Analysis of Deep Neural Networks for Detection of Coronary Artery Stenosis

2021 ◽  
Vol 47 (3) ◽  
pp. 153-160
Author(s):  
V. V. Danilov ◽  
O. M. Gerget ◽  
K. Yu. Klyshnikov ◽  
A. F. Frangi ◽  
E. A. Ovcharenko
Keyword(s):  
Neural Networks ◽  
Coronary Artery ◽  
Deep Neural Networks ◽  
Coronary Artery Stenosis ◽  
Artery Stenosis

scholarly journals Real-time coronary artery stenosis detection based on modern neural networks

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Viacheslav V. Danilov ◽  
Kirill Yu. Klyshnikov ◽  
Olga M. Gerget ◽  
Anton G. Kutikhin ◽  
Vladimir I. Ganyukov ◽  
...  
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Coronary Artery ◽  
Coronary Angiography ◽  
Real Time ◽  
Coronary Artery Stenosis ◽  
Artery Stenosis ◽  
Speed Ratio ◽  
Prediction Rate ◽  
Stenosis Detection

AbstractInvasive coronary angiography remains the gold standard for diagnosing coronary artery disease, which may be complicated by both, patient-specific anatomy and image quality. Deep learning techniques aimed at detecting coronary artery stenoses may facilitate the diagnosis. However, previous studies have failed to achieve superior accuracy and performance for real-time labeling. Our study is aimed at confirming the feasibility of real-time coronary artery stenosis detection using deep learning methods. To reach this goal we trained and tested eight promising detectors based on different neural network architectures (MobileNet, ResNet-50, ResNet-101, Inception ResNet, NASNet) using clinical angiography data of 100 patients. Three neural networks have demonstrated superior results. The network based on Faster-RCNN Inception ResNet V2 is the most accurate and it achieved the mean Average Precision of 0.95, F1-score 0.96 and the slowest prediction rate of 3 fps on the validation subset. The relatively lightweight SSD MobileNet V2 network proved itself as the fastest one with a low mAP of 0.83, F1-score of 0.80 and a mean prediction rate of 38 fps. The model based on RFCN ResNet-101 V2 has demonstrated an optimal accuracy-to-speed ratio. Its mAP makes up 0.94, F1-score 0.96 while the prediction speed is 10 fps. The resultant performance-accuracy balance of the modern neural networks has confirmed the feasibility of real-time coronary artery stenosis detection supporting the decision-making process of the Heart Team interpreting coronary angiography findings.

scholarly journals Real-Time Coronary Artery Stenosis Detection Based on Modern Neural Networks

2020 ◽  
Author(s):  
Viacheslav Danilov ◽  
Kirill Klyshnikov ◽  
Olga Gerget ◽  
Anton Kutikhin ◽  
Vladimir Ganyukov ◽  
...  
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Coronary Artery ◽  
Coronary Angiography ◽  
Real Time ◽  
Coronary Artery Stenosis ◽  
Artery Stenosis ◽  
Speed Ratio ◽  
Prediction Rate ◽  
Stenosis Detection

Abstract Invasive coronary angiography remains the gold standard for diagnosing coronary artery disease, which may be complicated by both, patient-specific anatomy and image quality. Deep learning techniques aimed at detecting coronary artery stenoses may facilitate the diagnosis. However, previous studies have failed to achieve superior accuracy and performance for real-time labeling. Our study is aimed at confirming the feasibility of real-time coronary artery stenosis detection using deep learning methods. To reach this goal we trained and tested eight promising detectors based on different neural network architectures (MobileNet, ResNet-50, ResNet-101, Inception ResNet, NASNet) using clinical angiography data of 100 patients.Three neural networks have demonstrated superior results. The network based on Faster-RCNN Inception ResNet V2 is the most accurate and it achieved the mean Average Precision of 0.954 and the prediction rate of 3 fps on the validation subset. The relatively lightweight SSD MobileNet V2 network proved itself as the fastest one with an mAP of 0.830 and a mean prediction rate of 38 fps. The model based on RFCN ResNet-101 V2 has demonstrated an optimal accuracy-to-speed ratio. Its mAP makes up 0.94, while the prediction speed is 10 fps.The resultant performance-accuracy balance of the modern neural networks has confirmed the feasibility of real-time coronary artery stenosis detection supporting the decision-making process of the Heart Team interpreting coronary angiography findings.

Two cases of unstable angina with proximal coronary artery stenosis diagnosed by transthoracic Doppler echocardiography

Choonpa Igaku ◽  
2008 ◽  
Vol 35 (4) ◽  
pp. 443-449 ◽  
Author(s):  
Yuko SUGIYAMA ◽  
Masayo SUZUKI ◽  
Keiichi HIRANO ◽  
Keijirou NAKAMURA ◽  
Mao TAKAHASHI ◽  
...  
Keyword(s):  
Coronary Artery ◽  
Unstable Angina ◽  
Doppler Echocardiography ◽  
Coronary Artery Stenosis ◽  
Artery Stenosis ◽  
Transthoracic Doppler Echocardiography ◽  
Proximal Coronary Artery

Relationship between Whole Blood Viscosity and Lesion Severity in Coronary Artery Disease

2021 ◽  
Author(s):  
Gökhan Ceyhun ◽  
Oğuzhan Birdal
Keyword(s):  
Coronary Artery ◽  
Shear Rate ◽  
Confidence Interval ◽  
Odds Ratio ◽  
Coronary Artery Stenosis ◽  
Artery Stenosis ◽  
Significant Stenosis ◽  
Whole Blood Viscosity ◽  
Significant Coronary Artery Stenosis ◽  
Coronary Syndrome

Abstract Objective This article investigates the relationship of fractional flow reserve (FFR) with whole blood viscosity (WBV) in patients who were diagnosed with chronic coronary syndrome and significant stenosis in the major coronary arteries and underwent the measurement of FFR. Material and Method In the FFR measurements performed to evaluate the severity of coronary artery stenosis, 160 patients were included in the study and divided into two groups as follows: 80 with significant stenosis and 80 with nonsignificant stenosis. WBVs at low shear rate (LSR) and high shear rate (HSR) were compared between the patients in the significant and nonsignificant coronary artery stenosis groups. Results In the group with FFR < 0.80 and significant coronary artery stenosis, WBV was significantly higher compared with the group with nonsignificant coronary artery stenosis in terms of both HSR (19.33 ± 0.84) and LSR (81.19 ± 14.20) (p < 0.001). In the multivariate logistic regression analysis, HSR and LSR were independent predictors of significant coronary artery stenosis (HSR: odds ratio: 1.67, 95% confidence interval: 1.17–2.64; LSR: odds ratio: 2.46, 95% confidence interval: 2.19–2.78). In the receiver operating characteristic (ROC) curve analysis, when the cutoff value of WBV at LSR was taken as 79.23, it had 58.42% sensitivity and 62.13% specificity for the prediction of significant coronary artery stenosis (area under the ROC curve: 0.628, p < 0.001). Conclusion WBV, an inexpensive biomarker that can be easily calculated prior to coronary angiography, was higher in patients with functionally severe coronary artery stenosis, and thus could be a useful marker in predicting the hemodynamic severity of coronary artery stenosis in patients with chronic coronary syndrome.

Sign in / Sign up

Export Citation Format

Share Document