UCSF

Background

The University of California, San Francisco, Cancer of the Prostate Risk Assessment Postsurgical (CAPRA-S) score uses pathologic data from radical prostatectomy (RP) to predict prostate cancer recurrence and mortality. However, this clinical tool has never been validated externally.

Objective

To validate CAPRA-S in a large, multi-institutional, external database.

Design, setting, and participants

The Shared Equal Access Regional Cancer Hospital (SEARCH) database consists of 2892 men who underwent RP from 2001 to 2011. With a median follow-up of 58 mo, 2670 men (92%) had complete data to calculate a CAPRA-S score.

Intervention

RP.

Outcome measurements and statistical analysis

The main outcome was biochemical recurrence. Performance of CAPRA-S in detecting recurrence was assessed and compared with a validated postoperative nomogram by concordance index (c-index), calibration plots, and decision curve analysis. Prediction of cancer-specific mortality was assessed by Kaplan-Meier analysis and the c-index.

Results and limitations

The mean age was 62 yr (standard deviation: 6.3), and 34.3% of men had recurrence. The 5-yr progression-free probability for those patients with a CAPRA-S score of 0-2, 3-5, and 6-10 (defining low, intermediate, and high risk) was 72%, 39%, and 17%, respectively. The CAPRA-S c-index was 0.73 in this validation set, compared with a c-index of 0.72 for the Stephenson nomogram. Although CAPRA-S was optimistic in predicting the likelihood of being free of recurrence at 5 yr, it outperformed the Stephenson nomogram on both calibration plots and decision curve analysis. The c-index for predicting cancer-specific mortality was 0.85, with the caveat that this number is based on only 61 events.

Conclusions

In this external validation, the CAPRA-S score predicted recurrence and mortality after RP with a c-index >0.70. The score is an effective prognostic tool that may aid in determining the need for adjuvant therapy.