We have previously examined the impact of incorporating CC and IDC in CAPRA and NCCN pretreatment stratification tools and our findings indicated a significantly improved patient stratification for CAPRA scores 3-5 and NCCN scores 4-6 in a cohort from Toronto, Canada.1 In this current study, we validated our findings in two additional independent retrospective cohorts; one from North America (Wisconsin, n=542) and another from Europe (Rotterdam, n=154). The ability of each risk stratification method to predict BCR-free survival (FS) and event-free survival (EFS), defined as time to metastases and/or death of prostate cancer, was assessed using Cox proportional hazards models, Harrell’s concordance index (c-index) and 1000 bootstrap resampling. ANOVA test was then utilized to compare the log-likelihood statistic of models with and without CC/IDC (P-values).
In both validation cohorts, CAPRA significantly outperformed NCCN for predicting BCR- FS and EFS (P < .001). Further, the addition of CC/IDC significantly improved CAPRA prognostication, most notably in scores 3-5 (intermediate risk) for BCR-FS as well as scores 3-5 and 6-10 (high risk) for EFS in both validation cohorts (P < .001). Similarly, the NCCN prognostication tool was significantly enhanced by the addition of CC/IDC in favorable (score 3), unfavorable intermediate (score 4), and the high/very high risk (scores 5-6) classes for both BCR-FS and EFS (P < .001).
In a sub-analysis focused exclusively on Grade group 2 biopsies (Gleason score 7/10, 3 + 4), which constitute the most common prostate cancer patient subset, similar results were found. The addition of CC/IDC improved BCR-FS prognostication in the CAPRA 3-5 risk score (P < .001) and in all NCCN risk categories, most notably for scores 4 and 5-6 (P < .001). For EFS, CAPRA 3-5 and 6-10 scores showed improved prognostication after the addition of CC/IDC status (P < .001). For NCCN, risk categories 4, 5-6 showed significant improvement of addition of CC/IDC (P = .043). This finding is particularly notable and highlights the importance of identifying and reporting CC/IDC in prostate cancer biopsy pathology reports.
Lastly, decision curve analysis for BCR-FS at 5 years demonstrated that the addition of CC/IDC to the NCCN tool has a greater net benefit in identifying patients for adjuvant treatment between threshold probabilities 15% to 30% while for CAPRA this improvement was noted between threshold probabilities 25% to 30%. Similarly, for EFS at 5 years, the addition of CC/IDC improved both the CAPRA between threshold probabilities 10%- 25% and NCCN between threshold probabilities 10%-20%.
The findings of this study validate the importance of incorporating CC/IDC into pretreatment prostate cancer stratification tools. Future studies incorporating larger cohorts from diverse geographic areas are essential to ascertain the potential broader applicability of our results across different patient populations.
Written by: Ekaterina Olkhov-Mitsel & Michelle R. Downes
Department of Anatomic Pathology, Sunnybrook Health Sciences Centre, Precision Diagnostic & Therapeutic Program, Toronto, Ontario, Canada
- Yu Y, Lajkosz K, Finelli A, Fleshner N, van der Kwast TH, Downes MR. Impact of cribriform pattern 4 and intraductal prostatic carcinoma on National Comprehen-sive Cancer Network (NCCN) and Cancer of Prostate Risk Assessment (CAPRA) patient stratification. Mod Pathol. 2022;35:1695–1701 .