Automatic Segmentation of Coronary Lumen and External Elastic Membrane in IVUS Images Using 8-layer U-Net

Abstract Background: Intravascular ultrasound (IVUS) is the golden standard in accessing the coronary lesions, stenosis, and atherosclerosis plaques. In this paper, a fully-automatic approach by an 8-layer U-Net is developed to segment the coronary artery lumen and the area bounded by external elastic membrane, i.e. EEM cross section area (EEM-CSA). The database comprises of single-vendor and single-frequency IVUS data. Particularly, the proposed data augmentation of MeshGrid combined with flip and rotation operations is implemented, improving the model performance without pre- or post-processing of the raw IVUS images.Results: The mean intersection of union (mIoU) of 0.941 and 0.750 for the lumen and EEM-CSA respectively were achieved, which exceeded the manual labeling accuracy of the clinician. Conclusion: The accuracy shown by the proposed method is sufficient for subsequent reconstruction of 3D IVUS images, which is essential for doctors’ diagnosis in the tissue characterization of coronary artery walls and plaque compositions, qualitatively and quantitatively.

Download Full-text

Automatic Segmentation of Coronary Lumen and External Elastic Membrane in Intravascular Ultrasound Images Using 8-layer U-Net

10.21203/rs.3.rs-42239/v3 ◽

2020 ◽

Author(s):

Liang Dong ◽

Wei Lu ◽

Jun Jiang ◽

Ya Zhao ◽

Xiangfen Song ◽

...

Keyword(s):

Coronary Artery ◽

Intravascular Ultrasound ◽

Tissue Characterization ◽

Data Augmentation ◽

Automatic Segmentation ◽

Model Performance ◽

Single Frequency ◽

Elastic Membrane ◽

Ultrasound Images ◽

External Elastic Membrane

Abstract Background: Intravascular ultrasound (IVUS) is the golden standard in accessing the coronary lesions, stenosis, and atherosclerosis plaques. In this paper, a fully-automatic approach by an 8-layer U-Net is developed to segment the coronary artery lumen and the area bounded by external elastic membrane (EEM), i.e. cross section area (EEM-CSA). The database comprises of single-vendor and single-frequency IVUS data. Particularly, the proposed data augmentation of MeshGrid combined with flip and rotation operations is implemented, improving the model performance without pre- or post-processing of the raw IVUS images.Results: The mean intersection of union (mIoU) of 0.937 and 0.804 for the lumen and EEM-CSA respectively were achieved, which exceeded the manual labeling accuracy of the clinician.Conclusion: The accuracy shown by the proposed method is sufficient for subsequent reconstruction of 3D IVUS images, which is essential for doctors’ diagnosis in the tissue characterization of coronary artery walls and plaque compositions, qualitatively and quantitatively.

Download Full-text

Automatic Segmentation of Coronary Lumen and External Elastic Membrane in Intravascular Ultrasound Images Using 8-layer U-Net

10.21203/rs.3.rs-42239/v5 ◽

2021 ◽

Author(s):

Liang Dong ◽

Wenbing Jiang ◽

Wei Lu ◽

Jun Jiang ◽

Ya Zhao ◽

...

Keyword(s):

Coronary Artery ◽

Intravascular Ultrasound ◽

Tissue Characterization ◽

Data Augmentation ◽

Automatic Segmentation ◽

Model Performance ◽

Single Frequency ◽

Elastic Membrane ◽

Ultrasound Images ◽

External Elastic Membrane

Abstract Background: Intravascular ultrasound (IVUS) is the golden standard in accessing the coronary lesions, stenosis, and atherosclerosis plaques. In this paper, a fully-automatic approach by an 8-layer U-Net is developed to segment the coronary artery lumen and the area bounded by external elastic membrane (EEM), i.e. cross section area (EEM-CSA). The database comprises of single-vendor and single-frequency IVUS data. Particularly, the proposed data augmentation of MeshGrid combined with flip and rotation operations is implemented, improving the model performance without pre- or post-processing of the raw IVUS images.Results: The mean intersection of union (MIoU) of 0.937 and 0.804 for the lumen and EEM-CSA respectively were achieved, which exceeded the manual labeling accuracy of the clinician. Conclusion: The accuracy shown by the proposed method is sufficient for subsequent reconstruction of 3D IVUS images, which is essential for doctors’ diagnosis in the tissue characterization of coronary artery walls and plaque compositions, qualitatively and quantitatively.

Download Full-text

Automatic Segmentation of Coronary Lumen and External Elastic Membrane in Intravascular Ultrasound Images Using 8-layer U-Net

10.21203/rs.3.rs-42239/v2 ◽

2020 ◽

Author(s):

Liang Dong ◽

Wei Lu ◽

Jun Jiang ◽

Ya Zhao ◽

Xiangfen Song ◽

...

Keyword(s):

Coronary Artery ◽

Intravascular Ultrasound ◽

Tissue Characterization ◽

Data Augmentation ◽

Automatic Segmentation ◽

Model Performance ◽

Single Frequency ◽

Elastic Membrane ◽

Ultrasound Images ◽

External Elastic Membrane

Download Full-text

Automatic Segmentation of Coronary Lumen and External Elastic Membrane in Intravascular Ultrasound Images Using 8-layer U-Net

10.21203/rs.3.rs-42239/v4 ◽

2021 ◽

Author(s):

Liang Dong ◽

Wei Lu ◽

Jun Jiang ◽

Ya Zhao ◽

Xiangfen Song ◽

...

Keyword(s):

Coronary Artery ◽

Intravascular Ultrasound ◽

Tissue Characterization ◽

Data Augmentation ◽

Automatic Segmentation ◽

Model Performance ◽

Single Frequency ◽

Elastic Membrane ◽

Ultrasound Images ◽

External Elastic Membrane

Abstract Background: Intravascular ultrasound (IVUS) is the golden standard in accessing the coronary lesions, stenosis, and atherosclerosis plaques. In this paper, a fully-automatic approach by an 8-layer U-Net is developed to segment the coronary artery lumen and the area bounded by external elastic membrane (EEM), i.e. cross section area (EEM-CSA). The database comprises of single-vendor and single-frequency IVUS data. Particularly, the proposed data augmentation of MeshGrid combined with flip and rotation operations is implemented, improving the model performance without pre- or post-processing of the raw IVUS images.Results: The mean intersection of union (mIoU) of 0.937 and 0.804 for the lumen and EEM-CSA respectively were achieved, which exceeded the manual labeling accuracy of the clinician. Conclusion: The accuracy shown by the proposed method is sufficient for subsequent reconstruction of 3D IVUS images, which is essential for doctors’ diagnosis in the tissue characterization of coronary artery walls and plaque compositions, qualitatively and quantitatively.

Download Full-text

Automatic segmentation of coronary lumen and external elastic membrane in intravascular ultrasound images using 8-layer U-Net

BioMedical Engineering OnLine ◽

10.1186/s12938-021-00852-0 ◽

2021 ◽

Vol 20 (1) ◽

Author(s):

Liang Dong ◽

Wenbing Jiang ◽

Wei Lu ◽

Jun Jiang ◽

Ya Zhao ◽

...

Keyword(s):

Coronary Artery ◽

Intravascular Ultrasound ◽

Tissue Characterization ◽

Data Augmentation ◽

Model Performance ◽

Single Frequency ◽

Elastic Membrane ◽

Ultrasound Images ◽

Cross Sectional ◽

External Elastic Membrane

Abstract Background Intravascular ultrasound (IVUS) is the golden standard in accessing the coronary lesions, stenosis, and atherosclerosis plaques. In this paper, a fully automatic approach by an 8-layer U-Net is developed to segment the coronary artery lumen and the area bounded by external elastic membrane (EEM), i.e., cross-sectional area (EEM-CSA). The database comprises single-vendor and single-frequency IVUS data. Particularly, the proposed data augmentation of MeshGrid combined with flip and rotation operations is implemented, improving the model performance without pre- or post-processing of the raw IVUS images. Results The mean intersection of union (MIoU) of 0.937 and 0.804 for the lumen and EEM-CSA, respectively, were achieved, which exceeded the manual labeling accuracy of the clinician. Conclusion The accuracy shown by the proposed method is sufficient for subsequent reconstruction of 3D-IVUS images, which is essential for doctors’ diagnosis in the tissue characterization of coronary artery walls and plaque compositions, qualitatively and quantitatively.

Download Full-text

Abstract 1165: Impact of Gender on Serial Atherosclerotic Changes in Response to Established Medical Therapies

Circulation ◽

10.1161/circ.116.suppl_16.ii_235-c ◽

2007 ◽

Vol 116 (suppl_16) ◽

Author(s):

Srinivasa R Kalidindi ◽

Amy Hsu ◽

Keon-Woong Moon ◽

E. Murat Tuzcu ◽

Steven E Nissen ◽

...

Keyword(s):

Coronary Artery ◽

Arterial Wall ◽

Elastic Membrane ◽

Single Coronary Artery ◽

Plaque Progression ◽

Ultrasound Evaluation ◽

External Elastic Membrane ◽

Remodeling Index ◽

Medical Therapies ◽

The Impact

Background: While the importance of coronary artery disease in females has become increasingly recognized, little is known regarding the impact of gender with regard to changes in arterial wall dimensions with progression and regression of atherosclerosis. This study investigated the remodeling response of the artery wall accompanying changes in atheroma burden in response to use of medical therapies, stratified according to gender. Methods: 1533 patients (27.5% female) underwent serial intravascular ultrasound evaluation of a single coronary artery in the context of clinical trials that assess the impact of medical therapies on plaque progression. The relationship between gender and remodeling of the arterial wall at baseline and its serial change in association with plaque progression and regression were studied. Results: Females were older (59 v 57 years, p<0.01), had a higher body mass index (31.5 v 29.5 kg/m 2 , p<0.01), were more likely to have hypertension (86 v 71.5%, p<0.01) and metabolic syndrome (57 v 49%, p<0.01) and less likely to have a history of smoking (57.5 v 73.5%, p=0.01) and myocardial infarction (27.5 v 35.5%, p<0.01). After adjusting for body surface area, females demonstrated a trend towards smaller external elastic membrane (EEM) (226.3 v 234.3 mm 3 , p=0.09) and larger lumen (143.7 v 137.7 mm 3 , p=0.01) volumes. The remodeling index at the most diseased site did not differ between genders (0.95 v 0.95, p=0.95). No differences were observed between genders with regard to changes in EEM (−5.6 v −6.2 mm 3 , p=0.29) and lumen (−4.9 v −4.5 mm 3 , p=0.82) volumes and remodeling index (−0.02 v −0.03, p=0.43) in response to use of medical therapies. Similarly, there were no differences between genders with regard to the percentage of patients undergoing expansion (34.7 v 35.5%, p=0.86) or contraction (20.4 v 21.8%, p=0.69) of lumen volume in association with regression of atherosclerotic plaque. Conclusion: A similar pattern of remodeling of the arterial wall was observed between genders in association with serial changes in atheroscle-rotic plaque. This further highlights our understanding of the pathological interactions between atherosclerosis and the arterial wall in females.

Download Full-text

Intravascular ultrasound

State of the Art Surgical Coronary Revascularization ◽

10.1093/med/9780198758785.003.0016 ◽

2021 ◽

pp. 60-63

Author(s):

Lorenz Räber

Keyword(s):

Coronary Artery ◽

Intravascular Ultrasound ◽

Coronary Intervention ◽

Elastic Membrane ◽

Arterial Grafts ◽

Anatomical Structures ◽

Percutaneous Coronary ◽

External Elastic Membrane ◽

Artery Disease ◽

Global And Local

Intravascular ultrasound represents the most established, well-validated, and widely used intracoronary imaging technology worldwide and was introduced approximately 25 years ago. Intravascular ultrasound enables the visualization of key anatomical structures of the coronary artery and saphenous or arterial grafts including the lumen, external elastic membrane, and adventitia and thereby provides the basis for the assessment of the degree of coronary artery stenosis, extent of atherosclerosis burden (i.e. global and local disease burden), and plaque composition. All this represents key information to plan and perform percutaneous coronary intervention procedures in native coronary artery disease or graft lesions and to estimate the risk for future cardiovascular events.

Download Full-text

Structural Thickening of Medial Layer in Coronary Artery With Spasm in Patients With Myocardial Infarction

Journal of the American Heart Association ◽

10.1161/jaha.120.018028 ◽

2021 ◽

Vol 10 (2) ◽

Author(s):

Takamitsu Nakamura ◽

Takeo Horikoshi ◽

Kiyotaka Kugiyama

Keyword(s):

Myocardial Infarction ◽

Optical Coherence Tomography ◽

Coronary Artery ◽

Coronary Artery Spasm ◽

Elastic Membrane ◽

Optical Coherence ◽

Unique Identifier ◽

External Elastic Membrane ◽

Relationship Of ◽

Medial Area

Background The underlying pathophysiology of coronary artery spasm (CAS) remains unclear. We aim to determine whether coronary artery medial layer thickness is associated with CAS using optical coherence tomography. Methods and Results A total of 50 patients with previous myocardial infarction underwent optical coherence tomography of the left anterior descending artery: 20 with CAS and 30 without CAS. Intimal and medial layer areas were measured by planimetric analysis of optical coherence tomography images. The medial area/external elastic membrane (EEM) area was significantly greater in patients with than without CAS (0.13±0.01 versus 0.09±0.01, respectively, P <0.01), whereas the intimal area/EEM area was similar in the 2 groups. In patients without CAS, the relationship of intimal area/EEM area with medial area/EEM area and coronary diameter response to intracoronary injection of acetylcholine was characterized by an inverted U‐shaped curve ( y =−1.85 x 2 +0.81 x +0.01, R 2 =0.43, P <0.001) and a U‐shaped curve ( y =2993.2 x 2 −1359.6 x +117.1, R 2 =0.53, P <0.001), respectively. Thus, the medial layer became thin and the contractile response became weak in coronary arteries with greater intimal area in the non‐CAS patients. In contrast, in patients with CAS, the intimal area/EEM area had no significant relationship with the medial area/EEM area in either linear correlation analysis or quadratic regression analysis. Thus, even when the intimal layer thickened, the medial layer did not thin in patients with CAS. Conclusions The structural thickness of the coronary medial layer was increased in patients with CAS, which may provide mechanistic insight into the pathogenesis of CAS. Registration URL: https://www.upload.umin.ac.jp ; Unique identifier: UMIN000018432.

Download Full-text

Data Augmentation Using Generative Adversarial Network for Automatic Machine Fault Detection Based on Vibration Signals

Applied Sciences ◽

10.3390/app11052166 ◽

2021 ◽

Vol 11 (5) ◽

pp. 2166

Author(s):

Van Bui ◽

Tung Lam Pham ◽

Huy Nguyen ◽

Yeong Min Jang

Keyword(s):

Fault Detection ◽

Data Augmentation ◽

Model Performance ◽

Original Data ◽

Fault Classification ◽

Training Process ◽

Generative Adversarial Network ◽

Data Set ◽

Adversarial Network ◽

Machine Fault

In the last decade, predictive maintenance has attracted a lot of attention in industrial factories because of its wide use of the Internet of Things and artificial intelligence algorithms for data management. However, in the early phases where the abnormal and faulty machines rarely appeared in factories, there were limited sets of machine fault samples. With limited fault samples, it is difficult to perform a training process for fault classification due to the imbalance of input data. Therefore, data augmentation was required to increase the accuracy of the learning model. However, there were limited methods to generate and evaluate the data applied for data analysis. In this paper, we introduce a method of using the generative adversarial network as the fault signal augmentation method to enrich the dataset. The enhanced data set could increase the accuracy of the machine fault detection model in the training process. We also performed fault detection using a variety of preprocessing approaches and classified the models to evaluate the similarities between the generated data and authentic data. The generated fault data has high similarity with the original data and it significantly improves the accuracy of the model. The accuracy of fault machine detection reaches 99.41% with 20% original fault machine data set and 93.1% with 0% original fault machine data set (only use generate data only). Based on this, we concluded that the generated data could be used to mix with original data and improve the model performance.

Download Full-text

Effects of sample size and data augmentation on U-Net-based automatic segmentation of various organs

Radiological Physics and Technology ◽

10.1007/s12194-021-00630-6 ◽

2021 ◽

Author(s):

Takafumi Nemoto ◽

Natsumi Futakami ◽

Etsuo Kunieda ◽

Masamichi Yagi ◽

Atsuya Takeda ◽

...

Keyword(s):

Sample Size ◽

Data Augmentation ◽

Automatic Segmentation

Download Full-text