Creating precision in metastatic hormone sensitive prostate cancer: Using tumor genetics and patient characteristics to predict outcomes and improve treatment selection

Background: Metastatic prostate cancer is a disease of older veterans and has a wide range of behavior as many patients have indolent disease. Even in metastatic disease, it is common for men to die WITH prostate cancer not FROM prostate cancer, especially in patients with significant comorbid disease. Methods to determine the risk of prostate cancer-specific death compared to non-cancer death are essential to appropriately guide therapy. Comorbidity evaluation in cancer patients can inform treatment selection and the risk of adverse events to provide personalized care. Furthermore, tumor genomics and response to therapy show significant promise in the prediction of prostate cancer survival and time to castration resistance. Additional data elements including volume and site of disease, treatments over time, and cause of death in the VA facilitate the analysis of prostate cancer outcomes. Significance: There are is a critical unmet need for reliable, patient-specific prostate-cancer risk assessment, which could guide therapeutic intensity in order to maximize treatment while minimizing risk of adverse events and non-cancer deaths. Our study findings will create the a prognostic model, the VA-PREMA (Precision Risk/Effectiveness Multi-omic Assessment) that will estimate risk of prostate cancer-specific mortality and non- prostate cancer death. The ultimate goal of the model is to inform decision making in Veterans and civilians with cancer and comorbid disease. Innovation & Impact: Our study is innovative both conceptually and technically. We will leverage the VA database, include somatic tumor sequencing from the National Precision Oncology Program, and use novel methods to assess the response to therapy. Due to the large amount of veterans available, we can test the impact of different treatment strategies on cancer treatment and outcomes. Our proposal fills a critical gap in prostate cancer care as no model of metastatic hormone sensitive prostate cancer (mHSPC) exists. Specific Aims: In Aim 1, we will evaluate the association of clinical prostate cancer features, patient factors, tumor genomics, and treatment efficacy features with prostate cancer-specific mortality, overall survival, and time to castration-resistance, accounting for competing risk of non-prostate cancer mortality to accurately prognosticate outcomes in de novo mHSPC. In Aim 2, we will develop and validate a predictive model, the VA- PREMA (Precision Risk/Effectiveness Multi-omic Assessment), that will predict the prostate cancer-specific mortality compared to non-cancer mortality with different initial treatments. Methodology: We will compile a cohort of Veterans (2014-2021) diagnosed with mHSPC in the VA cancer registry and collect data on patient characteristics as well as prostate cancer features. Imaging data will be reviewed for volume and site of metastatic prostate cancer as well as specific treatments. Data from NPOP will be categorized into genomic pathways of specific pathological alteration to organize the alteration of interest. Next, we will fit distinct models to compare the independent effect of prostate cancer features, patient factors, tumor genomics, and treatment efficacy on prostate cancer-specific mortality, non-prostate cancer death, and castration resistance. The model will then be validated in a cohort of veterans with metachronous metastatic prostate cancer to determine concordance/performance. Next Steps/Implementation: The VA-PREMA model could be used in a prospective fashion as a shared decision making tool as well as a selection criteria for clinical trials. There is considerable potential for VA- PREMA to identify patients who have low prostate cancer risk by high rates of comorbid disease that could have therapy de-escalated in a randomized trial. Project Number: 1I01CX002946-01 | Fiscal Year: 2025 | NIH Institute/Center: Veterans Affairs (VA) | Principal Investigator: Martin Schoen (+3 co-PIs) | Institution: ST. LOUIS VA MEDICAL CENTER, St. Louis, MO | Activity Code: I01 | Study Section: Special Emphasis Panel[ZRD1 SPLP-Y (01)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11111899