Assessment of Deep Learning Segmentation for Real-Time Free-Breathing Cardiac Magnetic Resonance Imaging at Rest and Under Exercise Stress

TL;DR

This study evaluates whether deep-learning segmentation methods trained on cine CMR can reliably segment the left and right ventricles in real-time free-breathing CMR at rest and during exercise. It compares a commercial cine-oriented method (comDL) with nnU-Net, using manually corrected contours as reference and Dice coefficient ($DC$) as the primary accuracy metric, while also deriving EDV, ESV, and EF. Cine CMR results show both methods achieving high $DC$ values, with nnU-Net generally outperforming comDL in real-time scenarios; at rest, nnU-Net segmentations yield errors within inter-/intra-observer variability, while under exercise stress the performance improves the potential for automation though not fully automatic yet. Overall, cine-trained DL methods demonstrate strong potential for automatic real-time CMR segmentation, especially at rest, with real-time performance under stress encouraging further development toward complete automation and clinical adoption.

Abstract

In recent years, a variety of deep learning networks for cardiac MRI (CMR) segmentation have been developed and analyzed. However, nearly all of them are focused on cine CMR under breathold. In this work, accuracy of deep learning methods is assessed for volumetric analysis (via segmentation) of the left ventricle in real-time free-breathing CMR at rest and under exercise stress. Data from healthy volunteers (n=15) for cine and real-time free-breathing CMR at rest and under exercise stress were analyzed retrospectively. Segmentations of a commercial software (comDL) and a freely available neural network (nnU-Net), were compared to a reference created via the manual correction of comDL segmentation. Segmentation of left ventricular endocardium (LV), left ventricular myocardium (MYO), and right ventricle (RV) is evaluated for both end-systolic and end-diastolic phases and analyzed with Dice's coefficient (DC). The volumetric analysis includes LV end-diastolic volume (EDV), LV end-systolic volume (ESV), and LV ejection fraction (EF). For cine CMR, nnU-Net and comDL achieve a DC above 0.95 for LV and 0.9 for MYO, and RV. For real-time CMR, the accuracy of nnU-Net exceeds that of comDL overall. For real-time CMR at rest, nnU-Net achieves a DC of 0.94 for LV, 0.89 for MYO, and 0.90 for RV; mean absolute differences between nnU-Net and reference are 2.9mL for EDV, 3.5mL for ESV and 2.6% for EF. For real-time CMR under exercise stress, nnU-Net achieves a DC of 0.92 for LV, 0.85 for MYO, and 0.83 for RV; mean absolute differences between nnU-Net and reference are 11.4mL for EDV, 2.9mL for ESV and 3.6% for EF. Deep learning methods designed or trained for cine CMR segmentation can perform well on real-time CMR. For real-time free-breathing CMR at rest, the performance of deep learning methods is comparable to inter-observer variability in cine CMR and is usable or fully automatic segmentation.

Figures (8)

• Figure 1: (Top) Mid-ventricular slices in ED phase of the same volunteer for cine and real-time free-breathing at different heart rates and (bottom) corresponding manually corrected segmentation in a short axis view are shown. Image quality decreases and reconstruction artifacts increase with an increasing heart rate. The left ventricular endocard (red), the left ventricular myocardium (green), and the right ventricle (blue) are segmented.
• Figure 2: Representative segmentations of manually corrected contours and deep learning methods. (First row) Mid-ventricular slices in ES phase of a volunteer for cine and real-time free-breathing at different heart rates with (second row) corresponding manually corrected, (third row) comDL, and (fourth row) nnU-Net segmentation. Accuracy of segmentation is measured with Dice's coefficient (DC). DC for left ventricular endocard (LV), left ventricular myocardium (MYO), and right ventricle (RV) are given for each segmentation.
• Figure 3: Example segmentation failures of nnU-Net for real-time CMR under exercise stress. Incomplete segmentation of the right ventricle in the apical region (first and second column). Anatomically incoherent segmentation of the myocardium and right ventricle in the basal region (third and fourth column).
• Figure 4: Dice's coefficient of nnU-Net and comDL segmentation for real-time CMR measurements plotted against heart rate. DC values of LV, MYO, and RV are calculated for (a) nnU-Net and (b) comDL segmentation in respect to manually corrected contours. Each data point presents the average DC of a segmentation class for a single real-time measurement of a volunteer. Real-time CMR at rest, under exercise stress and maximal exercise stress are presented by their average calculated heart rate.
• Figure 5: Bland-Altman plots for cardiac function parameters of nnU-Net. The cardiac function parameters of (a-c) the left ventricular end-diastolic volume (EDV), (d-f) the left ventricular end-systolic volume (ESV), and (g-i) the left ventricular ejection fraction (EF) derived from the nnU-Net segmentation and the manually corrected contours (mcc) are compared with each other in Bland-Altman plots. The parameters are compared for cine CMR (a, d, g), real-time free-breathing CMR at rest (b, e, h), and under exercise stress (c, f, i). $P$ values of a paired two-sample t-test are shown in the top left corner of each plot.