Diagnostic value of mammography density of breast masses by using deep learning

Abstract

ObjectiveIn 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.MethodsThis 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.ResultsIn 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).ConclusionsDeep 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.

A Systematic Literature Review on Mammography: Deep Learning in Redefining Breast Cancer Diagnosis for the Asian Perspective

“…After region growth, the unwanted pixel values are removed using the thresholding technique. U-Net (Chen et al, 2023b) Adapted-Black Widow Optimization used to increase the accuracy of segmentation and maximize convergence speed (Sivamurugan and Sureshkumar, 2023). For ResUnet rather than skip connection, res path used.…”

“…In the weakly supervised phase, different loss terms are activated using image-wise ground truth to train abnormality detection and segmentation tasks. (Chen et al, 2023b) C2FTrans is a multiscale segmentation framework built on coarse-to-fine transformers. It is made up of a cross-scale global transformer and a boundary-aware local transformer, that helps in the segmentation of medical images of various shapes and sizes.…”

“…This architecture uses TL to enhance the adaptability and efficiency of a CNN that has already been trained. (Chen et al, 2023b) (Alloqmani et al, 2023Al-Antari et al, 2020;Boudouh and Bouakkaz, 2023b,c;Jahangeer and Rajkumar, 2021;Cai et al, 2021;Rajakumari and Kalaivani, 2022;Quintana et al, 2023;Sivakrithika and Dinakaran, 2018) It is composed of layers, with each neuron connecting to all the neurons in the next layer. Trained on images, it learns to identify features and arrive at predictions, with the output layer employing softmax activation for class probability.…”