TY - JOUR A1 - Ankenbrand, Markus Johannes A1 - Lohr, David A1 - Schlötelburg, Wiebke A1 - Reiter, Theresa A1 - Wech, Tobias A1 - Schreiber, Laura Maria T1 - Deep learning-based cardiac cine segmentation: Transfer learning application to 7T ultrahigh-field MRI JF - Magnetic Resonance in Medicine N2 - Purpose Artificial neural networks show promising performance in automatic segmentation of cardiac MRI. However, training requires large amounts of annotated data and generalization to different vendors, field strengths, sequence parameters, and pathologies is limited. Transfer learning addresses this challenge, but specific recommendations regarding type and amount of data required is lacking. In this study, we assess data requirements for transfer learning to experimental cardiac MRI at 7T where the segmentation task can be challenging. In addition, we provide guidelines, tools, and annotated data to enable transfer learning approaches by other researchers and clinicians. Methods A publicly available segmentation model was used to annotate a publicly available data set. This labeled data set was subsequently used to train a neural network for segmentation of left ventricle and myocardium in cardiac cine MRI. The network is used as starting point for transfer learning to 7T cine data of healthy volunteers (n = 22; 7873 images) by updating the pre-trained weights. Structured and random data subsets of different sizes were used to systematically assess data requirements for successful transfer learning. Results Inconsistencies in the publically available data set were corrected, labels created, and a neural network trained. On 7T cardiac cine images the model pre-trained on public imaging data, acquired at 1.5T and 3T, achieved DICE\(_{LV}\) = 0.835 and DICE\(_{MY}\) = 0.670. Transfer learning using 7T cine data and ImageNet weight initialization improved model performance to DICE\(_{LV}\) = 0.900 and DICE\(_{MY}\) = 0.791. Using only end-systolic and end-diastolic images reduced training data by 90%, with no negative impact on segmentation performance (DICE\(_{LV}\) = 0.908, DICE\(_{MY}\) = 0.805). Conclusions This work demonstrates and quantifies the benefits of transfer learning for cardiac cine image segmentation. We provide practical guidelines for researchers planning transfer learning projects in cardiac MRI and make data, models, and code publicly available. KW - 7T KW - ultrahigh-field KW - transfer learning KW - segmentation KW - neural networks KW - deep learning KW - cardiac magnetic resonance KW - cardiac function Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-257604 VL - 86 IS - 4 ER - TY - JOUR A1 - Hock, Michael A1 - Terekhov, Maxim A1 - Stefanescu, Maria Roxana A1 - Lohr, David A1 - Herz, Stefan A1 - Reiter, Theresa A1 - Ankenbrand, Markus A1 - Kosmala, Aleksander A1 - Gassenmaier, Tobias A1 - Juchem, Christoph A1 - Schreiber, Laura Maria T1 - B\(_{0}\) shimming of the human heart at 7T JF - Magnetic Resonance in Medicine N2 - Purpose Inhomogeneities of the static magnetic B\(_{0}\) field are a major limiting factor in cardiac MRI at ultrahigh field (≥ 7T), as they result in signal loss and image distortions. Different magnetic susceptibilities of the myocardium and surrounding tissue in combination with cardiac motion lead to strong spatio‐temporal B\(_{0}\)‐field inhomogeneities, and their homogenization (B0 shimming) is a prerequisite. Limitations of state‐of‐the‐art shimming are described, regional B\(_{0}\) variations are measured, and a methodology for spherical harmonics shimming of the B\(_{0}\) field within the human myocardium is proposed. Methods The spatial B\(_{0}\)‐field distribution in the heart was analyzed as well as temporal B\(_{0}\)‐field variations in the myocardium over the cardiac cycle. Different shim region‐of‐interest selections were compared, and hardware limitations of spherical harmonics B\(_{0}\) shimming were evaluated by calibration‐based B0‐field modeling. The role of third‐order spherical harmonics terms was analyzed as well as potential benefits from cardiac phase–specific shimming. Results The strongest B\(_{0}\)‐field inhomogeneities were observed in localized spots within the left‐ventricular and right‐ventricular myocardium and varied between systolic and diastolic cardiac phases. An anatomy‐driven shim region‐of‐interest selection allowed for improved B\(_{0}\)‐field homogeneity compared with a standard shim region‐of‐interest cuboid. Third‐order spherical harmonics terms were demonstrated to be beneficial for shimming of these myocardial B\(_{0}\)‐field inhomogeneities. Initial results from the in vivo implementation of a potential shim strategy were obtained. Simulated cardiac phase–specific shimming was performed, and a shim term‐by‐term analysis revealed periodic variations of required currents. Conclusion Challenges in state‐of‐the‐art B\(_{0}\) shimming of the human heart at 7 T were described. Cardiac phase–specific shimming strategies were found to be superior to vendor‐supplied shimming. KW - 7 T KW - B KW - cardiac MRI KW - shimming KW - ultrahigh field Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-218096 VL - 85 IS - 1 SP - 182 EP - 196 ER - TY - JOUR A1 - Lohr, David A1 - Terekhov, Maxim A1 - Veit, Franziska A1 - Schreiber, Laura Maria T1 - Longitudinal assessment of tissue properties and cardiac diffusion metrics of the ex vivo porcine heart at 7 T: Impact of continuous tissue fixation using formalin JF - NMR in Biomedicine N2 - In this study we aimed to assess the effects of continuous formalin fixation on diffusion and relaxation metrics of the ex vivo porcine heart at 7 T. Magnetic resonance imaging was performed on eight piglet hearts using a 7 T whole body system. Hearts were measured fresh within 3 hours of cardiac arrest followed by immersion in 10% neutral buffered formalin. T\(_{2}\)* and T\(_{2}\) were assessed using a gradient multi‐echo and multi‐echo spin echo sequence, respectively. A spin echo and a custom stimulated echo sequence were employed to assess diffusion time‐dependent changes in metrics of cardiac diffusion tensor imaging. SNR was determined for b = 0 images. Scans were performed for 5 mm thick apical, midcavity and basal slices (in‐plane resolution: 1 mm) and repeated 7, 15, 50, 100 and 200 days postfixation. Eigenvalues of the apparent diffusion coefficient (ADC) and fractional anisotropy (FA) decreased significantly (P < 0.05) following fixation. Relative to fresh hearts, FA values 7 and 200 days postfixation were 90% and 80%, while respective relative ADC values at those fixation stages were 78% and 92%. Statistical helix and sheetlet angle distributions as well as respective mean and median values showed no systematic influence of continuous formalin fixation. Similar to changes in the ADC, values for T\(_{2}\), T\(_{2}\)* and SNR dropped initially postfixation. Respective relative values compared with fresh hearts at day 7 were 64%, 79% and 68%, whereas continuous fixation restored T\(_{2}\), T\(_{2}\)* and SNR leading to relative values of 74%, 100%, and 81% at day 200, respectively. Relaxation parameters and diffusion metrics are significantly altered by continuous formalin fixation. The preservation of microstructure metrics following prolonged fixation is a key finding that may enable future studies of ventricular remodeling in cardiac pathologies. KW - cardiovascular MR methods KW - diffusion tensor imaging KW - heart structure KW - relaxometry Y1 - 2020 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-215536 VL - 33 IS - 7 ER - TY - JOUR A1 - Lohr, David A1 - Terekhov, Maxim A1 - Weng, Andreas Max A1 - Schroeder, Anja A1 - Walles, Heike A1 - Schreiber, Laura Maria T1 - Spin echo based cardiac diffusion imaging at 7T: An ex vivo study of the porcine heart at 7T and 3T JF - PLoS ONE N2 - Purpose of this work was to assess feasibility of cardiac diffusion tensor imaging (cDTI) at 7 T in a set of healthy, unfixed, porcine hearts using various parallel imaging acceleration factors and to compare SNR and derived cDTI metrics to a reference measured at 3 T. Magnetic resonance imaging was performed on 7T and 3T whole body systems using a spin echo diffusion encoding sequence with echo planar imaging readout. Five reference (b = 0 s/mm\(^2\)) images and 30 diffusion directions (b = 700 s/mm\(^2\)) were acquired at both 7 T and 3 T using a GRAPPA acceleration factor R = 1. Scans at 7 T were repeated using R = 2, R = 3, and R = 4. SNR evaluation was based on 30 reference (b = 0 s/mm\(^2\)) images of 30 slices of the left ventricle and cardiac DTI metrics were compared within AHA segmentation. The number of hearts scanned at 7 T and 3 T was n = 11. No statistically significant differences were found for evaluated helix angle, secondary eigenvector angle, fractional anisotropy and apparent diffusion coefficient at the different field strengths, given sufficiently high SNR and geometrically undistorted images. R≥3 was needed to reduce susceptibility induced geometric distortions to an acceptable amount. On average SNR in myocardium of the left ventricle was increased from 29±3 to 44±6 in the reference image (b = 0 s/mm\(^2\)) when switching from 3 T to 7 T. Our study demonstrates that high resolution, ex vivo cDTI is feasible at 7 T using commercial hardware. KW - Heart KW - Diffusion tensor imaging KW - Eigenvectors KW - Cardiac ventricles KW - Tractography KW - Magnetic resonance imaging KW - Data acquisition KW - Swine Y1 - 2019 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-201376 VL - 14 IS - 3 ER -