Enrichment of "Cribriform" morphologies (intraductal and cribriform adenocarcinoma) and genomic alterations in radiorecurrent prostate cancer.

Locally relapsed prostate cancer (PCa) after radiation therapy (RT) is associated with substantial morbidity and mortality. Morphological and molecular consequences that may contribute to RT resistance and local recurrence remain poorly understood. Locally recurrent PCa tissue from 53 patients with clinically localized PCa who failed with primary RT and subsequently underwent salvage radical prostatectomy (RP) was analyzed for tumor focality, clinicopathological, molecular, and genomic characteristics. Targeted next-generation sequencing with full exon coverage of 1,425 cancer-related genes was performed on 10 representative radiorecurrent PCas exhibiting no RT effect with matched adjacent benign prostate tissue. At RP, 37 (70%) of PCas had no RT effect with the following characteristics: grade group (GG) ≥ 3 (70%), unifocal tumor (75%), extraprostatic disease (78%), lymph node metastasis (8%), and "cribriform" morphologies (84%) [cribriform PCa (78%) or intraductal carcinoma (IDC-P) (61%)] at a median percentage of approximately 80% of tumor volume. In the setting of multifocal tumors (25%) at RP, the cribriform morphologies were restricted to index tumors. Of 32 patients with available pre-RT biopsy information, 16 had GG1 PCa, none had cribriform morphologies at baseline but 81% demonstrated cribriform morphologies at RP. Notable alterations detected in the sequenced tumors included: defects in DNA damage response and repair (DDR) genes (70%) (TP53, BRCA2, PALB2, ATR, POLQ), PTEN loss (50%), loss of 8p (80%), and gain of MYC (70%). The median tumor mutational burden was 4.18 mutations/Mb with a range of 2.16 to 31.86. Our findings suggest that most radiorecurrent PCas are enriched in cribriform morphologies with potentially targetable genomic alterations. Understanding this phenotypic and genotypic diversity of radiorecurrent PCa is critically important to facilitate optimal patient management.