Effects of Inferior Myocardial Infarction Sizes and Sites on Simulated Electrocardiograms Based on a Torso-Heart Model

Chengyu Liu, Zhipeng Cai, Jianqing Li, Kan Luo, Bor-Shyh Lin, Zhigang Wang, Xiangyu Zhang, Chengyu Liu 0001IEEE Access 2019 · DOI

Introduction

Myocardial infarction (MI), or heart attack, is a critical global health challenge. The standard 12-lead electrocardiogram (ECG) is the primary diagnostic tool, yet interpreting how different MI sizes and locations manifest in ECG waveforms remains complex. Computational modeling offers a way to bridge clinical observation with physiological theory. However, existing models often struggle to balance anatomical accuracy with computational feasibility. This study presents a refined 3D torso-heart model designed to simulate inferior MIs of varying sizes and locations, providing a quantitative framework for optimizing ECG-based diagnosis.

Methodology

The researchers developed a refined 3D "bidomain" model encompassing the heart, lungs, and torso as volume conductors. The electrical activity of the heart was simulated using modified FitzHugh-Nagumo equations, while the spread of electrical potential to the body surface was calculated via the Laplace formulation of Maxwell’s equations.

Heart model geometry showing myocardial layers and simulated MI regions.

The study specifically modeled nine inferior MI scenarios, combining three anatomical sites (basal, middle, and apical) with three sizes (small, moderate, and large). To capture a comprehensive view of the electrical changes, the model included the standard 12 leads plus an additional 308 measuring points across the torso to generate Body Surface Potential Maps (BSPMs). The simulation is computationally efficient, with a one-second cardiac cycle taking approximately five hours on a standard high-end workstation.

The 3D torso model illustrating lead placement and measuring points.

A key innovation in this work is the Normalized Variation Coefficient (NVC), a metric used to quantify the impact of an MI on specific ECG components (Q, R, S, ST, and T waves) relative to a healthy baseline.

Key Insights and Results

The simulation results were validated against clinical data from the 2007 PhysioNet/Computers in Cardiology Challenge, achieving a Mean Structural Similarity (MSSIM) index of 0.739. This high correlation suggests the model accurately reflects human cardiac physiology.

Lead Sensitivity and the aVF Lead

The study mathematically confirmed that the augmented unipolar left lower limb lead (aVF) is the most sensitive indicator for inferior MIs. Lead aVF exhibited the highest average NVC (0.56) across all tested scenarios. Notably, the Q-wave in lead aVF was exceptionally responsive to basal-site infarctions, reaching an NVC of 1.00.

Complementary Lead Analysis

Analysis of the torso maps revealed that leads V4 and V6 provide nearly complementary information. This finding suggests that monitoring these leads together can significantly improve the ability to triangulate both the size and specific location of myocardial damage.

Detection of Small Infarcts

While large transmural MIs are typically obvious on an ECG, the model demonstrated that even small infarcts localized to the inner endocardium produce detectable changes, particularly in lead V5. This underscores the potential for high-density ECG systems to improve early-stage MI detection.

Conclusion and Future Directions

This research provides a validated, computationally efficient framework for studying the relationship between MI pathology and ECG signals. By quantifying lead sensitivity through the NVC metric, the study offers actionable insights for clinical diagnosis.

Future work aims to enhance the model by incorporating irregular MI geometries derived from CT or MRI scans and accounting for the anisotropic nature of heart tissue, where electrical conductivity varies by direction. These refinements will further improve the precision of simulated ECGs in predicting real-world clinical outcomes.

Summary updated

BibTeX
@article{liu2019effects,
  author = {Chengyu Liu and Zhipeng Cai and Jianqing Li and Kan Luo and Bor-Shyh Lin and Zhigang Wang and Xiangyu Zhang and Chengyu Liu},
  title = {{Effects of Inferior Myocardial Infarction Sizes and Sites on Simulated Electrocardiograms Based on a Torso-Heart Model}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2904707},
}
Figure