Prognostic heterogeneity of metastatic patterns in prostate cancer relative to bone-only disease.

Abstract

e17108 Background: Metastatic prostate cancer exhibits heterogeneous patterns of spread, but prior studies have not distinguished distant lymph node from visceral organ metastases when characterizing non-bone disease. Methods: Using the National Cancer Database (2016-2022), 56,112 patients with documented metastatic prostate cancer at diagnosis were identified. Patients were classified into five mutually exclusive groups based on sites of metastatic involvement: bone-only, distant lymph node-only, visceral organ-only (liver, lung, or brain), distant lymph node plus visceral organs without bone, and bone plus any non-bone site. Multivariable Cox proportional hazards regression was used to estimate associations between metastatic pattern and overall survival, adjusting for demographics, comorbidities, tumor characteristics, and treatment facility; complete-case analysis was performed. Results: Among 56,112 patients, 64% had bone-only metastases, 6% distant lymph node-only, 3% visceral organ-only, 1% distant lymph node plus visceral organs, and 26% bone plus non-bone metastases. Median follow-up ranged from 22 to 32 months across groups. Compared to bone-only disease (reference; 53% mortality), distant lymph node-only metastases conferred substantially lower mortality (36% deaths; adjusted hazard ratio [HR] 0.64, 95% confidence interval [CI] 0.60-0.68), while visceral organ-only metastases showed similar survival to bone-only disease (56% deaths; HR 1.06, 95% CI 1.00-1.13, p = 0.09). Patients with bone plus non-bone metastases had the highest mortality (61% deaths; HR 1.34, 95% CI 1.31-1.38). Among 6,853 patients with non-bone-only disease, 60% had isolated distant lymph node involvement, explaining the apparent survival advantage of aggregated non-bone disease. These associations persisted after adjusting for treatments received and were consistent across major histology subtypes. Conclusions: Distant lymph node and visceral organ metastases represent distinct prognostic entities in metastatic prostate cancer. Distinguishing these patterns enables more precise risk stratification relative to traditional bone-only classification and may inform treatment selection. Adjusted hazard ratios for overall survival by metastatic phenotype. Metastatic phenotype Adjusted HR 95% CI Bone-only (reference) 1 — Distant LN-only 0.64 0.60–0.68 Visceral-only 1.06 1.00–1.13 LN + visceral (no bone) 1.18 1.08–1.29 Bone + non-bone 1.34 1.31–1.38 Hazard ratios derived from multivariable Cox proportional hazards models adjusted for demographics, comorbidity burden, tumor characteristics, year of diagnosis, and treatment facility type. The better prognosis of isolated distant lymph node metastases warrants prospective validation and consideration in clinical trial design.