Enhancing Myocardial Ischemia Evaluation in CZT SPECT PerfusionImaging Using Neural Network

Introduction: The medical centre in Kaohsiung, Chang Gung Medical Centre, collects CZT SPECT perfusion myocardial pictures, which are analyzed in this study. Methods: The present research uses convolutional neural network methods to classify perfusion of the myocardium pictures for Coronary Artery Disease (CAD). The most significant characteristics in these grayscale images are related to the distribution of blood flow in the heart. Consequently, the development of data-driven preprocessing aims to gather information about the relevant areas. Results: The patient's classification of coronary heart disease is performed using a 3-D CNN model after background noise is removed. They are 81.58%, 92.16%, and 87.64% for prediction accuracy, sensitivity, and specificity, respectively. Discussion: The working prototype will significantly reduce the time required for physicians to write reports and interpret images. It can help medical professionals correctly diagnose cardiovascular conditions in real-world situations. Conclusion: Diagnosing myocardial conditions efficiently using CZT SPECT images with CNN approaches is a unique application of the model.