ABSTRACT
Although programmed cell death-ligand 1 (PD-L1) is a prognostic biomarker for nivolumab therapy, it is not very reliable due to its low accuracy. The pharmacological effect of nivolumab involves the cancer immunity cycle, a process that involves T cells, which are strongly associated with nutrition. In this study, we retrospectively investigated whether two measures of nutrition, namely, the Glasgow Prognostic Score (GPS) and the Prognostic Nutritional Index (PNI), could predict response to nivolumab as evaluated in terms of survival time. The subjects were 37 patients treated with nivolumab in the Department of Respiratory Medicine at our hospital between January 2017 and December 2018. Patients were classified into 2 PNI risk groups (low and high risk) and 3 GPS groups (0, 1, and 2), with lower GPS indicating better nutrition. Kaplan-Meier analysis was used to compare differences between the PNI groups and between each possible pairing of GPS groups. Survival time was significantly longer for the low-risk PNI group compared with the high-risk PNI group and for a GPS score of 0 versus 2 and 1 versus 2, but there was no significant difference for a GPS score of 0 versus 1. These results show that GPS and PNI may be potential predictors of response to nivolumab in non-small cell lung cancer.