An increasing number of studies have examined the ability of programmed death-ligand 1 (PD-L1) to function as a marker for tumor prognosis. However, whether PD-L1 expression is a prognostic factor for the poor outcomes in many human cancers remains controversial. This study aims to investigate the prognostic role of PD-L1 expression through a meta-analysis update of 60 studies.Methods:
The studies were identified by searching PubMed, Embase, Google Scholar, and Cochrane Library, and were assessed by further quality evaluation. The pooled hazard ratios (HRs) with 95% confidence intervals (CIs) for total and stratified analyses were calculated to investigate the association between the PD-L1 expression and the overall (OS) and disease-free (DFS) or progression-free survivals (PFS) of cancer patients. Heterogeneity and publication bias were also investigated.Results:
The results indicated that PD-L1 overexpression can predict a poor OS (HR = 1.58, 95% CI = 1.38–1.81, P <.000) and DFS/PFS (HR = 1.72, 95% CI = 1.26–2.33, P = .001). Subgroup analyses showed that PD-L1 overexpression was significantly related to the poor OS in patients with breast (HR = 1.98, 95% CI = 1.15–3.41, P = .014), urothelial (HR = 2.24, 95% CI = 1.61–3.12, P <.000), renal (HR = 3.30, 95% CI = 2.23–4.86, P <.000), and gastric cancers (HR = 1.56, 95% CI = 1.02–2.37, P = .040). Furthermore, PD-L1 overexpresion was significantly associated with poor DFS/PFS in patients with hepatocellular carcinoma (HCC) (HR = 1.72, 95% CI = 1.21–2.46, P = .003), melanoma (HR = 3.39, 95% CI = 2.02–5.69, P <.000), and renal carcinoma, (HR = 5.04, 95% CI = 2.87–8.86, P <.000). The adverse prognostic impact of PD-L1 was observed in patients of different ethnicities.Conclusions:
The findings of this meta-analysis suggest the correlation of PD-L1 overexpression with worse OS in patients with solid tumors. However, the correlations differed according to tumor types.