The impact of segmentation on whole-lung functional MRI quantification: Repeatability and reproducibility from multiple human observers and an artificial neural network

Neural network segmentation of disc volume from magnetic resonance images and the effect of degeneration and spinal level

Background

Magnetic resonance imaging (MRI) noninvasively quantifies disc structure but requires segmentation that is both time intensive and susceptible to human error. Recent advances in neural networks can improve on manual segmentation. The aim of this study was to establish a method for automatic slice-wise segmentation of 3D disc volumes from subjects with a wide range of age and degrees of disc degeneration. A U-Net convolutional neural network was trained to segment 3D T1-weighted spine MRI.

Methods

Lumbar spine MRIs were acquired from 43 subjects (23-83 years old) and manually segmented. A U-Net architecture was trained using the TensorFlow framework. Two rounds of model tuning were performed. The performance of the model was measured using a validation set that did not cross over from the training set. The model version with the best Dice similarity coefficient (DSC) was selected in each tuning round. After model development was complete and a final U-Net model was selected, performance of this model was compared between disc levels and degeneration grades.

Results

Performance of the final model was equivalent to manual segmentation, with a mean DSC = 0.935 ± 0.014 for degeneration grades I-IV. Neither the manual segmentation nor the U-Net model performed as well for grade V disc segmentation. Compared with the baseline model at the beginning of round 1, the best model had fewer filters/parameters (75%), was trained using only slices with at least one disc-labeled pixel, applied contrast stretching to its input images, and used a greater dropout rate.

Conclusion

This study successfully trained a U-Net model for automatic slice-wise segmentation of 3D disc volumes from populations with a wide range of ages and disc degeneration. The final trained model is available to support scientific use.

Lung structural and functional impairments in young children with cystic fibrosis diagnosed following newborn screening - A nationwide observational study

BACKGROUND

Non-invasive and sensitive clinical endpoints are needed to monitor onset and progression of early lung disease in children with cystic fibrosis (CF). We compared lung clearance index (LCI), FEV1, functional and structural lung magnetic resonance imaging (MRI) outcomes in Swiss children with CF diagnosed following newborn screening.

METHODS

Lung function (LCI, FEV1) and unsedated functional and structural lung MRI was performed in 79 clinically stable children with CF (3 - 8 years) and 75 age-matched healthy controls. Clinical information was collected throughout childhood.

RESULTS

LCI, ventilation and perfusion defects, and structural MRI scores were significantly higher in children with CF compared with controls, but FEV1 was not different between groups. Lung MRI outcomes correlated significantly with LCI (morphology score (r = 0.56, p < 0.001); ventilation defects (r = 0.43, p = 0.001); perfusion defects (r = 0.64, p < 0.001), but not with FEV1. Lung MRI outcomes were more sensitive to detect impairments in children with CF (abnormal ventilation and perfusion outcomes in 47 %, morphology score in 30 %) compared with lung function (abnormal LCI in 21 % and FEV1 in 4.8 %). Pulmonary exacerbations, respiratory hospitalizations, and increase in patient-reported cough was associated with higher LCI and higher structural and functional MRI outcomes.

CONCLUSIONS

The LCI and lung MRI outcomes non-invasively detect even mild early lung disease in young children with CF diagnosed following newborn screening. Pulmonary exacerbations and early respiratory symptoms were risk factors for structural and functional impairment in childhood.

Long-term pulmonary outcome of children with congenital diaphragmatic hernia: functional lung MRI using matrix-pencil decomposition enables side-specific assessment of lung function

OBJECTIVES

In patients with congenital diaphragmatic hernia (CDH) the exact functional outcome of the affected lung side is still unknown, mainly due to the lack of spatially resolved diagnostic tools. Functional matrix-pencil decomposition (MP-) lung MRI fills this gap as it measures side-specific ventilation and perfusion. We aimed to assess the overall and side-specific pulmonary long-term outcomes of patients with CDH using lung function tests and MP-MRI.

METHODS

Thirteen school-aged children with CDH (seven with small and six with large defect-sized CDH, defined as > 50% of the chest wall circumference being devoid of diaphragm tissue) and thirteen healthy matched controls underwent spirometry, multiple-breath washout, and MP-MRI. The main outcomes were forced expiratory volume in 1 second (FEV1), lung clearance index (LCI2.5), ventilation defect percentage (VDP), and perfusion defect percentage (QDP).

RESULTS

Patients with a large CDH showed significantly reduced overall lung function compared to healthy controls (mean difference [95%-CIadjusted]: FEV1 (z-score) -4.26 [-5.61, -2.92], FVC (z-score) -3.97 [-5.68, -2.26], LCI2.5 (TO) 1.12 [0.47, 1.76], VDP (%) 8.59 [3.58, 13.60], QDP (%) 17.22 [13.16, 21.27]) and to patients with a small CDH. Side-specific examination by MP-MRI revealed particularly reduced ipsilateral ventilation and perfusion in patients with a large CDH (mean difference to contralateral side [95%-CIadjusted]: VDP (%) 14.80 [10.50, 19.00], QDP (%) 23.50 [1.75, 45.20]).

CONCLUSIONS

Data indicate impaired overall lung function with particular limitation of the ipsilateral side in patients with a large CDH. MP-MRI is a promising tool to provide valuable side-specific functional information in the follow-up of patients with CDH.

CLINICAL RELEVANCE STATEMENT

In patients with congenital diaphragmatic hernia, easily applicable MP-MRI allows specific examination of the lung side affected by the hernia and provides valuable information on ventilation and perfusion with implications for clinical practice, making it a promising tool for routine follow-up.

KEY POINTS

• Functional matrix pencil decomposition (MP) MRI data from a small sample indicate reduced ipsilateral pulmonary ventilation and perfusion in children with large congenital diaphragmatic hernia (CDH). • Easily applicable pencil decomposition MRI provides valuable side-specific diagnostic information on lung ventilation and perfusion. This is a clear advantage over conventional lung function tests, helping to comprehensively follow up patients with congenital diaphragmatic hernia and monitor therapy effects.

How many specimens make a sufficient training set for automated three-dimensional feature extraction?

Deep learning has emerged as a robust tool for automating feature extraction from three-dimensional images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including assessing whether artficial expansion by data augmentation can achieve consistent results in less time and how consistent these benefits are across different types of traits. In this study, we manually segmented 50 planktonic foraminifera specimens from the genus Menardella to determine the minimum number of training images required to produce accurate volumetric and shape data from internal and external structures. The results reveal unsurprisingly that deep learning models improve with a larger number of training images with eight specimens being required to achieve 95% accuracy. Furthermore, data augmentation can enhance network accuracy by up to 8.0%. Notably, predicting both volumetric and shape measurements for the internal structure poses a greater challenge compared with the external structure, owing to low contrast differences between different materials and increased geometric complexity. These results provide novel insight into optimal training set sizes for precise image segmentation of diverse traits and highlight the potential of data augmentation for enhancing multivariate feature extraction from three-dimensional images.

Interactive Segmentation of Lung Tissue and Lung Excursion in Thoracic Dynamic MRI Based on Shape-guided Convolutional Neural Networks

Purpose

Lung tissue and lung excursion segmentation in thoracic dynamic magnetic resonance imaging (dMRI) is a critical step for quantitative analysis of thoracic structure and function in patients with respiratory disorders such as Thoracic Insufficiency Syndrome (TIS). However, the complex variability of intensity and shape of anatomical structures and the low contrast between the lung and surrounding tissue in MR images seriously hamper the accuracy and robustness of automatic segmentation methods. In this paper, we develop an interactive deep-learning based segmentation system to solve this problem.

Material & Methods

Considering the significant difference in lung morphological characteristics between normal subjects and TIS subjects, we utilized two independent data sets of normal subjects and TIS subjects to train and test our model. 202 dMRI scans from 101 normal pediatric subjects and 92 dMRI scans from 46 TIS pediatric subjects were acquired for this study and were randomly divided into training, validation, and test sets by an approximate ratio of 5:1:4. First, we designed an interactive region of interest (ROI) strategy to detect the lung ROI in dMRI for accelerating the training speed and reducing the negative influence of tissue located far away from the lung on lung segmentation. Second, we utilized a modified 2D U-Net to segment the lung tissue in lung ROIs, in which the adjacent slices are utilized as the input data to take advantage of the spatial information of the lungs. Third, we extracted the lung shell from the lung segmentation results as the shape feature and inputted the lung ROIs with shape feature into another modified 2D U-Net to segment the lung excursion in dMRI. To evaluate the performance of our approach, we computed the Dice coefficient (DC) and max-mean Hausdorff distance (MM-HD) between manual and automatic segmentations. In addition, we utilized Coefficient of Variation (CV) to assess the variability of our method on repeated dMRI scans and the differences of lung tidal volumes computed from the manual and automatic segmentation results.

Results

The proposed system yielded mean Dice coefficients of 0.96±0.02 and 0.89±0.05 for lung segmentation in dMRI of normal subjects and TIS subjects, respectively, demonstrating excellent agreement with manual delineation results. The Coefficient of Variation and p-values show that the estimated lung tidal volumes of our approach are statistically indistinguishable from those derived by manual segmentations.

Conclusions

The proposed approach can be applied to lung tissue and lung excursion segmentation from dynamic MR images with high accuracy and efficiency. The proposed approach has the potential to be utilized in the assessment of patients with TIS via dMRI routinely.

Automatic lung segmentation of magnetic resonance images: A new approach applied to healthy volunteers undergoing enhanced Deep-Inspiration-Breath-Hold for motion-mitigated 4D proton therapy of lung tumors

Background and purpose

Respiratory suppression techniques represent an effective motion mitigation strategy for 4D-irradiation of lung tumors with protons. A magnetic resonance imaging (MRI)-based study applied and analyzed methods for this purpose, including enhanced Deep-Inspiration-Breath-Hold (eDIBH). Twenty-one healthy volunteers (41-58 years) underwent thoracic MR scans in four imaging sessions containing two eDIBH-guided MRIs per session to simulate motion-dependent irradiation conditions. The automated MRI segmentation algorithm presented here was critical in determining the lung volumes (LVs) achieved during eDIBH.

Materials and methods

The study included 168 MRIs acquired under eDIBH conditions. The lung segmentation algorithm consisted of four analysis steps: (i) image preprocessing, (ii) MRI histogram analysis with thresholding, (iii) automatic segmentation, (iv) 3D-clustering. To validate the algorithm, 46 eDIBH-MRIs were manually contoured. Sørensen-Dice similarity coefficients (DSCs) and relative deviations of LVs were determined as similarity measures. Assessment of intrasessional and intersessional LV variations and their differences provided estimates of statistical and systematic errors.

Results

Lung segmentation time for 100 2D-MRI planes was ∼ 10 s. Compared to manual lung contouring, the median DSC was 0.94 with a lower 95 % confidence level (CL) of 0.92. The relative volume deviations yielded a median value of 0.059 and 95 % CLs of -0.013 and 0.13. Artifact-based volume errors, mainly of the trachea, were estimated. Estimated statistical and systematic errors ranged between 6 and 8 %.

Conclusions

The presented analytical algorithm is fast, precise, and readily available. The results are comparable to time-consuming, manual segmentations and other automatic segmentation approaches. Post-processing to remove image artifacts is under development.

Effects of elexacaftor/tezacaftor/ivacaftor therapy in children with cystic fibrosis - a comprehensive assessment using lung clearance index, spirometry, and functional and structural lung MRI

BACKGROUND

With improvement in supportive therapies and the introduction of cystic fibrosis transmembrane conductance regulator (CFTR)-modulator treatment in patients with cystic fibrosis (CF), milder disease courses are expected. Therefore, sensitive parameters are needed to monitor disease course and effects of CFTR-modulators. Functional lung MRI using matrix-pencil decomposition (MP-MRI) is a promising tool for assessing ventilation and perfusion quantitatively. This study aimed to assess the treatment effect of elexacaftor/tezacaftor/ivacaftor combination regimen (ELX/TEZ/IVA) on measures of structural and functional lung abnormalities.

METHODS

24 children with CF underwent lung function tests (multiple breath washout, spirometry), functional and structural MRI twice (one year apart) before and once after at least two weeks (mean 4.7 ± 2.6 months) on ELX/TEZ/IVA. Main outcomes were changes (Δ) upon ELX/TEZ/IVA in lung function, defect percentage of ventilation (VDP) and perfusion (QDP), defect distribution index of ventilation and perfusion (DDIV, DDIQ), and Eichinger score. Statistical analyses were performed using paired t-tests and multilevel regression models with bootstrapping.

RESULTS

We observed a significant improvement in lung function, structural and functional MRI parameters upon ELX/TEZ/IVA treatment (mean; 95%-CI): ΔLCI2.5 (TO) -0.84 (-1.62 to -0.06); ΔFEV1 (z-score) 1.05 (0.56 to 1.55); ΔVDP (% of impairment) -6.00 (-8.44 to -3.55); ΔQDP (% of impairment) -3.90 (-5.90 to -1.90); ΔDDIV -1.38 (-2.22 to -0.53); ΔDDIQ -0.31 (-0.73 to 0.12); ΔEichinger score -3.89 (-5.05 to -2.72).

CONCLUSIONS

Besides lung function tests, functional and structural MRI is a suitable tool to monitor treatment response of ELX/TEZ/IVA therapy, and seems promising as outcome marker in the future.

Reproducibility of non-contrast enhanced multi breath-hold ultrashort echo time functional lung MRI

PURPOSE

To evaluate the intraindividual reproducibility of functional lung imaging using non-contrast enhanced multi breath-hold 3D-UTE MRI.

METHODS

Ten healthy volunteers underwent non-contrast enhanced 3D-UTE MRI at three time points for same-day and different-day measurements employing a stack-of-spirals trajectory at 3 T. At each time point, inspiratory and expiratory breathing states were acquired for tidal and deep breathing, each within a single breath-hold. For functional image analysis, fractional ventilation (FV) was calculated pixelwise after image registration from the MR signal change. To decouple FV from breathing depth, the individual lung volume was used for volume adjustment (rFV). Reproducibility evaluation was performed in eight lung segments. Statistical analyses included two way mixed intraclass correlation (ICC), sign-test, Friedman-test and modified Bland-Altman analyses.

RESULTS

FV from tidal breathing showed an ICC of 0.81, a bias of 1.3% and an interval of confidence (CI) ranging from -67.1 to 69.6%. FV from deep breathing was higher reproducible with an ICC of 0.92 (bias, -0.2%; CI, -34.2 to 33.7%). Following volume adjustment, reproducibility of rFV for tidal breathing improved (ICC, 0,86; bias, 2.0%; CI, -34.3 to 38.3%), whereas it did not bear significant benefits for deep breathing (ICC, 0.89; bias, 2.8%; CI, -24.9 to 30.5%). Reproducibility was independent from the examination day.

CONCLUSION

Non-contrast-enhanced multi breath-hold 3D-UTE MRI allows for highly reproducible ventilation imaging.

Ventilation and perfusion MRI at a 0.35 T MR-Linac: feasibility and reproducibility study

BACKGROUND

Hybrid devices that combine radiation therapy and MR-imaging have been introduced in the clinical routine for the treatment of lung cancer. This opened up not only possibilities in terms of accurate tumor tracking, dose delivery and adapted treatment planning, but also functional lung imaging. The aim of this study was to show the feasibility of Non-uniform Fourier Decomposition (NuFD) MRI at a 0.35 T MR-Linac as a potential treatment response assessment tool, and propose two signal normalization strategies for enhancing the reproducibility of the results.

METHODS

Ten healthy volunteers (median age 28 ± 8 years, five female, five male) were repeatedly scanned at a 0.35 T MR-Linac using an optimized 2D+t balanced steady-state free precession (bSSFP) sequence for two coronal slice positions. Image series were acquired in normal free breathing with breaks inside and outside the scanner as well as deep and shallow breathing. Ventilation- and perfusion-weighted maps were generated for each image series using NuFD. For intra-volunteer ventilation map reproducibility, a normalization factor was defined based on the linear correlation of the ventilation signal and diaphragm position of each scan as well as the diaphragm motion amplitude of a reference scan. This allowed for the correction of signal dependency on the diaphragm motion amplitude, which varies with breathing patterns. The second strategy, which can be used for ventilation and perfusion, eliminates the dependency on the signal amplitude by normalizing the ventilation/perfusion maps with the average ventilation/perfusion signal within a selected region-of-interest (ROI). The position and size dependency of this ROI was analyzed. To evaluate the performance of both approaches, the normalized ventilation/perfusion-weighted maps were compared and the deviation of the mean ventilation/perfusion signal from the reference was calculated for each scan. Wilcoxon signed-rank tests were performed to test whether the normalization methods can significantly improve the reproducibility of the ventilation/perfusion maps.

RESULTS

The ventilation- and perfusion-weighted maps generated with the NuFD algorithm demonstrated a mostly homogenous distribution of signal intensity as expected for healthy volunteers regardless of the breathing maneuver and slice position. Evaluation of the ROI's size and position dependency showed small differences in the performance. Applying both normalization strategies improved the reproducibility of the ventilation by reducing the median deviation of all scans to 9.1%, 5.7% and 8.6% for the diaphragm-based, the best and worst performing ROI-based normalization, respectively, compared to 29.5% for the non-normalized scans. The significance of this improvement was confirmed by the Wilcoxon signed rank test with [Formula: see text] at [Formula: see text]. A comparison of the techniques against each other revealed a significant difference in the performance between best ROI-based normalization and worst ROI ([Formula: see text]) and between best ROI-based normalization and scaling factor ([Formula: see text]), but not between scaling factor and worst ROI ([Formula: see text]). Using the ROI-based approach for the perfusion-maps, the uncorrected deviation of 10.2% was reduced to 5.3%, which was shown to be significant ([Formula: see text]).

CONCLUSIONS

Using NuFD for non-contrast enhanced functional lung MRI at a 0.35 T MR-Linac is feasible and produces plausible ventilation- and perfusion-weighted maps for volunteers without history of chronic pulmonary diseases utilizing different breathing patterns. The reproducibility of the results in repeated scans significantly benefits from the introduction of the two normalization strategies, making NuFD a potential candidate for fast and robust early treatment response assessment of lung cancer patients during MR-guided radiotherapy.

Artificially-generated consolidations and balanced augmentation increase performance of U-net for lung parenchyma segmentation on MR images

PURPOSE

To improve automated lung segmentation on 2D lung MR images using balanced augmentation and artificially-generated consolidations for training of a convolutional neural network (CNN).

MATERIALS AND METHODS

From 233 healthy volunteers and 100 patients, 1891 coronal MR images were acquired. Of these, 1666 images without consolidations were used to build a binary semantic CNN for lung segmentation and 225 images (187 without consolidations, 38 with consolidations) were used for testing. To increase CNN performance of segmenting lung parenchyma with consolidations, balanced augmentation was performed and artificially-generated consolidations were added to all training images. The proposed CNN (CNNBal/Cons) was compared to two other CNNs: CNNUnbal/NoCons-without balanced augmentation and artificially-generated consolidations and CNNBal/NoCons-with balanced augmentation but without artificially-generated consolidations. Segmentation results were assessed using Sørensen-Dice coefficient (SDC) and Hausdorff distance coefficient.

RESULTS

Regarding the 187 MR test images without consolidations, the mean SDC of CNNUnbal/NoCons (92.1 ± 6% (mean ± standard deviation)) was significantly lower compared to CNNBal/NoCons (94.0 ± 5.3%, P = 0.0013) and CNNBal/Cons (94.3 ± 4.1%, P = 0.0001). No significant difference was found between SDC of CNNBal/Cons and CNNBal/NoCons (P = 0.54). For the 38 MR test images with consolidations, SDC of CNNUnbal/NoCons (89.0 ± 7.1%) was not significantly different compared to CNNBal/NoCons (90.2 ± 9.4%, P = 0.53). SDC of CNNBal/Cons (94.3 ± 3.7%) was significantly higher compared to CNNBal/NoCons (P = 0.0146) and CNNUnbal/NoCons (P = 0.001).

CONCLUSIONS

Expanding training datasets via balanced augmentation and artificially-generated consolidations improved the accuracy of CNNBal/Cons, especially in datasets with parenchymal consolidations. This is an important step towards a robust automated postprocessing of lung MRI datasets in clinical routine.

MRI lung lobe segmentation in pediatric cystic fibrosis patients using a recurrent neural network trained with publicly accessible CT datasets

PURPOSE

To introduce a widely applicable workflow for pulmonary lobe segmentation of MR images using a recurrent neural network (RNN) trained with chest CT datasets. The feasibility is demonstrated for 2D coronal ultrafast balanced SSFP (ufSSFP) MRI.

METHODS

Lung lobes of 250 publicly accessible CT datasets of adults were segmented with an open-source CT-specific algorithm. To match 2D ufSSFP MRI data of pediatric patients, both CT data and segmentations were translated into pseudo-MR images that were masked to suppress anatomy outside the lung. Network-1 was trained with pseudo-MR images and lobe segmentations and then applied to 1000 masked ufSSFP images to predict lobe segmentations. These outputs were directly used as targets to train Network-2 and Network-3 with non-masked ufSSFP data as inputs, as well as an additional whole-lung mask as input for Network-2. Network predictions were compared to reference manual lobe segmentations of ufSSFP data in 20 pediatric cystic fibrosis patients. Manual lobe segmentations were performed by splitting available whole-lung segmentations into lobes.

RESULTS

Network-1 was able to segment the lobes of ufSSFP images, and Network-2 and Network-3 further increased segmentation accuracy and robustness. The average all-lobe Dice similarity coefficients were 95.0 ± 2.8 (mean ± pooled SD [%]) and 96.4 ± 2.5, 93.0 ± 2.0; and the average median Hausdorff distances were 6.1 ± 0.9 (mean ± SD [mm]), 5.3 ± 1.1, 7.1 ± 1.3 for Network-1, Network-2, and Network-3, respectively.

CONCLUSION

Recurrent neural network lung lobe segmentation of 2D ufSSFP imaging is feasible, in good agreement with manual segmentations. The proposed workflow might provide access to automated lobe segmentations for various lung MRI examinations and quantitative analyses.

Do clinimetric properties of LCI change after correction of signal processing?

BACKGROUND

The recently described sensor-crosstalk error in the multiple-breath washout (MBW) device Exhalyzer D (Eco Medics AG) could highly influence clinimetric properties and the current interpretation of MBW results. This study reanalyzes MBW data from clinical routine in the corrected software version Spiroware® 3.3.1 and evaluates the effect on outcomes.

METHODS

We included nitrogen-MBW data from healthy children and children with cystic fibrosis (CF) from previously published trials and ongoing cohort studies. We specifically compared lung clearance index (LCI) analyzed in Spiroware 3.2.1 and 3.3.1 with regard to (i) feasibility, (ii) repeatability, and (iii) validity as outcome parameters in children with CF.

RESULTS

(i) All previously collected measurements could be reanalyzed and resulted in unchanged feasibility in Spiroware 3.3.1. (ii) Short- and midterm repeatability of LCI was similar in both software versions. (iii) Clinical validity of LCI remained similar in Spiroware 3.3.1; however, this resulted in lower values. Discrimination between health and disease was comparable between both software versions. The increase in LCI over time was less pronounced with 0.16 LCI units/year (95% confidence interval [CI] 0.08; 0.24) versus 0.30 LCI units/year (95% CI 0.21; 0.38) in 3.2.1. Response to intervention in children receiving CF transmembrane conductance-modulator therapy resulted in a comparable improvement in LCI, in both Spiroware versions.

CONCLUSION

Our study confirms that clinimetric properties of LCI remain unaffected after correction for the cross-sensitivity error in Spiroware software.

MRI Shows Lung Perfusion Changes after Vaping and Smoking

Background Evidence regarding short-term effects of electronic nicotine delivery systems (ENDS) and tobacco smoke on lung ventilation and perfusion is limited. Purpose To examine the immediate effect of ENDS exposure and tobacco smoke on lung ventilation and perfusion by functional MRI and lung function tests. Materials and Methods This prospective observational pilot study was conducted from November 2019 to September 2021 (substudy of randomized controlled trial NCT03589989). Included were 44 healthy adult participants (10 control participants, nine former tobacco smokers, 13 ENDS users, and 12 active tobacco smokers; mean age, 41 years ± 12 [SD]; 28 men) who underwent noncontrast-enhanced matrix pencil MRI and lung function tests before and immediately after the exposure to ENDS products or tobacco smoke. Baseline measurements were acquired after 2 hours of substance abstinence. Postexposure measurements were performed immediately after the exposure. MRI showed semiquantitative measured impairment of lung perfusion (R_Q) and fractional ventilation (R_FV) impairment as percentages of affected lung volume. Lung clearance index (LCI) was assessed by nitrogen multiple-breath washout to capture ventilation inhomogeneity and spirometry to assess airflow limitation. Absolute differences were calculated with paired Wilcoxon signed-rank test and differences between groups with unpaired Mann-Whitney test. Healthy control participants underwent two consecutive MRI measurements to assess MRI reproducibility. Results MRI was performed and lung function measurement was acquired in tobacco smokers and ENDS users before and after exposure. MRI showed a decrease of perfusion after exposure (R_Q, 8.6% [IQR, 7.2%-10.0%] to 9.1% [IQR, 7.8%-10.7%]; P = .03) and no systematic change in R_FV (P = .31) among tobacco smokers. Perfusion increased in participants who used ENDS after exposure (R_Q, 9.7% [IQR, 7.1%-10.9%] to 9.0% [IQR, 6.9%-10.0%]; P = .01). R_FV did not change (P = .38). Only in tobacco smokers was LCI elevated after smoking (P = .02). Spirometry indexes did not change in any participants. Conclusion MRI showed a decrease of lung perfusion after exposure to tobacco smoke and an increase of lung perfusion after use of electronic nicotine delivery systems. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Kligerman in this issue.

School-age structural and functional MRI and lung function in children following lung resection for congenital lung malformation in infancy

BACKGROUND

The management of asymptomatic congenital lung malformations is debated. Particularly, there is a lack of information regarding long-term growth and development of the remaining lung in children following lung resection for congenital lung malformations. In addition to conventional pulmonary function tests, we used novel functional magnetic resonance imaging (MRI) methods to measure perfusion and ventilation.

OBJECTIVE

To assess functionality of the remaining lung expanded into the thoracic cavity after resection of congenital lung malformations.

MATERIALS AND METHODS

A prospective, cross-sectional pilot study in five children who had surgery for congenital lung malformations during infancy. Participants had structural and functional MRI as well as spirometry, body plethysmography and multiple breath washout at school age.

RESULTS

Structural MRI showed an expansion of the remaining lung in all cases. Fractional ventilation and relative perfusion of the expanded lung were locally decreased in functional MRI. In all other parts of the lungs, fractional ventilation and relative perfusion were normal in all children. There was an association between overall impairment of perfusion and elevated lung clearance index. The results of spirometry and body plethysmography varied between patients, including normal lung function, restriction and obstruction.

CONCLUSION

Fractional ventilation and relative perfusion maps from functional MRI specifically locate impairment of the remaining lung after lung resection. These changes are not captured by conventional measures such as structural MRI and standard pulmonary function tests. Therefore, following lung resection for congenital lung malformation, children should be investigated more systematically with functional lung MRI.

Effect of Salbutamol on Lung Ventilation in Children with Cystic Fibrosis: Comprehensive Assessment Using Spirometry, Multiple-Breath Washout, and Functional Lung Magnetic Resonance Imaging

BACKGROUND

Inhalation therapy is one of the cornerstones of the daily treatment regimen in patients with cystic fibrosis (CF). Recommendations regarding the addition of bronchodilators, especially salbutamol are conflicting due to the lack of evidence. New diagnostic measures such as multiple-breath washout (<underline>MBW)</underline> and functional magnetic resonance imaging (MRI) have the potential to reveal new insights into bronchodilator effects in patients with CF.

OBJECTIVE

The objective of the study was to comprehensively assess the functional response to nebulized inhalation with salbutamol in children with CF.

METHODS

Thirty children aged 6-18 years with stable CF performed pulmonary function tests, MBW, and matrix pencil-MRI before and after standardized nebulized inhalation of salbutamol.

RESULTS

Lung clearance index decreased (improved) by -0.24 turnover (95% confidence interval [CI]: -0.53 to 0.06; p = 0.111). Percentage of the lung volume with impaired fractional ventilation and relative perfusion decreased (improved) by -0.79% (CI: -1.99 to 0.42; p = 0.194) and -1.31% (CI: -2.28 to -0.35; p = 0.009), respectively. Forced expiratory volume (FEV1) increased (improved) by 0.41 z-score (CI: 0.24-0.58; p < 0.0001). We could not identify specific clinical factors associated with a more pronounced effect of salbutamol.

CONCLUSIONS

There is a positive short-term effect of bronchodilator inhalation on FEV1 in patients with CF, which is independent of ventilation inhomogeneity. Heterogeneous response between patients suggests that for prediction of a therapeutic effect this should be tested by spirometry in every patient individually.

Defect distribution index: A novel metric for functional lung MRI in cystic fibrosis

Purpose

Lung impairment from functional MRI is frequently assessed as defect percentage. The defect distribution, however, is currently not quantified. The purpose of this work was to develop a novel measure that quantifies how clustered or scattered defects in functional lung MRI appear, and to evaluate it in pediatric cystic fibrosis.

Theory

The defect distribution index (DDI) calculates a score for each lung voxel categorized as defected. The index increases according to how densely and how far an expanding circle around a defect voxel contains more than 50% defect voxels.

Methods

Fractional ventilation and perfusion maps of 53 children with cystic fibrosis were previously acquired with matrix pencil decomposition MRI. In this work, the DDI is compared to a visual score of 3 raters who evaluated how clustered the lung defects appear. Further, spearman correlations between DDI and lung function parameters were determined.

Results

The DDI strongly correlates with the visual scoring (r = 0.90 for ventilation; r = 0.88 for perfusion; P < .0001). Although correlations between DDI and defect percentage are moderate to strong (r = 0.61 for ventilation; r = 0.75 for perfusion; P < .0001), the DDI distinguishes between patients with comparable defect percentage.

Conclusion

The DDI is a novel measure for functional lung MRI. It provides complementary information to the defect percentage because the DDI assesses defect distribution rather than defect size. The DDI is applicable to matrix pencil MRI data of cystic fibrosis patients and shows very good agreement with human perception of defect distributions.

Click here for author‐reader discussions

Deep learning-based segmentation of the lung in MR-images acquired by a stack-of-spirals trajectory at ultra-short echo-times

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 Sørensen-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.