Search

PURPOSE Nearly all men with prostate cancer treated with androgen receptor (AR) signaling inhibitors (ARSIs) develop resistance via diverse mechanisms including constitutive activation of the AR pathway, driven by AR genomic structural alterations, expression of AR splice variants (AR-Vs), or loss of AR dependence and lineage plasticity termed neuroendocrine prostate cancer. Understanding these de novo acquired ARSI resistance mechanisms is critical for optimizing therapy. MATERIALS AND METHODS A novel liquid biopsy technology was used to collect mRNA from circulating tumor cells (CTCs) to measure expression of AR-Vs, AR targets, and neuroendocrine prostate cancer markers. An institutional review board–approved prospective cohort (N = 99) was used to identify patterns of gene expression. Two prospective multicenter phase II clinical trials of ARSIs for men with castration-resistant prostate cancer (ClinicalTrials.gov: NCT01942837 [enzalutamide, N = 21] and NCT02025010 [abiraterone, N = 27]) were used to further validate these findings. RESULTS Hierarchical clustering of CTC transcripts identified two distinct clusters. Cluster 2 (C2) exhibited increased expression of AR-regulated genes and was associated with worse overall survival (median 8.6 v 22.4 months; P < .01; hazard ratio [HR] = 3.45 [1.9 to 6.14]). In multivariable analysis, C2 was prognostic independent of other clinicopathologic variables. AR-V status was not significant when accounting for C2. Upon further validation in pooled multicenter phase II trials, C2 was associated with worse overall survival (15.2 months v not reached; P < .01; HR = 8.43 [2.74 to 25.92]), prostate-specific antigen progression-free survival (3.6 v 12 months; P < .01; HR = 4.64 [1.53 to 14.11]), and radiographic progression-free survival (2.7 v 40.6 months; P < .01; HR = 4.64 [1.82 to 17.41]). CONCLUSION We demonstrate that a transcriptional profile detectable in CTCs obtained from liquid biopsies can serve as an independent prognostic marker beyond AR-V7 in patients with metastatic prostate cancer and can be used to identify the emergence of multiple ARSI resistance mechanisms. This is currently being investigated in additional prospective trials.

238 Background: Patients with high-risk prostate cancer have an increased risk of recurrence and death from prostate cancer. Intense NHT prior to RP has been associated with favorable pathologic responses in a subset of patients with high-risk disease. We conducted a pooled analysis of post-RP outcomes of four, multicenter, randomized, clinical trials of intense NHT prior to RP. Methods: All patients included were enrolled on trials evaluating intense NHT followed by RP. The primary endpoint was time to biochemical recurrence (BCR), defined as the time from RP to prostate specific antigen (PSA) >0.1 ng/mL or start of first post-RP therapy. Secondary endpoints included metastasis-free survival (MFS), overall survival (OS), and time to testosterone recovery. We evaluated outcomes in the overall population, favorable responders (defined as a pathologic complete response or minimum residual disease ≤ 5 mm at RP), and non-responders. We correlated pathologic parameters, including RP pathologic response, PTEN status, ERG status, and presence of intraductal carcinoma with BCR. Results: 121 patients were included of whom 78.5% (n=95) had high-risk disease by NCCN risk categories. Following NHT, 20.7% (n=25) had 0-5 mm of residual tumor (including 9.1%, n=11, with pathologic complete response). Overall, 52 patients (43.0%) experienced BCR, of whom two were exceptional responders (Table). Testosterone recovery was observed in 94.0% of patients (n=109/116). PTEN loss and presence of intraductal carcinoma were associated with shorter time to BCR. Conclusions: In this pooled analysis of clinical trials, we demonstrate that patients who achieve a favorable pathologic response to NHT prior to RP experienced a decreased PSA relapse rate and prolonged time to BCR. Larger residual tumor volumes and tissue-based parameters, including PTEN loss and intraductal carcinoma, were associated with BCR. Additional follow-up is warranted to evaluate the impact on long-term outcomes. Clinical trial information: NCT00298155; NCT00924469; NCT01547299; NCT02268175 . [Table: see text]

1507 Background: Approximately 10% of men with advanced prostate cancer (PC) have pathogenic/likely pathogenic variants (PV) in cancer susceptibility genes and their identification may lead to targeted therapy. Genetic testing (GT) can also guide cancer surveillance and prevention for family members. While GT is recommended for men with potentially lethal PC, traditional testing models are strained, and access limited. The ProGen study examined a novel pretest model aimed at providing access to GT while promoting informed consent. Methods: Inclusion criteria were: potentially lethal PC (metastatic, localized with Gleason score ≥8, rising/persistent PSA after local therapy), diagnosis age ≤ 55 years, prior malignancy, family history suggestive of a PV and/or at oncologist’s discretion. Consented subjects from 3 sites were randomized 3:1 to video education (VE) or in-person genetic counseling (GC). Subjects who consented to GT had 67 genes analyzed (Ambry, USA) with results disclosed by telephone by a genetic counselor. Outcomes included GT uptake, PV prevalence, and survey measures of satisfaction, distress, genetics knowledge, family communication, and impact on cancer care (obtained at the time of intervention, and at 1, 4, and 12 months after result disclosure). Two-sided Fischer exact tests were used for between-arm comparisons. Results: Over a 2-year period: 662 subjects were randomized, VE or GC were completed by 604 subjects (VE: 93.1%, GC: 88.8%) of whom 596 subjects (VE:98.9%, GC:97.9%) consented to GT. To date, 591 subjects have completed GT (VE: 99.3%, GC: 98.6%). At the time of intervention, most subjects agreed or strongly agreed that their assigned arm was useful (VE: 95%, GC: 88%). Differences were not statistically significant. Notably, 84 PV were identified in 78 subjects (13.2%), with BRCA1/2 PV accounting for 32% of subjects with a positive result ( BRCA2:21, BRCA1:4). Conclusions: In this randomized trial, both novel VE and traditional GC yielded high GT uptake without significant differences in outcome measures of acceptability and satisfaction. VE enabled access to critical GT results while maintaining the core tenants of informed consent. PV were found in 13.2% of subjects, 32% of whom had BRCA1/2 PV. Analysis of collected survey data to inform strengths and limitations of VE as compared with pretest GC will be presented. Clinical trial information: NCT03328091.

223 Background: Guidelines for optimal sequencing of approved mCRPC therapies do not exist. This study evaluated treatment patterns and overall survival (OS) among patients (pts) with mCRPC treated with Ra-223 in an academic clinical setting. Methods: A retrospective chart review study was conducted at Dana-Farber Cancer Institute of bone metastases-predominant mCRPC pts treated with Ra-223. Treatment patterns from 2013-2018 were evaluated including Ra-223 initiation pre- vs. post-chemotherapy (chemo). OS was examined using Kaplan-Meier medians and 95% confidence intervals (CIs). Results: 220 pts were treated with Ra-223 (64 pre-chemo, 83 post-chemo, 73 no chemo). Median Ra-223 injections per pt was 6.0 and 5.0 in the pre- vs. post-chemo cohorts, respectively (p

43 Background: PARP inhibitors (PARPi) and platinum chemotherapy cause synthetic lethality and result in clinical benefit in men with mCRPC whose tumors harbor deleterious alterations in DDR genes. BRCA2 is the strongest predictor, but data supports a role for BRCA1, ATM, and other genes. PARPi are more expensive than platinum chemotherapy; whether they are more effective is not known. We evaluated outcomes in men with DDR-deficient mCRPC treated with PARPi or platinum chemotherapy. Methods: We identified men treated at Dana-Farber Cancer Institute who received olaparib or carboplatin for mCRPC and underwent tumor mutation profiling. Progression-free survival (PFS) was determined using Prostate Cancer Working Group 3 guidelines. Pathogenic germline or somatic single-nucleotide variants, insertions/deletions, and copy number loss in BRCA2, BRCA1, and ATM, were identified. Log-rank test was utilized to compare differences in PFS. Results: We identified 27 men who received olaparib: 19/27 (70.4%) received prior taxane chemotherapy and 3/27 (11.1%) prior carboplatin; 16 had mono- or bi-allelic alterations in BRCA2 and 2 had alterations in BRCA1 or ATM. We identified 41 men treated with carboplatin: 39/41 (95.1%) received carboplatin in combination with a taxane; 36/41 (87.8%) received prior taxane chemotherapy and 7/41 (17.1%) prior PARPi; 16 had BRCA2 alterations and 5 had alterations in BRCA1 or ATM. Among men with BRCA2 alterations, there was no significant difference in PFS (HR 0.71, 95% CI 0.45-1.11; p = 0.13) between those treated with olaparib (4.9 months [m]) versus carboplatin (5.4 m). Similarly, no difference in PFS (HR 0.80, 95% CI 0.54-1.16; p = 0.24) was observed among men with BRCA2, BRCA1, or ATM alterations treated with olaparib (3.8 m) versus carboplatin (3.6 m). Conclusions: In a small real-world cohort of men with mCRPC harboring deleterious alterations in BRCA2, BRCA1, and ATM, there was no difference in PFS between those treated with olaparib versus carboplatin. These findings support further investigation into the comparative effectiveness and cost-effectiveness of PARPi and platinum chemotherapy in men with DDR-deficient mCRPC.