Integrating Bulk and Single-Cell Transcriptomics with Machine Learning Reveals a Heme Metabolism-Based Panel for Lung Adenocarcinoma Chemotherapy Resistance.

Lung adenocarcinoma (LUAD) is a leading cause of cancer-related mortality, with heme metabolism playing a critical role in tumor progression and treatment resistance. This study investigates the clinical implications of heme metabolism in LUAD, focusing on its link to ferroptosis and drug sensitivity. Using multi-omics data from TCGA-LUAD, GEO databases, and a single-cell RNA-seq cohort, we identified two molecular subtypes based on heme metabolism-related genes. We further developed a prognostic panel, termed the heme metabolism risk score (HMRS), using LASSO and multivariate Cox regression analyses. The HMRS panel effectively stratified patients into high- and low-risk groups, with high-risk patients showing enhanced tumor proliferation, suppressed ferroptosis, and resistance to chemotherapy. Single-cell analysis revealed elevated heme metabolism risk in epithelial cells correlated with tumor progression. Drug sensitivity predictions were validated in platinum-based chemotherapy cohorts, confirming HMRS as a robust prognostic tool. ABCC2 was identified as a key regulator of ferroptosis and cisplatin resistance, with in vitro experiments demonstrating that ABCC2 knockdown enhanced cisplatin-induced ferroptosis. These findings highlight HMRS as a critical tool for patient stratification and ABCC2 as a promising therapeutic target to overcome cisplatin resistance.