Nomograms for stratified prognosis prediction of gastric cancer by integrating programmed death ligand 1 and tumor-infiltrating immune cells including CD4+/CD8+ TILs and CD163+ TAMs.

PURPOSE

To develop nomograms for predicting disease-free survival (DFS) and overall survival (OS) of gastric cancer (GC) by integrating programmed death ligand 1 (PD-L1) and CD4+/CD8+ tumor-infiltrating lymphocytes (TILs) and CD163+ tumor-associated macrophages (TAMs).

MATERIALS AND METHODS

Immunohistochemistry for PD-L1, CD4+/CD8+ TILs and CD163+TAMs was performed on 126 surgically-resected GC specimens between January 2016 and May 2018. Subsequently, the expression of PD-L1 and these tumor-infiltrating immune cells(TIICs), in combination with multiple clinicopathologic features, was used to formulate nomograms for predicting DFS or OS based on the results of multivariate Cox regression analysis. The performance of the nomograms for DFS or OS was verified in the 10-fold cross-validation of the study cohort and measured by Harrell's concordance-index (C-index).

RESULTS

After multivariable Cox regression analyses, high PD-L1 expression (hazard ratio[HR]=2.17, 95% confidence interval [CI] 1.37-3.43), low CD8+ TILs density(HR=0.35, 95% CI 0.15-0.81), high CD163+ macrophages density (HR=1.84, 95% CI 1.17-2.89), TNM stage (stage III vs stage I+II, HR=1.37, 95% CI 1.06-2.23) and microsatellite instability-high(MSI-H) ( MSI-H VS microsatellite stability (MSS), HR=0.41, 95% CI 0.20-0.83) were found to be independent risk factors for DFS. Similarly, high PD-L1 expression (HR=2.64, 95% CI 1.61-4.34), high CD4+ TILs density (HR=1.98, 95% CI 1.21-3.24), low CD8+ TILs density (HR=0.23 95% CI 0.07-0.73), high CD163+ TAMs density (HR=2.31, 95% CI 1.43-3.74), MSI-H (MSI-H VS MSS, HR=0.26, 95% CI 0.12-0.60) and TNM stage (stage III vs stage I +II, HR=1.61, 95% CI 1.01-2.56) were independently associated with OS. These actors were then selected to establish nomograms for DFS and OS individually. The established nomogram for DFS yielded a corrected C-index of 0.679 by 10- fold cross-validation. Similarly, the established nomogram for OS yielded a corrected C-index of 0.755.These results suggest that PD-L1 and high density of CD4+ TILsas well as CD163+ TAMs are risk factors for poor prognosis in GC patients.On the contrary, MSI-H and high density of CD8+ TILsare associated with good prognosis in GC patients.

CONCLUSIONS

The developed prognostic nomograms for GC integrating PD-L1 and CD4+/CD8+ TILs as well as CD163+TAMs offer a more personalized and precise prediction of DFS and OS for patients, which can help to improve prognostic stratification.