Inverse Electrocardiographic Source Localization of Ischemia: An Optimization Framework and Finite Element Solution

Area of Science:

• Computational biology
• Biomedical engineering
• Medical physics

Background:

• Cardiac ischemic disease diagnosis often relies on invasive methods.
• Non-invasive localization of cardiac ischemia is a significant clinical challenge.
• Body-surface potential recordings offer a non-invasive data source.

Purpose of the Study:

• To reconstruct transmembrane potential (TMP) throughout the myocardium using the bidomain heart model.
• To develop a robust method for non-invasively localizing cardiac ischemic disease.
• To solve the inverse source problem within a physically-constrained optimization framework.

Main Methods:

• Utilized a partial differential equation (PDE)-constrained optimization framework for inverse problem solving.
• Formulated optimality conditions in the continuum before finite element discretization.
• Employed L2-norm Tikhonov regularization and total variation minimization.
• Applied a primal-dual interior-point method for numerical optimization.
• Simulations used detailed, fiber-included heart models with up to 18,000 nodes.

Main Results:

• Localized synthetic ischemic regions with low false-negative (10%) and false-positive (20%) rates under 5% noise.
• Reconstructed TMP from animal experimental data with a 0.9 correlation to ground truth.
• Demonstrated feasibility of TMP reconstruction during the ST interval for ischemia localization.

Conclusions:

• The PDE-constrained optimization approach is effective for solving the cardiac inverse source problem.
• This method enables non-invasive localization of cardiac ischemic disease using body-surface potential recordings.
• Accurate TMP estimation remains challenging, but feasibility for ischemia localization is established.