Dorottya Papp¹, Jose M. Castillo T.¹, Piotr A. Wielopolski¹, Pierluigi Ciet¹, Gyula Kotek¹, Jifke F Veenland¹, and Juan Antonio Hernandez-Tamames^2
1Radiology and Nuclear Medicine, Erasmus Medical Center, Rotterdam, Netherlands, ²Erasmus Medical Center, Rotterdam, Netherlands

Although  fully convolutional neural networks (FCNNs) have been widely used for MR imaging, they have not been extended for improving free-breathing lung imaging yet. Our aim was to improve the image quality of retrospective respiratory gated version of a Zero Echo Time (ZTE) MRI sequence (4D-ZTE) in free-breathing using a FCNN so enabling free-breathing acquisition in those patients who cannot perform breath-hold imaging. Our model obtained a MSE of 0.08% on the validation set. When tested on unseen data (4D-ZTE) the predicted images from our model had improved visual image quality and artifacts were reduced in free-breathing 4D-ZTE.   

Fabien J Bertin¹, Guilhem J Collier², Paul JC Hughes², Laurie Smith², James Aeden², Helen Marshall², Jim M Wild^2,3, and Alberto M Biancardi^2
1Télécom SudParis, Paris, France, ²POLARIS, Department of Infection, Immunity and Cardiovascular Disease, The University of Sheffield, Sheffield, United Kingdom, ³Insigneo Institute for in silico Medicine, The University of Sheffield, Sheffield, United Kingdom

The assessment of pulmonary hyperpolarised(HP)-gas MR images is instrumental in identifying potential pathologies, directing treatment, or monitoring disease progression. HP-gas images quantify the amount of gas concentration, whose distribution within the lungs is analysed. A key role in this process is played the main airways that are identified for quality control and excluded for the analysis. Currently, this task is performed manually and existing deep-learning (DL) applications do not provide an explicit labelling of the airways. A specific tailoring of a well-known DL approach was developed with the aim of replacing the manual editing thanks to its good performance.

Sarah H. Needleman¹, Jamie R. McClelland², Björn Eiben², and Geoff J. M. Parker^1,3
1Centre for Medical Image Computing, Quantitative Imaging Group, Department of Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, ²Centre for Medical Image Computing, Radiotherapy Image Computing Group, Department of Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, ³Bioxydyn Limited, Manchester, United Kingdom

We describe a flexible framework incorporating an anatomically realistic digital thorax phantom with physics-based simulation, respiratory motion and functional contrasts. We demonstrate the framework in the context of dynamic oxygen-enhanced MRI (OE-MRI).

 

The framework is designed to provide ground-truth for assessment of novel scan protocols and analysis methods, of utility for OE-MRI as derived lung function measures are limited in accuracy due to motion artefacts, blurring and poor signal-to-noise ratio. The framework was applied to a 2D inversion-prepared spoiled gradient echo dynamic OE-MRI readout. The simulated series displayed respiratory motion; quantitative measures describing hyperoxia-induced contrast enhancement agreed with experimental literature.

Brandon Zanette¹, Yonni Friedlander^1,2, Marcus J Couch^3,4,5, and Giles Santyr^1,2
1Translational Medicine, The Hospital for Sick Children, Toronto, ON, Canada, ²Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada, ³Siemens Healthcare Limited, Montreal, QC, Canada, ⁴McConnell Brain Imaging Centre, Montreal Neurological Institute, Montreal, QC, Canada, ⁵Department of Neurology and Neurosurgery, McGill University, Montreal, QC, Canada

Inert fluorinated (¹⁹F) gas MRI is an emergent technology that has potential to be a lower cost, reduced infrastructure alternative to hyperpolarized gas MRI for lung imaging. In this work we present an approach for non-Cartesian spiral imaging in a phantom. Compared to typical gradient echo imaging with the same spatial resolution, the proposed approach increased SNR by 44% in approximately a third of the acquisition time. Further simulation and extrapolation to clinical imaging experiments indicate significant potential for this approach to improve SNR and image quality.   

Bochao Li¹, Nam G. Lee¹, and Krishna Nayak^2
1Biomedical Engineering, University of Southern California, Los Angeles, CA, United States, ²Ming Hsieh Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, CA, United States

MRI of lung parenchyma is limited by low proton density and susceptibility effects at air-tissue interfaces. Recent high-performance low-field MRI systems have provided a new opportunity to mitigate the latter issue. In this work, we demonstrate a framework for simulating lung MRI across B0 field strengths. We estimate apparent transverse relaxation due to proximity to A) sub-voxel alveoli, and B) bronchial tree. Alveoli are modeled using face-centered cubic close packing. The bronchial tree is modeled using a recent XCAT phantom.

Efe Ilicak¹, Jascha Zapp¹, Safa Ozdemir¹, Lothar R. Schad¹, and Frank G. Zöllner^1,2
1Computer Assisted Clinical Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany, ²Mannheim Institute for Intelligent Systems in Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany

Functional lung imaging is of great importance for diagnosis of pulmonary diseases. Previously, a non-contrast-enhanced method called Fourier Decomposition was proposed for assessing pulmonary functions based on balanced steady-stated free precession pulse sequence. However, this pulse sequence is known to be sensitive to magnetic field inhomogeneities. Here, we propose a phase-cycled acquisition for improved robustness against field inhomogeneities. In vivo results from 1.5 T and 3 T scanners are provided to demonstrate the performance of phase-cycled acquisitions for functional lung imaging. Preliminary results indicate that phase cycling can be a viable option for reducing limitations arising from magnetic field inhomogeneities.

Grzegorz Bauman^1,2 and Oliver Bieri^1,2
1Department of Radiology, Division of Radiological Physics, University of Basel Hospital, Basel, Switzerland, ²Department of Biomedical Engineering, University of Basel, Allschwil, Switzerland

In this work, we demonstrate the application of free-breathing respiratory self-gated thoracic MRI with balanced steady-state free precession half-radial dual-echo imaging technique (bSTAR) in human subjects. The technique combines an efficient minimal-TR readout sampling with an interleaved randomly tilted Archimedean spiral trajectory. The methodological improvements result in high-quality visualization of pulmonary parenchyma and vessel structure. The proposed k-space sampling scheme allows reconstruction of multi-volume data sets at different respiratory phases.

Zekang Ding¹, Huajun She¹, and Yiping Du^1
1School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China

    In original iMoCo algorithm, a single frame image was reconstructed by solving the iMoCo reconstruction model including the estimated motion fields and TGV sparse constraint. Since motion fields is critical in iMoCo algorithm, errors in motion estimation would deteriorate the final reconstructed image. In this study, we improved the performance of iMoCo through (1) reconstructing the full resolution dynamic images for motion estimation, (2) estimating motion fields through nonrigid group-wise registration, and (3) using a motion state weighted iMoCo reconstruction model. Residual streaking artifacts and certain image blurring were suppressed using the proposed algorithm in comparison with the original iMoCo.

Vadim Malis¹, Won Bae¹, Asako Yamamoto¹, Yoshimori Kassai², Andrew Yen¹, Susan Hopkins¹, Yoshiharu Ohno³, and Mitsue Miyazaki^1
1Radiology, UC San Diego, San Diego, CA, United States, ²Canon Medical, Tochigi, Japan, ³Radiology, Fujita Health University, Toyoake, Japan

We have developed both the acquisition methodology and post-processing algorithms for 3D volume MRI. Using 3D MRI,  T₂*, lung volume changes between inspiration and expiration as a measure of regional specific ventilation (δV/V0), and pulmonary vasculature can be obtained using 3D ultra-short TE (UTE) imaging during free breathing.

Marta Tibiletti¹, Jo Naish^1,2, John C Waterton^1,3, Paul JC Hughes⁴, James A Eaden⁴, Jim M Wild⁴, and Geoff JM Parker^1,5
1Bioxydyn Ltd, Manchester, United Kingdom, ²MCMR, Manchester University NHS Foundation Trust, Manchester, United Kingdom, ³Centre for Imaging Sciences,, University of Manchester, Manchester, United Kingdom, ⁴POLARIS, Imaging Sciences, Department of Infection, Immunity and Cardiovascular Disease, The University of Sheffield, Sheffield, United Kingdom, ⁵Centre for Medical Image Computing, University College London, London, United Kingdom

In this work, we explore the possibility of extracting a population Arterial Input Function (AIF) to be used in the quantification of lung perfusion using T1-weighted contrast agent-based perfusion imaging. The population AIF averages the shape of first pass of the CA bolus at high temporal resolution in the pulmonary arteries from 90 scans acquired in 50 patients with interstitial lung disease. The population AIF is then scaled by the patient weight. The results of the analysis using a measured AIF and the population AIF are compared.

Anke Balasch¹, Hao Li², Patrick Metze¹, Alireza Abaei², and Volker Rasche^1
1Department of Internal Medicine II, Ulm University Medical Center, Ulm, Germany, Ulm, Germany, ²Core Facility Small Animal Imaging (CF-SANI), Ulm University, Ulm, Germany, Ulm, Germany

Lung imaging in small animals is particular difficult due to small anatomy dimensions, high respiratory and heart rate and the very short T2* value in high magnetic fields. A 2D tiny golden angle UTE sequence was tested in free-breathing healthy mice to generate high quality lung images. The functional ventilation and proton density was evaluated.

Anne Slawig¹, Andreas Max Weng¹, Bernhard Petritsch¹, Simon Veldhoen¹, and Herbert Köstler^1
1Department of Diagnostic and Interventional Radiology, University Hospital Würzburg, Würzburg, Germany

MR imaging of the lung is a challenge due to the low proton density in lung parenchyma and the very short relaxation times. To overcome low signal and SNR problems UTE-sequences have successfully been applied to lung imaging. Here, the benefit of a 2D UTE-acquisition for the determination of ventilation and perfusion, as determined by SENCEFUL-MRI, is evaluated. It is shown that SENCEFUL-MRI clearly benefits from the use of a UTE-sequence.

Ya-Jun Ma¹, Michael Carl², Hyungseok Jang¹, Saeed Jerban¹, Eric Y Chang^1,3, Seth Kligerman¹, and Jiang Du^1
1UC San Diego, San Diego, CA, United States, ²GE Healthcare, San Diego, CA, United States, ³VA Health system, San Diego, CA, United States

MR imaging of lung is challenging due to its short T2* and low proton density. In this study, we proposed a free breathing phase-resolved lung imaging technique using a 3D UTE sequence with an efficient Cones trajectory. Five different respiration phases were resolved using a self-navigator technique, and the corresponding lung images were reconstructed with a combined parallel imaging and compressed sensing algorithm. Small pulmonary vessels were well-displayed in these images.

Ahsan Javed¹, Rajiv Ramasawmy¹, Pan Su², Thomas Benkert³, Waqas Majeed², and Adrienne E Campbell-Washburn^1
1Cardiovascular Branch, Division of Intramural Research, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, United States, ²Siemens Medical Solutions USA Inc., Malvern, PA, United States, ³Siemens Healthcare GmbH, Erlangen, Germany

Recently, low-field MRI has been shown to improve the sensitivity of oxygen-enhanced (OE) MRI due to the increased relaxivity of oxygen at lower field strengths. In this work, we demonstrate that using self-gated 3D stack-of-spiral OE-MRI we can further improve the sensitivity and repeatability by: 1) using longer free-breathing acquisitions to increase the signal to noise ratio (SNR) of the acquired images, and 2) improving the consistency of respiratory position between two oxygen states. We also demonstrate that these two improvements give us higher temporal SNR which is a measure of sensitivity in quantitative MRI.

Fei Tan¹, Xucheng Zhu², and Peder E.Z. Larson^1,3
1Bioengineering, UC Berkeley - UCSF, San Francisco, CA, United States, ²GE Healthcare, Menlo Park, CA, United States, ³Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States

In this abstract, we conducted a reproducibility study on the Jacobian determinant-based regional ventilation quantification of 1H phase-resolved ultrashort-echo (UTE) lung MRI. We evaluated its feasibility on five healthy adult volunteers and the influence of three registration approaches, cyclic registration, multi-b-spline, and symmetric image normalization (SyN). Regional ventilation maps of two scans, split violin plots of each individual’s regional ventilation distribution, within-subject coefficient of variation, and Bland-Altman plot and linear regression of total ventilation are presented. We demonstrated that Jacobian determinant ventilation quantification and the registration methods are reproducible and may be applied to patient studies in the future.

Anke Balasch¹, Hao Li², Patrick Metze¹, Alireza Abaei², and Volker Rasche^1
1Department of Internal Medicine II, Ulm University Medical Center, Ulm, Germany, Ulm, Germany, ²Core Facility Small Animal Imaging (CF-SANI), Ulm University, Ulm, Germany, Ulm, Germany

In this study, a two dimensional ultra-short echo-time technique was combined with tiny golden angle angular ordering for investigating its feasibility of qualitative assessment of contrast agent dynamics in the lungs after systemic injection.

Yu Deng¹, Qiuxi Lin², Xinchun Li², Weiyin Vivian Liu³, Lei Zhang⁴, Qi Wan², and Chongpeng Sun^2
1Radiology, The first affiliated hospital of Guangzhou Medical University, Guangzhou, China, ²The first affiliated hospital of Guangzhou Medical University, Guangzhou, China, ³MR Research, GE Healthcare, Beijing, China, ⁴GE Healthcare, Beijing, China

The study assessed the feasibility and image quality of free-breathing 3D isotropic whole-lung zero echo time (ZTE) imaging. The overall image quality of the lung in most healthy subjects was good to excellent. The visibility of pulmonary arteries (up to the 7th generation) was better than the bronchi (up to the 5th generation). The demonstration of fissures was poor. In addition, one volunteer with azygos fissure was discovered in our study, which encourage us promote ZTE as a routine lung examination tool. The whole lung ZTE with free-breathing is feasible and can serve as an alternative method of chest imaging.

Anke Balasch¹, Patrick Metze¹, Kilian Stumpf¹, Meinrad Beer², Wolfgang Rottbauer¹, and Volker Rasche^1
1Department of Internal Medicine II, Ulm University Medical Center, Ulm, Germany, Ulm, Germany, ²Department of Radiology, Ulm University Medical Centre, Ulm, Germany, Ulm, Germany

Lung MRI is challenging due to the short T2* and respiratory and cardiac motion. In this study an UTE stack-of-stars sequence has been combined with tiny golden angle (tyGA SoS) for imaging of lung morphology (proton fraction, PF) and function (fractional ventilation, FV). Application of the techniques to smoker and non-smokers revealed significant differences in the PF thus indicating its sensitivity.

Anne-Christianne Kentgens¹, Kathryn Ramsey¹, Grzegorz Bauman^2,3, Francesco Santini^2,3, Christoph Corin Willers¹, Philipp Latzin¹, Oliver Bieri^2,3, and Orso Pusterla^2,3,4
1Division of Respiratory Medicine, Department of Pediatrics, Inselspital, University of Bern, Bern, Switzerland, ²Division of Radiological Physics, University Hospital Basel, Basel, Switzerland, ³Department of Biomedical Engineering, University Hospital Basel, Basel, Switzerland, ⁴Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland

Simultaneous acquisition of oxygen-enhanced MRI and Multiple Breath Washout (MBW) testing is an unexplored and innovative methodology for lung function assessment. Until date, these techniques have been conducted separately. Combined use provides complementary and simultaneous information on pulmonary ventilation, diffusion and perfusion, and has several advantages compared to separate measurements. In this study, we performed oxygen-enhanced relaxometry and non-contrast enhanced ventilation and perfusion mapping using matrix pencil MRI, simultaneously with the MBW.

Patrick Metze¹, Tobias Speidel¹, Fabian Straubmüller¹, and Volker Rasche^1,2
1Internal Medicine II, Ulm University Medical Center, Ulm, Germany, ²Core Facility Small Animal Imaging (CF-SANI), Ulm University, Ulm, Germany

In this work we evaluate the proposed nonuniform self-gating (nuSG) method for lung imaging in highly irregular respiratory patterns and compare its performance to image-based and k-space-based gating methods and to a breath-hold reference. Functional parameters (proton fraction, fractional ventilation) and image sharpness are evaluated, with the nuSG algorithm showing a superior performance in case of fractional ventilation and image sharpness. In more uniform motion patterns its performance is similar to image-based SG and both techniques outperform k-space based gating.