MP45-08: A Comparison of Biopsy Genomic Test to Predict Adverse Disease on Prostatectomy

Joon Yau Leong, Alex Uhr*, Benjamin Rudnick, Andrea Quinn, Brian Calio, Edouard Trabulsi, Tingting Zhan, Leonard Gomella, Mark Mann, Collaborators of PURC, Philadelphia, PA

Introduction: Despite progressively refined algorithms, it remains a challenge to accurately stratify patients with localized prostate cancer (PCa) into categories that correctly predict disease aggression and progression. Genomic analysis tests have attempted to remedy this with the addition of personalized cancer genomic risk categories which add proprietary genomic information into existing stratification algorithms. Currently, the three most commonly utilized tests on biopsy pathology are Oncotype DX, Decipher, and Prolaris. This study aimed to identify which genomic analysis test best predicts favorable or adverse pathology during radical prostatectomy.

Methods: Using data from the Pennsylvania Urologic Regional Collaborative (PURC), all patients who underwent genomic analysis on prostate biopsy with one of the three genomic tests, Oncotype DX, Decipher, or Prolaris were identified. Patients were divided into genomic risk categories and information on biopsy and prostatectomy pathology, treatment, and PSA was analyzed.

Results: No genomic analysis test was independently predictive of spread of disease outside the prostate. Patients with low genomic risk PCa as determined by Oncotype DX were statistically more likely to have a Grade Group (GG) upgrade compared to those in the intermediate risk group (p=0.045). When compared to patients with high Decipher genomic risk, patients with low risk on Decipher biopsy were statistically more likely to have a GG upgrade (p=0.027). For those who had Prolaris genomic analysis, patients with both low and intermediate genomic risk PCa were more likely to have a GG upgrade than patients with high genomic risk disease (p=0.001 and p<0.001 respectively). To examine GG upgrade risk, patients with either GG upgrade or pT3 or greater disease and PSA were compared across risk groups of each genomic analysis test in a multivariable logistic regression. There was no significant difference in upgrading risk between genomic risk levels for either Oncotype DX or Decipher. When compared to the high genomic risk patients, patients with low or intermediate genomic risk based on the Prolaris test were more likely to have a GG upgrade (p=0.001 and p=0.042 respectively).

Conclusions: In our study, patients in the low-risk category for all three genetic test groups demonstrated significantly higher risk for GG upgrading than patients designated in each test group’s high-risk cohort. This gives providers heightened awareness of the persistent threat of more aggressive disease and possible progression despite a seemingly reassuring report from genomic testing.

Source of Funding: None