Abstract

2564 Background: Immune checkpoint blockade (ICB) has changed the treatment landscape of non-small cell lung cancer (NSCLC). Despite being the mainstay of treatment for advanced NSCLC, the development of resistance to ICB is common. Hence, identifying biomarkers that can be used to select patients (pts) who will benefit from ICB is crucial. Here, we interrogate the circulating metabolome to identify the metabolic parameters associated with ICB effectiveness. Methods: This single-center retrospective analysis used a mass-spectrometry (MS) targeted metabolomics approach to assess the relative abundance of 115 metabolites in baseline (pre-ICB) plasma of 55 pts with advanced NSCLC. Pts treated with ICB at the Princess Margaret Cancer Centre from 2018-2020 were included. Electronic medical records were reviewed to collect clinicopathological and treatment data. All pts had tumour next-generation sequencing (NGS) by a targeted gene panel. Descriptive statistics were used to summarize the patient characteristics, treatment modalities, and outcomes. Time-to-event outcomes were analyzed using the Kaplan-Meier method. Progression-free survival (rwPFS) and overall survival (rwOS) were defined as the time from the first treatment dose to radiographic or clinical progression or death from any cause and time from the first dose of treatment to death from any cause, respectively. The response rate was assessed using RECIST 1.1. Statistical significance was determined as a p-value <0.05. Results: Amongst the 55 pts profiled, 42 (76.3%) had adenocarcinoma. The most common molecular alterations included KRAS (n=15), BRAF non-V600 (n=5), and METex14 (n=3). The median PD-L1 score was 65% (IQR 22.5 – 59%). All pts were treated with a single-agent PD-1 inhibitor, and 67.2% of pts received ICB as the first line of treatment. Metabolomic analysis of the pre-ICB initiation plasma samples identified 8 metabolites whose relative abundance significantly differed between responding (CR/PR) and non-responding pts (SD/PD). Amongst these metabolites was the endogenous danger signal Glucosylceramide (d 18:1/24:0), whose abundance was increased in the plasma of the responding pts. When we stratified pts by circulating levels of glucosylceramide, we found a statistically significant higher rwPFS (8.5 vs. 1.6 months, p=0.0001) and rwOS (13.5 vs. 6.1 months, p=0.0001) in pts who had high baseline plasma levels of Glucosylceramide (d 18:1/24:0) (top 50%) versus those with lower levels (bottom 50%). Conclusions: Utilizing a targeted metabolomics approach, we identified the metabolite and endogenous danger signal glucosylceramide (d 18:1/24:0) as a potential metabolic biomarker of response to ICB therapy. Future work will aim to validate this finding in a larger cohort and to understand the biological mechanism underpinning this correlation.