Clinically relevant metabolic subtypes of lung cancer for immunotherapy

The metabolic pathway subtypes (MPSs) showed distinct differences in outcomes from anti-PD-(L)1 treatment. Using metabolic pathway–based clustering, we identified clinically relevant metabolic signatures that may be used to determine the efficacy of immuno-oncology drugs in non-small cell lung cancer (NSCLC). These findings suggest that metabolic features play a critical role in shaping the response to immuno-oncology therapies. To establish these relationships, we integrated individual clinical data, including clinicopathologic and genomic variables, as well as survival information, into the transcriptional analysis. This study was designed as a post hoc analysis of a prospective observational cohort conducted at the National Cancer Center Hospital in Korea between November 2017 and February 2021. Patients with stage IV non-small cell lung cancer (NSCLC) who had measurable lesions and were scheduled to receive anti–PD-(L)1 therapy were prospectively enrolled. Whole blood samples and archival tumor tissues were collected prior to treatment, and tumor responses were assessed by computed tomography or other imaging every 6 to 9 weeks until disease progression. Transcriptomic profiling of tumor tissues was performed using the NanoString Metabolic Pathways Panel, which measures the expression of 768 metabolism-related genes, to investigate metabolic heterogeneity. Unsupervised clustering with non-negative matrix factorization (NMF) was applied to identify distinct metabolic subtypes, followed by differential gene expression and pathway enrichment analyses, including GSVA and GSEA. The reproducibility and robustness of the identified subtypes were validated using an independent cohort from the GSE161537 dataset. Clinical, pathological, and genomic characteristics were integrated with transcriptomic data. Survival analyses using Kaplan–Meier curves and log-rank tests were conducted to compare progression-free survival and overall survival across subtypes, while Chi-squared and Fisher’s exact tests were applied to assess associations with categorical variables. Collectively, this experimental design enabled the identification and validation of clinically relevant metabolic pathway subtypes of NSCLC with potential implications for immunotherapy response.