1116 - Discordance of Adverse Molecular Features between the 22-Gene Genomic Classifier Score, Histologic Grade, and NCCN Risk Groups: Analysis of Over 200,000 Patients
Presenter(s)
M. J. Zelefsky1, H. M. Ryu2, P. T. Tran3, J. Proudfoot2, Y. Hao4, M. Alshalalfa4, E. Davicioni5, P. Febbo4, N. B. Desai6, J. C. Hu7, A. Briganti8, E. M. Schaeffer9, A. E. Ross10, A. Y. Jia11, D. R. Wise12, and D. E. Spratt11; 1NYU Langone Health, New York, NY, 2Veracyte, San Diego, CA, 3University of Maryland School of Medicine, Baltimore, MD, 4Veracyte Inc, San Diego, CA, 5Veracyte Inc., San Diego, CA, 6Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, 7NewYork-Presbyterian/Weill Cornell Medical Center, New York, NY, 8Vita Salute San Raffaele University, Milan, Italy, 9Northwestern University, Evanston, IL, 10Northwestern University, Chicago, IL, 11Department of Radiation Oncology, University Hospitals Cleveland Medical Center/ Seidman Cancer Center, Cleveland, OH, 12Department of Medicine, Perlmutter Cancer Center at NYU Langone Medical Center, New York, NY
Purpose/Objective(s): The 22-gene genomic classifier (Decipher prostate (GC)) improves risk stratification independently of traditional clinicopathologic tools and NCCN risk groups. There is considerable heterogeneity of GC scores within a given grade group (GG) or NCCN risk group, and it is unclear if for a given GC score, irrespective of NCCN risk group and GG, there are similar adverse molecular features (AMF). We aim in the present study to determine the concordance of AMF based on GC score, GG, and NCCN risk group to understand what best captures biologically defined aggressive disease.
Materials/Methods: De-identified records of 206,412 patients who underwent GC biopsy testing (Veracyte, Inc., San Diego, CA) between October 2016 to February 2024 were retrieved from the GRID registry (NCT02609269). Distribution of GC scores were assessed by NCCN risk group and by GG. The five AMFs included expression signatures for TP53 mutations (p53+), PTEN inactivation, RB loss, neuroendocrine-like (NE-like), and low AR-activity, all of which correlate with worse oncologic outcomes. Prevalence of AMFs were examined among GC low, intermediate, high, and very high (>0.85) risk groups, and further assessed within NCCN and GG categories.
Results: Of 38,329 NCCN low-risk patients, the frequency of GC low/intermediate, high, and very high-risk GC scores were 86%, 12% and 2%, respectively. Among 136,420 patients with GG1-2 tumors, the incidence of high and very high-risk GC scores were 20% and 6%, respectively. Of 20,504 NCCN high-risk patients, the frequency of low/intermediate, high, and very high-risk GC were 33%, 29% and 38%, respectively. Among 28,045 patients with GG4-5 disease, the incidence of low/intermediate-risk GC scores were 26%. The presence of AMFs most closely correlated with GC scores, even within low grade tumors. We observed a nearly 20-fold increase in the presence of AMFs when comparing GC low to GC very high-risk. Among patients with GG1-2 disease, the prevalence of harboring an AMF was observed in 8% of GC low, 23% GC intermediate, 45% GC high, and 75% GC very high-risk. Interestingly, a similar distribution and relationship to GC of AMF frequency was observed for those with GG4-5 disease; the prevalence of harboring an AMFs was observed in 13% GC low, 25% GC intermediate, 47% GC high, and 82% GC very high-risk.
Conclusion: Discordance of the GC score with the histologic grade and NCCN risk grouping was observed. Adverse molecular features had greatest concordance with GC scores and had a similar distribution based on GC score within low grade versus high grade tumors. This study gives novel insights into the heterogeneous biologic landscape of prostate cancer that is insufficiently captured by current clinicopathologic tools, and clinical trials are warranted based on molecular phenotype rather than tumor grade.