BioFactors, Год журнала: 2024, Номер unknown

Опубликована: Окт. 11, 2024

Abstract The proliferation, metastasis, and drug resistance of cancer cells pose significant challenges to the treatment lung squamous cell carcinoma (LUSC). However, there is a lack optimal predictive models that can accurately forecast patient prognosis guide selection targeted therapies. extensive multi‐omic data obtained from multi‐level molecular biology provides unique perspective for understanding underlying biological characteristics cancer, offering potential prognostic indicators sensitivity biomarkers LUSC patients. We integrated diverse datasets encompassing gene expression, DNA methylation, genomic mutations, clinical patients achieve consensus clustering using suite 10 multi‐omics integration algorithms. Subsequently, we employed commonly used machine learning algorithms, combining them into 101 configurations design an performing model. then explored high‐ low‐risk groups in terms tumor microenvironment response immunotherapy, ultimately validating functional roles model genes through vitro experiments. Through application identified two prognostically relevant subtypes, with CS1 exhibiting more favorable prognosis. constructed subtype‐specific model, signature (LMS) based on seven key hub genes. Compared previously published biomarkers, our LMS score demonstrated superior performance. Patients lower scores had higher overall survival rates better responses immunotherapy. Notably, high group was inclined toward “cold” tumors, characterized by immune suppression exclusion, but drugs like dasatinib may represent promising therapeutic options these also validated SERPINB13 experiments, confirming its role as oncogene influencing progression target. Our research new insights refining classification further optimizing immunotherapy strategies.

Язык: Английский