scholarly journals Selective ensemble methods for deep learning segmentation of major vessels in invasive coronary angiography

2021 ◽  
Author(s):  
Jeeone Park ◽  
Jihoon Kweon ◽  
Hyehyeon Bark ◽  
Young In Kim ◽  
Inwook Back ◽  
...  
Keyword(s):  
Coronary Angiography ◽  
Imaging Modality ◽  
Invasive Coronary Angiography ◽  
External Validation ◽  
Ensemble Methods ◽  
Ground Truth ◽  
Diagnostic Methods ◽  
Dice Similarity Coefficient ◽  
Major Vessel ◽  
Selective Ensemble

Invasive coronary angiography is a primary imaging modality that visualizes the lumen area of coronary arteries for the diagnosis of coronary artery diseases and guidance for interventional devices. In the current practice of quantitative coronary analysis (QCA), semi-automatic segmentation tools require labor-intensive and time-consuming manual correction; this limits their application in the catheterization room. For a more automated QCA, it is necessary to minimize operator intervention through robust segmentation methods with improved predictability. In this study, we introduced two selective ensemble methods that integrated the weighted ensemble approach with per-image quality estimation. In our selective ensemble methods, the segmentation outcomes from five base models with different loss functions were ranked by mask morphology or estimated dice similarity coefficient (DSC). The final output was determined by imposing different weights according to the ranking. The ranking criteria based on mask morphology were determined empirically to avoid frequent types of segmentation errors, whereas the estimation of DSCs was performed by comparing the pseudo-ground truth generated from a meta-learner. In the assessment with 7,426 frames from 2,924 patients, the selective ensemble methods improved segmentation performance with DSCs of up to 93.11\% and provided a better delineation of lumen boundaries near the coronary lesion with local DSCs of up to 94.04\%, outperforming all individual models and hard voting ensembles. The probability of mask disconnection at the most narrowed region could be minimized to <1\%. The robustness of the proposed methods was evident in the external validation. Inference time for major vessel segmentation was approximately one-third, indicating that our selective ensemble methods may allow the real-time application of QCA-based diagnostic methods in routine clinical settings.

scholarly journals Deep learning segmentation of major vessels in X-ray coronary angiography

2019 ◽  
Vol 9 (1) ◽  
Author(s):  
Su Yang ◽  
Jihoon Kweon ◽  
Jae-Hyung Roh ◽  
Jae-Hwan Lee ◽  
Heejun Kang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Coronary Angiography ◽  
Coronary Arteries ◽  
Coronary Vessels ◽  
Quantitative Coronary Angiography ◽  
Vessel Segmentation ◽  
Diagnostic Methods ◽  
X Ray ◽  
Fully Convolutional Networks ◽  
Major Vessel

AbstractX-ray coronary angiography is a primary imaging technique for diagnosing coronary diseases. Although quantitative coronary angiography (QCA) provides morphological information of coronary arteries with objective quantitative measures, considerable training is required to identify the target vessels and understand the tree structure of coronary arteries. Despite the use of computer-aided tools, such as the edge-detection method, manual correction is necessary for accurate segmentation of coronary vessels. In the present study, we proposed a robust method for major vessel segmentation using deep learning models with fully convolutional networks. When angiographic images of 3302 diseased major vessels from 2042 patients were tested, deep learning networks accurately identified and segmented the major vessels in X-ray coronary angiography. The average F1 score reached 0.917, and 93.7% of the images exhibited a high F1 score > 0.8. The most narrowed region at the stenosis was distinctly captured with high connectivity. Robust predictability was validated for the external dataset with different image characteristics. For major vessel segmentation, our approach demonstrated that prediction could be completed in real time with minimal image preprocessing. By applying deep learning segmentation, QCA analysis could be further automated, thereby facilitating the use of QCA-based diagnostic methods.

scholarly journals Deep Learning for Automated Measurement of Hemorrhage and Perihematomal Edema in Supratentorial Intracerebral Hemorrhage

Stroke ◽  
2020 ◽  
Vol 51 (2) ◽  
pp. 648-651 ◽  
Author(s):  
Rajat Dhar ◽  
Guido J. Falcone ◽  
Yasheng Chen ◽  
Ali Hamzehloo ◽  
Elayna P. Kirsch ◽  
...  
Keyword(s):  
Deep Learning ◽  
Intracerebral Hemorrhage ◽  
External Validation ◽  
Ground Truth ◽  
Interquartile Range ◽  
Secondary Outcome ◽  
Dice Similarity Coefficient ◽  
Disease Biology ◽  
Underlying Mechanisms ◽  
Perihematomal Edema

Background and Purpose— Volumes of hemorrhage and perihematomal edema (PHE) are well-established biomarkers of primary and secondary injury, respectively, in spontaneous intracerebral hemorrhage. An automated imaging pipeline capable of accurately and rapidly quantifying these biomarkers would facilitate large cohort studies evaluating underlying mechanisms of injury. Methods— Regions of hemorrhage and PHE were manually delineated on computed tomography scans of patients enrolled in 2 intracerebral hemorrhage studies. Manual ground-truth masks from the first cohort were used to train a fully convolutional neural network to segment images into hemorrhage and PHE. The primary outcome was automated-versus-human concordance in hemorrhage and PHE volumes. The secondary outcome was voxel-by-voxel overlap of segmentations, quantified by the Dice similarity coefficient (DSC). Algorithm performance was validated on 84 scans from the second study. Results— Two hundred twenty-four scans from 124 patients with supratentorial intracerebral hemorrhage were used for algorithm derivation. Median volumes were 18 mL (interquartile range, 8–43) for hemorrhage and 12 mL (interquartile range, 5–30) for PHE. Concordance was excellent (0.96) for automated quantification of hemorrhage and good (0.81) for PHE, with DSC of 0.90 (interquartile range, 0.85–0.93) and 0.54 (0.39–0.65), respectively. External validation confirmed algorithm accuracy for hemorrhage (concordance 0.98, DSC 0.90) and PHE (concordance 0.90, DSC 0.55). This was comparable with the consistency observed between 2 human raters (DSC 0.90 for hemorrhage, 0.57 for PHE). Conclusions— We have developed a deep learning-based imaging algorithm capable of accurately measuring hemorrhage and PHE volumes. Rapid and consistent automated biomarker quantification may accelerate powerful and precise studies of disease biology in large cohorts of intracerebral hemorrhage patients.

scholarly journals Spontaneous Coronary Dissection: “Live Flash” Optical Coherence Tomography Guided Angioplasty

2016 ◽  
Vol 2016 ◽  
pp. 1-4
Author(s):  
Angela Pimenta Bento ◽  
Renato Gil dos Santos Pinto Fernandes ◽  
David Cintra Henriques Silva Neves ◽  
Lino Manuel Ribeiro Patrício ◽  
José Eduardo Chambel de Aguiar
Keyword(s):  
Optical Coherence Tomography ◽  
Coronary Angiography ◽  
Diagnostic Yield ◽  
Imaging Modality ◽  
Invasive Coronary Angiography ◽  
Artery Dissection ◽  
Optical Coherence ◽  
Coronary Artery Dissection ◽  
Coronary Syndrome ◽  
Coronary Imaging

Optical Coherence tomography (OCT) is a light-based imaging modality which shows tremendous potential in the setting of coronary imaging. Spontaneous coronary artery dissection (SCAD) is an infrequent cause of acute coronary syndrome (ACS). The diagnosis of SCAD is made mainly with invasive coronary angiography, although adjunctive imaging modalities such as computed tomography angiography, IVUS, and OCT may increase the diagnostic yield. The authors describe a clinical case of a young woman admitted with the diagnosis of ACS. The ACS was caused by SCAD detected in the coronary angiography and the angioplasty was guided by OCT. OCT use in the setting of SCAD has been already described and the true innovation in this case was this unique use of OCT. The guidance of angioplasty with live and short images was very useful as it allowed clearly identifying the position of the guidewires at any given moment without the use of prohibitive amounts of contrast.

scholarly journals Implementation of CT Coronary Angiography as an Alternative to Invasive Coronary Angiography in the Diagnostic Work-Up of Non-Coronary Cardiac Surgery, Cardiomyopathy, Heart Failure and Ventricular Arrhythmias

2021 ◽  
Vol 10 (11) ◽  
pp. 2374
Author(s):  
Thomas P. W. van den Boogert ◽  
Bimmer E. P. M. Claessen ◽  
Adrienne van Randen ◽  
Joost van Schuppen ◽  
S. Matthijs Boekholdt ◽  
...  
Keyword(s):  
Heart Failure ◽  
Cardiac Surgery ◽  
Coronary Angiography ◽  
Ventricular Arrhythmias ◽  
Imaging Modality ◽  
Invasive Coronary Angiography ◽  
Coronary Lesions ◽  
Diagnostic Work ◽  
Work Up ◽  
Diagnostic Work Up

To assess the need for additional invasive coronary angiography (CAG) after initial computed tomography coronary angiography (CTCA) in patients awaiting non-coronary cardiac surgery and in patients with cardiomyopathy, heart failure or ventricular arrhythmias, and to determine differences between patients that were referred to initial CTCA or direct CAG, consecutive patients were included between August 2017 and January 2020 and categorized as those referred to initial CTCA (conform protocol), and to direct CAG (non-conform protocol). Out of a total of 415 patients, 78.8% (327 patients, mean age: 57.9 years, 67.3% male) were referred to initial CTCA, of whom 260 patients (79.5%) had no obstructive lesions (<50% DS). A total of 55 patients (16.8%) underwent additional CAG after initial CTCA, which showed coronary lesions of >50% DS in 21 patients (6.3% of 327). Eighty-eight patients (mean age: 66.0 years, 59.1% male) were directly referred to CAG (non-conform protocol). These patients were older and had more cardiovascular risk factors compared to patients that underwent initial CTCA (conform protocol), and coronary lesions of >50% DS were detected in 16 patients (17.2%). Revascularization procedures were infrequently performed in both groups: initial CTCA (3.0%), direct CAG (3.4%). The use of CTCA as a gatekeeper CAG in the diagnostic work-up of non-coronary cardiac surgery, cardiomyopathy, heart failure and ventricular arrhythmias is feasible, and only 17% of these patients required additional CAG after initial CTCA. Therefore, CTCA should be considered as the initial imaging modality to rule out CAD in these patients.

Quantification of Coronary Artery Stenosis with 16-Slice MSCT in Patients before CABG Surgery: Comparison to Standard Invasive Coronary Angiography

2005 ◽  
Vol 8 (1) ◽  
pp. 42 ◽  
Author(s):  
C. Probst ◽  
A. Kovacs ◽  
C. Schmitz ◽  
W. Schiller ◽  
H. Schild ◽  
...  
Keyword(s):  
Coronary Artery ◽  
Coronary Angiography ◽  
Coronary Arteries ◽  
Artery Bypass ◽  
Invasive Coronary Angiography ◽  
Bypass Graft ◽  
Heart Association ◽  
Artery Stenosis ◽  
Vessel Disease ◽  
Selective Coronary Angiography

Objective: Invasive, selective coronary angiography is the gold standard for evaluation of coronary artery disease (CAD) and degree of stenosis. The purpose of this study was to compare 3-dimensional (3D) reconstructed 16-slice multislice computed tomographic (MSCT) angiography and selective coronary angiography in patients before elective coronary artery bypass graft (CABG) procedure. Methods: Sixteen-slice MSCT scans (Philips Mx8000 IDT) were performed in 50 patients (42 male/8 female; mean age, 64.44 8.66 years) scheduled for elective CABG procedure. Scans were retrospectively electrocardiogram-gated 3D reconstructed. The images of the coronary arteries were evaluated for stenosis by 2 independent radiologists. The results were compared with the coronary angiography findings using the American Heart Association segmental classification for coronary arteries. Results: Four patients (8%) were excluded for technical reasons. Thirty-eight patients (82.6%) had 3-vessel disease, 4 (8.7 %) had 2-vessel disease, and 4 (8.7%) had an isolated left anterior descending artery stenosis. In the proximal segments all stenoses >50% (56/56) were detected by MSCT; medial segment sensitivity was 97% (73/75), specificity 90.3%; distal segment sensitivity was 90.7% (59/65), specificity 77%. Conclusion: Accurate quantification of coronary stenosis greater than 50% in the proximal and medial segments is possible with high sensitivity and specificity using the new generation of 16-slice MSCTs. There is still a tendency to overestimate stenosis in the distal segments. MSCT seems to be an excellent diagnostic tool for screening patients with possible CAD.

Sign in / Sign up

Export Citation Format

Share Document