Abstract
Building personalized cardiac electrophysiology digital twins requires identifying the appropriate model structure for each patient, not merely fitting parameters. Traditional methods rely on experts to manually design hybrid physics-neural architectures, which requires deep domain expertise and does not transfer across patients. Recent works have applied large language models (LLMs) to generate as hybrid models. However, despite their promising generalization capacity, these LLMbased methods lack the structural priors needed for stable cardiac simulations. Hence, we propose LEADS, a framework that formulates cardiac EP domain knowledge as a structured action space and utilizes an LLM agenttodiscoverhybridmodels.Theagentfollowsaniterativereasoningand-actionloop to select, combine, and refine hybrid models, whilst gradient descent handles parameter fitting. The proposed LEADS designs every candidate model towards physically grounded, interpretable, and numerically stable, while leaving room for open-ended architectural edits beyond the catalog. We validate LEADS on synthetic data with three ground-truth reaction models and on real experimental cardiac EP data, where it outperforms human-designed hybrid models on synthetic data and remains competitive on real data, while an unconstrained LLMbased baseline fails to produce cardiac activation in either setting.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/DT4H_007.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=PWlvsCqZjp
BibTex
@InProceedings{ZhoZiq_Learning_MICCAISAT2026,
author = { Zhou, Ziqi AND Ye, Yubo AND Vadhavka, Sumeet Atul AND Wang, Linwei AND Tao, Zhiqiang},
title = { { Learning Cardiac Electrophysiology Digital Twins Through Agentic Discovery of Hybrid Structure } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17275},
month = {pending},
page = {pending}
}