3266 - Clinical Impact and Concordance of Three Advanced Risk Stratification Tools for ADT Decision-Making In Salvage Radiotherapy for Biochemically Recurrent Prostate Cancer
Presenter(s)
S. Feinstein1, S. Zaaijer2, G. K. Harada3, O. Yazdanpanah2, A. Rezazadeh2, and M. Shahait4; 1Department of Radiation Oncology, University of California, Irvine, Orange, CA, 2Division of Hematology and Oncology, University of California, Irvine, Orange, CA, 3Department of Radiation Oncology, University of California - Irvine, Orange, CA, 4Department of Urology, University of California, Irvine, Orange, CA
Purpose/Objective(s): The addition of androgen deprivation therapy (ADT) to salvage radiation therapy (RT) for biochemically persistent/recurrent prostate cancer after radical prostatectomy (RP) remains an individualized decision. Multiple tests are available to guide this clinical decision. Decipher is currently the only molecular classifier supported by NCCN in the post-RP setting. ArteraAI recently introduced a post-RP predictive assay, and PAM50 is the only prospectively validated predictive biomarker in this context. Whether these assays provide concordant classification of potential ADT benefit remains unknown. We evaluated concordance between these three classifiers using validated assay-specific cutoffs.
Materials/Methods: We conducted a retrospective single-institution study of patients with biochemically persistent or recurrent prostate cancer who underwent Decipher, ArteraAI, and PAM50 testing between September 2025 and February 2026. We adopted assay-specific cutoffs (Decipher High, ArteraAI High, PAM50 Luminal B) to dichotomize patients as more or less likely to benefit from ADT with salvage RT. All patients were staged with PSMA PET/CT and prostate mpMRI and had no clinical or pathologic evidence of regional or distant metastases. Agreement between classifiers was assessed using observed concordance.
Results: Eleven patients were included (median age 67 years) with a median PSA of 0.64 ng/mL (range 0.09–1.1) at the time of genomic testing. At RP, 82% had Grade Group 2–3 disease, 82% were pT2–3a, 64% had positive surgical margins, and 91% had a detectable post-operative PSA (>0.02 ng/mL). A greater likelihood of ADT benefit with salvage RT was identified in 91% (10/11) by Decipher, 45% (5/11) by ArteraAI, and 36% (4/11) by PAM50. Observed agreement (ADT benefit vs no ADT benefit) was 54.5% between ArteraAI and Decipher (6/11) and between ArteraAI and PAM50 (6/11), compared with 27.3% between Decipher and PAM50 (3/11). Concordance across all three assays was observed in 2 of 11 patients, both classified as likely to benefit from ADT by all tests. Three patients had only a single classifier indicate ADT benefit (two by Decipher and one by PAM50).
Conclusion: In this preliminary single-institution cohort, we observed discordance in classifications of potential ADT benefit across the three assays, with Decipher more frequently classifying patients as more likely to benefit from ADT compared with ArteraAI and PAM50. The limited sample size reflects the recent clinical availability of the post-prostatectomy ArteraAI assay. A larger study is planned to define how best to integrate these tools into individualized salvage ADT decision-making, with the goal of minimizing unnecessary treatment and improving quality of life.