@article{WechAnkenbrandBleyetal.2022, author = {Wech, Tobias and Ankenbrand, Markus Johannes and Bley, Thorsten Alexander and Heidenreich, Julius Frederik}, title = {A data-driven semantic segmentation model for direct cardiac functional analysis based on undersampled radial MR cine series}, series = {Magnetic Resonance in Medicine}, volume = {87}, journal = {Magnetic Resonance in Medicine}, number = {2}, doi = {10.1002/mrm.29017}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-257616}, pages = {972-983}, year = {2022}, abstract = {Purpose Image acquisition and subsequent manual analysis of cardiac cine MRI is time-consuming. The purpose of this study was to train and evaluate a 3D artificial neural network for semantic segmentation of radially undersampled cardiac MRI to accelerate both scan time and postprocessing. Methods A database of Cartesian short-axis MR images of the heart (148,500 images, 484 examinations) was assembled from an openly accessible database and radial undersampling was simulated. A 3D U-Net architecture was pretrained for segmentation of undersampled spatiotemporal cine MRI. Transfer learning was then performed using samples from a second database, comprising 108 non-Cartesian radial cine series of the midventricular myocardium to optimize the performance for authentic data. The performance was evaluated for different levels of undersampling by the Dice similarity coefficient (DSC) with respect to reference labels, as well as by deriving ventricular volumes and myocardial masses. Results Without transfer learning, the pretrained model performed moderately on true radial data [maximum number of projections tested, P = 196; DSC = 0.87 (left ventricle), DSC = 0.76 (myocardium), and DSC =0.64 (right ventricle)]. After transfer learning with authentic data, the predictions achieved human level even for high undersampling rates (P = 33, DSC = 0.95, 0.87, and 0.93) without significant difference compared with segmentations derived from fully sampled data. Conclusion A 3D U-Net architecture can be used for semantic segmentation of radially undersampled cine acquisitions, achieving a performance comparable with human experts in fully sampled data. This approach can jointly accelerate time-consuming cine image acquisition and cumbersome manual image analysis.}, language = {en} } @article{AnkenbrandLohrSchloetelburgetal.2021, author = {Ankenbrand, Markus Johannes and Lohr, David and Schl{\"o}telburg, Wiebke and Reiter, Theresa and Wech, Tobias and Schreiber, Laura Maria}, title = {Deep learning-based cardiac cine segmentation: Transfer learning application to 7T ultrahigh-field MRI}, series = {Magnetic Resonance in Medicine}, volume = {86}, journal = {Magnetic Resonance in Medicine}, number = {4}, doi = {10.1002/mrm.28822}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-257604}, pages = {2179-2191}, year = {2021}, abstract = {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.}, language = {en} } @unpublished{HeidenreichGassenmaierAnkenbrandetal.2021, author = {Heidenreich, Julius F. and Gassenmaier, Tobias and Ankenbrand, Markus J. and Bley, Thorsten A. and Wech, Tobias}, title = {Self-configuring nnU-net pipeline enables fully automatic infarct segmentation in late enhancement MRI after myocardial infarction}, edition = {accepted version}, doi = {10.1016/j.ejrad.2021.109817}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-323418}, year = {2021}, abstract = {Purpose To fully automatically derive quantitative parameters from late gadolinium enhancement (LGE) cardiac MR (CMR) in patients with myocardial infarction and to investigate if phase sensitive or magnitude reconstructions or a combination of both results in best segmentation accuracy. Methods In this retrospective single center study, a convolutional neural network with a U-Net architecture with a self-configuring framework ("nnU-net") was trained for segmentation of left ventricular myocardium and infarct zone in LGE-CMR. A database of 170 examinations from 78 patients with history of myocardial infarction was assembled. Separate fitting of the model was performed, using phase sensitive inversion recovery, the magnitude reconstruction or both contrasts as input channels. Manual labelling served as ground truth. In a subset of 10 patients, the performance of the trained models was evaluated and quantitatively compared by determination of the S{\o}rensen-Dice similarity coefficient (DSC) and volumes of the infarct zone compared with the manual ground truth using Pearson's r correlation and Bland-Altman analysis. Results The model achieved high similarity coefficients for myocardium and scar tissue. No significant difference was observed between using PSIR, magnitude reconstruction or both contrasts as input (PSIR and MAG; mean DSC: 0.83 ± 0.03 for myocardium and 0.72 ± 0.08 for scars). A strong correlation for volumes of infarct zone was observed between manual and model-based approach (r = 0.96), with a significant underestimation of the volumes obtained from the neural network. Conclusion The self-configuring nnU-net achieves predictions with strong agreement compared to manual segmentation, proving the potential as a promising tool to provide fully automatic quantitative evaluation of LGE-CMR.}, language = {en} } @article{WengHeidenreichMetzetal.2021, author = {Weng, Andreas M. and Heidenreich, Julius F. and Metz, Corona and Veldhoen, Simon and Bley, Thorsten A. and Wech, Tobias}, title = {Deep learning-based segmentation of the lung in MR-images acquired by a stack-of-spirals trajectory at ultra-short echo-times}, series = {BMC Medical Imaging}, volume = {21}, journal = {BMC Medical Imaging}, doi = {10.1186/s12880-021-00608-1}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-260520}, year = {2021}, abstract = {Background Functional lung MRI techniques are usually associated with time-consuming post-processing, where manual lung segmentation represents the most cumbersome part. The aim of this study was to investigate whether deep learning-based segmentation of lung images which were scanned by a fast UTE sequence exploiting the stack-of-spirals trajectory can provide sufficiently good accuracy for the calculation of functional parameters. Methods In this study, lung images were acquired in 20 patients suffering from cystic fibrosis (CF) and 33 healthy volunteers, by a fast UTE sequence with a stack-of-spirals trajectory and a minimum echo-time of 0.05 ms. A convolutional neural network was then trained for semantic lung segmentation using 17,713 2D coronal slices, each paired with a label obtained from manual segmentation. Subsequently, the network was applied to 4920 independent 2D test images and results were compared to a manual segmentation using the S{\o}rensen-Dice similarity coefficient (DSC) and the Hausdorff distance (HD). Obtained lung volumes and fractional ventilation values calculated from both segmentations were compared using Pearson's correlation coefficient and Bland Altman analysis. To investigate generalizability to patients outside the CF collective, in particular to those exhibiting larger consolidations inside the lung, the network was additionally applied to UTE images from four patients with pneumonia and one with lung cancer. Results The overall DSC for lung tissue was 0.967 ± 0.076 (mean ± standard deviation) and HD was 4.1 ± 4.4 mm. Lung volumes derived from manual and deep learning based segmentations as well as values for fractional ventilation exhibited a high overall correlation (Pearson's correlation coefficent = 0.99 and 1.00). For the additional cohort with unseen pathologies / consolidations, mean DSC was 0.930 ± 0.083, HD = 12.9 ± 16.2 mm and the mean difference in lung volume was 0.032 ± 0.048 L. Conclusions Deep learning-based image segmentation in stack-of-spirals based lung MRI allows for accurate estimation of lung volumes and fractional ventilation values and promises to replace the time-consuming step of manual image segmentation in the future.}, language = {en} } @article{StichPfaffWechetal.2020, author = {Stich, Manuel and Pfaff, Christiane and Wech, Tobias and Slawig, Anne and Ruyters, Gudrun and Dewdney, Andrew and Ringler, Ralf and K{\"o}stler, Herbert}, title = {The temperature dependence of gradient system response characteristics}, series = {Magnetic Resonance in Medicine}, volume = {83}, journal = {Magnetic Resonance in Medicine}, doi = {10.1002/mrm.28013}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-206212}, pages = {1519-1527}, year = {2020}, abstract = {Purpose: The gradient system transfer function (GSTF) characterizes the frequency transfer behavior of a dynamic gradient system and can be used to correct non-Cartesian k-space trajectories. This study analyzes the impact of the gradient coil temperature of a 3T scanner on the GSTF. Methods: GSTF self- and B\(_0\)-cross-terms were acquired for a 3T Siemens scanner (Siemens Healthcare, Erlangen, Germany) using a phantom-based measurement technique. The GSTF terms were measured for various temperature states up to 45°C. The gradient coil temperatures were measured continuously utilizing 12 temperature sensors which are integrated by the vendor. Different modeling approaches were applied and compared. Results: The self-terms depend linearly on temperature, whereas the B0-cross-term does not. Effects induced by thermal variation are negligible for the phase response. The self-terms are best represented by a linear model including the three gradient coil sensors that showed the maximum temperature dependence for the three axes. The use of time derivatives of the temperature did not lead to an improvement of the model. The B\(_0\)-cross-terms can be modeled by a convolution model which considers coil-specific heat transportation. Conclusion: The temperature dependency of the GSTF was analyzed for a 3T Siemens scanner. The self- and B0-cross-terms can be modeled using a linear and convolution modeling approach based on the three main temperature sensor elements.}, language = {en} } @article{RichterWechWengetal.2020, author = {Richter, Julian A. J. and Wech, Tobias and Weng, Andreas M. and Stich, Manuel and Weick, Stefan and Breuer, Kathrin and Bley, Thorsten A. and K{\"o}stler, Herbert}, title = {Free-breathing self-gated 4D lung MRI using wave-CAIPI}, series = {Magnetic Resonance in Medicine}, volume = {84}, journal = {Magnetic Resonance in Medicine}, number = {6}, doi = {10.1002/mrm.28383}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-218075}, pages = {3223 -- 3233}, year = {2020}, abstract = {Purpose The aim of this study was to compare the wave-CAIPI (controlled aliasing in parallel imaging) trajectory to the Cartesian sampling for accelerated free-breathing 4D lung MRI. Methods The wave-CAIPI k-space trajectory was implemented in a respiratory self-gated 3D spoiled gradient echo pulse sequence. Trajectory correction applying the gradient system transfer function was used, and images were reconstructed using an iterative conjugate gradient SENSE (CG SENSE) algorithm. Five healthy volunteers and one patient with squamous cell carcinoma in the lung were examined on a clinical 3T scanner, using both sampling schemes. For quantitative comparison of wave-CAIPI and standard Cartesian imaging, the normalized mutual information and the RMS error between retrospectively accelerated acquisitions and their respective references were calculated. The SNR ratios were investigated in a phantom study. Results The obtained normalized mutual information values indicate a lower information loss due to acceleration for the wave-CAIPI approach. Average normalized mutual information values of the wave-CAIPI acquisitions were 10\% higher, compared with Cartesian sampling. Furthermore, the RMS error of the wave-CAIPI technique was lower by 19\% and the SNR was higher by 14\%. Especially for short acquisition times (down to 1 minute), the undersampled Cartesian images showed an increased artifact level, compared with wave-CAIPI. Conclusion The application of the wave-CAIPI technique to 4D lung MRI reduces undersampling artifacts, in comparison to a Cartesian acquisition of the same scan time. The benefit of wave-CAIPI sampling can therefore be traded for shorter examinations, or enhancing image quality of undersampled 4D lung acquisitions, keeping the scan time constant.}, language = {en} } @article{TranGiaWechBleyetal.2015, author = {Tran-Gia, Johannes and Wech, Tobias and Bley, Thorsten and K{\"o}stler, Herbert}, title = {Model-Based Acceleration of Look-Locker T1 Mapping}, series = {PLoS One}, volume = {10}, journal = {PLoS One}, number = {4}, doi = {10.1371/journal.pone.0122611}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-126436}, pages = {e0122611}, year = {2015}, abstract = {Mapping the longitudinal relaxation time \(T_1\) has widespread applications in clinical MRI as it promises a quantitative comparison of tissue properties across subjects and scanners. Due to the long scan times of conventional methods, however, the use of quantitative MRI in clinical routine is still very limited. In this work, an acceleration of Inversion-Recovery Look-Locker (IR-LL) \(T_1\) mapping is presented. A model-based algorithm is used to iteratively enforce an exponential relaxation model to a highly undersampled radially acquired IR-LL dataset obtained after the application of a single global inversion pulse. Using the proposed technique, a \(T_1\) map of a single slice with 1.6mm in-plane resolution and 4mm slice thickness can be reconstructed from data acquired in only 6s. A time-consuming segmented IR experiment was used as gold standard for \(T_1\) mapping in this work. In the subsequent validation study, the model-based reconstruction of a single-inversion IR-LL dataset exhibited a \(T_1\) difference of less than 2.6\% compared to the segmented IR-LL reference in a phantom consisting of vials with \(T_1\) values between 200ms and 3000ms. In vivo, the \(T_1\) difference was smaller than 5.5\% in WM and GM of seven healthy volunteers. Additionally, the \(T_1\) values are comparable to standard literature values. Despite the high acceleration, all model-based reconstructions were of a visual quality comparable to fully sampled references. Finally, the reproducibility of the \(T_1\) mapping method was demonstrated in repeated acquisitions. In conclusion, the presented approach represents a promising way for fast and accurate \(T_1\) mapping using radial IR-LL acquisitions without the need of any segmentation.}, language = {en} }