MyMedR

Displaying all 7 publications

Abstract:

Sort:

Convolutional neural network regression for short-axis left ventricle segmentation in cardiac cine MR sequences

Tan LK, Liew YM, Lim E, McLaughlin RA

Med Image Anal, 2017 Apr 12;39:78-86.
PMID: 28437634 DOI: 10.1016/j.media.2017.04.002

Automated left ventricular (LV) segmentation is crucial for efficient quantification of cardiac function and morphology to aid subsequent management of cardiac pathologies. In this paper, we parameterize the complete (all short axis slices and phases) LV segmentation task in terms of the radial distances between the LV centerpoint and the endo- and epicardial contours in polar space. We then utilize convolutional neural network regression to infer these parameters. Utilizing parameter regression, as opposed to conventional pixel classification, allows the network to inherently reflect domain-specific physical constraints. We have benchmarked our approach primarily against the publicly-available left ventricle segmentation challenge (LVSC) dataset, which consists of 100 training and 100 validation cardiac MRI cases representing a heterogeneous mix of cardiac pathologies and imaging parameters across multiple centers. Our approach attained a .77 Jaccard index, which is the highest published overall result in comparison to other automated algorithms. To test general applicability, we also evaluated against the Kaggle Second Annual Data Science Bowl, where the evaluation metric was the indirect clinical measures of LV volume rather than direct myocardial contours. Our approach attained a Continuous Ranked Probability Score (CRPS) of .0124, which would have ranked tenth in the original challenge. With this we demonstrate the effectiveness of convolutional neural network regression paired with domain-specific features in clinical segmentation.
Fulltext Imaging of skin birefringence for human scar assessment using polarization-sensitive optical coherence tomography aided by vascular masking

Gong P, Chin L, Es'haghian S, Liew YM, Wood FM, Sampson DD, et al.

J Biomed Opt, 2014 Dec;19(12):126014.
PMID: 25539060 DOI: 10.1117/1.JBO.19.12.126014

We demonstrate the in vivo assessment of human scars by parametric imaging of birefringence using polarization-sensitive optical coherence tomography (PS-OCT). Such in vivo assessment is subject to artifacts in the detected birefringence caused by scattering from blood vessels. To reduce these artifacts, we preprocessed the PS-OCT data using a vascular masking technique. The birefringence of the remaining tissue regions was then automatically quantified. Results from the scars and contralateral or adjacent normal skin of 13 patients show a correspondence of birefringence with scar type: the ratio of birefringence of hypertrophic scars to corresponding normal skin is 2.2 ± 0.2 (mean ± standard deviation ), while the ratio of birefringence of normotrophic scars to normal skin is 1.1 ± 0.4 . This method represents a new clinically applicable means for objective, quantitative human scar assessment.
Automatic localization of the left ventricular blood pool centroid in short axis cardiac cine MR images

Tan LK, Liew YM, Lim E, Abdul Aziz YF, Chee KH, McLaughlin RA

Med Biol Eng Comput, 2018 Jun;56(6):1053-1062.
PMID: 29147835 DOI: 10.1007/s11517-017-1750-7

In this paper, we develop and validate an open source, fully automatic algorithm to localize the left ventricular (LV) blood pool centroid in short axis cardiac cine MR images, enabling follow-on automated LV segmentation algorithms. The algorithm comprises four steps: (i) quantify motion to determine an initial region of interest surrounding the heart, (ii) identify potential 2D objects of interest using an intensity-based segmentation, (iii) assess contraction/expansion, circularity, and proximity to lung tissue to score all objects of interest in terms of their likelihood of constituting part of the LV, and (iv) aggregate the objects into connected groups and construct the final LV blood pool volume and centroid. This algorithm was tested against 1140 datasets from the Kaggle Second Annual Data Science Bowl, as well as 45 datasets from the STACOM 2009 Cardiac MR Left Ventricle Segmentation Challenge. Correct LV localization was confirmed in 97.3% of the datasets. The mean absolute error between the gold standard and localization centroids was 2.8 to 4.7 mm, or 12 to 22% of the average endocardial radius. Graphical abstract Fully automated localization of the left ventricular blood pool in short axis cardiac cine MR images.
Fulltext Linear-regression convolutional neural network for fully automated coronary lumen segmentation in intravascular optical coherence tomography

Yong YL, Tan LK, McLaughlin RA, Chee KH, Liew YM

J Biomed Opt, 2017 12;22(12):1-9.
PMID: 29274144 DOI: 10.1117/1.JBO.22.12.126005

Intravascular optical coherence tomography (OCT) is an optical imaging modality commonly used in the assessment of coronary artery diseases during percutaneous coronary intervention. Manual segmentation to assess luminal stenosis from OCT pullback scans is challenging and time consuming. We propose a linear-regression convolutional neural network to automatically perform vessel lumen segmentation, parameterized in terms of radial distances from the catheter centroid in polar space. Benchmarked against gold-standard manual segmentation, our proposed algorithm achieves average locational accuracy of the vessel wall of 22 microns, and 0.985 and 0.970 in Dice coefficient and Jaccard similarity index, respectively. The average absolute error of luminal area estimation is 1.38%. The processing rate is 40.6 ms per image, suggesting the potential to be incorporated into a clinical workflow and to provide quantitative assessment of vessel lumen in an intraoperative time frame.
Fully automated segmentation of the left ventricle in cine cardiac MRI using neural network regression

Tan LK, McLaughlin RA, Lim E, Abdul Aziz YF, Liew YM

J Magn Reson Imaging, 2018 07;48(1):140-152.
PMID: 29316024 DOI: 10.1002/jmri.25932

BACKGROUND: Left ventricle (LV) structure and functions are the primary assessment performed in most clinical cardiac MRI protocols. Fully automated LV segmentation might improve the efficiency and reproducibility of clinical assessment.
PURPOSE: To develop and validate a fully automated neural network regression-based algorithm for segmentation of the LV in cardiac MRI, with full coverage from apex to base across all cardiac phases, utilizing both short axis (SA) and long axis (LA) scans.
STUDY TYPE: Cross-sectional survey; diagnostic accuracy.
SUBJECTS: In all, 200 subjects with coronary artery diseases and regional wall motion abnormalities from the public 2011 Left Ventricle Segmentation Challenge (LVSC) database; 1140 subjects with a mix of normal and abnormal cardiac functions from the public Kaggle Second Annual Data Science Bowl database.
FIELD STRENGTH/SEQUENCE: 1.5T, steady-state free precession.
ASSESSMENT: Reference standard data generated by experienced cardiac radiologists. Quantitative measurement and comparison via Jaccard and Dice index, modified Hausdorff distance (MHD), and blood volume.
STATISTICAL TESTS: Paired t-tests compared to previous work.
RESULTS: Tested against the LVSC database, we obtained 0.77 ± 0.11 (Jaccard index) and 1.33 ± 0.71 mm (MHD), both metrics demonstrating statistically significant improvement (P < 0.001) compared to previous work. Tested against the Kaggle database, the signed difference in evaluated blood volume was +7.2 ± 13.0 mL and -19.8 ± 18.8 mL for the end-systolic (ES) and end-diastolic (ED) phases, respectively, with a statistically significant improvement (P < 0.001) for the ED phase.
DATA CONCLUSION: A fully automated LV segmentation algorithm was developed and validated against a diverse set of cardiac cine MRI data sourced from multiple imaging centers and scanner types. The strong performance overall is suggestive of practical clinical utility.
LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018.
Motion corrected LV quantification based on 3D modelling for improved functional assessment in cardiac MRI

Liew YM, McLaughlin RA, Chan BT, Abdul Aziz YF, Chee KH, Ung NM, et al.

Phys Med Biol, 2015 Apr 7;60(7):2715-33.
PMID: 25768708 DOI: 10.1088/0031-9155/60/7/2715

Cine MRI is a clinical reference standard for the quantitative assessment of cardiac function, but reproducibility is confounded by motion artefacts. We explore the feasibility of a motion corrected 3D left ventricle (LV) quantification method, incorporating multislice image registration into the 3D model reconstruction, to improve reproducibility of 3D LV functional quantification. Multi-breath-hold short-axis and radial long-axis images were acquired from 10 patients and 10 healthy subjects. The proposed framework reduced misalignment between slices to subpixel accuracy (2.88 to 1.21 mm), and improved interstudy reproducibility for 5 important clinical functional measures, i.e. end-diastolic volume, end-systolic volume, ejection fraction, myocardial mass and 3D-sphericity index, as reflected in a reduction in the sample size required to detect statistically significant cardiac changes: a reduction of 21-66%. Our investigation on the optimum registration parameters, including both cardiac time frames and number of long-axis (LA) slices, suggested that a single time frame is adequate for motion correction whereas integrating more LA slices can improve registration and model reconstruction accuracy for improved functional quantification especially on datasets with severe motion artefacts.
Regional assessment of LV wall in infarcted heart using tagged MRI and cardiac modelling

Jahanzad Z, Liew YM, Bilgen M, McLaughlin RA, Leong CO, Chee KH, et al.

Phys Med Biol, 2015 May 21;60(10):4015-31.
PMID: 25919317 DOI: 10.1088/0031-9155/60/10/4015

A segmental two-parameter empirical deformable model is proposed for evaluating regional motion abnormality of the left ventricle. Short-axis tagged MRI scans were acquired from 10 healthy subjects and 10 postinfarct patients. Two motion parameters, contraction and rotation, were quantified for each cardiac segment by fitting the proposed model using a non-rigid registration algorithm. The accuracy in motion estimation was compared to a global model approach. Motion parameters extracted from patients were correlated to infarct transmurality assessed with delayed-contrast-enhanced MRI. The proposed segmental model allows markedly improved accuracy in regional motion analysis as compared to the global model for both subject groups (1.22-1.40 mm versus 2.31-2.55 mm error). By end-systole, all healthy segments experienced radial displacement by ~25-35% of the epicardial radius, whereas the 3 short-axis planes rotated differently (basal: 3.3°; mid: -1° and apical: -4.6°) to create a twisting motion. While systolic contraction showed clear correspondence to infarct transmurality, rotation was nonspecific to either infarct location or transmurality but could indicate the presence of functional abnormality. Regional contraction and rotation derived using this model could potentially aid in the assessment of severity of regional dysfunction of infarcted myocardium.