Obesity, dyslipidemia, and diabetes as determinants of response to immune checkpoint inhibitors in non-small cell lung cancer.

Abstract

e21125 Background: Immune checkpoint inhibitors (IOs) are the standard of care in the treatment of advanced non-small cell lung cancer (NSCLC). Predictive biomarkers of response to IOs are being studied extensively. Obesity has been correlated with a better response to IOs in NSCLC and other types of cancer. This may be related to high circulating leptin or impaired T-cell fatty acid oxidation within the tumor microenvironment. We aim to further explore the effects of other components of the metabolic syndrome (diabetes, dyslipidemia) on treatment outcomes in NSCLC patients receiving IOs. Methods: We performed a retrospective cohort study in patients with NSCLC treated with pembrolizumab, nivolumab, durvalumab, cemiplimab, or atezolizumab as monotherapy or combined with chemotherapy at Mount Auburn Hospital, Beth Israel Deaconess Medical Center, and Lahey Hospital & Medical Center between January 2016 and June 2021. Obesity was defined as body mass index (BMI) > 30 kg/m2; dyslipidemia and type 2 diabetes were defined by ICD-10 diagnosis codes. Overall survival (OS) was evaluated. Kaplan-Meier survival analysis and cox proportional hazards regression were used to assess correlation between OS and obesity, dyslipidemia, type 2 diabetes, metformin use, and statin use. These were adjusted for age, biological sex, cancer stage, and ECOG performance status. Patients with BMI < 18.5 were excluded. Results: We identified 502 patients with NSCLC who were treated with IOs. Of these, 68.5% had adenocarcinoma, 26.7% had squamous, and 4.8% had either adenosquamous, large cell or poorly differentiated histologies. Among these patients, 22% were classified as obese, 45% had dyslipidemia, and 22% had type 2 diabetes. Of these, 66% of patients received IO monotherapy. Obesity was associated with improved OS (adjusted HR 0.70, CI 0.49-0.99, p = 0.042). Conversely, dyslipidemia was associated with worse OS (adjusted HR 1.44, CI 1.11-1.89, p = 0.003). Interestingly, glucose level > 126 mg/dL at the time of IO initiation was also associated with worse OS (HR 1.42, CI 1.10-1.83, p = 0.008). OS was not significantly different between those with type 2 diabetes, metformin use, or statin use. In this cohort, none of these metabolic parameters were significantly associated with immune-related adverse events. Conclusions: Obesity is associated with improved OS in patients with NSCLC treated with IO. Conversely, dyslipidemia and hyperglycemia were associated with worse OS. This represents an important disconnect within the metabolic milieu governing differential responses to IOs. Further studies are needed to explore the molecular mediators that may be unique to obesity and provide a favorable microenvironment for IO efficacy.