Search Results
The literature screening process for this study was showed by flow diagram (Figure 1). A total of 1674 records were retrieved. 274 records were excluded for duplicates, 1391 records were excluded across exclusion criteria, 2 studies were excluded due to lack of the comparative data, and 7 randomized clinical studies which compared the efficacy and adverse events of PD-1 inhibitor plus chemotherapy with placebo plus chemotherapy as a first-line treatment in advanced or metastatic EC/GEJC[21-25, 28, 29].
The main characteristics of the included trials were summarized in Table 1. All literatures were randomized controlled trials (RCT) and published in 2021. The trials included a total of 3754 EC/GEJC patients. 4 trials enrolled patients with ESCC[22-25] , and 1 trial enrolled patients with ESCC, EAC and GEJC[21]. We also extracted the patients’ data with EAC and GEJC from 2 trials[28, 29].
Efficacy outcomes of PD-1 inhibitor plus chemotherapy
The pooled HR of OS and PFS, and the pooled OR of ORR were used to assess the efficacy of PD-1 inhibitor plus chemotherapy in first-line treating EC. In term of OS benefit, the PD-1 inhibitor plus chemotherapy led to a 28% reduction in the risk of death compared with chemotherapy (HR=0.72; 95% CI: 0.66-0.78, P<0.01), and there was no obvious heterogeneity (I2 = 0%, P =0.435) (Figure 2A). The pooled HR of PFS showed that PD-1 inhibitor plus chemotherapy significantly reduced the risk of disease progression compared with chemotherapy (HR=0.62; 95% CI: 0.57-0.68, P<0.01; and heterogeneity: I2 = 46.6%, P =0.112) (Figure 2B). In addition, the difference of ORR benefit was significantly between the PD-1 inhibitor plus chemotherapy group and chemotherapy group (OR=2.07; 95% CI: 1.76-2.43, P<0.01; and heterogeneity: I2 = 24.5%, P =0.264) (Figure 2C).
Association of histology and PD-L1 expression status with OS
Five studies had the result of OS for squamous cell carcinoma, and three studies had the result of OS for adenocarcinoma. The difference of OS benefit across histology subgroups obtained a near-significant trend (P =0.054) (Figure 3A). There were two studies assessed PD-L1 combined positive score (CPS). In the subgroup with PD-L1 CPS less than 10%, the pooled HR of OS was 0.76, and the OS benefit was the higher in patients with a PD-L1 CPS of at least 10% (HR=0.69). However, there was no statistically difference in terms of PD-L1 expression level (P = 0.188) (Figure 3B).
Subgroup analyses by clinical features
We performed subgroup analyses according to some basic information, including age, sex, ECOG PS. There was no significant interaction between treatment effect in terms of OS and clinical features (age: P=0.236; sex: P=0.340; ECOG: P=0.593) (Figure 4A). Same as before, the PFS benefit of PD-1 inhibitor plus chemotherapy compared to chemotherapy did not vary significantly across subgroups (age: P=0.922, sex: P = 0.390, ECOG PS: P= 0.319) (Figure 4B).
The Safety Evaluation of PD-1 inhibitor plus chemotherapy
The pooled OR of TrAEs was 1.85 (95% CI: 1.21-2.84, P<0.01; and heterogeneity: I2 = 9.9%, P=0.350), which showed that PD-1 inhibitor plus chemotherapy can increase the incidence of TrAEs, compared with chemotherapy (Figure 5A). In term of grade 3 or higher TrAEs, there was no statistical difference, but a near-significant trend (OR=1.24; 95% CI: 1.00-1.55, P=0.05) (Figure 5B).
Assessment of study quality and sensitivity analysis
All trials included in this study were multicenter and randomized clinical trial, and five trials were double blinded. The Jadad score ranged from 3 to 5, indicating that the quality was high (Table 1). The bias risk of the included studies was shown in Figure 6F. All trials included random sequence generation. Five trials were double blinded. Two trials were open label, and therefore these studies are at performance and detection bias. Two studies didn’t report all data, so that they had risk for reporting bias. Nevertheless, all studies were felt to be low risk for attrition and other bias.
Sensitivity analysis was applied to evaluate the stability of our meta-analysis. The results showed that our meta-analysis were robust in term of pooled HR for OS (Figure 6A) and PFS (Figure 6B), pooled OR for ORR (Figure 6C), TrAEs (Figure 6D) and grade 3 or higher TrAEs (Figure 6E). No significant deviation from the overall results was detected.