Stability of Prognostic Estimation Using the CAPRA Score Incorporating Imaging-based vs Physical Exam-based Staging.

ABSTRACT

PURPOSE: Although official T-staging criteria for prostate cancer are based on digital rectal examination findings, providers increasingly rely on transrectal US and MRI to define pragmatic clinical stage to guide management. We assessed the impact of incorporating imaging findings into T-staging on performance of a well-validated prognostic instrument. MATERIALS AND METHODS: Patients who underwent radical prostatectomy for prostate cancer diagnosed between 2000 and 2019 with stage ≤cT3a on both digital rectal examination and imaging (transrectal US/MRI) were included. The University of California, San Francisco CAPRA (Cancer of the Prostate Risk Assessment) score was computed 2 ways (1) incorporating digital rectal examination-based T stage and (2) incorporating imaging-based T stage. We assessed for risk changes across the 2 methods and associations of CAPRA (by both methods) with biochemical recurrence, using unadjusted and adjusted Cox proportional hazards models. Model discrimination and net benefit were assessed with time-dependent area under the curve and decision curve analysis, respectively. RESULTS: Of 2,222 men included, 377 (17%) increased in CAPRA score with imaging-based staging (P < .01). Digital rectal examination-based (HR 1.54; 95% CI 1.48-1.61) and imaging-based (HR 1.52; 95% CI 1.46-1.58) CAPRA scores were comparably accurate for predicting recurrence with similar discrimination and decision curve analyses. On multivariable Cox regression, positive digital rectal examination at diagnosis (HR 1.29; 95% CI 1.09-1.53) and imaging-based clinical T3/4 disease (HR 1.72; 95% CI 1.43-2.07) were independently associated with biochemical recurrence. CONCLUSIONS: The CAPRA score remains accurate whether determined using imaging-based staging or digital rectal examination-based staging, with relatively minor discrepancies and similar associations with biochemical recurrence. Staging information from either modality can be used in the CAPRA score calculation and still reliably predict risk of biochemical recurrence.