Body composition and mortality in men receiving prostate radiotherapy: A pooled analysis of NRG/RTOG 9406 and NRG/RTOG 0126.

Purpose: To validate the association between body composition and mortality in men treated with radiation for localized prostate cancer (PCa). Secondarily, to integrate body composition as a factor to classify patients by risk of all‐cause mortality. Materials and Methods: Participants of NRG/Radiation Therapy Oncology Group (RTOG) 9406 and NRG/RTOG 0126 with archived computed tomography were included. Muscle mass and muscle density were estimated by measuring the area and attenuation of the psoas muscles on a single slice at L4–L5. Bone density was estimated by measuring the attenuation of the vertebral body at mid‐L5. Survival analyses, including Cox proportional hazards models, assessed the relationship between body composition and mortality. Recursive partitioning analysis (RPA) was used to create a classification tree to classify participants by risk of death. Results: Data from 2066 men were included in this study. In the final multivariable model, psoas area, comorbidity score, baseline prostate serum antigen, and age were significantly associated with survival. The RPA yielded a classification tree with four prognostic groups determined by age, comorbidity, and psoas area. Notably, the classification among older (≥70 years) men into prognostic groups was determined by psoas area. Conclusions: This study strongly supports that body composition is related to mortality in men with localized PCa. The inclusion of psoas area in the RPA classification tree suggests that body composition provides additive information to age and comorbidity status for mortality prediction, particularly among older men. More research is needed to determine the clinical impact of body composition on prognostic models in men with PCa. This study strongly supports that body composition is related to mortality in men with localized prostate cancer. The inclusion of psoas area in the recursive partitioning analysis classification tree suggests that body composition provides additive information to age and comorbidity status for mortality prediction, particularly among older men. [ABSTRACT FROM AUTHOR]