A convolutional deep learning model for improving mammographic breast-microcalcification diagnosis

This study aimed to assess the diagnostic performance of deep convolutional neural networks (DCNNs) in classifying breast microcalcification in screening mammograms. To this end, 1579 mammographic images were collected retrospectively from patients exhibiting suspicious microcalcification in screening mammograms between July 2007 and December 2019. Five pre-trained DCNN models and an ensemble model were used to classify the microcalcifications as either malignant or benign. Approximately one million images from the ImageNet database had been used to train the five DCNN models. Herein, 1121 mammographic images were used for individual model fine-tuning, 198 for validation, and 260 for testing. Gradient-weighted class activation mapping (Grad-CAM) was used to confirm the validity of the DCNN models in highlighting the microcalcification regions most critical for determining the final class. The ensemble model yielded the best AUC (0.856). The DenseNet-201 model achieved the best sensitivity (82.47%) and negative predictive value (NPV; 86.92%). The ResNet-101 model yielded the best accuracy (81.54%), specificity (91.41%), and positive predictive value (PPV; 81.82%). The high PPV and specificity achieved by the ResNet-101 model, in particular, demonstrated the model effectiveness in microcalcification diagnosis, which, in turn, may considerably help reduce unnecessary biopsies.

Breast Cancer Diagnosis by Convolutional Neural Network and Advanced Thermal Exchange Optimization Algorithm

A common gynecological disease in the world is breast cancer that early diagnosis of this disease can be very effective in its treatment. The use of image processing methods and pattern recognition techniques in automatic breast detection from mammographic images decreases human errors and increments the rapidity of diagnosis. In this paper, mammographic images are analyzed using image processing techniques and a pipeline structure for the diagnosis of the cancerous masses. In the first stage, the quality of mammogram images and the contrast of abnormal areas in the image are improved by using image contrast improvement and a noise decline. A method based on color space is then used for image segmentation that is followed by mathematical morphology. Then, for feature image extraction, a combined gray-level cooccurrence matrix (GLCM) and discrete wavelet transform (DWT) method is used. At last, a new optimized version of convolutional neural network (CNN) and a new improved metaheuristic, called Advanced Thermal Exchange Optimizer, are used for the classification of the features. A comparison of the simulations of the proposed technique with three different techniques from the literature applied on the MIAS mammogram database is performed to show its superiority. Results show that the accuracy of diagnosing cancer cases for the proposed method and applied on the MIAS database is 93.79%, and sensitivity and specificity are obtained 96.89% and 67.7%, respectively.

Lesion Detection in Breast Tomosynthesis Using Efficient Deep Learning and Data Augmentation Techniques

Digital breast tomosynthesis (DBT) is one of the powerful breast cancer screening technologies. DBT can improve the ability of radiologists to detect breast cancer, especially in the case of dense breasts, where it beats mammography. Although many automated methods were proposed to detect breast lesions in mammographic images, very few methods were proposed for DBT due to the unavailability of enough annotated DBT images for training object detectors. In this paper, we present fully automated deep-learning breast lesion detection methods. Specifically, we study the effectiveness of two data augmentation techniques (channel replication and channel-concatenation) with five state-of-the-art deep learning detection models. Our preliminary results on a challenging publically available DBT dataset showed that the channel-concatenation data augmentation technique can significantly improve the breast lesion detection results for deep learning-based breast lesion detectors.

Classification of Mammographic Images by Openvino: A Proposal of use to Enhance More Effectivity in Cancer Diagnosis

Diseases that are characterized by the disordered growth of cells that, in many cases, have the property of invading tissues and organs are commonly called cancer. Such cells divide quickly and the invasion can be very aggressive and uncontrolled, resulting in formation of malignant tumors Mammographic images from libraries of the American digital database DDSM were used in this research for digital improvement and characteristic analysis using the OpenVino computer program This work has as main objective to analyze mammography images of breast nodules and to propose a method of classification by shape and texture using computer programs that can maximize the accuracy in the correct diagnosis regarding the malignancy or not of a tumor. It is a tool that it can be useful as a contribution in the interpretation of the results to mastologists who identify such nodules through the analyzed radiological images.

Automatic detection of various patterns of calcification distribution on mammograms: a graph convolution network approach (Preprint)

UNSTRUCTURED Purpose: To develop a novel artificial intelligence (AI) model algorithm focusing on automatic detection and classification of various patterns of calcification distribution in mammographic images using a unique graph convolution approach. Materials and methods: Images from 200 patients classified as Category 4 or 5 according to the American College of Radiology Breast Imaging Reporting and Database System, which showed calcifications according to the mammographic reports and diagnosed breast cancers. The calcification distributions were classified as either diffuse, segmental, regional, grouped, or linear. Excluded were mammograms with (1) breast cancer as a single or combined characterization such as a mass, asymmetry, or architectural distortion with or without calcifications; (2) hidden calcifications that were difficult to mark; or (3) incomplete medical records. Results: A graph convolutional network-based model was developed. 401 mammographic images from 200 cases of breast cancer were divided based on calcification distribution pattern: diffuse (n = 24), regional (n = 111), group (n = 201), linear (n = 8) or segmental (n = 57). The classification performances were measured using metrics including precision, recall, F1 score, accuracy and multi-class area under receiver operating characteristic curve. The proposed achieved precision of 0.483 ± 0.015, sensitivity of 0.606 (0.030), specificity of 0.862 ± 0.018, F1 score of 0.527 ± 0.035, accuracy of 60.642% ± 3.040% and area under the curve of 0.754 ± 0.019, finding method to be superior compared to all baseline models. The predicted linear and diffuse classifications were highly similar to the ground truth, and the predicted grouped and regional classifications were also superior compared to baseline models. Conclusion: The proposed deep neural network framework is an AI solution to automatically detect and classify calcification distribution patterns on mammographic images highly suspected of showing breast cancers. Further study of the AI model in an actual clinical setting and additional data collection will improve its performance.

Triage of 2D Mammographic Images Using Multi-view Multi-task Convolutional Neural Networks

With an aging and growing population, the number of women receiving mammograms is increasing. However, existing techniques for autonomous diagnosis do not surpass a well-trained radiologist. Therefore, to reduce the number of mammograms that require examination by a radiologist, subject to preserving the diagnostic accuracy observed in current clinical practice, we develop Man and Machine Mammography Oracle (MAMMO)—a clinical decision support system capable of determining whether its predicted diagnoses require further radiologist examination. We first introduce a novel multi-view convolutional neural network (CNN) trained using multi-task learning (MTL) to diagnose mammograms and predict the radiological assessments known to be associated with cancer. MTL improves diagnostic performance and triage efficiency while providing an additional layer of model interpretability. Furthermore, we introduce a novel triage network that takes as input the radiological assessment and diagnostic predictions of the multi-view CNN and determines whether the radiologist or CNN will most likely provide the correct diagnosis. Results obtained on a dataset of over 7,000 patients show that MAMMO reduced the number of diagnostic mammograms requiring radiologist reading by 42.8% while improving the overall diagnostic accuracy in comparison to readings done by radiologists alone.