Search

Background Deep learning (DL) algorithms are gaining extensive attention for their excellent performance in image recognition tasks. DL models can automatically make a quantitative assessment of complex medical image characteristics and achieve increased accuracy in diagnosis with higher efficiency. Purpose To determine the feasibility of using a DL approach to predict clinically negative axillary lymph node metastasis from US images in patients with primary breast cancer. Materials and Methods A data set of US images in patients with primary breast cancer with clinically negative axillary lymph nodes from Tongji Hospital (974 imaging studies from 2016 to 2018, 756 patients) and an independent test set from Hubei Cancer Hospital (81 imaging studies from 2018 to 2019, 78 patients) were collected. Axillary lymph node status was confirmed with pathologic examination. Three different convolutional neural networks (CNNs) of Inception V3, Inception-ResNet V2, and ResNet-101 architectures were trained on 90% of the Tongji Hospital data set and tested on the remaining 10%, as well as on the independent test set. The performance of the models was compared with that of five radiologists. The models' performance was analyzed in terms of accuracy, sensitivity, specificity, receiver operating characteristic curves, areas under the receiver operating characteristic curve (AUCs), and heat maps. Results The best-performing CNN model, Inception V3, achieved an AUC of 0.89 (95% confidence interval [CI]: 0.83, 0.95) in the prediction of the final clinical diagnosis of axillary lymph node metastasis in the independent test set. The model achieved 85% sensitivity (35 of 41 images; 95% CI: 70%, 94%) and 73% specificity (29 of 40 images; 95% CI: 56%, 85%), and the radiologists achieved 73% sensitivity (30 of 41 images; 95% CI: 57%, 85%; P = .17) and 63% specificity (25 of 40 images; 95% CI: 46%, 77%; P = .34). Conclusion Using US images from patients with primary breast cancer, deep learning models can effectively predict clinically negative axillary lymph node metastasis. Artificial intelligence may provide an early diagnostic strategy for lymph node metastasis in patients with breast cancer with clinically negative lymph nodes. Published under a CC BY 4.0 license. Online supplemental material is available for this article. See also the editorial by Bae in this issue.

Purpose: To determine whether carotid plaque neovascularization as assessed with contrast-enhanced ultrasonography (US) can help predict future coronary events in patients with stable coronary artery disease (CAD)., Materials and Methods: The study was approved by the hospital ethics committee, and informed consent was obtained from all patients. Three hundred twelve consecutive patients (228 men; mean age, 63 years ± 9; age range, 42-88 years) with both CAD and at least one carotid plaque thicker than 2.0 mm underwent both standard and contrast-enhanced carotid US. Patients with stable CAD were followed up for 8-47 months (mean, 33 months ± 9) or until a coronary event occurred. Statistical analysis was performed with the Student t test, χ(2) analysis, Kaplan-Meier method, and Cox proportional hazards regression models., Results: Contrast material enhancement of plaque was seen in 42 of 51 patients (82%) with acute coronary syndrome (ACS) and 114 of 261 patients (43.7%) with stable CAD (P < .001). Coronary events occurred during the follow-up period in 24 of 111 patients (21.6%) with contrast material enhancement of plaque and only seven of 137 patients (5.1%) without enhancement (P< .001). In 248 patients with stable CAD and follow-up, Kaplan-Meier analysis demonstrated a significantly higher probability of developing coronary events in patients with contrast material enhancement of plaque than in those without contrast material enhancement (P < .001). The presence of contrast material enhancement of plaque was a significant and independent predictor of future coronary events in patients with stable CAD (odds ratio: 3.90; 95% confidence interval: 1.60, 9.46; P = .003)., Conclusion: Contrast material enhancement of plaque is more common in patients with ACS than in those with stable CAD and is a significant and independent predictor of future coronary events in patients with stable CAD, suggesting that noninvasive contrast-enhanced carotid US may be used as a method for risk stratification of patients with stable CAD.

Purpose: To determine the correlation between the degree of plaque enhancement with contrast agent microbubbles and clinical symptoms in patients with carotid atherosclerotic plaque., Materials and Methods: The study was approved by the hospital ethical committee, and informed consent was obtained from all patients. One hundred four patients (83 men: mean age, 64 years +/- 9 [standard deviation]; 21 women: mean age, 61 years +/- 10) with carotid plaques were studied with standard and contrast material-enhanced ultrasonography (US). Contrast enhancement in the plaque was evaluated with visual interpretation and quantitative analysis., Results: Among the 104 patients, 35 (34%) had transient ischemic attack and/or cerebrovascular ischemic stroke. Plaque enhancement was found in 28 (80%) of 35 symptomatic patients and in 21 (30%) of 69 asymptomatic patients (P < .001). Enhanced intensity in the plaque (13.9 dB +/- 6.4) and the ratio of enhanced intensity in the plaque to that in the lumen of the carotid artery (0.54 +/- 0.23) in symptomatic patients were significantly greater than those in asymptomatic patients (8.8 dB +/- 5.2 [P < .001] and 0.33 +/- 0.19 [P < .001], respectively). Sensitivity and specificity were 74% and 62%, respectively, for enhanced intensity in the plaque (cutoff value, 10.0 dB) and 74% and 75%, respectively, for ratio of enhanced intensity in the plaque to that in the lumen of the carotid artery (cutoff value, 0.46)., Conclusion: Symptomatic patients had more intense contrast agent enhancement in the plaque than asymptomatic patients, suggesting that contrast-enhanced carotid US may be used for plaque risk stratification.