Research Questions:What distinct imaging-genetic signatures differentiate high-risk carotid atherosclerosis phenotypes associated with neurological/cardiovascular events? Can AI-driven multimodal integration of ultrasound morphology, genomic data, and clinical variables improve risk stratification beyond conventional paradigms? How do genotype-phenotype interactions underlying plaque vulnerability inform novel therapeutic targets? This study addresses these gaps by synergizing deep learning-based image analysis, GWAS, and network medicine to enable precision risk prediction and intervention.
Objectives:This study aims to: (1) identify high-risk imaging-genetic signatures of carotid atherosclerosis through multimodal integration of ultrasound-derived features and genomic data; (2) develop an artificial intelligence (AI)-driven risk stratification model by synthesizing demographic, clinical, imaging, and multi-omics data; and (3) explore novel therapeutic targets by deciphering the phenotype-genotype interplay underlying carotid plaque vulnerability.
Scientific Rationale: Ultrasonographic metrics, including carotid intima-media thickness (cIMT) and plaque burden, have demonstrated incremental prognostic value for ischemic events beyond traditional risk factors. Concurrently, recent advances in biomedical big data analytics-encompassing high-resolution imaging, genomic profiling, and computational biology-enable systematic interrogation of atherosclerosis pathophysiology. Specifically, AI frameworks exhibit unique potential to decode complex associations between plaque morphology, genetic variations, and clinical outcomes through multimodal data fusion, thereby addressing critical gaps in current risk prediction methodologies. The findings from this project will be disseminated through peer-reviewed scientific publications in open-access journals in accordance with UK Biobank’s policies.
