Background:
Sarcopenia, defined by the progressive loss of skeletal muscle mass and function, poses significant risks for falls, disability, metabolic dysfunction, and mortality in older adults. Current clinical diagnostics rely on static measures of muscle strength or mass, often missing early-stage or subclinical decline. Moreover, conventional interventions, such as resistance training and increased protein intake, show high inter-individual variability in outcomes due to factors like baseline muscle phenotype, metabolic status, genetics, and gut microbiome composition. Emerging technologies, including wearable sensors, high-throughput metabolic profiling, and AI/ML approaches, provide an opportunity to create predictive, individualized frameworks for sarcopenia risk assessment and management.
Objectives:
* Develop and validate an AI-driven model integrating muscle composition, functional performance, and metabolic biomarkers to predict sarcopenia risk.
* Implement a personalized, adaptive intervention combining exercise and nutrition, guided by AI predictions and real-time monitoring.
* Evaluate the effectiveness of this intervention on muscle mass, functional performance, and metabolic health in older adults.
Methods:
Phase 1: Retrospective analysis of multimodal data from 3,500 adults, including muscle composition (DXA, MRI), functional tests (grip strength, chair rise), metabolic markers, and microbiome profiles. AI/ML models will be trained to predict sarcopenia risk and identify key predictive features. Validation will occur using a subset of newly recruited participants under standard care.
Phase 2: A 12-week prospective intervention in 120 adults aged 50-70, stratified into sarcopenia risk groups based on Phase 1 predictions. Participants will receive AI-guided personalized exercise (resistance and aerobic) and nutrition plans, monitored via wearable sensors and a mobile app. Data collection includes MRI and DXA for muscle composition, functional performance tests, metabolic and inflammatory biomarkers, microbiome profiling, and self-reported outcomes. Intervention response will be analyzed using mixed-effects models and ML to identify predictors of efficacy.
Significance and Innovation:
This study integrates AI-driven risk prediction with personalized, real-time adaptive interventions, addressing current diagnostic and therapeutic gaps in sarcopenia care. By combining muscle structure, function, metabolic, behavioral, and microbiome data, it enables early detection of muscle decline, individualized management, and improved adherence. The framework has potential for broad clinical translation, digital health integration, and future commercialization as a scalable AI-based sarcopenia platform.
Anticipated Outcomes:
* AI-based sarcopenia screening tools for early detection and risk stratification.
* Personalized exercise and nutrition protocols tailored to individual risk and physiology.
* A scalable, data-driven intervention framework suitable for clinical or home-based deployment.
Enhanced understanding of heterogeneous responses to sarcopenia interventions.
