Shape constrained CNN for segmentation guided prediction of myocardial shape and pose parameters in cardiac MRI

TL;DR

This paper introduces a shape-constrained CNN that predicts myocardial shape and pose parameters in cardiac MRI, ensuring realistic segmentation and enabling direct quantification of regional shape properties.

Contribution

The novel integration of a statistical shape model with CNN-based segmentation, including new loss functions for consistency, improves shape and pose prediction accuracy in cardiac MRI.

Findings

Achieved 99% correlation for LV area across datasets

Attained 91-97% correlation for myocardial area

Reached 80-92% correlation for regional wall thickness

Abstract

Semantic segmentation using convolutional neural networks (CNNs) is the state-of-the-art for many medical image segmentation tasks including myocardial segmentation in cardiac MR images. However, the predicted segmentation maps obtained from such standard CNN do not allow direct quantification of regional shape properties such as regional wall thickness. Furthermore, the CNNs lack explicit shape constraints, occasionally resulting in unrealistic segmentations. In this paper, we use a CNN to predict shape parameters of an underlying statistical shape model of the myocardium learned from a training set of images. Additionally, the cardiac pose is predicted, which allows to reconstruct the myocardial contours. The integrated shape model regularizes the predicted contours and guarantees realistic shapes. We enforce robustness of shape and pose prediction by simultaneously performing pixel-wise semantic segmentation during training and define two loss functions to impose consistency between the two predicted representations: one distance-based loss and one overlap-based loss. We evaluated the proposed method in a 5-fold cross validation on an in-house clinical dataset with 75 subjects and on the ACDC and LVQuan19 public datasets. We show the benefits of simultaneous semantic segmentation and the two newly defined loss functions for the prediction of shape parameters. Our method achieved a correlation of 99% for left ventricular (LV) area on the three datasets, between 91% and 97% for myocardial area, 98-99% for LV dimensions and between 80% and 92% for regional wall thickness.