Prognostic value of baseline circulating tumor DNA (ctDNA) tumor fraction (TF) in metastatic hormone-sensitive prostate cancer (mHSPC).

5085 Background: In mHSPC, CHAARTED-defined disease volume and prior local definitive therapy are prognostic factors. These factors are imperfect and may not capture the clinical heterogeneity in mHSPC. Thus, new prognostic biomarkers are needed. ctDNA is a promising tool given the ease of obtaining genomic data and ctDNA TF. Although ctDNA has been well studied in metastatic castration-resistant prostate cancer (mCRPC), the prognostic value of ctDNA in mHSPC is unknown. Methods: We present a multi-center, non-interventional, prospective study of ctDNA in patients (pts) with newly diagnosed mHSPC. ctDNA was collected at baseline (prior to androgen deprivation therapy [ADT]), on mHSPC therapy (1 month [mo] and/or 6-7 mo after), and at mCRPC. Specimens were processed using the FoundationOne Liquid CDx assay (324 genes). Kaplan-Meier and Cox regression models were used to measure clinical endpoints including overall survival (OS) and time to mCRPC (TTCRPC). Results: Between January 2018 and August 2023, 73 pts were enrolled, of whom 56 had baseline ctDNA successfully processed. Median age was 72 years (interquartile range [IQR]: 66-76). 77% were White; 16% were Black. Most (68.5%) had de novo metastatic disease; 31.5% had recurrent disease. Median pre-ADT PSA was 38.0 ng/mL (IQR 15.2-175.0). CHAARTED high volume (HV) disease was present in 73%. Common metastatic sites were bone (80%) and lymph nodes (66%). All were treated with ADT, 19% of whom had ADT alone. Most (63%) had an androgen receptor signaling inhibitor (ARSI) or docetaxel (6%) added; 12% had triplet therapy (ADT with ARSI and docetaxel). Baseline detectable TF (TF+) was present in 35/56 (62.5%); 21/56 (37.5%) had baseline undetectable TF (TF-). Frequent gene alterations were TP53(39%), PTEN (23%), and TMPRSS2(16%). Pts with baseline TF+ had a median OS (mOS) of 37 mo vs not reached (NR) (HR 2.57, 95% CI 0.55-11.93, p=0.2) and median TTCRPC (mTTCRPC) of 15 mo vs NR (HR 2.07, 95% CI 0.75-5.71, p=0.2) compared to pts with baseline TF-. In subgroup analysis of pts with HV disease, mOS was 37 mo with baseline TF+ vs NR with TF- (HR 2.77, 95% CI 0.35-22.23, p=0.3); mTTCRPC was 12 mo vs 18 mo (HR 1.56, 95% CI 0.44-5.49, p=0.5). In pts with low volume disease, a similar observation with mTTCRPC was found in TF+ vs TF- (14 mo vs NR, HR 1.7, 95% CI 0.23-12.0, p=0.6). Among 30 pts with on-treatment samples, TF+ during therapy had a mOS of 37 mo vs 61 mo (HR 2.67, 95% CI 0.69-10.35, p=0.2) and mTTCRPC of 12 mo vs 52 mo (HR 1.75, 95% CI 0.68-4.47, p=0.2) compared to TF-. Conclusions: Pts with mHSPC and baseline TF+ appear to have a quantitative but not statistically significant worse OS and TTCRPC compared to pts with TF- regardless of disease volume. Inclusion of additional pts and longer follow-up may validate whether ctDNA can improve treatment decision making in mHSPC. Further work looking at TF threshold (>10% vs <10%) is ongoing.