Automated detection and classification of coronary atherosclerotic plaques on coronary CT angiography using deep learning algorithm

Abstract

Background: Coronary artery disease (CAD) is the leading cause of mortality worldwide. Recent advances in deep learning technology promise better diagnosis of CAD and improve assessment of CAD plaque buildup. The purpose of this study is to assess the performance of a deep learning algorithm in detecting and classifying coronary atherosclerotic plaques in coronary computed tomographic angiography (CCTA) images.

Methods: Between January 2019 and September 2020, CCTA images of 669 consecutive patients with suspected CAD from Nanjing Drum Tower Hospital Clinical College of Nanjing University of Chinese Medicine were included in this study. There were 106 patients included in the retrospective plaque detection analysis, which was evaluated by a deep learning algorithm and four independent physicians with varying clinical experience. Additionally, 563 patients were included in the analysis for plaque classification using the deep learning algorithm, and their results were compared with those of expert radiologists. Plaques were categorized as absent, calcified, non-calcified, or mixed.

Results: The deep learning algorithm exhibited higher sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy {92% [95% confidence interval (CI): 89.5-94.1%], 87% (95% CI: 84.2-88.5%), 79% (95% CI: 76.1-82.4%), 95% (95% CI: 93.4-96.3%), and 89% (95% CI: 86.9-90.0%)} compared to physicians with ≤5 years of clinical experience in CAD diagnosis for the detection of coronary plaques. The algorithm's overall sensitivity, specificity, PPV, NPV, accuracy, and Cohen's kappa for plaque classification were 94% (95% CI: 92.3-94.7%), 90% (95% CI: 88.8-90.3%), 70% (95% CI: 68.3-72.1%), 98% (95% CI: 97.8-98.5%), 90% (95% CI: 89.8-91.1%) and 0.74 (95% CI: 0.70-0.78), indicating strong performance.

Conclusions: The deep learning algorithm has demonstrated reliable and accurate detection and classification of coronary atherosclerotic plaques in CCTA images. It holds the potential to enhance the diagnostic capabilities of junior radiologists and junior intervention cardiologists in the CAD diagnosis, as well as to streamline the triage of patients with acute coronary symptoms.

Keywords: Coronary computed tomographic angiography (CCTA); atherosclerotic plaques; coronary artery disease (CAD); deep learning.

Figure 1

A flow diagram of study population. CCTA, coronary computed tomographic angiography.

Figure 2

Classification of coronary atherosclerotic plaques using CTA by two experienced cardiovascular radiologists on consensus as internal reference for deep learning. (A) Absence of plaques in the left anterior descending artery; (B) calcified plaques (arrow); (C) non-calcified plaque (arrow); and (D) mixed plaque (calcified and non-calcified, open arrow). CTA, computed tomographic angiography.

Figure 3

Schematic diagram of the 18 coronary artery segments. 1, proximal right coronary artery (pRCA); 2, middle right coronary artery (mRCA); 3, distal right coronary artery (dRCA); 4, right posterior descending artery (R-PDA); 5, left main (LM); 6, proximal left anterior descending (pLAD); 7, middle left anterior descending (mLAD); 8, distal left anterior descending (dLAD); 9, diagonal 1 (D1); 10, diagonal 2 (D2); 11, proximal circumflex (pCx); 12, obtuse marginal 1 (OM1); 13, mid and distal left circumflex (LCx); 14, obtuse marginal 2 (OM2); 15, posterior descending artery from LCx (L-PDA); 16, right posterior lateral branch (R-PLB); 17, ramus intermedius (RI); 18, posterior lateral branch from LCx (L-PLB).

Figure 4

Schematic diagram of the deep learning-based plaque detection algorithm workflow. (A) Stack of transverse images; (B) segmentations of aorta and coronary arteries (red area); (C) volume rending image of aorta and coronary arteries; (D) straightened rendering image of the left anterior descending artery; (E) architect of three-dimensional U-Net2 model; and (F) two probability curves corresponding to calcified and noncalcified plaques, respectively.

Figure 5

Boxplots and scatterplots of interpretation time by radiologists and deep learning. (A) There is significant reduction in interpretation time by deep learning for the classification of coronary artery atherosclerosis (mean of 298 s) as opposed to 658 s by radiologists; (B) there is only one plaque that shows longer interpretation time by deep learning than by radiologist. AI, artificial intelligence.