RESUMEN
BACKGROUND: To assess urinary symptoms and urine cytology as screening tools for cystoscopic detection of local recurrence after bladder-preserving trimodality treatment (TMT). METHODS: Patients with muscle-invasive bladder cancer receiving definitive TMT follow-up three monthly for 2 years, six monthly for the next 3 years and then yearly, with a clinical review, urine cytology and cystoscopy at each visit (triple assessment, TA). Grade 2+ cystitis/haematuria absent/present was scored 0/1, and urine cytology reported negative/suspicious or positive was scored 0/1, respectively. The performance of these two parameters for predicting local recurrence in cystoscopic biopsy was tested. Other hypothetical surveillance schedules included cystoscopy on alternate visits (COAV), or suspected recurrence (COSR), six-monthly COSR and six-monthly TA. RESULTS: A total of 630 follow-up visits in 112 patients with 19 recurrences (7 muscle invasive, 12 non-muscle invasive) at a median follow-up of 19 months were analysed. The sensitivity and specificity of clinical symptoms were 47.4% and 92%, and for urine cytology 58% and 85%, respectively. The combination of clinical symptoms and cytology (COSR) was 95% sensitive and 78% specific for local recurrence but 100% sensitive for muscle-invasive recurrence. Both COAV and COSV schedules showed a high area under the curve (AUC) for detecting local recurrence (COAV = 0.84, COSR = 0.83), muscle-invasive recurrence (AUC = 0.848 each) and non-muscle-invasive recurrence (COAV = 0.82, COSR = 0.81); reducing the need for TAs by 64% and 67% respectively, and overall cost by 18% and 33%, respectively. CONCLUSION: Cystoscopy at suspected recurrence during follow-up is safe and the most cost-effective for detecting muscle-invasive local recurrences, while cystoscopy at alternate visits may be more optimal for detecting any local recurrence.