Implementation of programmed death-ligand 1 (PD-L1) expression as a prognostic biomarker for patients with lung cancer.

We thank Li et al. for showing interest in our article entitled “The clinicopathological and prognostic significance of programmed death-ligand 1 (PD-L1) expression assessed by immunohistochemistry in lung cancer: a meta-analysis of 50 studies with 11,383 patients” (1). We would like to respond to the concerns one by one raised by Li et al. (2).

Firstly, we felt sorry for the carelessness in extracting hazard ratios (HRs) and 95% confident intervals (95% CIs) from two studies on lymphoepithelioma-like carcinoma (LELC) (3,4). We thus performed additional analysis with upper limit of 9.863 and converted HR and 95% CI, and the results did not alter the overall findings. After reanalysis, the pooled HRs and 95% CIs suggested that PD-L1 immunohistochemistry (IHC) expression was related to poor OS in lung cancer (HR =1.42, 95% CI: 1.22–1.65) (), consistent with results in the primary analysis. Subgroup analyses according to histology () revealed that high PD-L1 expression was significantly related to poor overall survival (OS) of non-small cell lung cancer (NSCLC) patients (HR =1.35, 95% CI: 1.13–1.61), adenocarcinoma (ADC) patients (HR =1.79, 95% CI: 1.22–2.64), squamous cell carcinoma (SCC) patients (HR =1.79, 95% CI: 1.39–2.32), but there was no association of PD-L1 expression with survival in small cell lung cancer (SCLC) patients (HR =1.05, 95% CI: 0.39–2.78) and LELC patients (HR =0.91, 95% CI: 0.10–8.16). In fact, the relationship between PD-L1 expression and prognosis in LELC has long been debated and remains controversial. A recent study by Sha et al. (5) showed that participants with positive PD-L1 expression tended to have longer progression-free survival (PFS) and OS times than those with negative PD-L1 expression, while Yu et al. (6) revealed that positive expression of PD-L1 in tumor cells had no association with OS in LELC. Different testing platforms and thresholds for defining positive PD-L1 expression may partly explained this discrepancy. More studies are needed to explore the prognostic significance of PD-L1 expression in LELC patients.

Figure 1

Forest plot describing the association between PD-L1 expression and OS of patients with lung cancer. PD-L1, programmed death-ligand 1; OS, overall survival.

Figure 2

Forest plot describing subgroup analysis of the association between PD-L1 expression and OS according to histology. PD-L1, programmed death-ligand 1; OS, overall survival; NSCLC, non-small cell lung cancer; ADC, adenocarcinoma; SCLC, small cell lung cancer; LELC, lymphoepithelioma-like carcinoma; SCC, squamous cell carcinoma.

Secondly, Fang and colleagues defined cases with more than 5% expression of PD-L1 as positive ones while PD-L1 H-score 30 was further determined as the best threshold to discriminate OS and analyzed in the Cox proportional hazard regression analysis. The authors did not clearly illustrate the definition of higher or positive PD-L1 expression as previous publications usually divided patients into high/low PD-L1 expression based on PD-L1 expression values (5%, etc.). We re-performed the subgroup analysis based on cutoff value (), which was comparable with results in our previous analysis.

Figure 3

Forest plot describing subgroup analysis of the association between PD-L1 expression and OS according to cutoff value. PD-L1, programmed death-ligand 1; OS, overall survival.

Thirdly, we would like to emphasize that P values in the Abstract indeed indicated values of heterogeneity. Although random-effect models were used to deal with significant heterogeneity in both primary and subgroup analyses in the meta-analysis, enough attention should be paid to and the findings need be cautiously interpreted (7).

Moreover, we agreed with Li et al. that the above two studies (3,4) may be overlapped. However, the two studies were conducted separately by two research teams. Jiang et al. enrolled 79 pulmonary LELC cases from January 2001 to December 2013 while Fang et al. enrolled 113 surgically resected pulmonary LELC cases from January 2008 to December 2012. In fact, it is hard to conclude whether patients in the two studies were overlapped. We further conducted sensitivity analysis by omitting one study each time and the results remain robust ().

Figure 4

Forest plot describing the association between PD-L1 expression and OS of patients with lung cancer. PD-L1, programmed death-ligand 1; OS, overall survival.

Figure 5

Forest plot describing the association between PD-L1 expression and OS of patients with lung cancer. PD-L1, programmed death-ligand 1; OS, overall survival.

The article’s supplementary files as