Bayesian Optimization of 2D Echocardiography Segmentation

TL;DR

This paper demonstrates that Bayesian Optimization effectively enhances deep learning models for echocardiography segmentation, leading to higher accuracy and better clinical metric predictions compared to previous methods.

Contribution

It introduces the application of Bayesian Optimization to optimize hyperparameters of a deep CNN for echocardiography segmentation, outperforming existing state-of-the-art models.

Findings

Achieved mean Dice overlaps of 0.95, 0.96, and 0.93 on key cardiac structures.

Significant reduction in errors for clinical indices like LV volumes and ejection fraction.

Model outperforms recent state-of-the-art on the CAMUS dataset.

Abstract

Bayesian Optimization (BO) is a well-studied hyperparameter tuning technique that is more efficient than grid search for high-cost, high-parameter machine learning problems. Echocardiography is a ubiquitous modality for evaluating heart structure and function in cardiology. In this work, we use BO to optimize the architectural and training-related hyperparameters of a previously published deep fully convolutional neural network model for multi-structure segmentation in echocardiography. In a fair comparison, the resulting model outperforms this recent state-of-the-art on the annotated CAMUS dataset in both apical two- and four-chamber echo views. We report mean Dice overlaps of 0.95, 0.96, and 0.93 on left ventricular (LV) endocardium, LV epicardium, and left atrium respectively. We also observe significant improvement in derived clinical indices, including smaller median absolute errors for LV end-diastolic volume (4.9mL vs. 6.7), end-systolic volume (3.1mL vs. 5.2), and ejection fraction (2.6% vs. 3.7); and much tighter limits of agreement, which were already within inter-rater variability for non-contrast echo. These results demonstrate the benefits of BO for echocardiography segmentation over a recent state-of-the-art framework, although validation using large-scale independent clinical data is required.