Combining Metabolomics and Machine Learning to Identify Diagnostic and Prognostic Biomarkers in Patients with Non-Small Cell Lung Cancer Pre- and Post-Radiation Therapy

Identifier: imarina:9380970

Handle: https://hdl.handle.net/20.500.11797/imarina9380970

Authors:
Murcia-Mejia, MauricioCanela-Capdevila, MartaGarcia-Pablo, RaquelJimenez-Franco, AndreaJimenez-Aguilar, Juan ManuelBadia, JoanBenavides-Villarreal, RocioAcosta, Johana CArguis, MonicaOnoiu, Alina-IulianaCastane, HelenaCamps, JordiArenas, MeritxellJoven, Jorge

Abstract:
Lung cancer is the leading cause of cancer-related deaths globally, with non-small cell lung cancer (NSCLC) accounting for over 85% of cases and poor prognosis in advanced stages. This study explored shifts in circulating metabolite levels in NSCLC patients versus healthy controls and examined the effects of conventionally fractionated radiation therapy (CFRT) and stereotactic body radiation therapy (SBRT). We enrolled 91 NSCLC patients (38 CFRT and 53 SBRT) and 40 healthy controls. Plasma metabolite levels were assessed using semi-targeted metabolomics, revealing 32 elevated and 18 reduced metabolites in patients. Key discriminatory metabolites included ethylmalonic acid, maltose, 3-phosphoglyceric acid, taurine, glutamic acid, glycocolic acid, and d-arabinose, with a combined Receiver Operating Characteristics curve indicating perfect discrimination between patients and controls. CFRT and SBRT affected different metabolites, but both changes suggested a partial normalization of energy and amino acid metabolism pathways. In conclusion, metabolomics identified distinct metabolic signatures in NSCLC patients with potential as diagnostic biomarkers. The differing metabolic responses to CFRT and SBRT reflect their unique therapeutic impacts, underscoring the utility of this technique in enhancing NSCLC diagnosis and treatment monitoring.