Multi-cohort validation of a lipid metabolism and ferroptosis-associated index for prognosis and immunotherapy response prediction in hormone receptor-positive breast cancer.

Abstract

Background: Hormone receptor-positive (HR+) breast cancer exhibits significant heterogeneity influenced by lipid metabolism and ferroptosis (LMF). While immune checkpoint inhibitors have shown promise in neoadjuvant therapy, as evidenced by the KEYNOTE-756 and CheckMate 7FL trials, identifying the optimal patient population remains challenging. This study aims to classify molecular clusters based on LMF-related genes and develop the LMF_index to predict prognosis and immunotherapy response in HR+ breast cancer. Methods: Transcriptome and clinical data of HR+ breast cancer were obtained from the Cancer Genome Atlas and Gene Expression Omnibus databases. Unsupervised clustering based on prognostic LMF-related genes identified molecular clusters, followed by tumor mutational burden (TMB) and immune microenvironment (TME) analysis. The LMF_index was constructed using least absolute shrinkage and selection operator and multivariate Cox regression analyses and validated across multiple internal and external cohorts. Its predictive value for neoadjuvant immunotherapy efficacy was assessed using GSE173839. Validation at the transcriptomic level was conducted in the Shanghai cohort, while protein-level validation was performed using multiplex immunohistochemistry (mIHC) on a tissue microarray comprising 113 breast cancer samples. Spatial analyses further examined the distribution of key panel genes within the TME. Results: Two molecular clusters were identified in this study. Cluster 1 exhibited higher TMB, tumor purity, and Ki-67, while Cluster 2 showed greater CD8+ T cells and elevated PD-1, PD-L1, and CTLA4 expression. The LMF_index, derived from a seven-gene panel (KRT5, CD209, KLRB1, MRC1, UGT2B4, FABP7, and BIRC3), effectively stratified patients into high and low LMF_index groups, with high LMF_index patients showing significantly shorter overall survival. Patients with a low LMF_index demonstrated elevated ACSL4 expression, enhanced immune activity, higher immunophenoscores, and increased pathological complete response rates following neoadjuvant immunotherapy, indicating a greater potential benefit from immunotherapy. The prognostic value of the LMF_index was validated at the transcriptomic level in the Shanghai cohort and at the protein level using mIHC on a tissue microarray. Spatial analysis further demonstrated KLRB1 enrichment in the tumor stroma, correlating with CD8+ T cell and M1 macrophage infiltration, and an enhanced response to immunotherapy. Conclusions: This study identified distinct LMF-related molecular clusters in HR+ breast cancer with unique prognostic and immune characteristics. The LMF_index shows potential as a prognostic biomarker and a guide for immunotherapy strategies in HR+ breast cancer.