Cardiovascular magnetic resonance images with susceptibility artifacts: artificial intelligence with spatial-attention for ventricular volumes and mass assessment

Abstract

Background Segmentation of cardiovascular magnetic resonance (CMR) images is an essential step for evaluating dimensional and functional ventricular parameters as ejection fraction (EF) but may be limited by artifacts, which represent the major challenge to automatically derive clinical information. The aim of this study is to investigate the accuracy of a deep learning (DL) approach for automatic segmentation of cardiac structures from CMR images characterized by magnetic susceptibility artifact in patient with cardiac implanted electronic devices (CIED). Methods In this retrospective study, 230 patients (100 with CIED) who underwent clinically indicated CMR were used to developed and test a DL model. A novel convolutional neural network was proposed to extract the left ventricle (LV) and right (RV) ventricle endocardium and LV epicardium. In order to perform a successful segmentation, it is important the network learns to identify salient image regions even during local magnetic field inhomogeneities. The proposed network takes advantage from a spatial attention module to selectively process the most relevant information and focus on the structures of interest. To improve segmentation, especially for images with artifacts, multiple loss functions were minimized in unison. Segmentation results were assessed against manual tracings and commercial CMR analysis software cvi42(Circle Cardiovascular Imaging, Calgary, Alberta, Canada). An external dataset of 56 patients with CIED was used to assess model generalizability. Results In the internal datasets, on image with artifacts, the median Dice coefficients for end-diastolic LV cavity, LV myocardium and RV cavity, were 0.93, 0.77 and 0.87 and 0.91, 0.82, and 0.83 in end-systole, respectively. The proposed method reached higher segmentation accuracy than commercial software, with performance comparable to expert inter-observer variability (bias ± 95%LoA): LVEF 1 ± 8% vs 3 ± 9%, RVEF − 2 ± 15% vs 3 ± 21%. In the external cohort, EF well correlated with manual tracing (intraclass correlation coefficient: LVEF 0.98, RVEF 0.93). The automatic approach was significant faster than manual segmentation in providing cardiac parameters (approximately 1.5 s vs 450 s). Conclusions Experimental results show that the proposed method reached promising performance in cardiac segmentation from CMR images with susceptibility artifacts and alleviates time consuming expert physician contour segmentation.

Cardiovascular Magnetic Resonance in Patients with Cardiac Electronic Devices: Evidence from a Multicenter Study

“…In the same study, in patients with left-sided ICDs, arm-raised imaging reduced the artifacts from 37.5% to 12.5% (p = 0.02). Once CMR images have been acquired, novel deep learning approaches have been developed for automatic segmentation of cardiac structures in patients with susceptibility to artifacts from CIEDs [25].…”

Kardiovaskuläre MRT in der Bewertung von Kardiomyopathien – ein aktueller Überblick

“…Algorithmen können Scanzeiten erheblich verkürzen, die Bildqualität erhöhen und durch automatisierte, genaue Interpretation von MRT-Daten die Diagnostik verbessern [60]. Sensitivere Techniken können so auch bei der Entwicklung von prädiktiven Modellen für verschiedene Kardiomyopathien helfen, was eine frühe Diagnose und personalisierte Behandlungsstrategien unterstützt [61,62,63,64,65] und, wie kürzlich publiziert, eine wichtige Rolle in der prognostischen Beurteilung spielen könnte [66].…”

Cardiac magnetic resonance in advanced heart failure

“…The second category comprehends all the artifacts arising from the deterioration of the RF pulses. 92 There are many strategies to mitigate these effects of CIEDs on CMR, like increasing the distance between the generator and the heart by lifting the left arm, 102 using deep learning techniques to overcome the artifacts, 103 or using smaller voxels, and smaller echo times with shorter RF pulses and larger receiver bandwidth in the frequency encoded direction. 98 A technique that showed the best results in correcting the artifacts is the use of wideband inversion pulses for IR or saturation recovery (SR) sequences.…”

“…There are many strategies to mitigate these effects of CIEDs on CMR, like increasing the distance between the generator and the heart by lifting the left arm, 102 using deep learning techniques to overcome the artifacts, 103 or using smaller voxels, and smaller echo times with shorter RF pulses and larger receiver bandwidth in the frequency encoded direction 98…”