Semi-automatic lung segmentation of DCE-MRI data sets of 2-year old children after congenital diaphragmatic hernia repair: Initial results

Zöllner, Frank G.; Daab, Markus; Weidner, Meike; Sommer, Verena; Zahn, Katrin; Schaible, Thomas; Weisser, G; Schoenberg, Stefan O.; Neff, K. W.; Schad, Lothar R.

doi:10.1016/j.mri.2015.08.003

Cited by 7 publications

(6 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nonetheless, the reported results are achieved on dual‐center data acquired on scanners from two manufacturers. We argue that this indicates a good generalization performance of the neural network, which implies that a comparison to the work of Kohlmann et al, Zöllner et al, and Böttger et al is valid.…”

Section: Discussionsupporting

confidence: 60%

“…In terms of DSC, our results show a better agreement with the ground truth segmentation than state‐of‐the‐art results reported by Kohlmann et al An extensive comparison to prior work is provided in Table . This includes the proposed methods by Zöllner et al and Böttger et al However, it has to be stressed that our results are reported on a different dataset and, therefore, a comparison of the results is not directly applicable. Nonetheless, the reported results are achieved on dual‐center data acquired on scanners from two manufacturers.…”

Section: Discussionmentioning

confidence: 98%

“…Ray et al 24 suggested an ensemble segmentation based on active contours. Additionally, Zöllner et al 25 propose a semiautomated segmentation based on calculating the cross correlation and the area under the curve for a reference region of interest in comparison to all voxels of the dataset.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Deep semantic lung segmentation for tracking potential pulmonary perfusion biomarkers in chronic obstructive pulmonary disease (COPD): The multi‐ethnic study of atherosclerosis COPD study

Winther

Gutberlet

Hundt³

et al. 2019

Magnetic Resonance Imaging

View full text Add to dashboard Cite

Background Chronic obstructive pulmonary disease (COPD) is associated with high morbidity and mortality. Identification of imaging biomarkers for phenotyping is necessary for future treatment and therapy monitoring. However, translation of visual analytic pipelines into clinics or their use in large‐scale studies is significantly slowed by time‐consuming postprocessing steps. Purpose To implement an automated tool chain for regional quantification of pulmonary microvascular blood flow in order to reduce analysis time and user variability. Study Type Prospective. Population In all, 90 MRI scans of 63 patients, of which 31 had a COPD with a mean Global Initiative for Chronic Obstructive Lung Disease status of 1.9 ± 0.64 (μ ± σ). Field Strength/Sequence 1.5T dynamic gadolinium‐enhanced MRI measurement using 4D dynamic contrast material‐enhanced (DCE) time‐resolved angiography acquired in a single breath‐hold in inspiration. [Correction added on August 20, 2019, after first online publication: The field strength in the preceding sentence was corrected.] Assessment We built a 3D convolutional neural network for semantic segmentation using 29 manually segmented perfusion maps. All five lobes of the lung are denoted, including the middle lobe. Evaluation was performed on 61 independent cases from two sites of the Multi‐Ethnic Study of Arteriosclerosis (MESA)‐COPD study. We publish our implementation of a model‐free deconvolution filter according to Sourbron et al for 4D DCE MRI scans as open source. Statistical Test Cross‐validation 29/61 (# training / # testing), intraclass correlation coefficient (ICC), Spearman ρ, Pearson r, Sørensen–Dice coefficient, and overlap. Results Segmentations and derived clinical parameters were processed in ~90 seconds per case on a Xeon E5‐2637v4 workstation with Tesla P40 GPUs. Clinical parameters and predicted segmentations exhibit high concordance with the ground truth regarding median perfusion for all lobes with an ICC of 0.99 and a Sørensen–Dice coefficient of 93.4 ± 2.8 (μ ± σ). Data Conclusion We present a robust end‐to‐end pipeline that allows for the extraction of perfusion‐based biomarkers for all lung lobes in 4D DCE MRI scans by combining model‐free deconvolution with deep learning. Level of Evidence: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:571–579.

show abstract

Section: Discussionsupporting

confidence: 60%

Section: Discussionmentioning

confidence: 98%

See 1 more Smart Citation

Deep semantic lung segmentation for tracking potential pulmonary perfusion biomarkers in chronic obstructive pulmonary disease (COPD): The multi‐ethnic study of atherosclerosis COPD study

Winther

Gutberlet

Hundt³

et al. 2019

Magnetic Resonance Imaging

View full text Add to dashboard Cite

show abstract

“…Our results are in line with a previous study that used TSP to assess perfusion deficits with dynamic susceptibility MRI [10]. Other non-deconvolution-based methods, such as crosscorrelation analysis and clustering, have been investigated for the use in acute stroke [24], although these were generally used for detection of perfusion deficits outside the brain [25,26]. Wissmuller et al also presented a method based on neural network clustering that was shown to be able to identify groups of voxels sharing common properties of signal dynamics and delineate perfusion deficits in stroke [27].…”

Section: Discussionsupporting

confidence: 90%

Identifying perfusion deficits on CT perfusion images using temporal similarity perfusion (TSP) mapping

et al. 2019

View full text Add to dashboard Cite

Objectives-Deconvolution-derived maps of CT perfusion (CTP) data may be confounded by transit delays. We propose temporal similarity perfusion (TSP) analysis to decrease CTP maps' dependence on transit times and investigate its sensitivity to detect perfusion deficits. Methods-CTP data of acute stroke patients obtained within 9 h of symptom onset was analyzed using a delay-insensitive singular value decomposition method and with TSP. The TSP method applies an iterative process whereby a pixel's highest Pearson's R value is obtained through comparison of a pixel's time-shifted signal density time-series curve and the average whole brain signal density time-series curve. Our evaluation included a qualitative and quantitative rating of deconvolution maps (MTT, CBV, and TTP), of TSP maps, and of follow-up CT.

show abstract

“…In another attempt, 57 the lung was segmented by deforming two meshes and this resulted in >20% volume errors and efficiency comparable to manual approaches. Statistical methods were also developed to segment 3D 1 H MRI by exploiting voxel-wise correlation coefficients 58 and this required 1.4 s to segment a 3D dataset with a spatial overlap ratio of ∼0.7 ± 0.1. An atlas-based approach 59 was also developed and while accurate, this required numerous registrations and label fusion (30 min) to generate target segmentation.…”

Section: Dice-similarity-coefficient (%)mentioning

confidence: 99%

Anatomical pulmonary magnetic resonance imaging segmentation for regional structure‐function measurements of asthma

et al. 2016

View full text Add to dashboard Cite

show abstract

Semi-automatic lung segmentation of DCE-MRI data sets of 2-year old children after congenital diaphragmatic hernia repair: Initial results

Cited by 7 publications

References 22 publications

Deep semantic lung segmentation for tracking potential pulmonary perfusion biomarkers in chronic obstructive pulmonary disease (COPD): The multi‐ethnic study of atherosclerosis COPD study

Deep semantic lung segmentation for tracking potential pulmonary perfusion biomarkers in chronic obstructive pulmonary disease (COPD): The multi‐ethnic study of atherosclerosis COPD study

Identifying perfusion deficits on CT perfusion images using temporal similarity perfusion (TSP) mapping

Anatomical pulmonary magnetic resonance imaging segmentation for regional structure‐function measurements of asthma

Contact Info

Product

Resources

About