Validation of an active surveillance threshold for the CCP score in conservatively managed men with localized prostate cancer.

Abstract

54 Background: Active surveillance (AS) is an increasingly popular treatment modality for men with localized prostate cancer. However, better risk stratification is needed to appropriately select men for AS. The cell cycle progression (CCP) score has proven to be a robust predictor of prostate cancer outcomes in various clinical settings, including conservatively managed cohorts. Here, we present a validation of an AS threshold for a predefined score that combines CCP with CAPRA (combined clinical CCP risk (CCR) score) for predicting prostate cancer mortality (PCM) in conservatively managed patients. Methods: We determined the CCR score distribution in 505 men who were tested in our clinical laboratory and, based on their clinical characteristics, might typically be considered for AS. Specifically, the training cohort consisted of men who had Gleason score ≤ 3+4; PSA < 10 ng/ml; < 25% cores positive; and clinical stage ≤ T2a. A threshold CCR score of 0.80 was selected such that 90% of the men in the training cohort had scores below the threshold. The performance characteristics of the threshold were then evaluated in two independent cohorts of conservatively managed men (TAPG1 [N= 180] and TAPG2 [N=585]). Survival data were censored at 10 years. Results: The primary pre−planned analysis called for evaluating the CCR threshold on TAPG2. There were 60 men (of 585) below the threshold in the validation cohort and the threshold validated, dichotomizing the cohort into high and low risk groups (log rank P−value = 0.0008). There were no deaths in patients below the threshold and the Cox proportional hazard estimate of 10−year PCM associated with the CCR threshold was 3.3%. The 10−year risk of PCM associated with the threshold in the combined cohort (TAPG1 and TAPG2) was 3.2% and there were no observed prostate cancer deaths in patients below the threshold. Conclusions: For patients considering deferred treatment, the CCR score provides significant prognostic information at disease diagnosis. The threshold presented here was derived from the ‘typical’ risk of PSM for AS patients and can be used to guide patient selection for AS based on an integrated view of risk assessment.