LAUSR

5058 Background: High-dose testosterone (HDT) is active in mCRPC pts and may allow successful re-sensitization to previously utilized androgen-axis targeted therapies. The relationship of genomic alterations in AR gene to HDT responsiveness is unclear. Methods: Analysis of consecutive pts treated with ≥1 dose of HDT (testosterone cypionate q 2-4 weeks n = 29; continuous gel n = 4). Baseline characteristics, ctDNA data (Guardant360), and clinical outcomes were assessed. Presence of genomic AR alterations included amplifications (amps) and mutations (muts); all muts had allele fraction ≥0.3%. PSA response rates included PSA declines of > 30% or ≥50%. PSA-progression-free survival (PSA-PFS) was defined as HDT start date to PSA ≥ 25% over baseline after a second confirmed PSA rise. Results: Between May 2016 and Feb 2018, 33 mCRPC pts had median age 73 (58-85), 39% Gleason 8-10, 100% bone mets, 24% nodes + bone, and median baseline PSA level 36.1 ng/mL (0.04-1290). HDT was given post-median of 2 (1-10) CRPC therapies. 73% (24/33) of pts previously received abiraterone (n = 14), enzalutamide (n = 4), or both sequentially (n = 6) prior to HDT for a median of 10.5 months (0.7-56.8). Baseline ctDNA showed 42% AR alterations (amps = 8, muts = 4, both = 2); 33% TP53, and 6% DNA repair (ATM n = 1; BRCA2 n = 1). With median follow-up 4.4 months, HDT given for median of 4.2 months (95% CI, 3.6-4.8); 29% had PSA ≥50% response and 45% PSA ≥30% response. Median PSA-PFS is immature at 5.5 months (95% CI, 1.5-9.5); 14 pts still on HDT treatment. Grade ≥3 AEs were observed in 6% of pts (G4 thrombocytopenia = 1; G4 asthenia = 1). For pts with baseline AR alterations and HDT treatment, repeated ctDNA assays (n = 7) showed that 100% had decreased AR alterations. No relationship between PSA response and baseline ctDNA AR characteristics are discerned at this time. Conclusions: HDT was safe and active in a subset of mCRPC. Responses were clearly noted for men receiving continuous daily testosterone gels, thus continuously high testosterone levels are active in addition to injection-induced bipolar changes. Further understanding of the genomic alterations predicting responsiveness to HDT in mCRPC is required.