Model-Based Method and Instrumentation for Noninvasive Identification of Local Ischemic Lesions in the Heart

Abstract

An inverse method for identification of possible local ischemic lesions is presented. Body surface potentials recorded in conditions with and without manifestation of ischemia were used to evaluate changes in time integrals of surface potentials caused by changed repolarization of ischemic myocardial cells. Surface distribution of integral differences together with information on torso volume conductor was used to find an equivalent electric generator with one or two dipoles representing the ischemic lesions.

Compact system for multichannel ECG measurement and evaluation with an intelligent battery powered data acquisition unit that enables simultaneous recording of up to 128 torso leads was developed. Active electrodes and optical connection to the controlling computer were used to achieve high signal quality and patient safety. Software for ECG acquisition and evaluation of isointegral surface maps was applied for location of one or two dipoles representing myocardium areas with changed repolarization.

Body surface potential maps from 64 leads were recorded in 10 patients at rest and during exercise on supine ergometer. Difference QRST integral maps computed by subtracting integral map at rest from the maps during exercise and common inhomogeneous torso model were used to find 2-dipoles generators that best represented the difference maps. In 8 patients the found dipoles were in agreement with SPECT: in 3 of them 2 distinct dipoles that represented 2 local lesions or borders of a large transmural lesion were found, for another 3 patients one dipole was identified and in 2 cases no changes representing local lesions were found. One patient was not evaluated because his map could not be satisfactorily represented by the dipolar generator and in 1 case location of the dipole was not confirmed by the SPECT image.

These results suggest that the method and device might be useful tools for non-invasive identification of local ischemia.