group and the benign group and evaluated the independent factors that predicted later TCC by use of a Cox proportional hazards regression model. RESULTS: Of the 58 patients, the mean follow-up was 12.7+/-17.3 months (range: 2-83 months), and 14 patients (23.7%) had a prior history of TCC. During follow-up, 9 patients (15.3%) had TCC and 1 patient had prostate cancer. In the later TCC group, the incidence of a prior history of TCC (p=0.03) and persistent abnormal cytology (p<0.001) were higher than in the benign group in univariate analysis. In the Cox proportional hazards regression model, persistent abnormal cytology (p=0.033, relative risk (RR): 17.380 [95% CI: 1.265-238.783]) was the only independent factor to predict later TCC. The mean follow-up duration of later TCC demonstration was 8.55 months (range: 2-32 months). CONCLUSIONS: Our results suggest that in the setting of persistent abnormal urine cytology with a negative initial evaluation, 53.3% of patients will later develop TCC. Patients with persistent abnormal cytology need intensive follow-up within 1 year.