Introduction

Radical prostatectomy (RP) is a primary treatment option for men with intermediate risk (IR) prostate cancer (PCa). Though many will be effectively cured with local therapy alone, these men are by definition at higher risk of disease recurrence. In this study, we evaluated whether a genomic signature of metastasis risk (Decipher PCa classifier) could improve pre-operative staging for predicting lymph node invasion (LNI).

Methods

We examined 263 NCCN intermediate men treated with RP and extended template pelvic lymph node dissection (ePLND) from 2007-2015 at MD Anderson Cancer Center, Houston, Texas. Patients were categorized into three risk groups: 1) men with N1 disease (N1), 2) men without N0, but who had either ≥pT3 stage, RP Gleason score ≥8, lymphovascular invasion or tertiary Gleason 5 pattern (N0 high-risk [HR]) and 3) men with no high-risk features at RP (N0IR). Decipher scores were obtained from 263 RP specimens and 25 matching biopsy specimens. Fisher's exact test was used to compare the difference in patient risk groups. Logistic regression analysis was used to evaluate performance of Decipher for prediction of LNI. Discrimination of the Partin tables (≥2%) and combined model of Partin tables (≥2%) and Decipher (>0.6) was assessed using c-index. Concordance of biopsy and RP Decipher (low- and intermediate- vs high-risk) was also assessed.

Results

Of the 263 men, 42 (16.0%), 98 (37.2%) and 123 (46.8%) men were categorized as N1, N0HR and N0IR risk groups, respectively. Partin tables classified 34/42 (81%) N1, 70/98 (71%) N0HR and 66/123 (54%) N0IR men as high clinical risk (≥2%) for LNI (p=0.0012). Decipher classified 23/42 (55%) N1, 34/98 (35%) N0HR and 35/123 (29%) N0IR as high genomic risk (>0.6) for metastasis (p=0.013). After adjusting for Partin Tables, Decipher high genomic risk had an odds ratio of 2.3 (95% CI 1.2-4.5) as a predictor of LNI (p=0.02). Addition of Decipher to Partin Tables improved the c-index from 0.60 (95%CI 0.53-0.67) to 0.66 (95%CI 0.57-0.75). The concordance of Decipher risk groups between matched Biopsy and RP specimens was 84%.

Conclusions

Decipher may be an important adjunct tool to improve preoperative staging that may be useful for prioritizing intermediate risk patients to ePLND. Further investigation of Decipher biopsy specimens is required to validate these findings.

Funding

GenomeDx Biosciences