BACKGROUND

The prognostic value of soluble programmed cell death ligand-1 (sPD-L1) in patients with cancer has been inconsistent across previous studies.

OBJECTIVE

This meta-analysis aimed to investigate the prognostic significance of sPD-L1 in human tumors.

METHODS

A comprehensive search of PubMed, Web of Science, Embase, and Cochrane databases from inception to January 6, 2020 was conducted. Studies of sPD-L1 measured by enzyme-linked immunosorbent assay (ELISA) that had available hazard ratios (HRs) for survival outcomes based on high or low sPD-L1 levels were included. The primary endpoint was long-term survival, namely, overall survival (OS), and the second endpoint was short-term survival, including progression-free survival (PFS), disease-free survival (DFS), recurrence-free survival (RFS), and cancer-specific survival (CSS).

RESULTS

A total of 21 studies, with 2413 patients, were included in this meta-analysis. Elevated sPD-L1 was associated with worse OS [HR = 2.46, 95% confidence interval (CI) 1.74-3.49, P < 0.001]. Moreover, high sPD-L1 was predictive of worse PFS/DFS/RFS/CSS (HR = 2.22, 95% CI 1.47-3.35, P < 0.001). High sPD-L1 was consistently correlated with poor OS and PFS/DFS/RFS/CSS irrespective of study design, sample, and cut-off value of sPD-L1. However, there was non-significant correlation between sPD-L1 and sex, age, clinical stage, Eastern Cooperative Oncology Group Performance Status, tumor differentiation, or serum lactate dehydrogenase.

CONCLUSIONS

This meta-analysis showed that sPD-L1 was correlated with poor prognosis in human tumors. In addition, sPD-L1 could be used as a predictive factor of inferior outcomes during multiple malignancy treatments.