AI Creates Multi-Organ Aging Clocks Using MRI Scans to Predict Disease and Mortality
Researchers developed 7 organ-specific aging clocks from MRI data, revealing molecular signatures of aging across brain, heart, liver, and other organs.
Summary
Scientists developed seven organ-specific biological aging clocks using MRI scans from over 313,000 people. These AI-powered clocks measure how fast different organs age compared to chronological age, revealing unique molecular signatures for brain, heart, liver, adipose tissue, spleen, kidney, and pancreas aging. The clocks successfully predicted disease risk, mortality, and cognitive decline, with some organs showing stronger aging signals than others. This multi-organ approach provides a comprehensive view of biological aging beyond traditional single-organ assessments.
Detailed Summary
This groundbreaking study represents the first systematic expansion of MRI-based aging clocks beyond the brain to include six additional organs, creating a comprehensive multi-organ aging assessment framework. Using data from 313,645 individuals across multiple cohorts, researchers developed AI-powered biological age gaps (MRIBAGs) for seven organs: brain, heart, liver, adipose tissue, spleen, kidney, and pancreas.
The research team employed machine learning algorithms to analyze organ-specific MRI features and predict biological versus chronological age. Performance varied significantly across organs, with brain aging clocks showing the strongest predictive power (correlation r=0.77) while abdominal organs like spleen showed weaker performance (r=0.23), partly due to limited imaging features and technical challenges.
Through comprehensive molecular profiling, the study linked these aging clocks to 2,923 plasma proteins and 327 metabolites, revealing organ-specific aging signatures. For example, kidney aging strongly associated with 301 proteins including NPDC1 and IGFBP6, while pancreas aging linked to digestive enzymes like PLA2G1B. Genetic analysis identified 53 significant genetic variants associated with organ aging and pinpointed 9 potentially druggable genes for anti-aging interventions.
The clinical validation demonstrated these aging clocks' ability to predict disease outcomes, all-cause mortality, and differential responses to Alzheimer's treatments over 240 weeks. Notably, the study revealed significant sex differences in aging patterns across multiple organ systems, manifesting at structural, molecular, and genetic levels.
This work establishes a new paradigm for aging research by providing a holistic, system-wide view of biological aging that could revolutionize personalized medicine approaches to age-related diseases and longevity interventions.
Key Findings
- Seven organ-specific MRI aging clocks developed with varying accuracy (brain r=0.77, spleen r=0.23)
- Kidney aging linked to 301 proteins, revealing strongest molecular aging signature
- 53 genetic variants and 9 druggable genes identified for potential anti-aging treatments
- Aging clocks successfully predicted disease risk, mortality, and Alzheimer's treatment response
- Significant sex differences found in aging patterns across multiple organ systems
Methodology
Cross-sectional study using machine learning (Lasso regression, support vector regression) on MRI data from 313,645 individuals across UK Biobank, BLSA, and A4 cohorts. Age prediction models validated using nested cross-validation with independent test datasets of 500 participants per organ.
Study Limitations
Abdominal organ clocks showed poor performance due to limited imaging features and high collinearity. Cross-sectional design limits causal inference. Domain shift issues when applying models to external datasets may affect generalizability across different populations and imaging protocols.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.