Cascaded Convolutional Neural Network for Automatic Myocardial Infarction Segmentation from Delayed-Enhancement Cardiac MRI

TL;DR

This paper introduces a cascaded CNN approach combining 2D and 3D U-Nets for automatic myocardial infarction segmentation from cardiac MRI, achieving state-of-the-art results on a public dataset.

Contribution

The novel cascaded CNN framework effectively leverages intra-slice and volumetric information for improved segmentation accuracy.

Findings

Achieved Dice scores of 0.8786 for myocardium

Outperformed all other teams in MICCAI 2020 EMIDEC challenge

Demonstrated the effectiveness of combining 2D and 3D U-Nets

Abstract

Automatic segmentation of myocardial contours and relevant areas like infraction and no-reflow is an important step for the quantitative evaluation of myocardial infarction. In this work, we propose a cascaded convolutional neural network for automatic myocardial infarction segmentation from delayed-enhancement cardiac MRI. We first use a 2D U-Net to focus on the intra-slice information to perform a preliminary segmentation. After that, we use a 3D U-Net to utilize the volumetric spatial information for a subtle segmentation. Our method is evaluated on the MICCAI 2020 EMIDEC challenge dataset and achieves average Dice score of 0.8786, 0.7124 and 0.7851 for myocardium, infarction and no-reflow respectively, outperforms all the other teams of the segmentation contest.