Developing Trustworthy Artificial Intelligence Models to Predict Vascular Disease Progression: the VASCUL-AID-RETRO...

Area of Science:

• Cardiovascular Medicine
• Artificial Intelligence
• Genomics and Proteomics

Background:

• Abdominal aortic aneurysms (AAA) and peripheral artery disease (PAD) pose significant risks for major adverse cardiovascular events and mortality.
• Current disease management faces challenges due to unpredictable individual patient disease progression.

Purpose of the Study:

• To develop trustworthy multimodal predictive artificial intelligence (AI) models for risk stratification of disease progression and cardiovascular events in patients with AAA and PAD.
• To enhance personalized medicine in vascular surgery through AI-driven predictions.

Main Methods:

• Retrospective collection of multimodal data (clinical, imaging, proteomics, genomics) from 5000 AAA and 6000 PAD patients across European centers (2015-2024).
• Development of AI models using integrated data, considering ethical guidelines and legal standards for trustworthy AI.
• Internal validation of models with prospective data in the VASCUL-AID-PRO study.

Main Results:

• The study will integrate diverse data sources including electronic health records, biobanks, and registries.
• AI models will be trained on segmented artery geometries for hemodynamic parameter estimation and disease progression quantification.
• Risk prediction models will be developed separately for each data modality before combining them into multimodal models.

Conclusions:

• The VASCUL-AID-RETRO study will leverage advanced AI and multimodal data integration to predict AAA and PAD progression and associated cardiovascular events.
• The findings aim to improve clinical practice by enabling more precise, individualized treatment plans for better patient outcomes.