AUTHOR=Tang Ling , Shen Chengchao 

TITLE=Multimodal AI-driven object detection with uncertainty quantification for cardiovascular risk assessment in autistic patients

JOURNAL=Frontiers in Cardiovascular Medicine

VOLUME=Volume 12 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/cardiovascular-medicine/articles/10.3389/fcvm.2025.1606159

DOI=10.3389/fcvm.2025.1606159

ISSN=2297-055X

ABSTRACT=IntroductionArtificial Intelligence (AI) has transformed medical diagnostics, offering enhanced precision and efficiency in detecting cardiovascular risks. However, traditional diagnostic approaches for cardiovascular risk assessment in autistic patients remain limited due to the complexity of medical data, inter-individual variability, and the challenges of integrating multi-modal clinical information. Conventional methods, relying heavily on manually extracted features and rule-based analysis, often fail to capture subtle cardiovascular abnormalities, leading to suboptimal clinical outcomes.MethodsTo address these limitations, we propose an AI-driven object detection framework that leverages advanced deep learning techniques for automated, accurate cardiovascular risk assessment in autistic patients. Our approach integrates multi-modal medical data, including imaging and electronic health records, through a novel feature fusion mechanism, enhancing diagnostic precision. Furthermore, an uncertainty quantification module is embedded to improve model interpretability and reliability, addressing concerns regarding AI-based medical decision-making.ResultsExperimental evaluations demonstrate that our method significantly outperforms traditional diagnostic techniques in sensitivity and specificity, making it a robust tool for clinical applications.DiscussionThe proposed framework represents a significant step towards personalized and data-driven cardiovascular care for autistic patients, aligning with the need for tailored diagnostic solutions in this specialized medical domain.