MOSCARD -- Causal Reasoning and De-confounding for Multimodal Opportunistic Screening of Cardiovascular Adverse Events

2025-06-23

AI Summary

This study addresses inaccurate risk assessment in opportunistic screening for major adverse cardiovascular events (MACE), caused by limitations of single-modality data and sampling bias. We propose a multimodal causal modeling framework integrating chest X-ray (CXR) and 12-lead electrocardiogram (ECG) data. Methodologically, we innovatively combine modality-aligned representation learning, cross-modal co-attention, and a causal inference network, augmented by a dual backward-propagation deconfounding graph to mitigate confounding bias and enhance model interpretability and generalizability. Evaluated on a real-world emergency department dataset and the external MIMIC validation set, our framework achieves AUCs of 0.83 and 0.71, respectively—significantly outperforming unimodal approaches and state-of-the-art baselines. The framework enables low-cost, non-invasive, early identification of MACE risk, supporting clinical decision-making and promoting health equity.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Reasoning under Uncertainty: Causality

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Major Adverse Cardiovascular Events (MACE) remain the leading cause of mortality globally, as reported in the Global Disease Burden Study 2021. Opportunistic screening leverages data collected from routine health check-ups and multimodal data can play a key role to identify at-risk individuals. Chest X-rays (CXR) provide insights into chronic conditions contributing to major adverse cardiovascular events (MACE), while 12-lead electrocardiogram (ECG) directly assesses cardiac electrical activity and structural abnormalities. Integrating CXR and ECG could offer a more comprehensive risk assessment than conventional models, which rely on clinical scores, computed tomography (CT) measurements, or biomarkers, which may be limited by sampling bias and single modality constraints. We propose a novel predictive modeling framework - MOSCARD, multimodal causal reasoning with co-attention to align two distinct modalities and simultaneously mitigate bias and confounders in opportunistic risk estimation. Primary technical contributions are - (i) multimodal alignment of CXR with ECG guidance; (ii) integration of causal reasoning; (iii) dual back-propagation graph for de-confounding. Evaluated on internal, shift data from emergency department (ED) and external MIMIC datasets, our model outperformed single modality and state-of-the-art foundational models - AUC: 0.75, 0.83, 0.71 respectively. Proposed cost-effective opportunistic screening enables early intervention, improving patient outcomes and reducing disparities.

Problem

• Develops multimodal causal model for cardiovascular risk prediction
• Integrates chest X-rays and ECG to reduce screening bias
• Improves early detection of major adverse cardiovascular events

Innovation

• Multimodal alignment of CXR with ECG guidance
• Integration of causal reasoning framework
• Dual back-propagation graph for de-confounding