Surveillance for Nonmuscle Invasive Bladder Cancer: Identifying the Point of Diminishing Returns

Abstract

To characterize first and second recurrence patterns using 26years of cohort-level follow-up and microsimulation modeling. Patients diagnosed with nonmuscle-invasive bladder cancer in Stockholm County between 1995 and 1996 were included. Clinical, pathological, and longitudinal follow-up data were gathered. Logistic regressions, Kaplan Meier curves, and Cox proportional hazards models were run to generate assumptions for a microsimulation model, simulating first and second recurrence and progression for 10,000 patients. Three hundred eighty-six patients were included: 67.4% were male; >50% were TaLG; and 37.5% were American Urological Association high-risk. Median time to recurrence was 300days. Three patients had missing data. Cohort follow-up has been carried out for 26years. For simulated first-recurrences, low-risk patients recurred at 56.6% over 15years of follow-up, with 2.2% muscle-invasive (MI) progression; intermediate-risk patients recurred at 62.8%, with 4.3% MI progression; high-risk patients recurred at 48.7% over 15years, with MI progression at 14.3%. For second recurrences, 70.7%, 75.7%, and 84.7% of low, medium, and high-risk patients recurred. No patients were seen to have first recurrences after 9years, with low, but notable, rates beyond 5years. These data suggest that low-, intermediate-, and high-risk patients without recurrence at 5years may be potentially transitioned to less invasive monitoring.