ABSTRACT
Background: Breast cancer is a major threat to women's health globally. Early detection of breast cancer is crucial for saving lives. One important early sign is the appearance of breast calcification in mammograms. Accurate segmentation and analysis of calcification can improve diagnosis and prognosis. However, small size and diffuse distribution make calcification prone to oversight. Purpose: This study aims to develop an efficient approach for segmenting and quantitatively analyzing breast calcification from mammograms. The goal is to assist radiologists in discerning benign versus malignant lesions to guide patient management. Methods: This study develops a framework for breast calcification segmentation and analysis using mammograms. A Pro_UNeXt algorithm is proposed to accurately segment calcification lesions by enhancing the UNeXt architecture with a microcalcification detection block, fused-MBConv modules, multiple-loss-function training, and data augmentation. Quantitative features are then extracted from the segmented calcification, including morphology, size, density, and spatial distribution. These features are used to train machine learning classifiers to categorize lesions as malignant or benign. Results: The proposed Pro_UNeXt algorithm achieved superior segmentation performance versus UNet and UNeXt models on both public and private mammogram datasets. It attained a Dice score of 0.823 for microcalcification detection on the public dataset, demonstrating its accuracy for small lesions. For quantitative analysis, the extracted calcification features enabled high malignant/benign classification, with AdaBoost reaching an AUC of 0.97 on the private dataset. The consistent results across datasets validate the representative and discerning capabilities of the proposed features. Conclusion: This study develops an efficient framework integrating customized segmentation and quantitative analysis of breast calcification. Pro_UNeXt offers precise localization of calcification lesions. Subsequent feature quantification and machine learning classification provide comprehensive malignant/benign assessment. This end-to-end solution can assist clinicians in early diagnosis, treatment planning, and follow-up for breast cancer patients.
ABSTRACT
Objective: In order to explore the relationship between mammographic density of breast mass and its surrounding area and benign or malignant breast, this paper proposes a deep learning model based on C2FTrans to diagnose the breast mass using mammographic density. Methods: This retrospective study included patients who underwent mammographic and pathological examination. Two physicians manually depicted the lesion edges and used a computer to automatically extend and segment the peripheral areas of the lesion (0, 1, 3, and 5 mm, including the lesion). We then obtained the mammary glands' density and the different regions of interest (ROI). A diagnostic model for breast mass lesions based on C2FTrans was constructed based on a 7: 3 ratio between the training and testing sets. Finally, receiver operating characteristic (ROC) curves were plotted. Model performance was assessed using the area under the ROC curve (AUC) with 95% confidence intervals (CI), sensitivity, and specificity. Results: In total, 401 lesions (158 benign and 243 malignant) were included in this study. The probability of breast cancer in women was positively correlated with age and mass density and negatively correlated with breast gland classification. The largest correlation was observed for age (r = 0.47). Among all models, the single mass ROI model had the highest specificity (91.8%) with an AUC = 0.823 and the perifocal 5mm ROI model had the highest sensitivity (86.9%) with an AUC = 0.855. In addition, by combining the cephalocaudal and mediolateral oblique views of the perifocal 5 mm ROI model, we obtained the highest AUC (AUC = 0.877 P < 0.001). Conclusions: Deep learning model of mammographic density can better distinguish benign and malignant mass-type lesions in digital mammography images and may become an auxiliary diagnostic tool for radiologists in the future.
ABSTRACT
Background: Architectural distortion (AD) is a common imaging manifestation of breast cancer, but is also seen in benign lesions. This study aimed to construct deep learning models using mask regional convolutional neural network (Mask-RCNN) for AD identification in full-field digital mammography (FFDM) and evaluate the performance of models for malignant AD diagnosis. Methods: This retrospective diagnostic study was conducted at the Second Affiliated Hospital of Guangzhou University of Chinese Medicine between January 2011 and December 2020. Patients with AD in the breast in FFDM were included. Machine learning models for AD identification were developed using the Mask RCNN method. Receiver operating characteristics (ROC) curves, their areas under the curve (AUCs), and recall/sensitivity were used to evaluate the models. Models with the highest AUCs were selected for malignant AD diagnosis. Results: A total of 349 AD patients (190 with malignant AD) were enrolled. EfficientNetV2, EfficientNetV1, ResNext, and ResNet were developed for AD identification, with AUCs of 0.89, 0.87, 0.81 and 0.79. The AUC of EfficientNetV2 was significantly higher than EfficientNetV1 (0.89 vs. 0.78, P=0.001) for malignant AD diagnosis, and the recall/sensitivity of the EfficientNetV2 model was 0.93. Conclusion: The Mask-RCNN-based EfficientNetV2 model has a good diagnostic value for malignant AD.
ABSTRACT
Purpose: To compare the mammographic malignant architectural distortion (AD) detection performance of radiologists who read mammographic examinations unaided versus those who read these examinations with the support of artificial intelligence (AI) systems. Material and Methods: This retrospective case-control study was based on a double-reading of clinical mammograms between January 2011 and December 2016 at a large tertiary academic medical center. The study included 177 malignant and 90 benign architectural distortion (AD) patients. The model was built based on the ResNeXt-50 network. Algorithms used deep learning convolutional neural networks, feature classifiers, image analysis algorithms to depict AD and output a score that translated to malignant. The accuracy for malignant AD detection was evaluated using area under the curve (AUC). Results: The overall AUC was 0.733 (95% CI, 0.673-0.792) for Reader First-1, 0.652 (95% CI, 0.586-0.717) for Reader First-2, and 0.655 (95% CI, 0.590-0.719) for Reader First-3. and the overall AUCs for Reader Second-1, 2, 3 were 0.875 (95% CI, 0.830-0.919), 0.882 (95% CI, 0.839-0.926), 0.884 (95% CI, 0.841-0.927),respectively. The AUCs for all the reader-second radiologists were significantly higher than those for all the reader-first radiologists (Reader First-1 vs. Reader Second-1, P= 0.004). The overall AUC was 0.792 (95% CI, 0.660-0.925) for AI algorithms. The combination assessment of AI algorithms and Reader First-1 achieved an AUC of 0.880 (95% CI, 0.793-0.968), increased than the Reader First-1 alone and AI algorithms alone. AI algorithms alone achieved a specificity of 61.1% and a sensitivity of 80.6%. The specificity for Reader First-1 was 55.5%, and the sensitivity was 86.1%. The results of the combined assessment of AI and Reader First-1 showed a specificity of 72.7% and sensitivity of 91.7%. The performance showed significant improvements compared with AI alone (p<0.001) as well as the reader first-1 alone (p=0.006). Conclusion: While the single AI algorithm did not outperform radiologists, an ensemble of AI algorithms combined with junior radiologist assessments were found to improve the overall accuracy. This study underscores the potential of using machine learning methods to enhance mammography interpretation, especially in remote areas and primary hospitals.