Prognostic and predictive factors in patients with metastatic gastric cancer treated with immune checkpoint inhibitors

Introduction. Microsatellite instability, PD-L1 CPS expression, high tumor mutational burden (TMB), and the presence of Epstein-Barr virus are the main tumor predictors of the response to immunotherapy in patients with metastatic gastric cancer (mGC). However, selecting patients for immunotherapy in mGC seems challenging due the lack of an optimal cut-off for PD-L1 CPS expression in microsatellite-stable gastric adenocarcinomas, significant benefit from anti-PD-L1 inhibitors in late-line treatment, and inaccessibility of Epstein-Barr virus and TMB determination in real clinical practice.

Aim. The aim of our study is to determine prognostic and predictive biomarkers of patients, who received ICIs for mGC.

Materials and methods. Our study included patients with mGC treated with anti-PD1 antibodies between 2018 and 2023 in five oncology centers in Moscow. Variables with p <0.05 obtained from a univariate analysis, were selected to perform multivariate analysis. According to the number of prognostic factors, patients were stratified into two groups with favorable and unfavorable prognosis. The optimal cut-off of the neutrophil-lymphocyte ratio (NLR) to predict of the efficacy of immunotherapy was determined using ROC analysis. The Kaplan–Meier method was performed to analyze survival curves of patients according to prognostic groups and NLR levels and the log-rank-test was used to compare the differences. Statistics was performed using the IBM SPSS v. 22 and PRISM 10.

Results. Between January 1, 2018 and February 28, 2023, 122 patients with mGC who received ICIs were included. NLR was analyzed in 71 (58 %) patients out of 122. The median NLR was 2.36 (0.41–10.00). The cut-off of NLR for predicting mortality was 1.8 (AUC 0.81, p <0.001). The median of PFS and OS in patients with high NLR (NLR ≥1.8) were 2 and 4 months, respectively; mOS and mPFS in the low NLR group were not achieved (p <0.001). Eight factors were statistically significant in univariate analysis of patients with MSS: ECOG status (0–1 and 2–3), signet-ring cell histology, primary tumor, the number of organs with metastases (1–2 and 3 or more), ascites, pain, the line of immunotherapy (I–II and III–IV) and N LR level. Multivariate analyses revealed the presence of ascites (p = 0.001), immunotherapy administration in III– IV lines (p = 0.02), and NLR≥1.8 (p = 0.004) were independent prognostic factors for OS. Each factor was assigned with a score from 1 to 2, depending on its significance: presence of ascites – 2 points, high NLR – 2 points, III–IV line of immunotherapy – 1 point. Patients were stratified into two prognostic groups according to the number of prognostic factors – the group with favorable prognosis (0–2 points, n = 20) and unfavorable prognosis (3–5 points, n = 22). The mOS of patients with favorable and unfavorable prognosis was 6 months and 3 months, respectively (p = 0.048).

Conclusion. Ascites, NLR level of ≥1.8 and administration of ICIs in late setting are associated with low efficacy of immunotherapy in patients with MSS mGC. Further research should be planned including more patients and those who did not receive ICIs to determine the prognostic significance of our model.