To continue to improve on risk calculators developed nearly 20 years ago, Dr. Spratt notes that we must (i) improve performance, (ii) optimize biomarkers for clinically meaningful outcomes, (iii) integrate clinical and genomic data simultaneously (iv) make biomarkers simple to use, and (v) develop algorithms that can be incorporated into existing NCCN guidelines to more accurately risk stratify patients. As such, the aim of this study was to develop a novel, 3-tiered, clinical-genomic risk system that can be incorporated into localized prostate cancer treatment guidelines for predicting metastasis.
To perform this study, the authors used four multi-center cohorts to identify and validate the clinical-genomic risk system in radical prostatectomy samples (training cohorts) and subsequently in pre-treatment biopsy samples (validation cohorts). Using microarray analysis, the expression values for 22 pre-specified biomarkers that constitute Decipher® were extracted. There were 5,937 prospective samples and 991 retrospective samples with long term follow-up for a total of 6,928 men followed for a median 8 years in the radical prostatectomy cohort. The 10-year distant metastasis rates for NCCN low, favorable-intermediate, unfavorable-intermediate, and high-risk were 7.8%, 9.4%, 40.1%, and 41.4%, respectively. When using the novel 3-tiered clinical-genomic risk group classifier, the 10-year distant metastasis rates were 3.7%, 30.7%, and 57.7%, for low, intermediate, and high-risk, respectively, compared to the NCCN guidelines incidence of 7.8%, 9.4%, 40.1% and 41.4% for low, favorable intermediate, unfavorable intermediate, and high risk. This grouping was validated in the pre-treatment biopsy cohort with 10-year rate of distant metastasis of 0%, 30.3%, and 63.2%, respectively. The authors used C-indices to compare the clinical-genomic system to NCCN and CAPRA, and found that the clinical-genomic system (c-index 0.84) outperformed NCCN (c-index 0.71) and CAPRA score (c-index 0.71). Importantly, 33% of men would be reclassified by the clinical-genomic system, including 13% of low risk, 48% of intermediate risk, and 27% of high risk patients. The strength of the current study is the large, multi-institutional sample size and long follow-up, both of which increase generalizability and external validity.
In conclusion, the authors suggest that a genomic classifier in combination with clinicopathologic variables can generate a simple to use 3-tier clinical-genomic risk system that is highly prognostic for distant metastasis. This risk calculator reclassifies 33% of patients, identifies an expanded and pure form of low risk patients, and identifies true intermediate and high risk groups. This appears to be more accurate than clinical risk and can be easily incorporated into NCCN guidelines to inform treatment decisions. Certainly including genetic risk factors into prognostication systems in combination with clinicopathologic factors is logical and better risk-stratifies patients for metastases at the onset of localized disease. Whether these results are similar across different racial profiles remains to be seen and should be a direction of future research.
Presented By: Daniel Eidelberg Spratt, University of Michigan, Ann Arbor, MI, USA
Co-Author(s): Jingbin Zhang, Maria Santiago-Jimenez, John W. Davis, Robert Benjamin Den, Adam Dicker, Christopher J. Kane, Alan Pollack, Ashley Ross, Marguerite du Plessis, Voleak Choeurng, Kasra Yousefi, Zaid Haddad, Elai Davicioni, Sheila Weinmann, Edward M. Schaeffer, Eric A. Klein, R. Jeffrey Karnes, Felix Yi-Chung Feng, Paul L. Nguyen
Written By: Zachary Klaassen, MD, Urologic Oncology Fellow, University of Toronto, Princess Margaret Cancer Centre
at the 2017 ASCO Annual Meeting - June 2 - 6, 2017 - Chicago, Illinois, USA