Abstract:Visual or manual characterization and classification of atherosclerotic plaque lesions are tedious, error-prone, and time-consuming. The purpose of this study is to develop and design an automated carotid plaque characterization and classification system into binary classes, namely symptomatic and asymptomatic types via the deep learning (DL) framework implemented on a supercomputer. We hypothesize that on ultrasound images, symptomatic carotid plaques have (a) a low grayscale median because of a histologically large lipid core and relatively little collagen and calcium, and (b) a higher chaotic (heterogeneous) grayscale distribution due to the composition. The methodology consisted of building a DL model of Artificial Intelligence (called Atheromatic 2.0, AtheroPoint, CA, USA) that used a classic convolution neural network consisting of 13 layers and implemented on a supercomputer. The DL model used a cross-validation protocol for estimating the classification accuracy (ACC) and area-under-the-curve (AUC). A sample of 346 carotid ultrasound-based delineated plaques were used (196 symptomatic and 150 asymptomatic, mean age 69.9 ± 7.8 years, with 39% females). This was augmented using geometric transformation yielding 2312 plaques (1191 symptomatic and 1120 asymptomatic plaques). K10 (90% training and 10% testing) cross-validation DL protocol was implemented and showed an (i) accuracy and (ii) AUC without and with augmentation of 86.17%, 0.86 (p-value < 0.0001), and 89.7%, 0.91 (p-value < 0.0001), respectively. The DL characterization system consisted of validation of the two hypotheses: (a) mean feature strength (MFS) and (b) Mandelbrot's fractal dimension (FD) for measuring chaotic behavior. We demonstrated that both MFS and FD were higher in symptomatic plaques compared to asymptomatic plaques by 64.15 ± 0.73% (p-value < 0.0001) and 6 ± 0.13% (p-value < 0.0001), respectively. The benchmarking results show that DL with augmentation (ACC: 89.7%, AUC: 0.91 (p-value < 0.0001)) is superior to previously published machine learning (ACC: 83.7%) by 6.0%. The Atheromatic runs the test patient in < 2 s. Deep learning can be a useful tool for carotid ultrasound-based characterization and classification of symptomatic and asymptomatic plaques.

Validation of a fully automated deep learning-enabled solution for CCTA atherosclerotic plaque and stenosis quantification in a diverse real-world cohort

Coronary Artery Stenosis and High-Risk Plaque Assessed With an Unsupervised Fully Automated Deep Learning Technique

Automated vessel specific coronary artery calcification quantification with deep learning in a large multi-center registry

Novel near-infrared spectroscopy–intravascular ultrasound-based deep-learning methodology for accurate coronary computed tomography plaque quantification and characterization

Automated coronary artery atherosclerosis detection and weakly supervised localization on coronary CT angiography with a deep 3-dimensional convolutional neural network

Domain adaptive and fully automated carotid artery atherosclerotic lesion detection using an artificial intelligence approach (LATTE) on 3D MRI

Ultrasound-based internal carotid artery plaque characterization using deep learning paradigm on a supercomputer: a cardiovascular disease/stroke risk assessment system

Automated Characterization of Stenosis in Invasive Coronary Angiography Images with Convolutional Neural Networks

Clinical Validation of a Machine-Learned, Point-of-Care System to IDENTIFY Functionally Significant Coronary Artery Disease

Automated Deep Learning Analysis of Angiography Video Sequences for Coronary Artery Disease

Automatic A-line coronary plaque classification using combined deep learning and textural features in intravascular OCT images

Automated Detection and Classification of Coronary Atherosclerotic Plaques on Coronary CT Angiography Using Deep Learning Algorithm

High-quality annotations for deep learning enabled plaque analysis in SCAPIS cardiac computed tomography angiography

Artificial Intelligence to Assist in Exclusion of Coronary Atherosclerosis during CCTA Evaluation of Chest-Pain in the Emergency Department: Preparing an Application for Real-World Use

Automated diagnosis of atherosclerosis using multi-layer ensemble models and bio-inspired optimization in intravascular ultrasound imaging

Prospective deep learning-based quantitative assessment of coronary plaque by computed tomography angiography compared with intravascular ultrasound: the REVEALPLAQUE study

Automatic coronary artery segmentation and diagnosis of stenosis by deep learning based on computed tomographic coronary angiography

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

An end-to-end framework for intima media measurement and atherosclerotic plaque detection in the carotid artery

Histopathology-Based Deep-Learning Predicts Atherosclerotic Lesions in Intravascular Imaging

Coronary Artery Plaque Characterization from CCTA Scans using Deep Learning and Radiomics