Clinical and Molecular Determinants of PSA Response to Bipolar Androgen Therapy in Prostate Cancer

BACKGROUND

Bipolar Androgen Therapy (BAT) is a novel therapy known to be effective in a subset of men with metastatic castrate resistant prostate cancer (mCRPC). A better understanding of responders and non-responders to BAT would be useful to clinicians considering BAT therapy for patients. Herein we analyze clinical and genetic factors in responders/non-responders to better refine our understanding regarding which patients benefit from this innovative therapy.

METHODS

mCRPC patients were assessed for response or no response to BAT. Patients with PSA declines of greater than 50% from baseline after 2 or more doses of testosterone were considered to be responders. Whereas, Non-responders had no PSA decline after 2 doses of testosterone and subsequently manifest a PSA increase of >50%. Differences between these two groups of patients were analyzed using clinical and laboratory parameters. All patients underwent genomic testing using circulating tumor DNA (ctDNA) and germline testing pre-BAT.

RESULTS

Twenty five patients were non-responders and 16 were responders. Baseline characteristics between non-responders and responders varied. Responders were more likely to have had a radical prostatectomy as definitive therapy and were more likely to have been treated with an androgen receptor (AR) antagonist (enzalutamide or apalutamide) immediately prior to BAT (compared to abiraterone). Duration of prior enzalutamide therapy was longer in responders. Non-responders were more likely to have bone-only metastases and responders were more likely to have nodal metastases. Assays detected ctDNA AR amplifications more often in responding patients. Responders trended toward having the presence of more TP53 mutations at baseline.

CONCLUSIONS

BAT responders are distinct from non-responders in several ways however each of these distinctions are imperfect. Patterns of metastatic disease, prior therapies, duration of prior therapies, and genomics each contribute to an understanding of patients that will or will not respond. Additional studies are needed to refine the parameters that clinicians can utilize prior to choosing among the numerous treatment alternatives available for CRPC patients. This article is protected by copyright. All rights reserved.

Characterization of ctDNA findings at the end of life in patients with prostate cancer

230

Background: Somatic alterations change over time in response to treatment and disease progression in patients with prostate cancer (PCa). We report ctDNA findings of patients who died from PCa within 3 months of a ctDNA assessment to better characterize patients with fatal disease. Methods: A total of 118 patients with PCa specific mortality who had been treated at Tulane Cancer Center and had ctDNA assessments between 2015 and 2022. The ctDNA was assessed by Guardant360 (Guardent Health, Inc) assays to identify alterations, pathogenic mutations and/or copy number alterations (CNAs) in cancer-related genes. Clinical annotation including treatment history, genetics, and staging were also obtained. Statistical analyses included Fischer’s Exact and Wilcoxon tests. Results: Of the 118 patients with PCa specific mortality, 42% (49/118) had a ctDNA assessment <3 months from death. Of 49 CRPC patients tested within 3 months of death, the median number of life extending therapies (LET) at death was 5 (2-9). Patients had a median of 2 (0-6) LET prior to first ctDNA screening and 3 (0-7) LET in between first and last ctDNA assessment. Of the total gene alterations detected on ctDNA analysis, within 3 months of death, the most common alterations detected were 65.3% (32/49) TP53, 44.9% (22/49) AR, 28.6% (14/49) EGFR, 24.5% (12/49) PIK3CA, 22.4% (11/49) MYC, and 20.4% (10/49) CDK6. In a paired analysis(n= 45) of first and last ctDNA screening, AR (OR= 2.35, 95% C.I. (0.99, 5.62), p= 0.05), CDK6 (OR= 4.00, 95% C.I. (1.02, 15.68), p= 0.04), FGFR1 (OR= 9.51, 95% C.I. (1.14, 79.61), p= 0.03), and EGFR (OR= 9.71, 95% C.I. (2.06, 45.83), p= 0.0009) were significantly more likely to be detected in ctDNA screening within 3 months of death. In addition, ctDNA alterations in general were significantly more likely to be detected at the end-of-life (p=< 0.00001). Other ctDNA gene mutations did not have statistically significant increases. Conclusions: An analysis of patients with PCa mortality showed most frequent gene alterations in TP53, AR, EGFR, PIK3CA, MYC, and CDK6. When comparing patients’ first and last ctDNA, alterations were significantly more likely by time of death in AR, CDK6, EGFR, FGFR and there was a significant increase in overall detection of somatic alterations in ctDNA. These analyses are limited to the genes assessed on the ctDNA panel and may or may not reflect all of the functionally relevant alterations. Further the CNAs reported herein may reflect broad genomic changes rather than specific gene alterations.

Comparison of ctDNA between African American and Caucasian patients with CRPC post abiraterone and/or enzalutamide

35

Background: Genetic differences between African American and Caucasian patients with advanced prostate cancer may contribute to racial disparities in terms of treatment outcomes and survival, hence further exploration is warranted. We assessed ctDNA differences between African American and Caucasian men in the setting of CRPC post treatment with abiraterone and/or enzalutamide. Methods: From 2015 through 2022 at Tulane Cancer Center, 250 patients with CRPC including 50 African Americans and 200 Caucasian with prior abiraterone and/or enzalutamide treatment were included. All patients had ctDNA assessed via Guardant360. Data including both gene mutations and types of mutations were for between 73-80 genes. Clinical annotation including initial staging, treatment history, and genetic testing were obtained. Statistical analyses included Fisher’s exact test and Wilcoxon rank-sum test. Results: The most common pathogenic/likely pathogenic (P/LP) alterations in both African Americans and Caucasians were TP53 (44% and 46%, respectively), AR (50% and 39%), and PIK3CA (14% and 9%). CDK12 (OR= 8.955, 95% C.I. [2.156, 37.192], p=0.003) and KIT (OR= 5.710, 95% C.I. [1.235, 26.397]. p=0.031) alterations were more frequently detected in African Americans. In terms of pathologic mutation type, frameshift mutations were significantly more frequent in African Americans (OR= 2.293, 95% C.I. [1.103, 4.769], p=0.035). All patients were CRPC at the time of testing and had prior abiraterone and/or enzalutamide, but there were no significant differences between African American and Caucasian patients with regards to prior life-extending therapies. Conclusions: African Americans with CRPC post treatment with abiraterone and/or enzalutamide had a higher frequency of P/LP CDK12 and KIT mutations, which have both been shown to lead to aggressive clinical features and treatment resistance. African Americans also had a higher incidence of frameshift mutations, a finding not previously noted. [Table: see text]

Evaluation of ctDNA in patients with CRPC with pathogenic germline mutations in BRCA2

253

Background: Approximately 3-5% of advanced prostate cancer patients have pathogenic BRCA2 mutations in germline tests. In this study, we examine the relationship between pathogenic germline BRCA2 mutation and somatic changes in ctDNA. Methods: Germline screenings were performed by Invitae multi-cancer gene panel which includes 50-84 genes. ctDNA alterations were detected by Guardant 360 assays which report somatic changes in 70-83 genes. All ctDNA samples were collected in post-abiraterone and/or enzalutamide (in CRPC patients). Any pathogenic/likely pathogenic somatic alterations in the ctDNAs with more than 0.1% of allelic fraction were included in this cohort. The type of mutation detected in ctDNA was also assessed (truncating, point, etc.). Statistical significance for comparison is calculated with Fischer Exact Probablity Test and Chi-Square Test. Results: A total of 11 patients had germline BRCA2 pathogenic mutations and ctDNA assays; 267 patients had no germline DNA pathogenic alterations and ctDNA assays. Compared to germline normal patients, the germline BRCA2 mutations were less likely to have AR alterations on ctDNA (OR=0.2133, 95% C.I. [0.087, 0.525], p-value = 0.0003). BRCA2 germline positive patients were also more likely to have a mutated BRCA1, BRCA2, and TP53 ctDNA (OR=7.899, 95% C.I. [1.2745, 48.9548], p=0.055), (OR=7.899, 95% C.I. [1.7529, 16.059], p=0.008), and (OR=6.442, 95% C.I. [2.449, 16.946], p=0.00001), respectively. All other ctDNA assessed genes were mutated at a similar frequency between germline BRCA2 mutated and “normal” germline patients. BRCA2 germline mutations patients are less likely to have copy number alterations (CNVs) (OR=0.3992, 95% C.I. [0.2168, 0.7352], p=0.0031) and more likely to have frameshift mutations (OR=2.3182, 95% C.I. [1.169, 4.5972], p=0.0183). Conclusions: The ctDNA testing in this CRPC population (after abiraterone and/or enzalutamide) was less likely to find AR alterations and more likely to find pathogenic mutations for BRCA1, BRCA2, and TP53 in BRCA2 germline positive patients. In addition, BRCA2 germline positive CRPC patients were more likely to have frameshift mutations and less likely to have CNVs than those with an intact germline. Characterizing the mutational landscape in BRCA2 germline mutated patients may help to better define underlying disease biology. Those with BRCA mutated germline may be less likely to have AR driven tumors. More study is needed to better understand patients with underlying DNA repair defects.

Analysis of TP53 gain of function mutations in metastatic castration-resistant prostate cancer

246

Background: TP53, an oncogene implicated in the development of many malignancies, is commonly altered in mCRPC. Gain-of-function mutations in TP53 confer increased oncogenic properties of this gene and play a role in mCRPC. The goal of this study was to characterize somatic TP53 mutations, specifically gain-of-function mutations, in the ctDNA of mCRPC patients in the context of prior therapies. Methods: A retrospective analysis of mCRPC patients at Tulane Cancer Center between 2015-2022 was performed. All patients had ctDNA testing performed with the Guardant360 multigene panel assay. Clinical annotation including initial diagnosis, staging, treatment history, and family history were obtained. TP53 mutations were classified based on existing published functional studies and/or in silico evaluation. Statistical analyses were performed with Fisher's exact and Chi-squared tests where appropriate. Results: 338 mCRPC patients with ctDNA testing were included in this analysis. 76 patients had no prior treatment with either abiraterone or enzalutamide, while 262 patients had been treated with abiraterone and/or enzalutamide. Somatic TP53 mutations were similar in frequency between those with or without abiraterone/enzalutamide pretreatment; 46% (35/76) of patients in the abiraterone/enzalutamide naïve subset had a somatic TP53 mutation, compared to 41% (108/262) of patients previously treated with abiraterone and/or enzalutamide. Only 9% (7/76) of abiraterone/enzalutamide naïve patients had a TP53 gain-of-function mutation, compared to 19% (49/262) of patients previously treated with one or both drugs ( p = 0.05). The most common type of TP53 mutation was loss-of-function. There were no significant associations between TP 53 mutations and occurrence of other common mutations. Conclusions: mCRPC patients with prior treatment of abiraterone and/or enzalutamide were significantly more likely to have a gain-of-function TP53 mutation. Further studies are needed to investigate therapeutic implications of these findings.

68Ga-PSMA-11 patients with newly diagnosed and recurrent prostate cancer (Firefly Study): Preliminary results

54

Background: A primary challenge facing oncologists is the accurate identification of the source of the rising PSA in the recurrent disease setting and the failure of proper staging at the time of initial therapy. PSMA imaging allows for progress in this regard, however access remains an issue. As such, we designed the Firefly study to offer PSMA imaging to patients between May 2021 and May 2022, to better characterize their disease. Methods: This is a phase II expanded access intermediate sized clinical trial using 68Ga-PSMA-11 (Telix Pharma cold kit) under Tulane’s own Investigational New Drug (IND). Clinical trial information: NCT04854369. Two cohorts are utilized, recurrent disease with a PSA > 0.2 or > 2.0 post-radiation, and newly diagnosed high risk or oligo-metastatic patients as assessed by conventional imaging. Between May 2021 and May 2022, a total of 90 patients were enrolled. The study was designed for a maximum of 300 patients but was terminated when access to insurance approved PSMA scans was available. Patients received 1.8–2.2 MBq/kg body weight with 68Ga-PSMA-11 (per EANM guidelines). The lower and upper limits of the dose were set to 3 to 7 mCi respectively. Primary Objective(s): Utilize 68Ga-PSMA-11 PET to define uptake location for localization of prostate cancer metastatic sites in patients prior to initial therapy, and in both local recurrences and metastatic sites in patients with recurrent disease after initial therapy. Secondary Objectives: To assess the therapeutic consequences of 68Ga-PSMA-11 PET/CT imaging in prostate cancer patients with and without prior treatment. Results: Of the 90 patients enrolled 81 were in the recurrent disease cohort and nine had newly diagnosed disease. Of the 81 patients with recurrent disease; 30 were known metastatic and were scanned with the intention of application for expanded access to 177Lu-PSMA-617, 8 received radiation for their oligometastases, 12 received a combination of radiation and initiation of hormone therapy, 13 continued surveillance, 5 were sent for salvage therapy, 13 were treated with novel hormones. Of the 9 patients with newly diagnosed disease all of them were negative for metastasis on conventional imaging. However, when scanned with 68Ga-PSMA-11, 7/9 had a positive PSMA scan. A lesion in the prostate was identified in 2/7 patients, 2/7 were found to have de novo metastatic disease to the lymph nodes and 3/7 were found to have bone metastasis. All these patients were treatment naïve and 5/7 had changes to their planned definitive treatment plan based on these results, as high-risk patients found to have previously unidentified de novo metastatic disease. Conclusions: Further analysis of the data is planned to complete secondary objectives; however, this study demonstrates the potentially meaningful impact of PSMA imaging for patient treatment planning and individualized care. Clinical trial information: NCT04854369 .

Genomic alterations in patients with prostate cancer with liver metastases

248

Background: mCRPC patients with liver metastases have a poor prognosis and often progress rapidly on a variety of treatments. Previously, preliminary ctDNA analyses of mCRPC patients with liver metastases showed a range of commonly altered genes in patients with liver metastases (Ranasinghe et al; 2019). In this follow-up, we evaluated ctDNA alterations in an expanded cohort of mCRPC patients with liver metastases. Methods: From Tulane Cancer Center, retrospective review of mCRPC patients was used to identify patients with confirmed liver metastasis. All liver metastases were confirmed based on imaging data. All patients included had ctDNA evaluated with a multi-gene cancer panel via Guardant 360 assay (Guardant Health, Inc). Additional clinical annotation including family history, germline testing, staging, imaging, and laboratory values. Statistical analyses were performed with Fisher’s Exact and Wilcoxon Rank Sum tests. Results: 158 mCRPC patients with appropriate diagnostic imaging as well as ctDNA testing. From this group, 8% (n= 12) had confirmed liver metastases. Among the patients with liver metastasis, the most common alterations detected were in AR (50%; 6/12) and PIK3CA (25%; 3/12). Patients with liver metastasis were more likely to have amplifications in FGFR1 detected in their ctDNA (OR= 14.40; 95% C.I. (1.83, 113.22); p= 0.03). In addition to ctDNA, germline data was assessed, and it was found that patients with liver metastasis were more likely to have a pathogenic germline mutation (OR= 7.61; 95% C.I. (2.85, 20.31); p<.0001). The most common germline mutations detected in patients with liver metastasis were in BRCA2 (n= 3) and TP53 (n= 2). Conclusions: Though liver metastasis are less common in prostate cancer, it often occurs following extensive treatment and results in a poor prognosis for patients. In patients with liver metastasis, FGFR1 amplification was more often detected in ctDNA. Importantly, patients with liver metastasis were significantly more likely to have a pathogenic germline alteration.

Utility of circulating tumor DNA in monitoring treatment response to immune checkpoint inhibitors in patients with advanced genitourinary cancers

721

Background: Current methodologies for monitoring treatment response are largely based on conventional scans and/or tumor biopsies, which may be limited in their ability to accurately assess disease burden at the molecular level. Circulating tumor DNA (ctDNA) detection in blood has emerged as a prognostic and predictive biomarker and has shown to better assess treatment response in patients receiving immune checkpoint inhibitors (ICI). We conducted a prospective, pilot study to investigate the concordance of serial ctDNA detection and dynamics with radiographic response in patients with advanced GU malignancies undergoing ICI-based treatment. Methods: Twenty patients with histologically confirmed advanced GU malignancies (renal, urothelial, and prostate) were enrolled in the prospective study. All eligible patients received ICI treatment for at least 12 weeks and were followed by serial collection of blood samples every 6-8 weeks until disease progression. Conventional scans were performed approximately every 12 weeks until disease progression. Overall response rate (ORR) by investigator was reported and associated with ctDNA detection. Results: ctDNA analysis was performed on 122 plasma samples obtained from 20 patients (N=15 renal cell carcinoma; N=4 urothelial carcinoma; N=1 prostate cancer). Prior therapies to ICI-based treatment included chemotherapy (10%), hormonal therapy (5%) and anti-VEGF (5%). After study enrollment, patients received anti-PD-1 (95%), anti-CTLA-4 (30%) or anti-PD-L1 (5%) with an ORR of 70% as best response. With a median follow-up of 19 months (range: 4-48), progressive disease was observed in 7 patients. Nineteen patients had longitudinal plasma samples available and ctDNA detection at any time point was 45% (9/20). The overall concordance between ctDNA dynamics and radiographic response at 12 weeks was observed in 89% (17/19) of patients. Of these 17 concordant patients, one patient showed transient ctDNA positivity followed by clearance at the last two timepoints on treatment. The two patients with discordant results included the ones with CNS-only metastasis (ctDNA negative). Of the 7 patients who progressed on ICI, ctDNA was detected in five (71%); the remaining two had CNS-only metastases. The last patient had a single time point available on treatment that showed ctDNA-positivity and passed away 7 weeks after molecular evidence of disease. Conclusions: In this study, serial collection of blood samples for ctDNA analysis to monitor response to ICI-based therapiesin patients with advanced GU tumors was feasible. There was a high concordance rate between radiological imaging and ctDNA data, especially in extra-CNS disease. Further studies are needed to validate ctDNA as a tool to aid disease monitoring in patients treated with ICI.

Evaluation of ctDNA in patients treated with lutetium-177-PSMA-617

243

Background: Lutetium-177- PSMA-617 (Lu-177) is a radioligand therapy that delivers radiation to PSMA-expressing cells in patients with advanced prostate cancer. In this study, we aimed to analyze ctDNA in responders and non-responders after Lu-177 treatments. Methods: Data was retrospectively collected on 31 heavily pre-treated metastatic CRPC patients who received Lu-177 treatment at Tulane Cancer Center. All patients fulfilled VISION criteria for treatment and all had ctDNA assessment with Guardant 360 within 30 days prior to first treatment with Lu-177. Of the 31 patients, 7 had paired ctDNA assessment both prior initiation of treatment and at the end of treatment. Clinical data such as PSA response (PSA decline 50% or more) to Lu-177, initial diagnosis, pathology, treatment history, and relevant germline genetic data were collected. Results: Of the 31 patients who received Lu-177 treatment, 18 had PSA response to Lu-177 (responders) and 13 did not (non-responders). In ctDNA mutational analyses, there were no significant differences detected prior to treatment between responders and non-responders. There was, however, a significant increase in the presence of copy number amplifications in non-responders (n= 11/13) when compared to responders (n= 7/18) (OR= 8.64, 95% C.I. [1.46, 51.25)], p = 0.0250). Amplification was detected in 10 genes in non-responders, whereas responders only had amplifications in AR (7/18) and EGFR (1/18). The most frequently amplified genes in non-responders were AR (8/13), CCNE1 (6/13), EGFR (4/13), and FGFR1 (4/13). Furthermore, in analysis of the 7 patients with paired ctDNA assessments, amplifications increased during treatment with Lu-177. Conclusions: Radioligand therapy with Lu-177 has been shown to prolong life and reduce disease progression. Our analysis of the cohort of Lu-177 treated patients showed that the presence of gene amplifications in ctDNA may play a role in predicting resistance to treatment with Lu-177.

Screening detects clinically significant hearing loss in patients with GU malignancies

e18624

Background: Four in 10 people aged 60-69 have hearing loss in frequencies that mediate speech comprehension. Untreated losses undermine medication adherence, increase risk for accidental injury, and generate higher total healthcare expenditures. Genitourinary (GU) malignancies are most frequently diagnosed in people over 60, and these patients are commonly exposed to ototoxic treatments like platinum-based chemotherapy. We hypothesized that a significant number of GU patients would fail hearing screenings at levels reflecting clinically significant hearing loss. Methods: This is a prospective, two-cohort pilot study conducted at Tulane Medical Center. Consecutive patients with active GU malignancies who did not use hearing aids were enrolled in cohort A. The primary objective of cohort A (screening) was to determine the prevalence of hearing loss among GU patients. Cohort B (interventional) will investigate the efficacy of a non-custom amplifier in improving communication for patients with hearing loss. The hearing screening consisted of a 9-item Self-Assessment of Communication (SAC) and a 25 decibel (dB) pure tone hearing screening at 1 kHz, 2 kHz, 4 kHz via an Earscan3 portable audiometer with headphones in a quiet exam room. The primary endpoint was the proportion of patients who failed the hearing screening (missed ≥ 2/6 tones). The SAC generates a global handicap score (none, slight, mild-to-moderate, severe). Surveys identifying any amount of handicap were considered positive. Patients who failed screenings were offered referrals to audiology. Here, we present results from cohort A. Results: From 8/21 to 1/22, 66 patients were invited to participate. Nine patients screen-failed (n = 8 hearing aids; n = 1 uninterested). Cohort A enrolled 57 patients [median age 68 (33-86); 91% men; 60% prostate, 24% kidney and 14% bladder cancer]. Most patients (82%) had evidence of metastatic disease and 12%, 19%, 36%, 21% of patients were receiving chemotherapy, immunotherapy, hormonal therapy, and no active therapy, respectively. Twenty (35%) patients had prior exposure to neurotoxic chemotherapy. Forty patients (70%) met the primary endpoint. Compared to the pure tone screening, the SAC had a specificity of 94%, and a sensitivity of 40%. The SAC asks patients for a situation where they want to hear better; 61% of patients who volunteered a scenario (17/28) identified a complex listening environment like an outpatient encounter. A minority of patients (n = 3) accepted audiology referrals after failing screenings. Cohort B is actively enrolling. Conclusions: Clinically significant hearing loss was prevalent among patients with GU malignancies. Patients reported struggling in listening situations with auditory demands like outpatient encounters but were unable to self-identify losses and did not pursue audiology referrals after failed screenings. Interventional studies addressing this unmet need are warranted.

Comparative ctDNA analyses of African-American and Caucasian patients with CRPC

24

Background: Somatic genetic analyses have indicated genetic distinctions in AA as compared to C patients. In the Mahal et al. study (1) evaluating a broad spectrum of pts with tissue based assays, FOXA1 mutations were more frequent in AA men and TP53 mutations were less frequent in AA men as compared to C men. In a separate analysis by Khashab et al. (2) conducted in prostate cancer pts receiving androgen deprivation therapy, using both tissue and ctDNA assays, the authors reported AR, TP53, SPOP, and BRCA2 were more frequently mutated in AA men as compared to C men. Herein we assessed the Guardant 360 platform in assessing ctDNA differences in AA and C men, all of whom had CRPC at the time assays were performed. Methods: Guardant 360 was used to analyze ctDNA with a cut-off of >0.5% for allelic fractions for ascertaining the presence or absence of pathogenic mutations and various amplifications. Lower allelic fractions were not analyzed given these may represent less relevant mutations. Depending on the timing of the assays (2015-2021), 70-83 genes were analyzed. All pts had CRPC and all patients were treated at Tulane Cancer Center. Chi Square analyses were used to determine statistical differences. AR, BRCA2, and TP53 were assessed but SPOP and FOXA1 were not assessable in the Guardant ctDNA assay. Both mutations and amplifications were evaluated. Results: Among men with CRPC, a total of 48 AA men and 179 C men were analyzed using ctDNA. Clear distinctions were found in the alteration reported in APC, TP53, and CDK12. TP53 was less frequently mutated and other genes were more frequently altered in the AA men. Conclusions: Using Guardant ctDNA assays in men with CRPC, clear distinctions were found in AA men as compared to C men. It is unclear why these results differ from that reported by others, however distinctions in both the assays and the populations are notable.[Table: see text]

Longitudinal ctDNA alterations in germline positive CRPC patients

275

Background: Circulating tumor-derived DNA (ctDNA) is an accessible method for characterizing somatic alterations. We report longitudinal ctDNA screenings of mCRPC patients (pts) who have had germline testing. Methods: Patients with both germline testing and ctDNA assessment were included. Germline testing was performed with a multi-gene cancer panel from Invitae (50-83 genes) and somatic alterations in ctDNA were identified by testing with Guardant 360 (70-83 genes). ctDNA alterations were characterized as deletions, frameshift, missense, nonsense, and other mutations. A total of 177 patients in various stages of therapy had both ctDNA and germline DNA tested. Results: From 2015-2021, 177 mCRPC patients were included and had an average of 3 ctDNA tests. 11.3% (20/177) had pathogenic or likely-pathogenic germline mutations. The common pathogenic germline mutations were in BRCA2 (25%; 5/20), ATM (10%; 2/20), and MSH2 (10%; 2/20). In ctDNA, missense mutations were the most prevalent type of gene alteration in germline negative (n = 539/790, 68%) and germline positive (n = 124/218, 57%) followed by frameshift mutations at 22% (n = 48/218) in germline positive and 10% (n = 80/790) in germline negative patients. Germline positive patients were more likely to have somatic frameshift mutations (OR = 2.09, 95% C.I. (1.3792, 3.1618), p = 0.001) and less likely to have missense mutations (OR = 0.61, 95% C.I. (0.4519, 0.8351), p = 0.002). Other alterations including deletions, nonsense, and other mutations were not significantly different. Of the germline positive pts, BRCA2 mutation was associated with the highest number of somatic alterations. Conclusions: Germline positive CRPC patients were more likely to have frameshift mutations and less likely to missense mutations compared to germline negative CRPC patients. Patients with germline mutations in BRCA2 and TP53 had the highest number of somatic alterations detected in ctDNA over the course of ctDNA evaluation.

Evaluation of ctDNA alterations in mCRPC patients with germline pathogenic mutations

177

Background: Germline alterations are found in approximately 12-17% of CRPC patients. Similarly, evaluating tumoral changes with circulating tumor DNA (ctDNA) has become an increasingly useful tool for understanding mechanisms underpinning disease progression. In this study, we evaluated both germline and somatic genetic changes in patients with mCRPC. Methods: Patients included had germline screening and ctDNA analyzed with the Guardant 360 assay. All patients were CRPC at the time of Guardant testing. Germline alterations were classified as pathogenic/likely-pathogenic or not pathogenic. Only ctDNA alterations with an alleleic fraction greater than 0.5% were included in analyses. Additional evaluation of CRPC status, treatment history, family history and other clinical covariates are ongoing. Chi-square and Fischer’s Exact tests were used for comparison of cohorts. Results: A total of 168 Caucasian CRPC patients had Guardant 360 testing at time of progression and germline testing between 2015-2021. 61% (n = 102/168) of patients have previously had treatment with abiraterone, 49% (n = 82/168) have had Enzalutamide and 40% (n = 68/168) have had treatment with taxanes. 12% (n = 20/168) of CRPC patients had a pathogenic/likely-pathogenic (P/LPv) germline alteration, 46% (n = 77/168) had a germline variant of unknown significance (VUS), 42% (n = 71/168) were germline negative. CRPC patients with pathogenic germline mutations were significantly more likely to have subsequent somatic alterations in BRCA2 (OR = 5.05, 95% C.I. (1.11, 23.01), p = 0.055), NF1 (OR = 7.89, 95% C.I. (2.15, 28.10), p = 0.004), and TP53 (OR = 3.52, 95% C.I. (1.28, 9.68), p = 0.015). In TP53, among germline positive patients, 45% (n = 9/20) had TP53 alterations compared to 30% (n = 45/148) of germline negative patients. Conclusions: Germline positive (P/LPv) CRPC patients were significantly more likely to have somatic alterations in BRCA2, NF1, and TP53. Understanding the totality of genetic changes, both germline and acquired somatic alterations is essential as the arsenal of targeted treatment for CRPC continues to expand. Additional studies including longitudinal assessment genetic changes and clinical correlates will be necessary to evaluate these findings in the context of treatment outcomes and disease progression.

68ga-PSMA-11 patients with newly diagnosed and recurrent prostate cancer (Firefly)

TPS189

Background: A primary challenge facing oncologists is the accurate identification of the source of the rising PSA in the recurrent disease setting and the failure of proper staging at the time of initial therapy. Current imaging modalities perform poorly in this regard. PSMA imaging is FDA approved but reimbursement remains a considerable issue for most patients. Methods: This is a phase II expanded access intermediate sized clinical trial using 68Ga-PSMA-11 (Telix Pharma cold kit) under our own Investigational New Drug (IND) with the FDA. Two cohorts are utilized, recurrent disease with a PSA > 0.2 or > 2.0 post-radiation, and newly diagnosed high risk localized disease and/or oligo-metastatic patients as assessed by conventional imaging. 300 patients will be allowed to enroll and 63 are enrolled to date. Patients will be injected with 1.8–2.2 MBq/kg body weight with 68Ga-PSMA-11 (per EANM guidelines). The lower and upper limits of the dose are set to 3 to 7 mCi respectively. Primary Objective(s): Utilize 68Ga-PSMA-11 PET images to define uptake location for localization of prostate cancer metastatic sites in patients prior to initial therapy and to utilize 68Ga-PSMA-11 PET images to define uptake location for localization of prostate cancer including both local recurrences and metastatic sites in patients with recurrent disease after initial therapy. Secondary Objectives: To assess the therapeutic consequences of 68Ga-PSMA-11 PET/CT imaging in prostate cancer patients with and without prior treatment. As such the planned therapies after the PSMA scan will be annotated and recorded. In addition, we aim to determine the number of patients with metastatic disease prior to initial therapy that was not diagnosed with conventional imaging and we aim to determine the locations of recurrent disease for those patients who have demonstrated PSA recurrence after curative therapy. Percentage of positive scans will be assessed for PSA values 0.2-0.5 ng/mL, > 0.5-1.0 ng/mL, > 1.0- > 2.0 ng/mL, and > 2 ng/mL. Results will be summarized descriptively with 95% confidence intervals. Localization of all lesions will be performed with regard to lymph nodes (pelvic versus non-pelvis), bones (by location), and other (by site). Treatment planning will be captured and quantified by radiation, surgery, hormonal therapy (type and duration), and/or observation at the time of the initial clinical visit after the PSMA scan and 6 months later. This trial is open and recruiting patients. Data analysis has not been performed. Clinical trial information: NCT04854369.

Circulating tumor DNA responses to high-dose testosterone injections in CRPC patients

173

Background: Some proportion of men with CRPC have favorable responses to high doses of testosterone (HDT) and some do not. Genomic determinants of these responders and non-responders are poorly understood. Herein ctDNA predictors of response are assessed using a responder/non-responder analysis. Methods: All men with CRPC had been pretreated with abiraterone, enzalutamide, or both. A ctDNA test was obtained prior to high doses of testosterone (administered as a dose of 400 mg IM testosterone cypionate every 3-4 weeks). Guardant 360 was used to analyze ctDNA (with an allelic fraction cut-off of > 0.5%) to ascertain the presence or absence of pathogenic mutations. Lower allelic fractions were not analyzed. Responders had a PSA decline of 50% or more (N = 16), non-responders received at least two doses of testosterone but never had any PSA decline at all (N = 20). All patients were treated at Tulane Cancer Center. Results: AR amplifications were more commonly detected (p = 0.036) pre-treatment in the ctDNA of responders (5/16) as compared to non-responders (1/20). No differences were found in those with common AR mutations; T878A was detected in 2/16 responders and 2/20 non-responders, L702H was detected in 1/16 responders and 2/20 non-responders. No differences were seen with regard to TP53 mutations, 6/20 non-responders and 7/16 responders. Non- AR/non- TP53 mutations were not distinct in the two groups but trended (P = 0.15) toward being more common in non-responders (5/16 in responders versus 11/20 non-responders). Pre- and post- ctDNA analyses were conducted in 5/6 patients for those with AR amplification at baseline. In all 5 of these patients, the degree of AR amplification diminished after testosterone injections. Conclusions: In this analysis of CRPC patients who were responders and non-responders to 400 mg testosterone cypionate q 3-4 weeks (post-abiraterone and/or enzalutamide), only AR amplifications in ctDNA were predictive of response. In all measured patients, the degree of AR amplification in the ctDNA diminished after testosterone injections.

ctDNA pathogenic variants (PVs) in homologous recombination repair (HRR) genes in patients with metastatic CRPC

138

Background: HRR PVs can serve as predictive biomarkers and two PARP inhibitors are approved for metastasic CRPC (mCRPC) pts. Published data are predominantly focused on tissue-based assays, but obtaining tissue from mCRPC pts is problematic. In a large tissue based series (PROfound), 4047 mCRPC pts had tumor samples submitted for genomic testing but only 69% had interpretable results. No data were published from PROfound enumerating pts without available tissue to submit. Herein we assess frequency of PVs from selected HRR genes using a ctDNA assay. Methods: 292 mCRPC pts at Tulane Cancer Center were assessed for detectable HRR ctDNA changes using the Guardant 360 assay (which assesses the HRR genes BRCA1, BRCA2, and ATM). Results: 20/292 (6.8%) pts had a PV in ATM. However only 4/292 (1.4%) had > 1% mutant allelic fraction. Germline testing occurred in 18/20 of the ctDNA ATM PV pts and 0/18 had a germline PV. The PROfound series had 6.3% somatic PVs in ATM. 18/292 pts (6.2%) had a PV in BRCA2 and 12/292 (4.1%) had a mutant allelic fraction of > 1%. Germline testing was performed in 17/18 with BRCA2 ctDNA PVs and 9/17 had germline PVs. The PROfound series had 9.7% somatic BRCA2 PVs. BRCA1 PVs were detected in 6/292 (2.1%) pts and 3/292 (1%) had a mutant allelic fraction > 1%. 6/6 of the ctDNA PVs has germline testing and 1/6 had a BRCA1 PV. The PROfound series had 1.3% somatic PVs in BRCA1. Conclusions: Using ctDNA essay, it is feasible to measure PVs in only a small subset of HRR genes in mCRPC pts. These assays fail to detect deep deletions, a known and important mechanism of HRR gene loss. The ctDNA mutant allelic fractions are often low. The ability of ctDNA PVs using this assay to predict treatment effects are unknown.

Baseline pathogenic mutations in non-AR/non-TP53 genes and prediction of response to high-dose testosterone

146

Background: The fact that high dose of testosterone (HDT) elicits positive responses in a subset of prostate cancer patients (pts) is surprising and puzzling. Genomics data differentiating responders (Rs) from non-responders (NRs) is sparse. Pts with mutations in DNA repair pathway genes may be particularly sensitive to HDT (see BA Teply et al. Eur Urol 71:499, 2017). Herein we perform exploratory analyses to better understand the role of pathogenic mutations (muts) in ctDNA as a predictive biomarker for patients treated with HDT. Methods: ctDNA essays were performed with the Guardant360 methodology pre-HDT. Point mutations were classified by cancervar (http://cancervar.wglab.org). Truncating mutations (frameshift and nonsense) were manually curated to assess for pathogenicity. All patients had CRPC and were pre-treated with abiraterone and/or enzalutamide. HDT was typically administered as 400 mg testosterone cypionate q 3-4 weeks. Rs were compared to NRs. All Rs had >3 more cycles of HDT and a >50% PSA decline (N = 17). Non-responders had <3 cycles and no PSA decline (N = 23). Only muts with an allelic fraction of >0.5% were analyzed given dubious importance of mutations with lower allelic fractions. Results: AR muts (4/17 vs 6/23 in Rs and NRs) and TP53 muts (10/17 vs 11/23 in Rs and NRs) were similar ( P= 0.85 and 0.49, respectively) but the number of pts with non-AR/non-TP53 muts was distinct (3/17 for Rs, and 12/23 for NRs; P= 0.026). The average number of non-AR/non-TP53 muts (Rs = 0.23 and NRs = 0.83) was higher in the NRs (P = 0.046). When analyzing DNA repair alterations, no differences were noted in those with BRCA1/BRCA2/ATM mutations in the Rs and NRs (1/17 vs 6/23 respectively; P= 0.09). Conclusions: AR and TP53 pathogenic mutations are common in both Rs and NRs but other pathogenic mutations are more common in non-responders. We hypothesize that genetic pathways outside of the AR/TP53 axis drive resistance to HDT. Additional studies are warranted to assess whether or not these pathways drive resistance to HDT. Limitations are acknowledged with regard to the Guardant assay gene selection for CRPC pts.

PD-L1 inhibition with avelumab plus abiraterone acetate or enzalutamide in African Americans with metastatic castrate-resistant prostate cancer (mCRPC)

87

Background: African Americans (AA) are at higher risk for prostate cancer death compared to other ethnic groups in the United States. Early immune-based therapy prolonged overall survival in men with mCRPC and appears to be more pronounced in AA compared to Caucasian men. We sought to explore efficacy of PD-L1 inhibition in AA men with mCRPC. Methods: AA men > 18 years of age and had developed mCRPC on next generation hormonal therapies (NHTs) were eligible for this study. Patients received avelumab 10mg/kg IV every 2 weeks while remaining on NHTs until clinical or radiographic disease progression. This pilot, single-arm phase 2 prospective study was designed to enroll 13 patients with primary endpoint to assess > 50% reduction in PSA (PSA50). Radiographic assessments were planned every 12 weeks until progression or two years, whichever came first. Results: Of the eight patients enrolled, there were three screen-failures: +hepatitis titer (n = 1), rapid clinical deterioration (n = 1) and not confirmed CRPC (n = 1). Median age at time of enrollment was 62 (54-73) with median PSA 7.2 (3.6 – 8.63). Five patients received at least one dose of avelumab and remained on abiraterone acetate with steroid (n = 4) or enzalutamide (n = 1). Median duration on NHTs prior to enrollment was 364 days (95% CI, 260.9-467.1). Median time from screening to cycle 1 was 8 days (3-14). One patient withdrew consent after a single dose of avelumab. Two patients had rapid clinical progression within 8 weeks of starting avelumab. One patient received 9 cycles of avelumab and progressed. One patient experienced a grade 4 adverse event after 2 doses of avelumab with clinical progression during treatment delay. None of the five enrolled patients achieved a PSA50. The median time on study to PSA progression was 35 days (95 CI%, 0-94.8) with median time to radiographic progression of 44 days (95 CI%, 0-118.5). Adverse events occurred in 80% (n = 4/5) of cases. One of the 11 reported grade 3/4 adverse events was related to study drug (G3 hyperglycemia). The study was closed to further accrual prior to the pre-planned completion due to safety concerns related to lack of efficacy of study intervention and rapid clinical progression. Conclusions: The addition of avelumab did not demonstrate clinical activity in AA men with mCRPC who progressed on abiraterone acetate or enzalutamide. The short time interval between PSA progression and radiographic and/or clinical progression is concerning and notable to be much shorter than previously reported pivotal phase III trials. This suggests potential for very significant clinical implications regarding not only future clinical trial designs but also clinical management in this underrepresented ethnicity. We will report updated analysis with longer follow-up (NCT03770455). Clinical trial information: NCT03770455.

18F-fluciclovine positron emission tomography (PET) in metastatic castration-resistant prostate cancer (mCRPC) treated with abiraterone acetate

TPS171

Background: Conventional imaging of prostate cancer has limitations in staging, restaging after biochemical relapse, and response assessment. Functional imaging with positron emission tomography (PET) can target various aspects of tumor biology and has shown to be superior in the detection of prostate cancer compared with conventional computed tomography (CT) and bone scans. 18F-Fluciclovine, a synthetic amino acid transported across mammalian cell membranes by amino acid transporters that is upregulated to a greater extent in prostate cancer cells than in surrounding tissue, is currently approved for PET imaging for patients with biochemical recurrence. The role of 18F-Fluciclovine PET scans in monitoring response to novel hormonal therapies such as abiraterone acetate is unclear. We hypothesize that 1) using 18F-Fluciclovine PET scanning will allow a more sensitive assessment of mCRPC patients at the initiation of systemic therapy with abiraterone acetate and 2) the changes observed in 18F-Fluciclovine PET will correlate better with the serologic changes in PSA, allowing superior disease monitoring, than conventional imaging modalities. Methods: This single-arm, pilot study (NCT04158245) will describe the changes in 18F-Fluciclovine PET scan and compare these results with PSA and conventional computerized tomography (CT) and bone scans, in mCRPC patients treated with abiraterone acetate plus prednisone. Patients must have a detectable baseline PSA of ≥ 2 ng/mL and metastatic disease detected on conventional CT and bone scans. The use of docetaxel in the hormone-sensitive setting is allowed. Twelve patients will be treated with abiraterone 1000 mg daily plus prednisone 5 mg (or dexamethasone 0.5 mg) daily for mCRPC and get 18F-Fluciclovine PET and conventional CT and bone scans at baseline and 12 weeks after starting abiraterone therapy or at disease progression. PSA progression will be defined as a repeated increase in PSA of at least 2 ng/dL and 25% from nadir values, at least 1 week apart, according to PCWG3 criteria and clinical or radiographic progression by RECIST version 1.1. The co-primary objectives of the study include the 18F-Fluciclovine PET changes at baseline and 12 weeks after abiraterone acetate for mCRPC and the comparison between 18F-Fluciclovine PET and conventional scans. Secondary and exploratory endpoints include PSA response, PSA progression and genomic alterations by next-generation sequencing. As of 12 September 2020, this trial is actively enrolling. Clinical trial information: NCT04158245.

Multi-institutional evaluation of the clinical outcomes and genomic correlates of African Americans with metastatic castration-sensitive prostate cancer (mCSPC)

17

Background: Prostate cancer incidence and mortality is higher in African American (AA) as compared with non-AA men. The outcomes of mCSPC have significantly improved through treatment intensification yet, AA representation in those studies was suboptimal. We aimed to report the clinical, treatment outcomes and genomic data of AA men with mCSPC. Methods: Retrospective analysis of consecutive AA men with mCSPC at six Academic Institutions. The primary objective was to report the baseline characteristics and treatment patterns of mCSPC AA patients. The secondary objectives included the germline and somatic data and the clinical outcomes including PSA response, progression-free survival and subsequent treatments. Results: A total of 71 patients, median age 63 years (range, 41-84) with 58% Gleason 8-10, initial PSA of 69.8 ng/mL (0.02-7650), 59% with de-novo and 55% with high-volume (CHAARTED criteria; 20% visceral) disease, were included in this analysis. Twenty-two patients (31%) were treated with androgen deprivation therapy (ADT; 67% prior to year 2017), while 24%, 45% and 3% received docetaxel (median 6 cycles), abiraterone acetate and enzalutamide, respectively. Two patients received triplet therapy with ADT/docetaxel plus abiraterone or enzalutamide. Undetectable PSA was achieved in 35% after a median of 8.9 months (1.8-22.3). Among patients with mCSPC who received radiation therapy to prostate (n = 8), 89% had low volume disease. At time of cut off, thirty-two patients developed CRPC and the estimated median time to CRPC was 2.9 years (95% CI, 1.6-4.2). Subsequent therapies (n = 29) included abiraterone acetate (41%), enzalutamide (24%), bicalutamide (10%), radium-223 (7%), chemotherapy (7%), sipuleucel-T (3%) and others (7%). Five patients (8%) had pathogenic germline alterations (n = 2 BRCA1; n = 1 HOXB13, PALB2 and PMS2). Additionally, the most common somatic alterations among tested patients (n = 27) included CDK12, SPOP, TMPRSS2-ERG fusion, and TP53, all in 11% frequency. Of note, n = 2 BRCA1 and n = 1 high MSI/TMB. Conclusions: In one of the largest reported cohorts to our knowledge, mCSPC AA presented with a high number of de-novo and high-volume disease and might harbor a different germline and somatic genomic profile. The outcomes were comparable to contemporary phase III trials with treatment intensification, yet 31% were treated with ADT. Despite the known limitations associated with retrospective analysis, these data support prior observations where AA might have better initial PSA responses to ADT-based strategies compared with Caucasians, requiring further validation.

Family history and pathogenic/likely pathogenic germline variants in prostate cancer patients

143

Background: Recent literature highlights the importance of germline genetic testing in prostate cancer (PCa) patients. Surprisingly, a literature review indicates that family history records are incomplete in published studies. Methods: Prospective and complete family history data were gathered from 496 men in a single institution with a personal history of PCa who underwent germline genetic testing using a panel of at least 79 genes (Invitae testing) from 2016-2020. Comprehensive FH were obtained in all PCa patients in this database and analysis of prevalent FH was assessed at the time of sample collection. Age, race, metastastes at any time, and Gleason score were also ascertained. MUTYH heterozygotes were not considered pathogenic. Results: Pathogenic/likely pathogenic variants (PV/LPVs) were not associated with age at diagnosis, race, or presence of metastasis. Men with Gleason scores 8-10 at time of diagnosis were more likely to have PV/LPV ( P= 0.004). One or more first degree relatives (FDR) with any cancer with was not predictive for germline PV/LPVs for men with PCa ( P= 0.96). Analysis of patients with one or more FDR with breast, prostate, ovarian, or pancreatic cancer revealed that only FDR with breast cancer ( P = 0.028) or ovarian cancer ( P = 0.015) was predictive for PV/LPVs. Though one or more FDR with prostate cancer did not predict a PV/LPV in the overall panel, further analysis indicate that a history of a FDR with PCa was predictive for PV/LPV in a DNA damage repair (DDR) gene ( P= 0.044). Conclusions: In men with a personal history of PCa, germline PV/LPVs were associated with a FDR with breast or ovarian cancer. A FDR with PCa was predictive for PV/LPV in DDR genes. These data emphasize the contribution of FH to germline genetic testing results in a cohort with complete ascertainment of cancer in first degree relatives.

First-line PD-1/PD-L1 inhibitor monotherapy for advanced renal cell carcinoma (aRCC): A multi-institutional cohort

e17109

Background: Combination regimens containing a programmed cell death protein 1 (PD-1) or a programmed cell death ligand 1 (PD-L1) inhibitor improve the clinical outcomes of previously untreated aRCC. These regimens are clinically active but are associated with adverse events (AEs). One prospective study (KEYNOTE-427) has shown clinical activity of pembrolizumab monotherapy in aRCC. We aimed to describe the utilization and outcomes of single-agent PD-1/PD-L1 in previously untreated aRCC patients (pts). Methods: Consecutive pts treated with front-line PD-1/PD-L1 monotherapy for aRCC were included. Descriptive statistics were used to analyze outcomes including overall response rate (ORR), progression-free survival (PFS) and safety. Results: A total of 28 pts (median age 60; 32% ECOG 2/3; intermediate/poor risk 86%/14%; 35% non-clear cell aRCC) were identified. Common sites of metastases included lung (57%), lymph nodes (39%), bone (21%) and brain (21%). Pts received pembrolizumab (n = 9), avelumab (n = 1) or nivolumab (n = 18) due to pt/physician’s decision (46%), poor performance status (29%) or clinical trial (21%). In evaluable pts (n = 24), the ORR was 33% in ccRCC and 30% in nccRCC. With a median follow up of 13 months (mo), the median PFS was 6.3 mo (CI 95%, 1.5-11.1) and 16 pts progressed on therapy. Frequent AEs (57%) included fatigue (21%), myalgias/arthralgias (11%) and hypothyroidism (7%). No significant grade AEs (G3+) were noted. After progression on PD-1/PD-L1 (n = 16), 39% received a subsequent anti-VEGF (cabozantinib n = 6; sunitinib n = 4; pazopanib n = 1), 42% went to hospice or died and 21% were lost to follow up. Conclusions: First-line single-agent treatment with PD-1/PD-L1 inhibitors showed a favorable safety profile and clinical activity in patients with clear cell and non-clear cell aRCC who were not considered for a combination regimen. [Table: see text]

Family history and germline alterations in metastatic prostate cancer patients

231

Background: Recent literature has highlighted the importance of germline genetic testing in metastatic prostate cancer (PCa)patients (pts). This retrospective study evaluated family history (FH), evaluate treatment outcomes in metastatic patients with pathogenic germline mutations, and compared family history in Caucasian (CA) and Africans Americans (AA). Methods: At Tulane Cancer Center, 428 metastatic PCa pts had germline testing in at least 79 genes. Comprehensive family histories were obtained in all. Analysis of prevalent FH, including breast, ovarian, prostate, and pancreatic cancers was assessed for all metastatic pts. Statistical analyses including chi square were performed. Results: 64 (64/428; 14.9%) pts had at least one or more pathogenic mutations, while 166 (166/428; 38.7%) had were normal. The remaining 199 (199/428; 46.4%) of pts had a variant of uncertain significance (VUS). Pts with a DNA repair pathogenic mutation were more likely to have >2 family members affected by cancer, regardless of cancer type or degree of relationship (p=0.0047); 6 pts without any family history of cancer had a pathogenic mutation. In CA, the presence of either a breast or PCa family history was associated with pathogenic germline findings (p=.022/.0367) However, in AAs neither breast nor prostate cancer predicted pathogenic germline alterations. Conclusions: The correlation between family history of breast and PCa cancer in CA pts with pathogenic mutations is notable as is the finding of >2 family members with cancer predicting germline pathogenic alterations. With AA pathogenic pts, there was no correlation of family history of breast and PCa, which highlights the need to learn more about the genetic factors which influence AA prostate cancer risk.[Table: see text]

Circulating tumor DNA (ctDNA) landscape in metastatic castrate-resistant prostate cancer patients with germline alterations

200

Background: Circulating tumor-DNA (ctDNA) in mCRPC patients (pts) provides a viable approach for examining the genetic landscape of prostate cancer. In this follow-up, we report ctDNA variants in germline tested mCRPC pts. Methods: ctDNA alterations in 73 genes were detected using Guardant360 (G360) assays. Alteration types assessed were missense, frameshift, insertions, splice variants, truncations, amplifications (amp), deletions, and other. Pts included in the analysis received germline genetic testing (Invitae Corporation, San Francisco, CA) and ctDNA assays at various treatment timepoints. Statistical analyses were performed using chi-square and fisher exact test with p-value <0.05 for significance. Results: Germline and ctDNA testing was completed in 270 mCRPC pts. 13% (35/270) of pts had pathogenic germline alterations. Germline alterations detected were BRCA2 (43%, n=15), ATM (8.5%, n=3), CHEK2 (8.5%, n=3), and BRCA1 (6%, n=2). Of the 673 alterations detected in G360 assays, TP53 (25%, n=167) and AR (17%, n=117) were most commonly observed. ctDNA alteration breakdown for germline negative/positive pts is summarized in Tables A/B. Germline negative pts had more AR alterations compared to germline positive (p = 0.023). Also, germline negative pts presented with more amps (p < 0.001) and germline positive pts with more frameshift alterations (p = 0.005). The association of ctDNA alteration to clinical outcomes in germline positive/negative pts was also assessed and is ongoing. Conclusions: Pts with germline positive alterations had few somatic AR alterations and higher frequency of deleterious mutation in comparison to their germline negative counterparts.[Table: see text][Table: see text]

Comparison of Caucasian and African-American DNA repair alterations in men with metastatic prostate cancer

199

Background: Data on germline DNA repair defects and VUS rates are sparse in African American (AA) men with metastatic prostate cancer (PCa). Methods: Germline testing data from two centers with a significant percentage of metastatic AA PCa patients were combined and compared to Caucasian American (CA) with metastatic PCa. Fourteen canonical DNA repair genes (ATM, BARD1, BRCA1, BRCA2, BRIP1, CHEK2, MLH1, MSH2, MSH6, NBN, PALB2, PMS2, RAD51C, RAD51D) were assessed in all tested patients (pts) using a pathogenic/likely pathogenic (P/LP) classification. Variants of unknown significance (VUS) were assessed in an Invitae-derived dataset with consistent VUS reporting. Results: A total of 105 AA men with metastatic disease were evaluated and 7/105 of these men (6.67%) had P/LP alteration. Among the AA pt alterations, there were 4 pts with BRCA2, 2 pts with BRCA1, and 1 pt with PALB2. A total of 39/417 (9.3%) of CA metastatic patients had P/LP alterations in the canonical 14 genes. No differences were detected in the AA vs CA metastatic comparison (p=0.39). A total of 1/105 (0.95%) AA pts and 23/418 (5.5%) CA had non-BRCA P/LP mutations. The number of non-BRCA P/LP mutations were lower in the AA as compared to the CA men (p=0.045). When evaluating VUS calls in the metastatic AAs using Invitae multi-gene panels, 28/92 (30.43%) pts had a VUS in the canonical 14 genes as compared to 67/366 (18.31%) of the CA men. AAs were more likely than CA to have a VUS (p=0.010). These data indicate that metastatic AA pts and CA are not significantly distinct in the P/LP alterations in 14 canonical DNA repair genes but that there were lower percentages of P/LP in the AA non-BRCA gene subset. Further, when assessing these genes, it is clear that a VUS is more likely to be called in the AA men. Conclusions: Among men with metastatic PCa, AAs have similar rates of inherited P/LP alterations in 14 well accepted DNA repair genes as compared to CA men, however the non-BRCA gene P/LP alterations were less frequent among the AAs. Variants classified as a VUS were clearly higher in these AA pts as compared to the CA pts.

Androgen receptor cfDNA longitudinal mutational analysis in metastatic castrate-resistant prostate cancer

197

Background: Androgen receptor (AR) mutations commonly occur in metastatic castrate resistant prostate cancer (mCRPC). Methods: Circulating tumor DNA (ctDNA) data were obtained from Guardant 360 assays throughout the clinical course of mCRPC patients (pts). Retrospective analysis for any pt with ≥ 3 Guardant assays at least 4 weeks apart were reviewed. Patients must have at least 1 AR mutation or amplification to qualify for inclusion. Statistical analyses, including chi-sq and longitudinal analyses, were conducted. Results: Of the 259 patients with Guardant testing, a total of 88 patients had at least 3 Guardant tests; of these, 59 (67%) had at least one AR alteration. Patients had a median of 4 Guardant assays (range 3-10). Patients with AR amplification, AR mutation or both were identified (23, 20, 16 respectively). The most common and clinically relevant AR mutations found alone or in combination with amplification were T878A (22%), L702H(19%), W742C (19%), and H875Y (10%). These particular functional AR mutations occurred alone in 16 patients. Only 3 patients had neither amplification nor common AR mutation. 17/59 patients were found to have at least one common AR mutation and amplification at some point (on same or different Guardant). One patient had seven different AR mutations with no amplification and two other patients had 3 AR mutations. Remainder of patients had either AR amplification or ≤ 2 alternative mutations. Patients with an AR amplification were 0.1138x (95% Cl 0.0289 - 0.4491) significantly less likely of having a common known functional mutation (p <0.002) at any point. Conclusions: Patients with the most frequently identified known functional AR mutations are less likely to have AR amplification in pts with mCRPC; these common AR mutations have been shown to be associated with resistance to 2nd generation androgen deprivation. Further clinical correlation between treatment regimen and %cfDNA of these and other non-AR driver mutations is planned.

AR changes in circulating-tumor DNA (ctDNA) in patients with metastatic castration-resistant prostate cancer (mCRPC) treated with high-dose testosterone

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.

TP53 mutations in circulating tumor DNA in men with metastastic castration-resistant prostate cancer mCRPC

249

Background: Genes involved in mCRPC, including the tumor suppressor gene TP53, can be monitored with ctDNA but implications are unclear. Methods: Between 8/6/15 to 9/19/18, patients with mCRPC with available ctDNA genomic data using Guardant360 (Guardant360, Redwood City, CA) were included. Relationships between various lines of therapy and TP53 mutational status were analyzed. Patients positive for TP53 mutants were categorized using mutant allelic fraction of > 0%, ≥0.5%, ≥1% and ≥5%. Results: This study included 215 patients with mCRPC, median age 70 (41-90), and 80% Caucasians who had ctDNA after one line treatment for CRPC that included abiraterone (55%), enzalutamide (22%), and sipuleucel T (21%). Frequent co-alterations with TP53 mutants after 1 line of CRPC therapy included AR (14%), NF1 (5%), EGFR (4%), and PIK3CA (4%). See Table for relationships between lines of therapy and mutant TP53 allelic fraction. Conclusions: In this cohort of men with mCRPC, a higher mutational burden of TP53 in ctDNA was frequent and particularly associated with multiple lines of therapy. [Table: see text]

Evaluation of circulating tumor DNA (ctDNA) with respect to germline alterations in metastatic castrate resistant prostate cancer patients

254

Background: Circulating tumor-derived DNA (ctDNA) is an accessible method for characterizing tumoral alterations. We report ctDNA screenings of mCRPC patients (pts) who have had germline testing. Methods: Guardant360 (Guardant Health, Inc.) assesses ctDNA using sequencing to identify genomic alterations in 73 cancer-related genes. Alterations were categorized by type which included amplifications, deletions, frameshift mutations, insertions, missense mutations, splice mutations, truncations, and other. A total of 186 PCa pts in various stages of therapy had both ctDNA and germline DNA tested. Results: Of the 186 pts tested for germline mutations, 26 (14%) were germline positive. The most common germline mutation was BRCA2 with 12 (46%) pts, followed by ATM with 3 (11%). Of the total gene alterations were detected on ctDNA analysis of germline positive pts, with the most common genes being TP53 (n = 14/73, 19%), NF1 (n = 6/73, 8%), PIK3CA (n = 6/73, 8%), and BRCA2 (n = 5/73, 7%). Of the total gene alteration were detected on ctDNA analysis of germline negative pts, with the most common genes being TP53 (n = 94/588, 16%), AR (n = 90/588, 15%), EGFR (n = 31/588, 5%), and BRAF (n = 29/588, 5%). Germline negative pts showed had more amplifications (p = 0.008) while germline positive patients had more frameshift mutations (p = 0.025). Other alterations (deletion, missense, insertion, other, splicing, and truncating) were not significantly different. Missense mutations were the most prevalent type of gene alteration in germline negative (n = 306/609, 44%) and germline positive (n = 45/77, 48%), followed by amplifications (n = 210/609, 25% germline negative and n = 15/45, 18% germline positive). The median percent ctDNA values for missense mutations in germline negative and positive patients were 0.5% and 0.3% respectively. Of the germline positive pts, BRCA2 mutation was associated with the highest number of genes with alterations (n = 39), followed by RECQL4 (n = 8), ATM (n = 5), and MSH2 (n = 5). Conclusions: Germline positive pts had a higher number of frameshift mutations compared to germline negative pts. Additionally, pts with BRCA2 had the highest number of genes altered in ctDNA.

Genomic changes of AR in ctDNA prior to enzalutamide in men with mCRPC after abiraterone acetate

320

Background: Approximately 20-30% of men with metastatic castration-resistant prostate cancer (mCRPC) seem to respond to enzalutamide (ENZA) after prior exposure to abiraterone acetate and steroids (AA). The impact of genomic changes in androgen receptor (AR) gene in ctDNA and subsequent response to ENZA after prior AA exposure, is not entirely known. Methods: We performed a retrospective analysis of 28 mCRPC who were treated with ENZA after prior exposure to AA. Patients had NGS testing of ctDNA (Guardant 360) done post-AA and prior to ENZA. AR changes were defined by either amplification (amp; copy number ≥1) and/or mutations (mut; allele fraction ≥0.3%) in the AR gene. The primary endpoint of this study was time to PSA progression on enzalutamide. Correlations between genomic data and clinical outcomes were evaluated. Results: The 28 mCRPC patients had a median age of 67 (47-86), and 66.7% had a Gleason score 8-10. Metastases were found in bone (93%), lymph nodes (35.7%) and viscera (35%). Patients received ENZA after a median of 1 (1-5) treatment for CRPC. Patients were treated with prior AA for a median of 11.0 months (1.2-51.9) and median initial PSA at time of ENZA initiation was 35 ng/mL (1.52-356). The median time from ctDNA testing to ENZA start date was 0.9 months (range 0-5.6). Somatic changes in AR genes were detected in 36% (5/10 AR amp, 4/10 AR mut, 1/10 both). Other common alterations included 39% TP53, 11% DNA repair genes (2/3 BRCA2, 1/3 ATM), 7% PTEN. With a median follow-up of 8.9 months, 32% achieved PSA response (≥50%) and median time to PSA progression was 1.6 months (95% CI 1.0-2.2). On univariate analysis, lower Gleason scores (p = 0.021) and lack of AR changes (p = 0.042) were associated with PSA response to ENZA. Conclusions: In this cohort, responses to ENZA were associated with the absence of AR changes in ctDNA and less aggressive phenotypes. Prospective validation of this genomic association is required.

Pembrolizumab (pembro) in heavily pretreated metastatic castrate-resistant prostate cancer (mCRPC)

255

Background: KEYNOTE-199 has shown PSA responses of 11% to pembro in mCRPC. This study aims to further evaluate pembro in heavily pre-treated mCRPC patients (pts) correlating clinical outcomes with somatic and germline mutational burden. Methods: Single-institution retrospective analysis of mCRPC pts treated with pembro with germline panel testing and baseline ctDNA analysis using Guardant360 (Redwood City, CA). Baseline clinical annotation was collected and correlated with ctDNA data and clinical outcomes; ctDNA annotations included amplification (amp) and somatic mutation (mut; allele fraction ≥0.3%). Clinical outcomes were assessed after 3 cycles defined as: PSA≥50% PSA decline) or PSA≥30 response. Results: 27 mCRPC pts were treated with pembro between Oct 2016 and June 2018, median age 69 (56-82), 70% Caucasian, 26% African-American, and 4% Other, 70% Gleason 8-10, 59% bone-only, 22% bone+tissue, and 19% bone+LN metastases, were included. Pembro was given after a median 5 CRPC therapies with a median initial PSA of 76.1 ng/mL (4.85-1160), median treatment duration of 1.4 months (0-24.3). Prior treatments include abiraterone (n = 27), enzalutamide (n = 17), docetaxel (n = 19), and provenge (n = 16). 18 pts had ctDNA testing both pre- and post-pembro with a median time from testing to pembro of 0.9 months (0-3.9) and a median time from pembro to NGS testing of 0.7 months (0-2.6). Pre-pembro NGS had the following cfDNA alterations: 74% AR (mut = 8, amp = 8, both = 4), 55% TP53, and 0% DNA repair. ctDNA allelic fraction decrease occurred in 50% (6/12) of pts with AR mutations. 55% (15/27) of pts completed ≥3 cycles of pembro with the following responses: PSA≥50 13% and PSA≥30 20%. Two PSA complete responders (CR) (PSA < 0.01) had germline pathogenic alterations (BLM or MSH2). One of these pts was bone only; the other had radiographic CR. The MSH2 germline tumor was MSI-H and lacked MSH2/MSH6 on immunohistochemistry. The other PSA CR was not assessed given limited tissue. Conclusions: 33% of pts evaluated, achieved a PSA decline of ≥30% al response or greater signifying there is a subset of mCRPC pts that can benefit from pembro. Further evaluation is necessary to determine predictive biomarkers for this patient population.

DNA repair germline pathogenic mutation associations with treatment duration and family history in prostate cancer

245

Background: Germline mutation testing for metastatic prostate cancer patients creates a potential opportunity to personalize targeted therapies to improve treatment outcomes. The goal of this study was to characterize cancer family history, and evaluate treatment outcomes, in mCRPC patients with DNA repair pathogenic germline alterations. Methods: A retrospective study of metastatic PCa patients at Tulane Cancer Center identified 246 patients undergoing germline testing using panels (30-80 genes) (Color.com or Inviate.com). Clinical annotations included family history, life-extending treatments, and treatment duration. Statistical analyses including chi-square and Wilcoxon Rank Sum. Results: In the 246 patients tested for germline mutations, 27 patients (11.0%) had ≥1 DNA repair germline pathogenic mutation (BRCA2 = 11, BRCA1 = 3, CHEK2 = 5, ATM = 3, NBN = 1, PMS2 = 2, MSH2 = 1, PALB2 = 1) while 219 patients (89.0%) possessed no pathogenic mutation in these genes. Patients with a DNA repair pathogenic mutation were more likely to have > 2 family members affected by cancer, regardless of cancer type or degree of relationship (p = 0.04). In the DNA repair population, 5 pathogenic patients had no family history of cancer (18.5%, n = 5). Patients were more likely to have a germline alteration if they had 1 or more first degree relatives affected with breast cancer (p = 0.00001). Median lines of life-extending treatments to date between the pathogenic and non-pathogenic population were equal at 2. There were no significant differences in treatment duration for abiraterone (p = 0.49), enzalutamide (p = 0.99), docetaxel (p = 0.28), cabazitaxel (p = 0.53), carboplatin+docetaxel (p = 0.41), or radium-223 (p = 0.59) between the two groups. Conclusions: In this study, DNA repair pathogenic germline mutations did not affect treatment durations or lines of therapy but these studies are underpowered. The relationship between a family history of breast cancer and a DNA repair pathogenic mutation has not previously been reported.

ctDNA and copy-number variant quantification in mCRPC at the time of progression

265

Background: Tumor burden is a prognostic biomarker in cancer; however, in metastatic castration resistant prostate cancer (mCRPC), quantification of tumor burden markers is difficult. Circulating ctDNA is an accessible biomarker which can be used to characterize somatic DNA. The goal of this study was to evaluate the utility of ctDNA quantification and characterization at time of disease progression in mCRPC. Methods: From 2015 to 2018, 221 mCRPC patients underwent ctDNA screening with Guardant 360 (Guardant Health, Inc. Redwood City, CA) at times of disease progression. This analysis consisted of exonic coverage of 70 genes as well as amplifications in 18 genes, with pathogenic mutations categorized by effect and mutant fraction ctDNA. Clinical annotation including lines of life-extending therapies (LET) before and after testing were collected. Statistical analyses included repeated measure ANOVA, logistic regression, and parameter estimations. Results: An increasing number of LETs resulted in a clinically significant increase in mutant ctDNA fraction (P < 0.001). Subjects that received ≥ 5 or 3-4 LETs had significantly higher total mutant ctDNA fraction compared to subjects exposed with either 0 (p = 0.0007 and 0.0196, respectively) or 1-2 LETs (p = 0.0013 and 0.0432, respectively). 1-2 LETs did not demonstrate a significantly different total mutant ctDNA compared to subjects with 0 LETs (p = 0.4578). The estimated mean total mutant ctDNA fraction demonstrated a positive trend with an increasing number of LETs, including 2.08% (95% CI 1.24-3.49) without a LET, 2.62% (95% CI 1.71–4.01%) with 1-2 LETs, 4.58% (95% CI 2.92–7.17) with 3–4 LETs, and 7.79% (95% CI 4.51-13.48) for subjects with ≥5 LETs. The presence of copy number gain variations (CNVs) demonstrated an increased frequency with exposure to more LETs (p = 0.0005). Subjects with ≥5 or 3-4 LETs were 5.192 (p = 0. 0.0011, 95% CI 1.934-13.942) and 4.808 (p = 0.0019, 1.780-12.983) times more likely to have CNVs present compared to those without LETs. Conclusions: As mCRPC patients progress through lines of therapy, both the mutant ctDNA fraction and frequency of copy number gain variations increase.

Continuous infusion 5-fluorouracil (5FU) as a novel treatment for heavily pretreated prostate cancer patients: An update

319

Background: Currently patients failing multiple conventional therapies enroll on clinical trials but in many cases, despite a good performance status, patients may not qualify for trial enrollment. CRPC tumor cells typically express prostate specific membrane antigen (PSMA), a folate hydrolase which increases cell folate uptake (Yao et al. Prostate 70:305, 2010). We hypothesize that 5FU, functioning as an anti-folate, antagonizes the downstream effects of PSMA over-expression. Herein we present an update to our 5FU experience in heavily pretreated mCRPC. Methods: Data was retrospectively collected at Tulane Cancer Center for patient treatment history, disease history, performance status and laboratory parameters. All patients were treated with continuous infusion of 5-fluorouracil at a dose of 200mg/m2-day. Results: 24 patients at Tulane Cancer Center were treated with 5FU between Oct/2013 and Oct/2018. The median age was 72, patients had an median of 6 lines of prior therapy, and 37.5% had liver metastatic disease. The 30% and 50% PSA response rates were 41.6% and 29% respectively. Of the PSA responders, 100% had some level of PSA decline after 7 days of therapy with a median time to best PSA response of 2.6 months (95% CI 0.2-5.1). Median time on treatment was 11 weeks for the entire cohort and 18.5 weeks for patients with a ≥30% PSA decline; 20.8% of patients remained progression free by PSA at 4 months. Improvement in performance status as well as transfusion independence were noted in a subset of patients with myelophthistic anemia due to prostate cancer. No grade > 3 AEs were attributable to 5FU. Conclusions: 5FU is a relatively inexpensive and well tolerated therapy that shows benefit in some heavily pretreated patients. Additionally, the ability to quickly assess the potential for response, and lack of > grade 3 adverse events makes it a reasonable option for this population.

Aberrations in androgen receptor ctDNA varies by race in metastatic castrate-resistant prostate cancer

247

Background: Characterization of circulating free DNA (ctDNA) may aid understanding of the pathophysiology of metastatic castrate resistant prostate cancer (mCRPC). The goal of this study was to evaluate and compare somatic alterations in ctDNA between African-American (AA) and Caucasian (C) mCRPC patients. Methods: 24 AA were retrospectively case-matched by prior treatment with 45 CA mCRPC; ctDNA was assessed with Guardant360 assay (Guardant Health, Redwood City, CA). Mutant allelic fraction, mutations, and gene amplification were compared. Results: Prior to testing, 6 AA pts and 12 CA pts had 0 lines of CRPC therapy, 7 AA pts and 10 CA pts had 1-2 lines, 5 AA and 17 CA had 3-4 lines, 5 AA and 2 CA had 5-6 lines, and 1 AA pt and 4 CA pts had >6 lines. The median Gleason score was 8, regardless of race. The median ctDNA mutant allelic fraction was 0.40% for AA pts and 0.60% for the CA pts. Mutations and/or amplifications in individually assessed genes are shown in the Table. No statistically significant differences were detected except for the androgen receptor (AR) gene where alterations (mutations and/or amplifications) were more frequent in AA as compared to C (p=0.04). Conclusions: AR alterations were more commonly detected in ctDNA in AA men as compared to C suggesting that AR driven pathophysiology may predominate in this setting. Additional analyses with a larger cohorts are warranted. Data from prospective trials in mCRPC using abiraterone/prednisone (George et al. LBA 5009, ASCO 2018) similarly suggest that mCRPC may be comparatively more AR driven in this racial setting. Alterations in ctDNA compared between AA and C men with mCRPC. [Table: see text]

cfDNA TP53 mutations with treatment history and disease progression in metastatic CRPC

263

Background: Cell free DNA (cfDNA) has made it possible to actively monitor alterations of genes during the course of therapy. Herein we evaluate cfDNA pathologic mutations in TP53 after treatment resistance to AR targeted therapy and taxanes in mCRPC. In addition to the presence or absence of pathologic mutations, the quantity of the mutations was characterized as present/absent, quantitatively as > or < 1%, or >or <10%. Methods: A retrospective study was done for 111 patients, with 226 separate cfDNA testing dates. Each cfDNA test was treated as an individual sample; of the 226 tests, 112 were positive for pathologic TP53 mutations, and 114 were negative. Treatment histories were collected and TP53 data analyzed in relation to resistance for abiraterone (abi), enzalutamide (enza), abi + enza, or abi + enza and a taxane. Treatment resistance was categorized as being present for those patients completing therapy with an agent. Results: Patients with pathologic TP53 mutations were more likely to have had progression after a novel hormone and a taxane (P = 0.005). Higher concentrations of TP53 mutational loads as measured by cfDNA concentration of >1% and >10% were more likely to be present in patients progressing after a taxane, abi, and enza. Common alterations associated with TP53 included AR amplifications and mutations, MYC amplifications, and BRCA2 mutations. Conclusions: TP53 mutations in cfDNA of mCRPC are progressively more likely to be present in patients, especially after patients receiving both taxanes and a novel hormonal agent. The predictive and prognostic significance of these changes are being further evaluated.[Table: see text]

Cell-free DNA (cfDNA) analysis and evaluation of BRAF amplifications and mutations in metastatic castration-resistant prostate cancer

255

Background: Cell-free DNA (cfDNA) is an accessible method for characterizing tumoral alterations. We report cfDNA screenings of prostate cancer pts positive for BRAF amplifications/mutations in pts with metastatic CRPC. Methods: Guardant360 testing (Guardant Health, Inc.) assesses cell-free DNA analysis using sequencing to identify genomic alterations in 73 cancer-related genes in circulation. A total of 133 metastatic castrate resistant prostate cancer (mCRPC) pts in various stages of therapy had Guardant cfDNA analyses. Treatment histories prior to testing and concurrent cfDNA alterations were analyzed. Results: BRAF amplifications were detected in 32 (24%) mCRPC pts; 5 pts had concurrent BRAF mutations. Of the mutations detected, only one (K601E, n = 2) was a known activating mutation while all others were variants of unknown significance (VUS). One K601E mutation pt had no other cfDNA alterations. Additionally, 4 pts without BRAF amplification had VUS BRAF mutations. BRAF amplification pts had ≥ 2 concurrent gene amplifications/alterations with the median being 8. The most common recurrent amplifications/alterations were AR (75%), p53 (59%), CDK6 (53%), MET (50%), and MYC (50%). Abiraterone (Abi) and/or Enzalutamide (Enza) resistance was associated with BRAF amplification (p = 0.0042). Non-Abi/Enza resistance pts were less likely to have BRAF amplification. The 2 pts with BRAF K601E mutation were treated with targeted protocol therapy without success however one K601E pts was subsequently treated with cabazitaxel+carboplatin which produced a positive clinical response and a 99.79% reduction in PSA. Conclusions: Pts resistant to Abi/Enza have an increased risk of developing BRAF amplifications. BRAF amplifications arise in the context of multiple additional detectable cfDNA alterations. Identification of actionable mutations, such as BRAF K601E, illustrates the potential for cfDNA testing to direct pt treatment. As cfDNA profiling continues to expand, the ability translate alterations into clinically actionable strategies is critical.

cfDNA analysis of mCRPC patients expressing mutations in Wnt signaling

256

Background: Alterations in Wnt signaling have been shown to play a role in the development of castrate resistant prostate cancer. Cell free DNA (cfDNA) isolated from patient plasma can provide a non-invasive way to further assess this role. The goal of this study was to identify patients with cfDNA alterations in major canonical Wnt signaling components (APC and/or beta-catenin), and relate the emergence of these mutations during treatment to alterations in other common mutations. Methods: 134 clinically progressive metastatic CRPC patients from Tulane Cancer Center underwent cfDNA analysis through Guardant360 test (Guardant Health, Redwood City, CA). This analysis consisted of exonic coverage of 70 genes as well as amplifications in 18 genes, with mutations categorized as either pathologic, non-pathologic or as variants of unknown significance (VUS). Clinical annotation of prior treatment history was recorded. Results: 21.6% (29/134) of the mCRPC patients evaluated had a canonical Wnt signaling (APC and/or CTNNB) alteration. Of these patients, 62.1% (18/29) had mutations identified as pathologic. 77.8% (14/18) of patients identified with a pathologic Wnt mutation were treated with abiraterone and/or enzalutamide prior to Guardant360 testing. To determine potential associations between Wnt signaling alterations and other detected changes in cfDNA, the relationship between Wnt mutations (APC and/or CTNNB1) with pathologic TP53 mutations, AR mutations, BRAF amplifications, and MYC amplifications was assessed using a patient’s latest Guardant360 test. A significant positive association was found between Wnt mutations (n = 18) and MYC amplifications (n = 22), (p = 0.0373). 33.3% (6/18) of patients with a pathologic Wnt mutation were found to have amplification in MYC. Other notable associations included Wnt mutations and AR mutations (n = 74), which approached statistical significance (p = 0.112). Conclusions: While the understanding of the role of Wnt signaling in the treatment of mCRPC is still evolving, co-segregation of Wnt-signaling alterations with other oncogenic alterations, particularly MYC, which has been identified as a target of canonical Wnt signaling, may provide insights with regards to future management of mCRPC.

Cosegregation of cfDNA AR amplifications and mutations in relation to BRAF and MYC in CRPC

261

Background: cfDNA can be utilized to evaluate androgen receptor (AR) alterations in metastatic castrate resistant prostate cancer (mCRPC) patients. Given the role of androgen receptors in the pathophysiology of PC, mutations in the AR gene can have important implications regarding prognosis and treatment. The goal of this research was to evaluate associations between certain PC treatments and AR mutations. Methods: Treatment and demographic data from 121 patients diagnosed with PC was collected including genetic data derived from the Guardant360 test (Guardant Health, Redwood City, CA). AR alterations were compared among race, treatment history, and other somatic alterations. Results: 52.9% (n = 64) of the mCRPC pts evaluated had an AR alteration. Of the pts with AR alterations, 43.8% (n = 28) had AR amplification (amps), 37.5% (n = 24) had AR mutations (muts), and 18.8% (n = 12) had both. AR muts included: T878A (n = 15), H875Y (n = 8), W742C (n = 11), AR L702H (n = 7), and others. AR alterations comprised on average about 7% of tumoral cfDNA. To better understand the relationship between AR alterations and other commonly detected cfDNA aberrations, associations between BRAF (27.3%), TP53 (44.6%), and MYC (19.0%) and AR were assessed. Among these genes, P53 alterations were all Muts. MYC (n = 21) and BRAF (n = 27) alterations were predominantly amps though muts were also detected in MYC (n = 3) and BRAF (n = 9). P53 muts were not significantly associated with AR alterations. BRAF and MYC alterations significantly associated with AR alterations (p = 0.0012 and p = 0.0223). Pts were re-tested upon disease progression; among these patients, 52.2% (n = 24) had an increase in overall mut burden. Conclusions: AR alterations in cfDNA impact both disease progression and response to therapy. Co-segregation of AR, BRAF, and MYC alterations may also have significant prognostic and therapeutic implications. Further research, a larger sample size, and longitudinal assessment are needed to further elucidate associations between disease progression and treatment response to the development of somatic alterations over time in mCRPC.

The impact of neoadjuvant weekly ixabepilone for high-risk prostate cancer: A phase I/II clinical trial

158

Background: Potential benefits of neoadjuvant chemotherapy are tumor downstaging and treatment of micrometastatic disease. A prior study using docetaxel yielded no pathologic complete responses and concerns for increased operative morbidity. In Phase II studies, ixabepilone has promising activity in metastatic prostate cancer. Our study is a Phase I/II clinical trial evaluating neoadjuvant, weekly ixabepilone in men with high-risk prostate cancer opting for radical prostatectomy.

Methods: Men with high risk prostate cancer defined as either Gleason 8-10, cT3 disease, high volume Gleason 4+3 and a palpable nodule or a PSA>20 ng/ml were eligible. Men received weekly ixabepilone 16-20/m2 for 12-16 weeks prior to surgery. Fifteen men underwent robotic prostatectomy; one patient who had an open prostatectomy. Initial PSA response, post-operative PSA values, pathology, and evaluation of adverse events were recorded.

Results: We enrolled 16 men with a mean follow-up of 15.25 months at time of review. All had pretreatment Gleason scores of 4+3 or higher. With neoadjuvant treatment, PSA values decreased in 14/16 men (mean 46.8%); increased in 2/16 men. None reached an undetectable pre-operative PSA. Nine men experienced an adverse event requiring dose modification or cessation of chemotherapy (neuropathy or allergic reaction). Only 5/16 men completed planned treatment. Mean operative time, EBL, and hospital stay were 189 minutes, 184mL, and1.5 days, respectively; all consistent with institutional and national norms. Post surgery 15/16 (94%) had pT3 disease, 8/16 (50%) had a positive surgical margin and 2/16 (12.5%) had positive regional lymph nodes. There were no pathologic complete responses. Only 1/16 (6.25%) had a biochemical relapse.

Conclusions: While a PSA response is achieved, there is substantial toxicity with neoadjuvant weekly ixabepilone. Men were able to undergo prostatectomy without increased morbidity after neoadjuvant therapy. Extracapsular extension and positive surgical margins remained common in this population with high-risk disease. Assessment of biochemical recurrence rates and time to treatment failure will require longer, planned follow-up.