scholarly journals Diagnostic Accuracy and Generalizability of a Deep Learning-Based Fully Automated Algorithm for Coronary Artery Stenosis Detection on CCTA: A Multi-Centre Registry Study

2021 ◽  
Vol 8 ◽  
Author(s):  
Lixue Xu ◽  
Yi He ◽  
Nan Luo ◽  
Ning Guo ◽  
Min Hong ◽  
...  
Keyword(s):  
Deep Learning ◽  
Coronary Artery ◽  
Diagnostic Accuracy ◽  
Invasive Coronary Angiography ◽  
Area Under The Curve ◽  
Geographic Area ◽  
Ct Scanner ◽  
Significant Group ◽  
Automated Algorithm ◽  
Higher Sensitivity

Aims: In this retrospective, multi-center study, we aimed to estimate the diagnostic accuracy and generalizability of an established deep learning (DL)-based fully automated algorithm in detecting coronary stenosis on coronary computed tomography angiography (CCTA).Methods and results: A total of 527 patients (33.0% female, mean age: 62.2 ± 10.2 years) with suspected coronary artery disease (CAD) who underwent CCTA and invasive coronary angiography (ICA) were enrolled from 27 hospitals from January 2016 to August 2019. Using ICA as a standard reference, the diagnostic accuracy of the DL algorithm in the detection of ≥50% stenosis was compared to that of expert readers. In the vessel-based evaluation, the DL algorithm had a higher sensitivity (65.7%) and negative predictive value (NPV) (78.8%) and a significantly higher area under the curve (AUC) (0.83, p < 0.001). In the patient-based evaluation, the DL algorithm achieved a higher sensitivity (90.0%), NPV (52.2%) and AUC (0.81). Generalizability analysis of the DL algorithm was conducted by comparing its diagnostic performance in subgroups stratified by sex, age, geographic area and CT scanner type. The AUCs of the DL algorithm in the aforementioned subgroups ranged from 0.79 to 0.86 and from 0.75 to 0.93 in the vessel-based and patient-based evaluations, both without significant group differences (p > 0.05). The DL algorithm significantly reduced post-processing time (160 [IQR:139–192] seconds), in comparison to manual work (p < 0.001).Conclusions: The DL algorithm performed no inferior to expert readers in CAD diagnosis on CCTA and had good generalizability and time efficiency.

scholarly journals Improved computer-aided detection of pulmonary nodules via deep learning in the sinogram domain

2019 ◽  
Vol 2 (1) ◽  
Author(s):  
Yongfeng Gao ◽  
Jiaxing Tan ◽  
Zhengrong Liang ◽  
Lihong Li ◽  
Yumei Huo
Keyword(s):  
Deep Learning ◽  
Area Under The Curve ◽  
Pulmonary Nodules ◽  
Ct Image ◽  
Computer Aided Detection ◽  
Ct Scanner ◽  
Human Visual Perception ◽  
Computer Aided ◽  
Detection And Diagnosis ◽  
Cade System

AbstractComputer aided detection (CADe) of pulmonary nodules plays an important role in assisting radiologists’ diagnosis and alleviating interpretation burden for lung cancer. Current CADe systems, aiming at simulating radiologists’ examination procedure, are built upon computer tomography (CT) images with feature extraction for detection and diagnosis. Human visual perception in CT image is reconstructed from sinogram, which is the original raw data acquired from CT scanner. In this work, different from the conventional image based CADe system, we propose a novel sinogram based CADe system in which the full projection information is used to explore additional effective features of nodules in the sinogram domain. Facing the challenges of limited research in this concept and unknown effective features in the sinogram domain, we design a new CADe system that utilizes the self-learning power of the convolutional neural network to learn and extract effective features from sinogram. The proposed system was validated on 208 patient cases from the publicly available online Lung Image Database Consortium database, with each case having at least one juxtapleural nodule annotation. Experimental results demonstrated that our proposed method obtained a value of 0.91 of the area under the curve (AUC) of receiver operating characteristic based on sinogram alone, comparing to 0.89 based on CT image alone. Moreover, a combination of sinogram and CT image could further improve the value of AUC to 0.92. This study indicates that pulmonary nodule detection in the sinogram domain is feasible with deep learning.

scholarly journals Prediction of coronary artery calcium scoring from surface electrocardiogram in atherosclerotic cardiovascular disease: a pilot study

2020 ◽  
Vol 1 (1) ◽  
pp. 51-61
Author(s):  
Peter D Farjo ◽  
Naveena Yanamala ◽  
Nobuyuki Kagiyama ◽  
Heenaben B Patel ◽  
Grace Casaclang-Verzosa ◽  
...  
Keyword(s):  
Cardiovascular Risk ◽  
Coronary Artery ◽  
Risk Stratification ◽  
Coronary Artery Calcium ◽  
Invasive Coronary Angiography ◽  
Area Under The Curve ◽  
Suspected Coronary Artery Disease ◽  
Major Adverse Cardiovascular Events ◽  
Atherosclerotic Cardiovascular Disease ◽  
Cardiovascular Risk Stratification

Abstract Aims Coronary artery calcium (CAC) scoring is an established tool for cardiovascular risk stratification. However, the lack of widespread availability and concerns about radiation exposure have limited the universal clinical utilization of CAC. In this study, we sought to explore whether machine learning (ML) approaches can aid cardiovascular risk stratification by predicting guideline recommended CAC score categories from clinical features and surface electrocardiograms. Methods and results In this substudy of a prospective, multicentre trial, a total of 534 subjects referred for CAC scores and electrocardiographic data were split into 80% training and 20% testing sets. Two binary outcome ML logistic regression models were developed for prediction of CAC scores equal to 0 and ≥400. Both CAC = 0 and CAC ≥400 models yielded values for the area under the curve, sensitivity, specificity, and accuracy of 84%, 92%, 70%, and 75%, and 87%, 91%, 75%, and 81%, respectively. We further tested the CAC ≥400 model to risk stratify a cohort of 87 subjects referred for invasive coronary angiography. Using an intermediate or higher pretest probability (≥15%) to predict CAC ≥400, the model predicted the presence of significant coronary artery stenosis (P = 0.025), the need for revascularization (P < 0.001), notably bypass surgery (P = 0.021), and major adverse cardiovascular events (P = 0.023) during a median follow-up period of 2 years. Conclusion ML techniques can extract information from electrocardiographic data and clinical variables to predict CAC score categories and similarly risk-stratify patients with suspected coronary artery disease.

scholarly journals Diagnostic Accuracy of 128-Slice Dual Source CT Coronary Angiography with Invasive Catheter Coronary Angiography in a Tertiary Care Teaching Hospital

2021 ◽  
Vol 15 (6) ◽  
pp. 2057-2062
Author(s):  
Vishram Singh ◽  
Suresh Babu Kottapalli ◽  
Rakesh Gupta ◽  
Nitin Agarwal ◽  
Yogesh Yadav
Keyword(s):  
Coronary Artery Disease ◽  
Coronary Artery ◽  
Coronary Angiography ◽  
Diagnostic Accuracy ◽  
Coronary Arteries ◽  
Ct Coronary Angiography ◽  
Dual Source Ct ◽  
Ct Scanner ◽  
Dual Source ◽  
Artery Disease

Background: Coronary artery disease (CAD) morbidity and mortality increasing day by day in India as well as worldwide. Coronary arteries visualization by using invasive catheterization angiography is still using as a front-line diagnostic tool to evaluate the patients with CAD. 128 slice dual source CT improves the cardiac imaging such as high scanning speed, good temporal resolution and low radiation dose. Objective: To assess the diagnostic accuracy of 128-slice dual source CT cardiac angiography with conventional catheter angiography to find common arteries involved in CAD. Methods: This is a prospective, comparative, cross sectional study conducted at cardiology OPD. Patients with complaint of chest pain and suspected CAD were evaluated by CT and conventional coronary angiography and results were compared. Serum creatinine and ECG status were analyzed before the angiography. SIEMENS 128-slice Dual Source Flash Definition CT Scanner was used as a CT coronary angiography. Severity distribution of coronary artery disease, artery wise distribution of non-significant, significant lesions and coronary artery dominance pattern were analyzed and compared. Results: A total of 70 suspected CAD patients were selected and analyzed. American Heart Association (AHA) model of 17-segment was used to assess the coronary arteries. Normal angiograms reported in 15.71% patients and 58.57% had significant disease. A total of 356 lesions were identified from 690 out of 720 segments. Right coronary artery (RCA) is the most common location of significant lesions which contributes 33.5% (n=55/164). Coronary circulation of right-sided dominance was most commonly reported (70.0%). CT angiography showed 96.13% of an overall sensitivity, 96.28% specificity, 89.72% positive predictive value and 98.49% negative predictive value. Conclusion: 128-slice dual source CT scanner has showed high accuracy and act as non-invasive assessment of coronary arteries in patients with CAD Keywords: Cardiac angiography, Catheter coronary angiography, CT coronary angiography, 128-slice MDCT, Conventional angiography

scholarly journals Using Slit-Lamp Images for Deep Learning–Based Identification of Bacterial and Fungal Keratitis

2021 ◽  
Author(s):  
Ning Hung ◽  
Eugene Yu-Chuan Kang ◽  
Andy Guan-Yu Shih ◽  
Chi-Hung Lin ◽  
Ming‐Tse Kuo ◽  
...  
Keyword(s):  
Deep Learning ◽  
Diagnostic Accuracy ◽  
Learning Algorithm ◽  
Model Performance ◽  
Area Under The Curve ◽  
Fungal Keratitis ◽  
Classification Model ◽  
Mapping Technique ◽  
Slit Lamp ◽  
Deep Learning Algorithm

In this study, we aimed to develop a deep learning model for identifying bacterial keratitis (BK) and fungal keratitis (FK) by using slit-lamp images. We retrospectively collected slit-lamp images of patients with culture-proven microbial keratitis between January 1, 2010, and December 31, 2019, from two medical centers in Taiwan. We constructed a deep learning algorithm, consisting of a segmentation model for cropping cornea images and a classification model that applies convolutional neural networks to differentiate between FK and BK. The model performance was evaluated and presented as the area under the curve (AUC) of the receiver operating characteristic curves. A gradient-weighted class activation mapping technique was used to plot the heatmap of the model. By using 1330 images from 580 patients, the deep learning algorithm achieved an average diagnostic accuracy of 80.00%. The diagnostic accuracy for BK ranged from 79.59% to 95.91% and that for FK ranged from 26.31% to 63.15%. DenseNet169 showed the best model performance, with an AUC of 0.78 for both BK and FK. The heat maps revealed that the model was able to identify the corneal infiltrations. The model showed better diagnostic accuracy than the previously reported diagnostic performance of both general ophthalmologists and corneal specialists.

scholarly journals Myocardial perfusion quantification by CMR for detection of obstructive coronary artery disease in patients with previous coronary artery bypass surgery

2021 ◽  
Vol 22 (Supplement_2) ◽  
Author(s):  
R Franks ◽  
X Milidonis ◽  
H Morgan ◽  
M Ryan ◽  
D Perera ◽  
...  
Keyword(s):  
Coronary Artery ◽  
Myocardial Perfusion ◽  
Diagnostic Accuracy ◽  
Artery Bypass ◽  
Invasive Coronary Angiography ◽  
Bypass Graft ◽  
Visual Assessment ◽  
Stress Perfusion ◽  
Artery Disease ◽  
Perfusion Quantification

Abstract Funding Acknowledgements Type of funding sources: Other. Main funding source(s): British Heart Foundation Background Coronary artery bypass grafting (CABG) is an established treatment for patients with advanced coronary artery disease (CAD). A subsequent recurrence of symptoms can cause the need for re-assessment of the coronary circulation. The accuracy of visually assessed stress perfusion cardiovascular magnetic resonance (CMR) for the detection of obstructive CAD is reduced in patients with prior CABG. In patients with complex multi-vessel CAD, myocardial perfusion quantification by CMR is superior to visual assessment (VA) for detection of obstructive disease however patients with CABG have been absent from previous studies. Purpose This study sought to assess the performance of myocardial perfusion quantification by CMR against invasive coronary angiography (ICA) for detecting obstructive CAD in patients with previous CABG. Methods Twenty-nine patients with a history of previous CABG and subsequent clinically indicated perfusion CMR study and invasive coronary angiography were recruited. Patients underwent a dual bolus stress perfusion CMR with late gadolinium enhancement (LGE) imaging at 3 Tesla. Stress myocardial blood flow (MBF) was estimated at the coronary territory level according to the AHA 16 segment model using Fermi function-constrained deconvolution. Segments with transmural LGE were excluded from MBF analysis. Stress perfusion images were analysed visually alongside LGE images and matched perfusion-LGE defects were considered negative. On ICA, coronary territories with lumen stenosis >70% without an unobstructed bypass graft (<70% stenosis) were considered positive. Results 86/87 coronary territories were suitable for analysis. Sixty-five territories had at least one bypass graft including 32 territories with arterial grafts. 28/86 territories (33%) had obstructive disease on angiography. Territories with obstructive CAD had significantly lower stress MBF than unobstructed territories (1.21 [IQR: 0.96–1.45] vs 1.58 [1.40–1.84] ml/g/min, p < 0.001, Figure 1). Stress MBF had good accuracy to detect coronary territories with obstructive CAD (sensitivity 71%, specificity 84%, area under the curve (AUC) 0.83, p < 0.001, Figure 2A). For visual assessment, sensitivity was 79%, specificity 78% and diagnostic accuracy 78%. When analysis was confined to only territories with bypass grafts, stress MBF had 78% sensitivity, 81% specificity and AUC of 0.85, p < 0.001 (Figure 2B).. In this subgroup, VA had a sensitivity of 78%, specificity of 76% and a 77% diagnostic accuracy. Conclusions In patients with previous surgical revascularisation, quantification of stress myocardial blood flow by CMR offers good diagnostic accuracy for the detection and localisation of anatomically significant stenoses. Accuracy is reduced compared with published data in patients without coronary grafts but remains comparable to expert visual assessment.

Abstract 724: The Addition of Coronary Artery Calcium Scoring (CAC) to Myocardial Perfusion Imaging (MPI) Improves Diagnostic Accuracy in the Detection of Obstructive CAD: Results of a Prospective Multicenter Trial

Circulation ◽  
2008 ◽  
Vol 118 (suppl_18) ◽  
Author(s):  
Gregory S Thomas ◽  
Matthew J Budoff ◽  
Jennifer H Mieres ◽  
Ella A Kazerooni ◽  
Leslee J Shaw
Keyword(s):  
Coronary Artery ◽  
Coronary Angiography ◽  
Diagnostic Accuracy ◽  
Coronary Artery Calcium ◽  
Roc Analysis ◽  
Invasive Coronary Angiography ◽  
Multicenter Trial ◽  
Diagnostic Sensitivity ◽  
Coronary Artery Calcium Scoring ◽  
Prospective Multicenter Trial

Background: The PICTURE Trial evaluated the accuracy of 64-CT coronary angiography (CCTA) & MPI to invasive coronary angiography (ICA) for the detection of obstructive CAD. CCTA included an evaluation of Coronary Artery Calcium. The objective of this substudy was to evaluate the combination of measures of ischemia & atherosclerosis, with MPI & CAC respectively. Methods: 230 patients [52% male, 56.9 yrs] referred for MPI for chest pain were prospectively enrolled at 12 sites. 22% of patients underwent ICA for either abnormal MPI or coronary CCTA. The efficacy was assessed on a per-patient basis. Results: Of the 47 patients prospectively enrolled that underwent ICA, the prevalence of a >50% stenosis by ICA was 47% (22/47). Mean CAC for all patients was 434 +/−926. The frequency of obstructive CAD based on MPI & CAC are shown in the Figure . The odds (95% CI) for obstructive CAD is elevated 12.73-fold (2.43– 66.55) for ≥5% myocardium with stress defects (p<0.0001). Adding stress EF to perfusion results in a diagnostic sensitivity & specificity of 76.0% & 90.9% (p<0.0001). For CAC ≥100 alone, the diagnostic sensitivity & specificity is 76% & 72.7% (p=0.001). Adding a CAC score >100 to perfusion + EF increases the odds of obstructive CAD by 2.5-fold (1.31– 4.77, p=0.006). In patients with <5% defects, the addition of a CAC score of >100 increased the odds by 4.8-fold (0.95–24.14, p=0.057). ROC analysis revealed the % myocardium with stress defects categorized obstructive CAD with an area of 0.76 (95% CI=0.62– 0.90, p=0.002). Adding CAC improved the area to 0.85 (95% CI=0.74–0.96, p<0.0001). Conclusion: The addition of CAC to MPI improves diagnostic accuracy in the detection of obstructive CAD. Prevalence of ≥50% Obstructive CAD by MPI & CAC

Sign in / Sign up

Export Citation Format

Share Document