Identifying risk factors for sarcopenia using machine learning: insights from multimodal data

Abstract

Aim: To identify the most relevant multimodal risk factors for sarcopenia and evaluate how SARC-F and SARC–CalF influence machine learning model performance and feature prioritization. Findings: Functional measures (chair stand, gait speed), nutritional indicators (protein, folate, copper, vitamin B7), clinical conditions (diabetes, comorbidities, low-density lipoprotein), anthropometric markers (body mass index, calf circumference), and genetic markers (MTHFR polymorphism, Sarcopenia Risk Genotype (SRG)) were among the strongest predictors of sarcopenia across models. Message: Machine learning approaches can uncover complex risk factor interactions and enhance sarcopenia screening by integrating functional, nutritional, clinical, and genetic data beyond traditional tools alone.