A robust and versatile method for the quantification of myocardial infarct size

Abstract

In order to determine the size of myocardial perfusion infarcts in clinical SPECT images, we have developed a software package iQuant that eliminates the need for a normal-heart database - a problem that exists in commercial products. Details of the iQuant method are presented in this paper together with results from preliminary tests of accuracy and reproducibility. iQuant uses different count thresholds to define viable myocardial tissue and to provide a rough outline of the complete myocardium. This outline together with the visible thickness of the viable myocardium is used by the operator as a guide to determine the location and size of infarcted tissue. Using noiseless data iQuant is shown to be accurate to within 5%; with more clinically realistic simulated data the measured accuracy is 14%. When testing reproducibility, independent operators consistently produced results within 10% of the truth. Clinical experience was shown to be an important factor in the reproducible accuracy of each individual observer. Measurements to date only involve infarcts considered clinically small, but further work is planned to investigate the reliability of this method over a large range of infarct sizes and locations. Initial analysis shows the iQuant software to be an objective, robust and versatile tool suitable for the determination of myocardial perfusion infarct size in the clinical research environment. Its elimination of the need for a normal-heart database is its main advantages over current, commercially availably software.