MyI-Net: Fully Automatic Detection and Quantification of Myocardial Infarction from Cardiovascular MRI Images

TL;DR

This paper introduces MyI-Net, an automated deep learning system for detecting and quantifying myocardial infarction in MRI images, aiming to improve accuracy and consistency over manual methods.

Contribution

It presents a fully automated, end-to-end deep learning approach combining ResNet, MobileNet, and ASPP for MI detection and quantification in MRI images, outperforming existing models.

Findings

Four-fold improvement in scar pixel matching accuracy

Favorable performance in global segmentation tasks

Outperforms state-of-the-art models and manual quantification

Abstract

A "heart attack" or myocardial infarction (MI), occurs when an artery supplying blood to the heart is abruptly occluded. The "gold standard" method for imaging MI is Cardiovascular Magnetic Resonance Imaging (MRI), with intravenously administered gadolinium-based contrast (late gadolinium enhancement). However, no "gold standard" fully automated method for the quantification of MI exists. In this work, we propose an end-to-end fully automatic system (MyI-Net) for the detection and quantification of MI in MRI images. This has the potential to reduce the uncertainty due to the technical variability across labs and inherent problems of the data and labels. Our system consists of four processing stages designed to maintain the flow of information across scales. First, features from raw MRI images are generated using feature extractors built on ResNet and MoblieNet architectures. This is followed by the Atrous Spatial Pyramid Pooling (ASPP) to produce spatial information at different scales to preserve more image context. High-level features from ASPP and initial low-level features are concatenated at the third stage and then passed to the fourth stage where spatial information is recovered via up-sampling to produce final image segmentation output into: i) background, ii) heart muscle, iii) blood and iv) scar areas. New models were compared with state-of-art models and manual quantification. Our models showed favorable performance in global segmentation and scar tissue detection relative to state-of-the-art work, including a four-fold better performance in matching scar pixels to contours produced by clinicians.