Introduction

The 2015 prostate cancer grading system was introduced to simplify pathologic stratification. We examine the performance of the Prognostic Grade Groups (PGG) in the Shared Equal Access Regional Cancer Hospital (SEARCH) database with respect to long term prostate cancer outcomes and whether associations vary by race within an equal access healthcare system.

Methods

We performed a retrospective review of men undergoing radical prostatectomy at one of six Veterans Affairs hospitals between 1988 and 2015. We identified 4,325 men with available data. The prognostic ability of PGG for multiple long term clinical endpoints was examined using Cox models. Interactions between PGG and race were tested.

Results

The cohort consisted of PGG 1 through 5, respectively: 2,077(48%), 1,171(27%), 521(12%), 409(10%), 147(3%). 1,596(38%) were African American. Median follow up was 86(IQR: 45 to 135) months. Higher PGG was associated with higher stage, older age, more recent year of surgery and surgical center (p<0.02). African American men had a lower PGG distribution (p=0.028). Higher PGG was associated with increased risk of all clinical endpoints on univariable and multivariable regression including biochemical recurrence(BCR), adjuvant therapy, castrate resistant prostate cancer(CRPC)[Figure 1, left], metastases, prostate cancer specific mortality(PCSM) and overall survival(OS)[Figure 1, right](all p<0.001). _x000D_ We found no significant interactions with race in predicting any of the measured outcomes. (BCR: p=0.78, adjuvant therapy: p=0.60, CRPC: p=0.91, metastases: p=0.61, PCSM: p=0.83, OS: p=0.21).

Conclusions

The 2015 Prognostic Grade Groups predicted multiple long term clinical endpoints after prostatectomy in a large, multiracial cohort of men. The predictive value for survival endpoints was similar in Caucasian and African American men. _x000D_

Funding

none