Dataset Information

A Dual-Channel Deep Learning Approach for Lung Cavity Estimation From Hyperpolarized Gas and Proton MRI.

ABSTRACT:

Background

Hyperpolarized gas MRI can quantify regional lung ventilation via biomarkers, including the ventilation defect percentage (VDP). VDP is computed from segmentations derived from spatially co-registered functional hyperpolarized gas and structural proton (¹ H)-MRI. Although acquired at similar lung inflation levels, they are frequently misaligned, requiring a lung cavity estimation (LCE). Recently, single-channel, mono-modal deep learning (DL)-based methods have shown promise for pulmonary image segmentation problems. Multichannel, multimodal approaches may outperform single-channel alternatives.

Purpose

We hypothesized that a DL-based dual-channel approach, leveraging both ¹ H-MRI and Xenon-129-MRI (¹²⁹ Xe-MRI), can generate LCEs more accurately than single-channel alternatives.

Study type

Retrospective.

Population

A total of 480 corresponding ¹ H-MRI and ¹²⁹ Xe-MRI scans from 26 healthy participants (median age [range]: 11 [8-71]; 50% females) and 289 patients with pulmonary pathologies (median age [range]: 47 [6-83]; 51% females) were split into training (422 scans [88%]; 257 participants [82%]) and testing (58 scans [12%]; 58 participants [18%]) sets.

Field strength/sequence

1.5-T, three-dimensional (3D) spoiled gradient-recalled ¹ H-MRI and 3D steady-state free-precession ¹²⁹ Xe-MRI.

Assessment

We developed a multimodal DL approach, integrating ¹²⁹ Xe-MRI and ¹ H-MRI, in a dual-channel convolutional neural network. We compared this approach to single-channel alternatives using manually edited LCEs as a benchmark. We further assessed a fully automatic DL-based framework to calculate VDPs and compared it to manually generated VDPs.

Statistical tests

Friedman tests with post hoc Bonferroni correction for multiple comparisons compared single-channel and dual-channel DL approaches using Dice similarity coefficient (DSC), average boundary Hausdorff distance (average HD), and relative error (XOR) metrics. Bland-Altman analysis and paired t-tests compared manual and DL-generated VDPs. A P value < 0.05 was considered statistically significant.

Results

The dual-channel approach significantly outperformed single-channel approaches, achieving a median (range) DSC, average HD, and XOR of 0.967 (0.867-0.978), 1.68 mm (37.0-0.778), and 0.066 (0.246-0.045), respectively. DL-generated VDPs were statistically indistinguishable from manually generated VDPs (P = 0.710).

Data conclusion

Our dual-channel approach generated LCEs, which could be integrated with ventilated lung segmentations to produce biomarkers such as the VDP without manual intervention.

Evidence level

Technical efficacy

Stage 1.

SUBMITTER: Astley JR

PROVIDER: S-EPMC10947587 | biostudies-literature | 2023 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A Dual-Channel Deep Learning Approach for Lung Cavity Estimation From Hyperpolarized Gas and Proton MRI.

Astley Joshua R JR Biancardi Alberto M AM Marshall Helen H Hughes Paul J C PJC Collier Guilhem J GJ Smith Laurie J LJ Eaden James A JA Hughes Rod R Wild Jim M JM Tahir Bilal A BA

Journal of magnetic resonance imaging : JMRI 20221114 6

<h4>Background</h4>Hyperpolarized gas MRI can quantify regional lung ventilation via biomarkers, including the ventilation defect percentage (VDP). VDP is computed from segmentations derived from spatially co-registered functional hyperpolarized gas and structural proton (<sup>1</sup> H)-MRI. Although acquired at similar lung inflation levels, they are frequently misaligned, requiring a lung cavity estimation (LCE). Recently, single-channel, mono-modal deep learning (DL)-based methods have shown ...[more]

PMID: 36373828

Similar Datasets

Project description:Respiratory diseases are leading causes of mortality and morbidity worldwide. Pulmonary imaging is an essential component of the diagnosis, treatment planning, monitoring, and treatment assessment of respiratory diseases. Insights into numerous pulmonary pathologies can be gleaned from functional lung MRI techniques. These include hyperpolarized gas ventilation MRI, which enables visualization and quantification of regional lung ventilation with high spatial resolution. Segmentation of the ventilated lung is required to calculate clinically relevant biomarkers. Recent research in deep learning (DL) has shown promising results for numerous segmentation problems. Here, we evaluate several 3D convolutional neural networks to segment ventilated lung regions on hyperpolarized gas MRI scans. The dataset consists of 759 helium-3 (3He) or xenon-129 (129Xe) volumetric scans and corresponding expert segmentations from 341 healthy subjects and patients with a wide range of pathologies. We evaluated segmentation performance for several DL experimental methods via overlap, distance and error metrics and compared them to conventional segmentation methods, namely, spatial fuzzy c-means (SFCM) and K-means clustering. We observed that training on combined 3He and 129Xe MRI scans using a 3D nn-UNet outperformed other DL methods, achieving a mean ± SD Dice coefficient of 0.963 ± 0.018, average boundary Hausdorff distance of 1.505 ± 0.969 mm, Hausdorff 95th percentile of 5.754 ± 6.621 mm and relative error of 0.075 ± 0.039. Moreover, limited differences in performance were observed between 129Xe and 3He scans in the testing set. Combined training on 129Xe and 3He yielded statistically significant improvements over the conventional methods (p < 0.0001). In addition, we observed very strong correlation and agreement between DL and expert segmentations, with Pearson correlation of 0.99 (p < 0.0001) and Bland-Altman bias of - 0.8%. The DL approach evaluated provides accurate, robust and rapid segmentations of ventilated lung regions and successfully excludes non-lung regions such as the airways and artefacts. This approach is expected to eliminate the need for, or significantly reduce, subsequent time-consuming manual editing.

Project description:The noninvasive assessment of regional lung ventilation is of critical importance in the quantification of the severity of disease and evaluation of response to therapy in many pulmonary diseases. This work presents, for the first time, the implementation of a hyperpolarized (HP) gas MRI technique to measure whole-lung regional fractional ventilation (r) in Yorkshire pigs (n = 5) through the use of a gas mixing and delivery device in the supine position. The proposed technique utilizes a series of back-to-back HP gas breaths with images acquired during short end-inspiratory breath-holds. In order to decouple the radiofrequency pulse decay effect from the ventilatory signal build-up in the airways, the regional distribution of the flip angle (α) was estimated in the imaged slices by acquiring a series of back-to-back images with no interscan time delay during a breath-hold at the tail end of the ventilation sequence. Analysis was performed to assess the sensitivity of the multislice ventilation model to noise, oxygen and the number of flip angle images. The optimal α value was determined on the basis of the minimization of the error in r estimation: α(opt) = 5-6º for the set of acquisition parameters in pigs. The mean r values for the group of pigs were 0.27 ± 0.09, 0.35 ± 0.06 and 0.40 ± 0.04 for the ventral, middle and dorsal slices, respectively (excluding conductive airways r 0.9). A positive gravitational (ventral-dorsal) ventilation gradient effect was present in all animals. The trachea and major conductive airways showed a uniform near-unity r value, with progressively smaller values corresponding to smaller diameter airways, and ultimately leading to lung parenchyma. The results demonstrate the feasibility of the measurement of the fractional ventilation in large species, and provide a platform to address the technical challenges associated with long breathing time scales through the optimization of acquisition parameters in species with a pulmonary physiology very similar to that of humans.

Dataset Information

A Dual-Channel Deep Learning Approach for Lung Cavity Estimation From Hyperpolarized Gas and Proton MRI.

Background

Purpose

Study type

Population

Field strength/sequence

Assessment

Statistical tests

Results

Data conclusion

Evidence level

Technical efficacy

Publications

A Dual-Channel Deep Learning Approach for Lung Cavity Estimation From Hyperpolarized Gas and Proton MRI.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets