Fully Automated On-Scanner Aortic Four Dimensional Flow Magnetic Resonance Imaging Processing and Hemodynamic Analysis.
Authors
Affiliations (4)
Affiliations (4)
- Northwestern Radiology, Chicago, IL, United States. Electronic address: [email protected].
- Northwestern Radiology, Chicago, IL, United States.
- Cardiovascular MR R&D, Siemens Medical Solutions USA, Inc.
- Northwestern Radiology, Chicago, IL, United States; Cardiovascular MR R&D, Siemens Medical Solutions USA, Inc.
Abstract
To develop an end-to-end 4D flow MRI analysis pipeline for automated hemodynamic analysis with full on-scanner deployment. The Framework for Image Reconstruction (FIRE) framework was used to integrated 4D flow processing tasks directly into the scanner reconstruction pipeline. The method builds containerized applications with standardized raw and image MRI data input/output in the open-source Magnetic Resonance Data (MRD) format. In this study, deep learning models and algorithms from previous work (4D flow preprocessing, 3D aorta segmentation, aorta velocity maps, quantification of aortic systolic peak velocities) were implemented in TensorFlow and executed within a containerized Python 3.6 environment on the MRI scanner, directly following the MRI data acquisition. All tasks were executed in-line using the MRI system's own computational resources. Analysis results were returned alongside the standard 4D flow MRI magnitude/phase images, available for review on-scanner console immediately after the MRI scan. In a study with 20 subjects (n=10 patients with aortic disease, n=10 healthy controls), FIRE performance was evaluated and compared to manual 4D flow analysis (reference standard). We successfully implemented on-scanner automated 4D flow hemodynamic analysis on a 1.5T MRI system. Total on-scanner computation time for 4D flow analysis was 220 ± 35seconds. Dice scores between manual vs. deep learning processing (eddy current static tissue selection: 0.84 ± 0.14; noise voxel detection: 0.92 ± 0.04; aortic 3D segmentation 0.92 ± 0.06) demonstrated good to excellent pipeline performance. Bland Altman analysis revealed a small but significant bias (0.04m/s, p = 0.01) for peak systolic velocities between manual and deep learning processing with good limits of agreement (-0.10, 0.18m/s) and a mean relative difference of 4%. An automated 4D flow processing workflow was successfully deployed for fully automated on-scanner hemodynamic analysis with good in-line vs. human performance, indicating its potential for increased workflow efficiency.