Melanoma patients showed higher serum concentrations of PD-1 (P = 0.0054) and PD-L1 (P < 0.0001) than healthy controls. Elevated serum PD-1 and PD-L1 levels at treatment baseline were associated with an impaired best overall response (BOR) to anti-PD-1 (P = 0.014, P = 0.041), but not to BRAF inhibition therapy. Baseline PD-1 and PD-L1 serum levels correlated with progression-free (PFS; P = 0.0081, P = 0.053) and overall survival (OS; P = 0.055, P = 0.0062) in patients who received anti-PD-1 therapy, but not in patients treated with BRAF inhibitors. By combining both markers, we obtained a strong discrimination between favorable and poor outcome of anti-PD-1 therapy, with elevated baseline serum levels of PD-1 and/or PD-L1 associated with an impaired BOR (P = 0.037), PFS (P = 0.048), and OS (P = 0.0098). This PD-1/PD-L1 combination serum biomarker was confirmed in an independent multicenter validation set of serum samples prospectively collected at baseline of PD-1 inhibition (BOR, P = 0.019; PFS, P = 0.038; OS, P = 0.022). Multivariable Cox regression demonstrated serum PD-1/PD-L1 as an independent predictor of PFS (P = 0.010) and OS (P = 0.003) in patients treated with PD-1 inhibitors.
Programmed cell death protein 1 (PD-1) checkpoint inhibition has recently advanced to one of the most effective treatment strategies in melanoma. Nevertheless, a considerable proportion of patients show upfront therapy resistance and baseline predictive biomarkers of treatment outcome are scarce. In this study we quantified PD-1 and programmed death-ligand 1 (PD-L1) in baseline sera from melanoma patients in relation to therapy response and survival.
Our findings indicate PD-1 and PD-L1 as useful serum biomarkers to predict the outcome of PD-1 inhibition therapy in melanoma patients and to select patients for PD-1-based versus BRAF-based therapy strategies.
Sera taken at therapy baseline from a total of 222 metastatic melanoma patients (two retrospectively selected monocentric discovery cohorts, n = 130; one prospectively collected multicentric validation cohort, n = 92) and from 38 healthy controls were analyzed for PD-1 and PD-L1 concentration by sandwich enzyme-linked immunosorbent assay.
This website uses cookies to ensure you get the best experience on our website.