Deep learning applied to two-dimensional color Doppler flow imaging ultrasound images significantly improves diagnostic performance in the classification of breast masses: a multicenter study / 中华医学杂志(英文版)

BACKGROUND@#The current deep learning diagnosis of breast masses is mainly reflected by the diagnosis of benign and malignant lesions. In China, breast masses are divided into four categories according to the treatment method: inflammatory masses, adenosis, benign tumors, and malignant tumors. These categorizations are important for guiding clinical treatment. In this study, we aimed to develop a convolutional neural network (CNN) for classification of these four breast mass types using ultrasound (US) images.@*METHODS@#Taking breast biopsy or pathological examinations as the reference standard, CNNs were used to establish models for the four-way classification of 3623 breast cancer patients from 13 centers. The patients were randomly divided into training and test groups (n = 1810 vs. n = 1813). Separate models were created for two-dimensional (2D) images only, 2D and color Doppler flow imaging (2D-CDFI), and 2D-CDFI and pulsed wave Doppler (2D-CDFI-PW) images. The performance of these three models was compared using sensitivity, specificity, area under receiver operating characteristic curve (AUC), positive (PPV) and negative predictive values (NPV), positive (LR+) and negative likelihood ratios (LR-), and the performance of the 2D model was further compared between masses of different sizes with above statistical indicators, between images from different hospitals with AUC, and with the performance of 37 radiologists.@*RESULTS@#The accuracies of the 2D, 2D-CDFI, and 2D-CDFI-PW models on the test set were 87.9%, 89.2%, and 88.7%, respectively. The AUCs for classification of benign tumors, malignant tumors, inflammatory masses, and adenosis were 0.90, 0.91, 0.90, and 0.89, respectively (95% confidence intervals [CIs], 0.87-0.91, 0.89-0.92, 0.87-0.91, and 0.86-0.90). The 2D-CDFI model showed better accuracy (89.2%) on the test set than the 2D (87.9%) and 2D-CDFI-PW (88.7%) models. The 2D model showed accuracy of 81.7% on breast masses ≤1 cm and 82.3% on breast masses >1 cm; there was a significant difference between the two groups (P < 0.001). The accuracy of the CNN classifications for the test set (89.2%) was significantly higher than that of all the radiologists (30%).@*CONCLUSIONS@#The CNN may have high accuracy for classification of US images of breast masses and perform significantly better than human radiologists.@*TRIAL REGISTRATION@#Chictr.org, ChiCTR1900021375; http://www.chictr.org.cn/showproj.aspx?proj=33139.

Carotid Plaque Stiffness Measured with Supersonic Shear Imaging and Its Correlation with Serum Homocysteine Level in Ischemic Stroke Patients

OBJECTIVE: To ascertain the feasibility of using shear wave velocity (SWV) in assessing the stiffness of carotid plaque by supersonic shear imaging (SSI) and explore preliminary clinical value for such evaluation. MATERIALS AND METHODS: Supersonic shear imaging was performed in 142 patients with ischemic stroke, including 76 males and 66 females with mean age of 66 years (range, 45–80 years). The maximum, minimum, and mean values of SWV were measured for 129 carotid plaques. SWVs were compared between echolucent and echogenic plaques. Correlations between SWVs and serum homocysteine levels were investigated. Based on neurological symptom, the surrogate marker of vulnerable plaque (VP), binary logistic regression was performed and area under curve (AUC) of homocysteine only and homocysteine combing SWVmean was calculated respectively. RESULTS: Echogenic plaques (n = 51) had higher SWVs than echolucent ones (n = 78) (SWVmin 3.91 [3.24–4.17] m/s vs. 1.51 [1.04–1.94] m/s; SWVmean, 4.29 [3.98–4.57] m/s vs. 2.09 [1.69–2.41] m/s; SWVmax, 4.67 [4.33–4.86] m/s vs. 2.62 [2.32–3.31] m/s all p values < 0.01). Pearson correlation analysis showed that stiffness of plaques was negatively correlated with homocysteine level. R values for SWVmin, SWVmean, and SWVmax were −0.205, −0.213, and −0.199, respectively. Binary logistic regression analysis showed that sex (p = 0.008), low-density lipoprotein (p = 0.015), triglycerides (p = 0.011), SWVmean (p = 0.004), and hyper-homocysteinemia (p = 0.010) were significantly associated with symptomatic ischemic stroke. Receiver operating characteristic curves revealed that SWVmean combing serum homocysteine level (AUC = 0.67) presented better diagnostic value than serum homocysteine only (AUC = 0.60) for symptomatic ischemic stroke. CONCLUSION: Supersonic shear imaging could be used to quantitatively evaluate stiffness of both echolucent and echogenic carotid plaques. More importantly, SWVs of plaques were not only correlated to serum homocysteine level, but also associated with symptomatic ischemic stroke, suggesting that SSI might be useful for understanding more about VP.