Background: We directly compared the mammography image processing results obtained with the help of the YOLOv4 convolutional neural network (CNN) model versus those obtained with the help of the NCA-based nested contours algorithm model. Method: We used 1080 images to train the YOLOv4, plus 100 images with proven breast cancer (BC) and 100 images with proven absence of BC to test both models. Results: the rates of true-positive, false-positive and false-negative outcomes were 60, 10 and 40, respectively, for YOLOv4, and 93, 63 and 7, respectively, for NCA. The sensitivities for the YOLOv4 and the NCA were comparable to each other for star-like lesions, masses with unclear borders, round- or oval-shaped masses with clear borders and partly visualized masses. On the contrary, the NCA was superior to the YOLOv4 in the case of asymmetric density and of changes invisible on the dense parenchyma background. Radiologists changed their earlier decisions in six cases per 100 for NCA. YOLOv4 outputs did not influence the radiologists’ decisions. Conclusions: in our set, NCA clinically significantly surpasses YOLOv4.

The amount of ultrasound (US) breast exams continues to grow because of the wider endorsement of breast cancer screening programs. When a solid lesion is found during the US the primary task is to decide if it requires a biopsy. Therefore, our goal was to develop a noninvasive US grayscale image analysis for benign and malignant solid breast lesion differentiation. We used a dataset consisting of 105 ultrasound images with 50 benign and 55 malignant non-cystic lesions. Features were extracted from the source image, the image of the gradient module after applying the Sobel filter, and the image after the Laplace filter. Subsequently, eight gray-level co-occurrence matrices (GLCM) were constructed for each lesion, and 13 Haralick textural features were calculated for each GLCM. Additionally, we computed the differences in feature values at different spatial shifts and the differences in feature values between the inner and outer areas of the lesion. The LASSO method was employed to determine the most significant features for classification. Finally, the lesion classification was carried out by various methods. The use of LASSO regression for feature selection enabled us to identify the most significant features for classification. Out of the 13 features selected by the LASSO method, four described the perilesional tissue, two represented the inner area of the lesion and five described the image of the gradient module. The final model achieved a sensitivity of 98%, specificity of 96%, and accuracy of 97%. Considering the perilesional area, Haralick feature differences, and the image of the gradient module can provide crucial parameters for accurate classification of US images. Features with a low AUC index (less than 0.6 in our case) can also be important for improving the quality of classification.

The incidence of asymmetries is 1-2% of all mammographic studies performed. However, there is no data in the literature on the prevalence of individual causes for mammographically determined asymmetries, as well as information about the opportunities of mammography and ultrasound examination (U/S) in their differential diagnosis. The aim of the study is to assess the incidence of individual causes for mammographically determined asymmetries and the opportunities of mammography and ultrasound of the mammary glands in their differential diagnosis. Material and methods. The results of mammographic and ultrasound examination of 202 patients were analyzed, in whom zones of asymmetrically distributed breast tissue were detected during mammography. Results. Technical asymmetries were detected in 14 out of 202 cases (6.93%) and could not be adequately analyzed. The true ones were asymmetries without space-occupying lesions caused by asymmetric uneven distribution of glandular (as a rule) or fibrous tissue (161 out of 202; 79.70%), as well as asymmetric edema of the breast parenchyma not associated with edematous infiltrative form of breast cancer (1 out of 202; 0.50%). These cases have always required performing breast ultrasound. The causes of true asymmetries caused by the presence of a space-occupying lesion were malignant neoplasms (breast cancer; 7 out of 202; 3.47%), benign formations (cysts, fibroadenomas, granulomas, abscesses, hematomas, chronic mastitis and other focal benign processes; 11 out of 202; 5.45%). These cases also required to perform breast ultrasound. The cause of the true asymmetries were also postoperative scars (8 out of 202; 3.96%), the differential diagnosis of which did not require ultrasound, but it and/or magnetic resonance mammography were required to assess the condition of the scar for its possible malignant transformation. Conclusions. The largest proportion of mammographically determined asymmetries in the density of mammary parenchyma (79.70%) are physiological variants of asymmetrically distributed tissues of this organ; their differential diagnosis and that of asymmetries caused by space-occupying lesions, according to mammography findings, is difficult, which requires additional ultrasound of the mammary glands. At the same time, mammography makes it possible to confidently identify technical asymmetries of the breast requiring no use of ultrasound, however, the interpretation of such sets of mammograms does not seem correct.

The lifetime probability of surgical interventions on mammary glands for benign pathology is approximately 4.8%, i.e. approximately they are performed in every 20th woman. At this, it is known that surgical interventions on organs containing an epithelial component significantly increase the risk of epithelial malignant tumors developing in them. The aim of the study was to assess the diagnostic value of magnetic resonance imaging (MRI) of the mammary glands with dynamic contrasting in identifying malignant transformation of breast scars after surgical interventions on it for benign pathology. Materials and methods. The study involved 63 patients aged 47.35±7.33 years who underwent breast surgery for benign pathology 1.5 years – 30 years before breast MRI was performed. Breast MRI with dynamic contrasting was performed using a Siemens Amira system (Germany) with a magnetic field intensity of 1.5 T. Study results. Morphological analysis: four types of morphological changes were identified during MRI in the area of previously performed intervention: 1) type 1 (thin-walled cavity / seroma; 5 out of 63; 7.94%; 0 cases of breast cancer); 2) type 2 (thin linear scar; 27 out of 63; 42.86%; 2 cases of breast cancer); 3) type 3 (mass lesion; 5 out of 63; 7.94%; 3 cases of breast cancer); 4) type 4 (fat necrosis; 26 out of 63; 41.30%; 1 case of breast cancer). Contrasting dynamics: 36 (57.14%) of 63 patients were noted to have contrast accumulation in the scar area. ACR type I of the kinetic curve (slow gradual accumulation of contrast) was observed in 10 out of 28 cases (35.71%) and in all cases corresponded to benign changes. ACR type II (rapid initial contrasting and plateau in the delayed phase) occurred in 10 out of 36 (27.78%) patients, of whom 2 had breast cancer. ACR type III (rapid initial contrasting and washout in the delayed phase) occurred in 13 of 36 (36.11%) patients, of whom 3 had breast cancer verified. Conclusions. The overall sensitivity of MRI with dynamic contrasting in detecting malignant transformation of scars in the mammary gland was 80%, specificity was 85.9%, the prognostic value of a positive result was 30.8%, the prognostic value of a negative result was 98.2%.