Functional Lung MRI Research Articles

Repeatability of dynamic 3D phase-resolved functional lung (PREFUL) ventilation MR Imaging in patients with chronic obstructive pulmonary disease andhealthy volunteers.

A previous study has demonstrated the feasibility of 3D phase-resolved functional lung (PREFUL) MRI in healthy volunteers and patients with chronic pulmonary disease. Before clinical use, the repeatability of the ventilation parameters derived from 3D PREFUL MRI must be determined. To evaluate repeatability of 3D PREFUL and to compare with pulmonary functional lung testing (PFT). Prospective. Fifty-three healthy subjects and 13 patients with chronic obstructive pulmonary disease (COPD). A prototype 3D stack-of-stars spoiled-gradient-echo sequence at 1.5 T. Study participants underwent repeated MRI examination (median time interval between scans COPD/healthy subjects [interquartile range]: 7/0 days [6-8/0-0 days]) and one PFT carried out at the time of the baseline MRI. For 3D PREFUL, regional ventilation (RVent) and flow-volume loops were computed and rated by cross-correlation (CC). Also, ventilation time-to-peak (VTTP) was computed. Ventilation defect percentage (VDP) maps were obtained for RVent and CC. Repeatability of 3D PREFUL parameters was evaluated using Bland-Altman analysis, coefficient of variation (COV) and intraclass correlation coefficient (ICC). The relation between 3D PREFUL and PFT measures (forced expiratory volume in 1 second (FEV1 ) and forced vital capacity (FVC) was assessed using the Pearson correlation coefficient (r). In healthy subjects and COPD patients, no significant bias (all P range: 0.09-0.77) and a moderate to good repeatability of RVent, VTTP, and VDPRVent were found (COV range: 0.1%-18.2%, ICC range: 0.51-0.88). For CC and VDPCC moderate repeatability was found (COV range: 0.6%-43.6%, ICC: 0.38-0.60). CC, VDPRVent , and VDPCC showed a good correlation with FEV1 (all |r| > 0.58, all P < 0.05) and FEV1 /FVC ratio (all |r| > 0.62, all P < 0.05). 3D PREFUL provided a good repeatability of RVent, VTTP, and VDPRVent and moderate repeatability of CC and VDPCC in healthy volunteers and COPD patients, and correlated well with FEV1 and FEV1 /FVC. 2 TECHNICAL EFFICACY STAGE: 2.

Flow Volume Loop and Regional Ventilation Assessment Using Phase-Resolved Functional Lung (PREFUL) MRI: Comparison With 129 Xenon Ventilation MRI and Lung Function Testing.

Regional flow volume loop ventilation-weighted noncontrast-enhanced proton lung MRI in free breathing has emerged as a novel technique for assessment of regional lung ventilation, but has yet not been validated with 129 Xenon MRI (129 Xe-MRI), a direct visualization of ventilation in healthy volunteers, cystic fibrosis (CF), and chronic obstructive pulmonary disease (COPD) patients. To compare regional ventilation and regional flow volume loops measured by noncontrast-enhanced ventilation-weighted phase-resolved functional lung MRI (PREFUL-MRI) with 129 Xe-MRI ventilation imaging and with lung function test parameters. Retrospective study. Twenty patients with COPD, eight patients with CF, and six healthy volunteers. PREFUL and 129 Xe-MRI gradient echo sequences were acquired at 1.5T. Coronal slices of PREFUL-MRI (free breathing) and 129 Xe-MRI (single breath-hold) were acquired on the same day, matched by their ventrodorsal position and coregistered for evaluation. Ventilation defect percentage (VDP) was calculated based on regional ventilation (RV), regional flow volume loops (RFVL), or 129 Xe-MRI with two different threshold methods. A combined VDP was calculated for RV and RFVL. Additionally, lung function testing was performed (such as the forced expiratory volume in 1 second [FEV1 ]) was used. The obtained parameters were compared using Wilcoxon tests, correlated using Spearman's correlation coefficient (r), and agreement between PREFUL and 129 Xe-MRI parameters was assessed using Bland-Altman analysis and Dice coefficients. VDP measured by PREFUL and 129 Xe were significantly correlated with both thresholding techniques (r = 0.62-0.69, P < 0.05 for all) and with lung function test parameters. Combined RV and RFVL PREFUL defect maps correlated with lung function testing (eg, with FEV1 r = -0.87 P < 0.05), and showed better regional agreement to 129 Xe-MRI ventilation defects (Dice coefficient defect 0.413) with significantly higher VDP values (10.2 ± 27.3, P = 0.04) than either PREFUL defect map alone. Combined RV and RFVL PREFUL defect maps likely increase sensitivity to mild airway obstruction with increased VDP values compared to 129 Xe-MRI, and correlate strongly with lung function test parameters. 3 TECHNICAL EFFICACY STAGE: 2.

3D phase-resolved functional lung ventilation MR imaging in healthy volunteers and patients with chronic pulmonary disease.

To test the feasibility of 3Dphase-resolved functional lung (PREFUL)MRIin healthy volunteers and patients withchronicpulmonary disease, to compare 3D to 2D PREFUL,and to investigatethe required temporal resolution to obtainstable3D PREFULmeasurement. Sixteen participants underwent MRIusing 2D and 3D PREFUL. Retrospectively, the spatial resolutionof 3D PREFUL (4 × 4 × 4 mm3 ) was decreased to match the spatial resolution of 2D PREFUL (4 × 4 × 15 mm3 ), abbreviated as 3Dlowres .In addition toregional ventilation (RVent),flow-volume loops were computed and rated by a cross-correlation (CC). Ventilation defect percentage (VDP) maps were obtained.RVent, CC, VDPRVent , and VDPCC were compared for systematic differences between 2D, 3Dlowres , and 3D PREFUL.Dividing the 3D PREFUL data into 4-(≈ 20 phases),8-(≈ 40 phases),and 12-min(≈ 60 phases)acquisition pieces, theventilation parameter maps, includingthe heterogeneity ofventilation time to peak,weretested regarding the requiredtemporal resolution. RVent, CC,VDPRVent , and VDPCC presented significant correlationsbetween 2D and 3D PREFUL (r = 0.64-0.94).CC and VDPCC of 2D and 3Dlowres PREFULwere significantlydifferent (P < .0113).Comparing3Dlowres and 3D PREFUL,all parameters were found to be statistically different (P < .0045). 3D PREFUL MRI depicts the whole lung volume and breathing cycle with superior image resolutionand with likely more precisioncompared to 2D PREFUL.Furthermore,3DPREFUL is more sensitive to detect regions of hypoventilation andventilation heterogeneitycompared to 3Dlowres PREFUL, which is important for earlydetectionand improved monitoring of patients with chronic lung disease.

Three-dimensional Ultrashort Echo Time MRI for Functional Lung Imaging in Cystic Fibrosis.

Background In cystic fibrosis (CF), recurrent imaging and pulmonary function tests (PFTs) are needed for the assessment of lung function during disease management. Purpose To assess the clinical feasibility of pulmonary three-dimensional ultrashort echo time (UTE) MRI at breath holding for quantitative image analysis of ventilation inhomogeneity and hyperinflation in CF compared with PFT. Materials and Methods In this prospective study from May 2018 to June 2019, participants with CF and healthy control participants underwent PFTs and functional lung MRI by using a prototypical single breath-hold three-dimensional UTE sequence. Fractional ventilation (FV) was calculated from acquired data in normal inspiration and normal expiration. FV of each voxel was normalized to the whole lung mean (FVN), and interquartile range of normalized ventilation (IQRN; as a measure of ventilation heterogeneity) was calculated. UTE signal intensity (SI) was assessed in full expiration (SIN, normalized to aortic blood). Obtained metrics were compared between participants with CF and control participants. For participants with CF, MRI metrics were correlated with the standard lung clearance index (LCI) and PFT. Mann-Whitney U tests and Spearman correlation were used for statistical analysis. Results Twenty participants with CF (mean age, 17 years ± 9 [standard deviation]; 12 men) and 10 healthy control participants (24 years ± 8; five men) were included. IQRN was higher for participants with CF than for control participants (mean, 0.66 ± 0.16 vs 0.50 ± 0.04, respectively; P = .007). In the 20 participants with CF, IQRN correlated with obstruction markers forced expiratory volume in 1 second-to-forced vital capacity ratio (r = -0.70; 95% confidence interval [CI]: -0.92, -0.28; P < .001), mean expiratory flow 25% (r = 0.78; 95% CI: -0.95, -0.39; P < .001), and with the ventilation inhomogeneity parameter LCI (r = 0.90; 95% CI: 0.69, 0.96; P < .001). Mean SIN in full expiration was lower in participants with CF than in control participants (0.34 ± 0.08 vs 0.39 ± 0.03, respectively; P = .03). Conclusion Three-dimensional ultrashort echo time MRI in the lungs allowed for functional imaging of ventilation inhomogeneity within a few breath holds in patients with cystic fibrosis. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Wielpütz in this issue.

Comparison of MRI and VQ-SPECT as a Screening Test for Patients With Suspected CTEPH: CHANGE-MRI Study Design and Rationale.

The diagnostic strategy for chronic thromboembolic pulmonary hypertension (CTEPH) is composed of two components required for a diagnosis of CTEPH: the presence of chronic pulmonary embolism and an elevated pulmonary artery pressure. The current guidelines require that ventilation–perfusion single-photon emission computed tomography (VQ-SPECT) is used for the first step diagnosis of chronic pulmonary embolism. However, VQ-SPECT exposes patients to ionizing radiation in a radiation sensitive population. The prospective, multicenter, comparative phase III diagnostic trial CTEPH diagnosis Europe - MRI (CHANGE-MRI, ClinicalTrials.gov identifier NCT02791282) aims to demonstrate whether functional lung MRI can serve as an equal rights alternative to VQ-SPECT in a diagnostic strategy for patients with suspected CTEPH. Positive findings are verified with catheter pulmonary angiography or computed tomography pulmonary angiography (gold standard). For comparing the imaging methods, a co-primary endpoint is used. (i) the proportion of patients with positive MRI in the group of patients who have a positive SPECT and gold standard diagnosis for chronic pulmonary embolism and (ii) the proportion of patients with positive MRI in the group of patients with negative SPECT and gold standard. The CHANGE-MRI trial will also investigate the performance of functional lung MRI without i.v. contrast agent as an index test and identify cardiac, hemodynamic, and pulmonary MRI-derived parameters to estimate pulmonary artery pressures and predict 6–12 month survival. Ultimately, this study will provide the necessary evidence for the discussion about changes in the recommendations on the diagnostic approach to CTEPH.

Effect of intravenously injected gadolinium-based contrast agents on functional lung parameters derived by PREFUL MRI.

To evaluate the influence of intravenously administered gadolinium-based contrast agents on functional ventilation and perfusion parameters derived by phase-resolved functional lung (PREFUL) MRI. Fourteen participants underwent functional MRI at 1.5T using a 2D spoiled gradient echo sequence during free breathing. Three data sets of PREFUL images were obtained-the 1st data set was acquired in mean 33:46 min (SD = 6:20 min) prior, the 2nd and 3rd data sets 43 and 91 s (both SD = 1.9 s), respectively, after i.v. application of gadobutrol. Full respiratory and cardiac cycles were reconstructed and functional parameters of regional ventilation (RV), perfusion (Q), and quantified perfusion (QQuant ) together with perfusion-defected percentages (QDP), ventilation-defected percentages (VDP), and ventilation-perfusion match (VQM) were calculated and compared for systematic differences between the acquired data sets. RV- and Q-values presented no significant alteration after gadobutrol administration. Consequently, QDP, VDP, and VQ maps were not significantly different. Sørensen-Dice coefficients of QDP and VDP maps between the different series varied up to ±9%. QQuant was significantly increased after the application of gadobutrol (1st vs. 2nd series, P = 0.0021; 1st vs. 3rd, P = 0.0188), which can be explained by the velocity-dependent signal in the completely blood-filled voxel (ROI of the aorta) after shortening of T1 relaxation time (1st vs. 2nd series, P = 0.0003; 1st vs. 3rd series, P = 0.0008). Except for quantified perfusion, all evaluated functional parameters including ventilation- and perfusion-weighted maps derived by PREFUL MRI were independent of gadolinium-based contrast agents, which is important for the design of MRI protocols in future studies.

MRI-derived regional flow-volume loop parameters detect early-stage chronic lung allograft dysfunction.

Chronic lung allograft dysfunction (CLAD) is a major cause for the low long-term survival rates after lung transplantation (LTx). Early detection of CLAD may enable providing medical treatment before a nonreversible graft dysfunction has occurred. MRI is advantageous to pulmonary function testing (PFT) in the ability to assess regional function changes, and thus have the potential in detecting very early stages of CLAD before changes in global forced expiratory volume during the first second (FEV1%) occur. To examine whether early stages of CLAD (diagnosed based on PFT values) could also be detected using MRI-derived parameters of regional flow-volume dynamics. Retrospective. 62 lung transplantation recipients were included in the study, 29 of which had been diagnosed with CLAD at various stages. MRI datasets were acquired with a 1.5T Siemens scanner using a spoiled gradient echo sequence. MRI datasets were retrospectively preprocessed and analyzed by a blinded radiologist according to the phase resolved functional lung MRI (PREFUL-MRI) approach, resulting in fractional ventilation (FV) maps and regional flow-volume loops (rFVL). FV- and rFVL-based parameters of regional lung ventilation were estimated. Differences between groups were compared by Mann-Whitney U-test with a Bonferroni correction for multiple comparisons (n = 2). rFVL-CC-based parameters discriminated significantly between the presence or absence of CLAD (P < 0.003). Using the contrast media-free PREFUL-MRI technique, parameters of ventilation dynamics and its regional heterogeneity were shown to be sensitive for the detection of early CLAD stages. 3 TECHNICAL EFFICACY: Stage 3 J. Magn. Reson. Imaging 2019;50:1873-1882.

Free-breathing quantification of regional ventilation derived by phase-resolved functional lung (PREFUL) MRI.

To test the feasibility of regional fully quantitative ventilation measurement in free breathing derived by phase-resolved functional lung (PREFUL) MRI in the supine and prone positions. In addition, the influence of T2 * relaxation time on ventilation quantification is assessed. Twelve healthy volunteers underwent functional MRI at 1.5T using a 2D triple-echo spoiled gradient echo sequence allowing for quantitative measurement of T2 * relaxation time. Minute ventilation (ΔV) was quantified by conventional fractional ventilation (FV) and the newly introduced regional ventilation (VR), which corrects volume errors due to image registration. ΔVFV versus ΔVVR and ΔVVR versus ΔVVR with T2 * correction were compared using Bland-Altman plots and correlation analysis. The repeatability and physiological plausibility of all measurements were tested in the supine and prone positions. On global and regional scales a strong correlation was observed between ΔVFV versus ΔVVR and ΔVVR versus ΔVVRT2* (r>0.93); however, regional Bland-Altman analysis showed systematic differences (p<0.0001). Unlike ΔVVRT2* , ΔVVR and ΔVFV showed expected physiologic anterior-posterior gradients, which decreased in the supine but not in the prone position at second measurement during 3min in the same position. For all quantification methods a moderate repeatability (coefficient of variation <20%) of ventilation was found. A fully quantified regional ventilation measurement using ΔVVR in free breathing is feasible and shows physiologically plausible results. In contrast to conventional ΔVFV, volume errors due to image registration are eliminated with the ΔVVR approach. However, correction for the T2 * effect remains challenging.

UTE-SENCEFUL: first results for 3D high-resolution lung ventilation imaging.

This study aimed to develop a 3D MRI technique to assess lung ventilation in free-breathing and without the administration of contrast agent. A 3D-UTE sequence with a koosh ball trajectory was developed for a 3 Tesla scanner. An oversampled k-space was acquired, and the direct current signal from the k-space center was used as a navigator to sort the acquired data into 8 individual breathing phases. Gradient delays were corrected, and iterative SENSE was used to reconstruct the individual timeframes. Subsequently, the signal changes caused by motion were eliminated using a 3D image registration technique, and ventilation-weighted maps were created by analyzing the signal changes in the lung tissue. Six healthy volunteers and 1 patient with lung cancer were scanned with the new 3D-UTE and the standard 2D technique. Image quality and quantitative ventilation values were compared between both methods. UTE-based self-gated noncontrast-enhanced functional lung (SENCEFUL) MRI provided a time-resolved reconstruction of the breathing motion, with a 49% increase of the SNR. Ventilation quantification for healthy subjects was in statistical agreement with 2D-SENCEFUL and the literature, with a mean value of 0.11 ± 0.08 mL/mL for the whole lung. UTE-SENCEFUL was able to visualize and quantify ventilation deficits in a patient with lung tumor that were not properly depicted by 2D-SENCEFUL. UTE-SENCEFUL represents a robust MRI method to assess both morphological and functional information of the lungs in 3D. When compared to the 2D approach, 3D-UTE offered ventilation maps with higher resolution, improved SNR, and reduced ventilation artifacts.

19 F MRI of the Lungs Using Inert Fluorinated Gases: Challenges and New Developments.

Fluorine-19 (19 F) MRI using inhaled inert fluorinated gases is an emerging technique that can provide functional images of the lungs. Inert fluorinated gases are nontoxic, abundant, relatively inexpensive, and the technique can be performed on any MRI scanner with broadband multinuclear imaging capabilities. Pulmonary 19 F MRI has been performed in animals, healthy human volunteers, and in patients with lung disease. In this review, the technical requirements of 19 F MRI are discussed, along with various imaging approaches used to optimize the image quality. Lung imaging is typically performed in humans using a gas mixture containing 79% perfluoropropane (PFP) or sulphur hexafluoride (SF6 ) and 21% oxygen. In lung diseases, such as asthma, chronic obstructive pulmonary disease (COPD), and cystic fibrosis (CF), ventilation defects are apparent in regions that the inhaled gas cannot access. 19 F lung images are typically acquired in a single breath-hold, or in a time-resolved, multiple breath fashion. The former provides measurements of the ventilation defect percent (VDP), while the latter provides measurements of gas replacement (ie, fractional ventilation). Finally, preliminary comparisons with other functional lung imaging techniques are discussed, such as Fourier decomposition MRI and hyperpolarized gas MRI. Overall, functional 19 F lung MRI is expected to complement existing proton-based structural imaging techniques, and the combination of structural and functional lung MRI will provide useful outcome measures in the future management of pulmonary diseases in the clinic. Level of Evidence: 3 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;49:343-354.

Convolutional Neural Networks with Template-Based Data Augmentation for Functional Lung Image Quantification

We propose an automated segmentation pipeline based on deep learning for proton lung MRI segmentation and ventilation-based quantification which improves on our previously reported methodologies in terms of computational efficiency while demonstrating accuracy and robustness. The large data requirement for the proposed framework is made possible by a novel template-based data augmentation strategy. Supporting this work is the open-source ANTsRNet-a growing repository of well-known deep learning architectures first introduced here. Deep convolutional neural network (CNN) models were constructed and trained using a custom multilabel Dice metric loss function and a novel template-based data augmentation strategy. Training (including template generation and data augmentation) employed 205 proton MR images and 73 functional lung MRI. Evaluation was performed using data sets of size 63 and 40 images, respectively. Accuracy for CNN-based proton lung MRI segmentation (in terms of Dice overlap) was left lung: 0.93 ± 0.03, right lung: 0.94 ± 0.02, and whole lung: 0.94 ± 0.02. Although slightly less accurate than our previously reported joint label fusion approach (left lung: 0.95 ± 0.02, right lung: 0.96 ± 0.01, and whole lung: 0.96 ± 0.01), processing time is <1 second per subject for the proposed approach versus ∼30 minutes per subject using joint label fusion. Accuracy for quantifying ventilation defects was determined based on a consensus labeling where average accuracy (Dice multilabel overlap of ventilation defect regions plus normal region) was 0.94 for the CNN method; 0.92 for our previously reported method; and 0.90, 0.92, and 0.94 for expert readers. The proposed framework yields accurate automated quantification in near real time. CNNs drastically reduce processing time after offline model construction and demonstrate significant future potential for facilitating quantitative analysis of functional lung MRI.

Functional lung MRI for regional monitoring of patients with cystic fibrosis.

PurposeTo test quantitative functional lung MRI techniques in young adults with cystic fibrosis (CF) compared to healthy volunteers and to monitor immediate treatment effects of a single inhalation of hypertonic saline in comparison to clinical routine pulmonary function tests.Materials and methodsSixteen clinically stable CF patients and 12 healthy volunteers prospectively underwent two functional lung MRI scans and pulmonary function tests before and 2h after a single treatment of inhaled hypertonic saline or without any treatment. MRI-derived oxygen enhanced T1 relaxation measurements, fractional ventilation, first-pass perfusion parameters and a morpho-functional CF-MRI score were acquired.ResultsCompared to healthy controls functional lung MRI detected and quantified significantly increased ventilation heterogeneity in CF patients. Regional functional lung MRI measures of ventilation and perfusion as well as the CF-MRI score and pulmonary function tests could not detect a significant treatment effect two hours after a single treatment with hypertonic saline in young adults with CF (p>0.05).ConclusionThis study shows the feasibility of functional lung MRI as a non-invasive, radiation-free tool for monitoring patients with CF.

Feasibility of quantitative regional ventilation and perfusion mapping with phase-resolved functional lung (PREFUL) MRI in healthy volunteers and COPD, CTEPH, and CF patients.

In this feasibility study, a phase-resolved functional lung imaging postprocessing method for extraction of dynamic perfusion (Q) and ventilation (V) parameters using a conventional 1H lung MRI Fourier decomposition acquisition is introduced. Time series of coronal gradient-echo MR images with a temporal resolution of 288 to 324 ms of two healthy volunteers, one patient with chronic thromboembolic hypertension, one patient with cystic fibrosis, and one patient with chronic obstructive pulmonary disease were acquired at 1.5 T. Using a sine model to estimate cardiac and respiratory phases of each image, all images were sorted to reconstruct full cardiac and respiratory cycles. Time to peak (TTP), V/Q maps, and fractional ventilation flow-volume loops were calculated. For the volunteers, homogenous ventilation and perfusion TTP maps (V-TTP, Q-TTP) were obtained. The chronic thromboembolic hypertension patient showed increased perfusion TTP in hypoperfused regions in visual agreement with dynamic contrast-enhanced MRI, which improved postpulmonary endaterectomy surgery. Cystic fibrosis and chronic obstructive pulmonary disease patients showed a pattern of increased V-TTP and Q-TTP in regions of hypoventilation and decreased perfusion. Fractional ventilation flow-volume loops of the chronic obstructive pulmonary disease patient were smaller in comparison with the healthy volunteer, and showed regional differences in visual agreement with functional small airways disease and emphysema on CT. This study shows the feasibility of phase-resolved functional lung imaging to gain quantitative information regarding regional lung perfusion and ventilation without the need for ultrafast imaging, which will be advantageous for future clinical translation. Magn Reson Med 79:2306-2314, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Non‐contrast‐enhanced perfusion and ventilation assessment of the human lung by means of fourier decomposition in proton MRI

Assessment of regional lung perfusion and ventilation has significant clinical value for the diagnosis and follow-up of pulmonary diseases. In this work a new method of non-contrast-enhanced functional lung MRI (not dependent on intravenous or inhalative contrast agents) is proposed. A two-dimensional (2D) true fast imaging with steady precession (TrueFISP) pulse sequence (TR/TE = 1.9 ms/0.8 ms, acquisition time [TA] = 112 ms/image) was implemented on a 1.5T whole-body MR scanner. The imaging protocol comprised sets of 198 lung images acquired with an imaging rate of 3.33 images/s in coronal and sagittal view. No electrocardiogram (ECG) or respiratory triggering was used. A nonrigid image registration algorithm was applied to compensate for respiratory motion. Rapid data acquisition allowed observing intensity changes in corresponding lung areas with respect to the cardiac and respiratory frequencies. After a Fourier analysis along the time domain, two spectral lines corresponding to both frequencies were used to calculate the perfusion- and ventilation-weighted images. The described method was applied in preliminary studies on volunteers and patients showing clinical relevance to obtain non-contrast-enhanced perfusion and ventilation data.

Plasma Oxidized Low-Density Lipoprotein, a Strong Predictor for Acute Coronary Heart Disease Events in Apparently Healthy, Middle-Aged Men From the General Population

We propose an automated segmentation pipeline based on deep learning for proton lung MRI segmentation and ventilation-based quantification which improves on our previously reported methodologies in terms of computational efficiency while demonstrating accuracy and robustness. The large data requirement for the proposed framework is made possible by a novel template-based data augmentation strategy. Supporting this work is the open-source ANTsRNet—a growing repository of well-known deep learning architectures first introduced here.Deep convolutional neural network (CNN) models were constructed and trained using a custom multilabel Dice metric loss function and a novel template-based data augmentation strategy. Training (including template generation and data augmentation) employed 205 proton MR images and 73 functional lung MRI. Evaluation was performed using data sets of size 63 and 40 images, respectively.Accuracy for CNN-based proton lung MRI segmentation (in terms of Dice overlap) was left lung: 0.93 ± 0.03, right lung: 0.94 ± 0.02, and whole lung: 0.94 ± 0.02. Although slightly less accurate than our previously reported joint label fusion approach (left lung: 0.95 ± 0.02, right lung: 0.96 ± 0.01, and whole lung: 0.96 ± 0.01), processing time is <1 second per subject for the proposed approach versus ∼30 minutes per subject using joint label fusion. Accuracy for quantifying ventilation defects was determined based on a consensus labeling where average accuracy (Dice multilabel overlap of ventilation defect regions plus normal region) was 0.94 for the CNN method; 0.92 for our previously reported method; and 0.90, 0.92, and 0.94 for expert readers.The proposed framework yields accurate automated quantification in near real time. CNNs drastically reduce processing time after offline model construction and demonstrate significant future potential for facilitating quantitative analysis of functional lung MRI.