Background: Several hypotheses proposed to explain the worse prognosis for older melanoma patients include different tumor biology and diminished host response. If the latter were true, then biologic frailty, and not age, should be an independent prognostic factor in melanoma.
Methods: Our prospective institutional review board (IRB)-approved database was queried for stage III patients with computed tomography (CT) scans at time of lymph node dissection (LND). Psoas area (PA) and density (PD) were determined in semi-automated fashion. Kaplan-Meier (K-M) survival estimates and Cox proportional-hazard models were used to determine PA and PD impact on survival and surgical complications.
Results: Among 101 stage III patients, PD was significantly associated with both disease-free survival (DFS) (P = 0.04) and distant disease-free survival (DDFS) (P = 0.0002). Cox multivariate modeling incorporating thickness, age, ulceration, and N stage showed highly significant association with PD and both DFS and DDFS. DDFS was significantly associated with Breslow thickness (P = 0.04), number of positive nodes (P = 0.001), ulceration (P = 0.04), and decreasing muscle density (P = 0.01), with hazard ratio of 0.55 [95% confidence interval (CI) 0.35-0.87]. PD also correlated with surgical complications, with odds ratio (OR) of 1.081 [95% CI 1.016-1.150, P = 0.01].
Conclusions: Decreased psoas muscle density on CT, an objective measure of frailty, was as important a predictor of outcome as tumor factors in a cohort of stage III melanoma patients. On multivariate analysis, frailty, not age, was associated with decreased disease-free survival and distant disease-free survival, and higher rate of surgical complications.