To develop a deep learning model that predicts age from echocardiographic data to assess cardiovascular health and biological aging, highlighting the importance of distinguishing biological aging.
Key Findings:
The model achieved a mean absolute error (MAE) of 6.76 years (range: 6.65–6.87) and a coefficient of determination (R²) of 0.732 (range: 0.72–0.74) on the internal test set.
Predicted age was associated with increased risk of coronary artery disease, heart failure, and stroke.
The model highlighted key echocardiographic features, particularly focusing on the mitral valve and basal inferior wall.
Interpretation:
The findings suggest that AI-based age prediction from echocardiographic data can serve as a valuable tool for assessing cardiovascular health, identifying biological aging, and potentially guiding clinical interventions.
Limitations:
The model's training excluded individuals with prior cardiac surgery, which may limit generalizability to surgical patients.
Variations in performance across different external validation cohorts indicate potential biases that need to be addressed.
Conclusion:
AI-driven age estimation from echocardiographic data holds promise for enhancing cardiovascular risk assessment and understanding biological aging.
by Meenal Rawlani, Hirotaka Ieki, Christina Binder, Victoria Yuan, I-Min Chiu, Ankeet Bhatt, Joseph E. Ebinger, Yuki Sahashi, Andrew P. Ambrosy, Hiroki Usuku, Kenichi Tsujita, Paul Cheng, Alan C. Kwan, Susan Cheng, David Ouyang
