Genetics-Driven, Intensity-Modulated Adaptive Management of Patients With Metastatic Prostate Cancer.

[BACKGROUND] Survival differs markedly between men with metastatic prostate cancer (mPC) confined to lymph nodes (LNM) versus bone (BM). We examined whether site-specific genomic alterations-and their combinations-explain this disparity and could inform intensity-modulated follow-up or therapy.

[METHODS] Clinical and targeted-sequencing data for 1011 men with mPC in the cBioPortal for Cancer Genomics registry were analyzed (LNM-only = 622; BM-only = 389). Genes altered in > 5% of tumors (two-sided p < 0.05) were assessed individually and in every possible multigene cluster for associations with overall survival (OS). Survival was evaluated with Kaplan-Meier curves and the difference in restricted mean survival time (dRMST) to pinpoint the first significant curve divergence. Synthetic-lethal (SL) interactions were explored via the SLOAD database.

[RESULTS] In total, 18 of 184 profiled genes (9.8%) exceeded the 5% alteration threshold. FOXA1 was enriched in BM, whereas TMPRSS2, ERG, PTEN, ZFHX3, CDK12, and KMT2C were enriched in LNM (p < 0.05). Among 9143 tested gene clusters, 65 were associated with inferior OS; 48 occurred in the LNM subgroup, 17 in the combined cohort, and none in the BM alone. High-risk clusters showed first OS divergence 10-60 months after diagnosis of metastasis. SLOAD identified 615 putative SL pairs involving these genes.

[CONCLUSIONS] We identified 65 site-specific multigene clusters-chiefly in lymph node-only mPC-that underlie the survival gap between nodal and bone metastases. These signatures suggest a 10-60-month interval that may lend itself for intensity-modulated follow-up. We also discovered hundreds of synthetic-lethal gene-alteration pairs, opening future research opportunities in combinatorial therapeutic targeting.