Immunotherapy for Esophageal and Gastric Cancer.

PD-L1 upregulation occurs in approximately 40% of gastroesophageal cancers. However, unlike other solid tumors, there is minimal PD-L1 expressed on the cancer cells; rather, expression occurs predominantly on infiltrating myeloid cells. Preliminary clinical data involving single-agent PD-1/PD-L1 inhibitors in metastatic gastroesophageal cancer have reported response rates of 22%-27% for patients with PD-L1+ tumors and 10%-17% for unselected patients. The phase III ONO-4538-12 (ATTRACTION 2) trial has demonstrated an improved overall survival for nivolumab compared with placebo for patients with heavily pretreated gastric cancer. In the future, we will need better biomarkers to select those most likely to respond and/or identify patients who may need combination immunotherapeutics or alternate strategies. A number of subsets of gastric cancer with different immune signatures, most notably tumors positive for Epstein-Barr virus and microsatellite instability, have been identified, with approximately 50% and 94% PD-L1+ staining seen on tumor cells and immune cells in the EBV subtype and approximately 33% and 45% PD-L1+ staining seen on tumor cells and immune cells in MSI high tumors. Both subtypes demonstrate PD-L1+ immune cells with tumor-infiltrating patterns, unlike the more commonly seen PD-L1+ immune cells at the invasive margin. PD-L2 expression has been reported in 52% of esophageal adenocarcinomas but little is known about the expression of other immune checkpoints. Additional factors that suggest gastroesophageal cancers may respond to checkpoint inhibition include the high somatic mutation burden and the link with chronic inflammation. Here we provide a comprehensive review of the checkpoint inhibitor data published to date in advanced esophagogastric cancers and rationalize how the immune microenvironment in these diverse tumors can explain response or resistance to immunotherapeutics.