Introduction

Adverse pathology (AP, defined as pathological Gleason grade > 4+3 and/or >pT3) is a strong predictor of biochemical recurrence, metastasis, and cancer specific mortality in men with prostate cancer (PCa). A 17 gene tissue-based RTPCR assay (Oncotype Dx® Genomic Prostate Score™, GPS) has been validated as a predictor of AP in multiple retrospective cohorts. In this study, we aim to prospectively validate GPS as a predictor of AP in men with clinically low-risk PCa treated with radical prostatectomy (RP).

Methods

A pre-specified analysis from a 1200-patient prospective study was performed on patients who elected RP as initial PCa management. The primary endpoint was AP. Descriptive statistics are reported on demographic and clino-pathological characteristics. Binary logistic regression was performed to determine the association between GPS and AP. The odds ratio (OR) per 20 GPS units and 95% confidence interval (CI) were calculated. All analyses were conducted using SAS 9.4.

Results

Of 1200 patients enrolled from 21 sites in the study, RP was selected as initial management by 150 patients; 122 (81%) had complete surgical pathology data available. Median age was 63 yrs. (range 50-79), with 38% > 65 yrs. In this cohort, 11 (9%), 39 (32%), and 72 (59%) had NCCN Very Low-, Low-, and Intermediate-Risk disease. Biological risk (GPS+NCCN) differed from NCCN risk in 28 cases (23%). At surgery, 41 (34%) of patients had AP. Among NCCN VL, Low-Risk and Intermediate-Risk men, 1 (10%), 11 (28%), and 29 (40%) had AP at RP. Combining GPS and NCCN led to redistribution of VL, Low-Risk, and Intermediate-Risk groups and modified rates of AP at RP (Table 1). GPS was a significant predictor of AP (OR per 20 GPS units: 2.4; 95% CI: [1.3, 4.4]; p= 0.004) and remained significant after adjusting for NCCN (OR per 20 units: 2.2; 95% CI: [1.2-4.1]; p=0.01).

Conclusions

We report the first prospective validation of a biopsy-based genomic marker in prostate cancer. GPS is a strong predictor of AP in contemporary PCa patients. Patients with GPS+NCCN intermediate-risk categorization have twice the rate of AP as those with GPS+NCCN very low/low risk. GPS refines risk stratification and may help inform decision-making for patients with clinically low risk PCa.

Funding

Genomic Health, Inc.