Large-scale Genomic Landscape and Clinical Outcomes of De Novo and Treatment-emergent Neuroendocrine Prostate Cancer

Background and objective Neuroendocrine prostate cancer (NEPC) is a rare, aggressive subtype of prostate cancer that arises either as de novo or as a treatment-emergent phenotype during androgen-deprivation therapy. Given its rarity, large-scale studies comparing genomic features and clinical outcomes between de novo NEPC and treatment-emergent NEPC (t-NEPC) are limited. We leveraged a nationwide comprehensive genomic profiling database to characterize the genomic landscape of NEPC and to describe the overall survival (OS) after initiation of NEPC treatment, with exploratory comparisons between de novo NEPC and t-NEPC. Design, outcome measurements and statistical analysis We retrospectively analyzed 302 patients with NEPC identified from among 5893 patients with prostate cancer registered in the Center for Cancer Genomics and Advanced Therapeutics database. Patients were classified as de novo NEPC ( n = 184) or t-NEPC ( n = 118). Genomic alterations detected by comprehensive profiling, including pathogenic/likely pathogenic mutations and oncogenic/likely oncogenic variants, were analyzed. OS was assessed using the Kaplan-Meier method, and prognostic factors were evaluated using Cox proportional hazards models. Results and limitations The most frequent genomic alterations in de novo NEPC and t-NEPC, respectively, were TP53 (50% and 49%), RB1 (44% and 44%), PTEN (21% and 24%), BRCA2 (15% and 23%), and MYC (17% and 14%), with no statistically significant differences across histology and specimen site. Median OS from initiation of NEPC treatment was 21.5 mo (95% confidence interval [CI], 15.6–32.1) in de novo NEPC and 13.6 mo (95% CI, 11.5–32.2) in t-NEPC, corresponding to a median difference of 7.9 mo (bootstrap 95% CI, 3.1–13.7); however, the between-group difference was not statistically significant (log-rank p = 0.3; hazard ratio, 1.29 [95% CI, 0.81–2.07]). In exploratory multivariable analysis, concurrent alterations in TP53 , RB1 , and PTEN were statistically significantly associated with a higher hazard of death after initiation of NEPC treatment compared with no alterations in these genes. Conclusions In this nationwide genomic analysis, de novo NEPC and t-NEPC showed largely similar genomic profiles after neuroendocrine differentiation; however, these findings should be interpreted with caution because of limitations in available clinical variables. These findings provide real-world insights into the clinical and genomic characteristics of NEPC.