Lymph Node-Positive (N1M0) Prostate Cancer: An Analysis of Prognostic and Predictive Factors for Men Treated With and Without Definitive External Beam Radiation Therapy in the United States Population

Abstract

For men with N1M0 prostate cancer (PCa), the benefits of definitive external beam radiation therapy (RT) and the prognostic and predictive value of standard clinical risk factors are not well characterized. Here, we evaluate the survival outcomes for patients with N1M0 PCa treated with RT vs no local therapy (NLT) and the impact of PSA, Gleason (G), T stage (T), and age in the United States population. The Surveillance, Epidemiology, and End-Results (SEER) database was queried for T1-4N1M0 PCa, managed with RT or NLT, G 6-10, diagnosed from 2004-2006 (with 2004 being the first year of PSA and individual G coding in SEER). RT included external beam RT with or without brachytherapy (BT). Patients undergoing prostatectomy or BT alone were excluded. Overall (OS) and cause-specific survival (CSS) were estimated using the Kaplan-Meier method. The survival impact of RT (yes/no), age (< 65, ≥ 65), PSA (continuous), G (6-10), and T stage (1-4) were evaluated by univariate (UVA) log-rank and multivariate (MVA) Cox proportional hazards comparisons. Forest plots with Cox model interaction tests were used to evaluate the predictive value of collected variables on the impact RT vs NLT. Interactions between RT and PSA were evaluated both continuously and in groups of < 10, 10-19.9, 20-49.9, 50-97.9, and ≥ 98 ng/mL. Five hundred twenty-five patients with complete datasets were evaluable. The median age was 65 and median follow-up was 57 months. Two hundred sixty-six patients (51%) received RT and 259 (49%) NLT. On UVA, RT was associated with an improved 6.5-year OS of 67% vs 48% (p < 0.001) and CSS of 79% vs 59% (p < 0.001), which remained highly significant on MVA for OS (HR = 0.54, p < 0.001) and CSS (HR = 0.51, p < 0.001). Increasing G (6-10) and T (1-4) were associated with inferior survival on UVA (all p ≤ 0.001), which remained significant on MVA for G (OS HR = 1.26, p = 0.001; CSS HR = 1.37, p < 0.001) and T (OS HR = 1.24, p = 0.011; CSS HR = 1.30, p = 0.008). Age ≥ 65 was associated with inferior OS on UVA (p = 0.003) and MVA (HR = 1.46, p = 0.013), but not CSS (all p > 0.05). Increasing PSA was not significantly associated with OS or CSS on UVA or MVA (all p > 0.05). No significant predictive interactions for OS or CSS were observed between PSA (continuous or groupings), G, T, or age and the survival benefits achieved with RT (all interaction p > 0.05). In this analysis, definitive RT was associated with an approximate 20% improvement in OS and CSS at 6.5 years in the United States population. G and T were observed to be prognostic for OS and CSS, while age was prognostic for OS alone, and PSA was not prognostic for OS or CSS. No predictive interactions for RT benefit were observed for PSA, G, T, or age. Clinically, this analysis suggests that high-risk features, notably including high PSA and advanced age, should not exclude N1M0 PCa patients from consideration of definitive RT.