Physiology and Anatomy Aware Inverse Inference of Myocardial Infarction for Cardiac Digital Twin

Mengxiao Wang, Yilin Lyu, Julia Camps, Ching Hui Sia, Mark Yan-Yee Chan, Yanrui Jin, Shuzhi Sam Ge, Chengliang Liu, Lei Li 🖂 · 2026

Read Physiology and Anatomy Aware Inverse Inference of Myocardial Infarction for Cardiac Digital Twin by Mengxiao Wang, Yilin Lyu, Julia Camps, Ching Hui Sia, Mark Yan-Yee Chan, Yanrui Jin, Shuzhi Sam Ge, Chengliang Liu, Lei Li 🖂 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-22044.txt.

Introduction

Mengxiao Wang and co-authors propose a physiology and anatomy aware method for spotting myocardial infarction, or heart attack damage, for a cardiac digital twin. Accurate infarct location guides risk stratification, yet late gadolinium enhancement MRI, the gold standard, is costly. Their framework combines cine MRI with multi-lead ECG data to spot infarct areas across varied locations, sizes, extents, and depths, while respecting realistic scar shape and cardiac repolarization. It beats existing inverse inference methods, scoring Dice 0.7391 for scar and 0.5503 for border zone segmentation.

Catalog record

CREATORMengxiao Wang, Yilin Lyu, Julia Camps, Ching Hui Sia, Mark Yan-Yee Chan, Yanrui Jin, Shuzhi Sam Ge, Chengliang Liu, Lei Li 🖂
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20980
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.22044

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Physiology and Anatomy Aware Inverse Inference of Myocardial Infarction for Cardiac Digital Twin Thanks: This arXiv version corresponds to the submitted version. The final version will be available on Springer Link. Mengxiao Wang, Yilin Lyu, Julia Camps, Ching Hui Sia, Mark Yan-Yee Chan, Yanrui Jin, Shuzhi Sam Ge, Chengliang Liu, Lei Li 🖂 Abstract Accurate localization of myocardial infarction is essential for risk stratification. While LGE-MRI remains the gold standard, it is resource-intensive. Integrating cine MRI with ECG enables a more detailed representation of infarct properties. Existing inverse MI inference methods overlook realistic scar morphology and cardiac repolarization, reducing sensitivity to subtle ECG variations and interpretability of infarct-induced electrophysiological changes. In this paper, we propose a novel framework for noninvasive MI localization using cardiac digital twins. To bridge the domain gap between simulation and reality, we introduce an anatomy-aware stochastic infarct synthesis strategy to synthesize realistic, irregular scars with border zones, mimicking ischemic transmural progression. We then construct a virtual cohort to simulate QRS-T waveforms, capturing both depolarization and repolarization of electrophysiological dynamics. Furthermore, we design a Physiology and Anatomy Aware Network (PAA-Net) that jointly encodes 3D myocardial geometry and multi-lead ECGs to infer infarct area with various localizations, sizes, spatial extents and transmuralities. Experimental results demonstrate that our framework significantly outperforms existing methods in inverse inference, achieving 0.7391 and 0.5503 in D ​ i ​ c ​ e S ​ c ​ a ​ r Dice_{Scar} and D ​ i ​ c ​ e B ​ Z Dice_{BZ} , respectively, and further enhances interpretability of the ECG–infarct relationship. Our code will be released upon acceptance. 1 Introduction Myocardial infarction (MI) is a major global cardiovascular disease, associated with substantial morbidity and mortality [ 9 ] . Post-infarction structural remodeling, particularly myocardial scar formation, constitutes a critical substrate for life-threatening arrhythmias. Consequently, accurate localization of the infarcted myocardium is essential for effective risk stratification and therapeutic planning [ 22 ] , and in cardiac digital twins, it is formulated as an inverse problem [ 6 ] . In out-of-hospital or resource-limited settings, accurate MI diagnosis and localization remain challenging due to the trade-off between accessibility and diagnostic resolution [ 10 , 24 ] . Although Late Gadolinium Enhancement Magnetic Resonance Imaging (LGE-MRI) is the gold standard for MI assessment, it is costly, time-consuming, and associated with potential safety risks [ 20 ] . Electrocardiography (ECG) is widely accessible and convenient, it offers only limited localization capability and cannot quantify detailed infarct characteristics, such as size, spatial extent, or transmurality [ 2 ] . Consequently, a joint framework that combines MRI and ECG is indispensable for comprehensive MI assessment. Recently, cine MRI, a standard contrast-free modality, has been explored for MI assessment by analyzing cardiac motion [ 23 , 25 ] . However, most studies rely on 2D single-view representations, limiting their ability to capture the complex 3D structure of infarcts. With the rise of cardiac digital twins, precise 3D anatomical and functional modeling [ 12 ] is increasingly needed, motivating the shift from planar imaging to volumetric modeling [ 3 ] . Studies show that 3D motion patterns derived from cine MRI improve MI detection [ 14 ] , and integrating computational cardiac models enables noninvasive identification of infarct location and extent [ 8 ] . Recent work has combined cine MRI with ECG to infer myocardial tissue properties [ 7 ] , and computational electrophysiology helps study infarct effects on simulated ECGs. Nevertheless, these approaches mainly address ventricular depolarization, neglecting T-wave repolarization. Moreover, irregular and stochastic scar synthesis within digital twins should account for real-world characteristics [ 17 ] , yet this aspect remains insufficiently addressed in existing methods. Furthermore, the inverse inference model can be enhanced to capture subtle variations in ECG morphology and to improve the interpretability of how infarcted regions influence the electrical signals. In this work, we propose a novel framework for inverse MI inference based on cardiac digital twins, enhancing both stochastic scar synthesis and the fusion network for infarct localization. To the best of our knowledge, this is the first study to simulate full QRS–T complexes and integrate them into an inverse MI inference framework. The main contributions are summarized as follows: 1. We propose an anatomy-aware stochastic strategy to generate realistic irregular scars with border zones, mimicking ischemic transmural progression and covering diverse locations, sizes, spatial extents and transmuralities. 2. We construct a virtual cohort of QRS-T waveforms, including both depolarization and repolarization phases of the ventricle under various MI scenarios. 3. We develop the Physiology and Anatomy Aware Network (PAA-Net) that jointly models spatial point cloud information and ECG signals for MI inverse inference, and also improves interpretability. 2 Method The overall pipeline comprises two stages as illustrated in Fig. 1 . The first stage generates diverse MI scenarios through stochastic infarct synthesis, modeling variations in transmurality and scar burden while capturing ventricular depolarization and repolarization. The second stage introduces the Physiology and Anatomy Aware Network (PAA-Net) for inverse MI inference, which incorporates both the physiological characteristics of ECG lead positions and the anatomical spatial localization of infarcted regions. Figure 1 : The pipeline of cohort simulation and inverse inference of infarcted area. 2.1 Anatomy-Aware Stochastic Infarct Synthesis We first construct simulation-ready biventricular meshes from multi-view cine MRI. The surface meshes are reconstructed from SAX and LAX segmentation labels following [ 4 ] , and volumetric meshes are then generated using Iso2Mesh [ 19 ] . Then, fields are assigned, including ventricular coordinates via Cobiveco [ 18 ] , as well as rule-based fibre orientations and electrode placements [ 5 ] . For scar synthesis, we considered the realistic scar distribution in Myocardial Infarction. Specifically, we applied clinical knowledge to split the scar area into 7 types [ 13 ] based on the 17-segment according to American Heart Association (AHA) standard. Different from [ 7 ] , we do not consider scars as regular ellipsoid, instead, we developed a stochastic generation framework based on ventricular coordinates. The scar generation process accounts for both the spatial heterogeneity and the transmural progression characteristic of ischemic infarct. First, a spatially correlated noise field N(x) was generated by applying a Gaussian filter to a random uniform distribution. Second, to mimic the transmural ischemic wavefront, the scar core was defined by a dynamic thresholding function dependent on the transmural coordinate T_{m} (where T_{m}=1 at the endocardium): Scar=\{x|N(x)>\tau_{base}+\lambda(1-T_{m}(x))\}, (1) where \tau_{base} controls overall scar volume and \lambda penalizes scar formation towards the epicardium. Finally, the border zone (BZ) was generated morphologically using a K-D tree search, reclassifying all viable nodes within a Euclidean radius r=2\,mm of the scar core as BZ tissue, thereby creating a transitional halo surrounding the dense core. The final synthetic scars contain 17 scenarios covering 7 infarct locations, 2 infarct sizes, and 2 transmurality types. 2.2 Electrophysiological Modeling of QRS-T Waveforms For electrophysiology simulation, we utilize the reaction-Eikonal model [ 11 ] , which combines the Eikonal-type model with an action potential shape model to recover transmembrane voltages. First, local activation times t_{a}(x) are computed by solving the Eikonal equation: \sqrt{\nabla t_{a}(x)^{T}\mathbf{V}\nabla t_{a}(x)}=1,\quad x\in\Omega (2) subject to t_{a}(y_{i})=t_{i} for N_{root} early activation sites, where \mathbf{V} represents the orthotropic conduction velocity tensor aligned with

Read the complete text