A Deep Learning Pipeline for Assessing Ventricular Volumes from a Cardiac MRI Registry of Patients with Single
Tina Yao1, Nicole St Clair1, Gabriel F Miller1
1From the Institutes of Health Informatics (T.Y.) and Cardiovascular Science (M.Q., J.A.S., V.M.), University College London, 20c Guilford Street, London WC1N 1DZ, England; Department of Cardiology, Boston Children's Hospital, Boston, Mass (N.S.C., G.F.M., S.G., D.S., R.H.R.); Department of Pediatrics, University of Michigan, Ann Arbor, Mich (A.L.D.); Division of Cardiology, The Children's Hospital of Philadelphia, Philadelphia, Pa (M.A.F.); Department of Radiology, Nationwide Children's Hospital, Columbus, Ohio (R.K.); Department of Diagnostic Imaging, Hospital for Sick Children, Toronto, Canada (C.Z.L.); Department of Pediatrics, Ann and Robert H Lurie Children's Hospital of Chicago, Chicago, Ill (J.D.R.); Department of Pediatric Cardiology, Emory University School of Medicine, Atlanta, Ga (T.C.S.); and Department of Cardiology, Texas Children's Hospital, Houston, Tex (J.W.).
Abstract:
Purpose To develop an end-to-end deep learning (DL) pipeline for automated ventricular segmentation of cardiac MRI data from a multicenter registry of patients with Fontan circulation (Fontan Outcomes Registry Using CMR Examinations [FORCE]). Materials and Methods This retrospective study used 250 cardiac MRI examinations (November 2007-December 2022) from 13 institutions for training, validation, and testing. The pipeline contained three DL models: a classifier to identify short-axis cine stacks and two U-Net 3+ models for image cropping and segmentation. The automated segmentations were evaluated on the test set (n = 50) by using the Dice score. Volumetric and functional metrics derived from DL and ground truth manual segmentations were compared using Bland-Altman and intraclass correlation analysis. The pipeline was further qualitatively evaluated on 475 unseen examinations. Results There were acceptable limits of agreement (LOA) and minimal biases between the ground truth and DL end-diastolic volume (EDV) (bias: -0.6 mL/m2, LOA: -20.6 to 19.5 mL/m2) and end-systolic volume (ESV) (bias: -1.1 mL/m2, LOA: -18.1 to 15.9 mL/m2), with high intraclass correlation coefficients (ICCs > 0.97) and Dice scores (EDV, 0.91 and ESV, 0.86). There was moderate agreement for ventricular mass (bias: -1.9 g/m2, LOA: -17.3 to 13.5 g/m2) and an ICC of 0.94. There was also acceptable agreement for stroke volume (bias: 0.6 mL/m2, LOA: -17.2 to 18.3 mL/m2) and ejection fraction (bias: 0.6%, LOA: -12.2% to 13.4%), with high ICCs (>0.81). The pipeline achieved satisfactory segmentation in 68% of the 475 unseen examinations, while 26% needed minor adjustments, 5% needed major adjustments, and in 0.4%, the cropping model failed. Conclusion The DL pipeline can provide fast standardized segmentation for patients with single ventricle physiology across multiple centers. This pipeline can be applied to all cardiac MRI examinations in the FORCE registry. Keywords: Cardiac, Adults and Pediatrics, MR Imaging, Congenital, Volume Analysis, Segmentation, Quantification Supplemental material is available for this article. © RSNA, 2023.
Related Concept Videos
01:30Radiological Investigation II: MRI and Ventilation Perfusion Scan
Magnetic Resonance Imaging (MRI) and Ventilation Perfusion Scans are two radiological investigations that offer detailed diagnostic images of the body, particularly lung structures.
MRI
MRI uses magnetic fields and radiofrequency signals to distinguish between normal and abnormal tissues. This technology provides a more detailed diagnostic image than CT scans, enabling it to characterize pulmonary nodules, stage bronchogenic carcinoma, and evaluate inflammatory activity in...
01:21Imaging Studies for Cardiovascular System IV: CMRI