Background
The prognostic significance of programmed cell death-ligand 1 (PD-L1) expression on circulating tumor cells (CTCs) has been explored but is still in controversy. We performed, for the first time, a meta-analysis to systematically evaluate its prognostic value in human cancers.
Conclusions
Our analysis demonstrated that PD-L1 positive expression on CTCs predicted better survival prognosis for ICI treatment but worse survival for other therapies, which thus can be potentially used as a prognostic marker of malignant tumor treatment. However, the prognostic value of PD-L1+ CTCs for ICI treatment needs validation by more large-scale studies in the future.
Methods
Literature databases were searched for eligible studies prior to June 30, 2021. The pooled hazard ratios (HRs) and 95% confidence intervals (95% CIs) were calculated for the associations of pre-treatment and post-treatment PD-L1+ CTCs with progression-free survival (PFS) and overall survival (OS). Subgroup analyses with regards to cancer type, treatment, CTC enrichment method, PD-L1 detection method, cut-off, and specifically the comparison model were performed.
Results
We included 30 eligible studies (32 cohorts, 1419 cancer patients) in our analysis. Pre-treatment PD-L1+ CTCs detected by immunofluorescence (IF) tended to predict better PFS (HR = 0.55, 95% CI 0.28-1.08, p = 0.084) and OS (HR = 0.61, 95% CI 0.36-1.04, p = 0.067) for immune checkpoint inhibitor (ICI) treatment, but were significantly associated with unfavorable survival for non-ICI therapies (PFS: HR = 1.85, 95% CI 1.21-2.85, p = 0.005; OS: HR = 2.44, 95% CI 1.69-3.51, p < 0.001). Post-treatment PD-L1+ CTCs predicted markedly worse PFS and OS. The prognostic value was obviously modulated by comparison models. Among patients with detectable CTCs, PD-L1+ individuals had comparable survival to PD-L1- individuals, except ICI treatment for which PD-L1+ may predict better PFS (HR = 0.42, 95% CI 0.17-1.06, p = 0.067). Patients with PD-L1+ CTCs had worse survival prognosis compared to those without PD-L1+ CTCs in overall analysis (PFS: HR = 2.10, 95% CI 1.59-2.77, p < 0.001; OS: HR = 2.55, 95% CI 1.70-3.81, p < 0.001) and in most subgroups. Conclusions: Our analysis demonstrated that PD-L1 positive expression on CTCs predicted better survival prognosis for ICI treatment but worse survival for other therapies, which thus can be potentially used as a prognostic marker of malignant tumor treatment. However, the prognostic value of PD-L1+ CTCs for ICI treatment needs validation by more large-scale studies in the future.