A new approach to develop computer-aided diagnosis scheme of breast mass classification using deep learning...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Mammography is crucial for breast cancer screening.
• Accurate classification of malignant and benign masses is essential for patient management.
• Computer-aided diagnosis (CAD) systems can aid radiologists in interpreting mammograms.

Purpose of the Study:

• To develop and evaluate a deep learning-based CAD scheme for classifying malignant and benign breast masses from mammograms.
• To assess the performance of the proposed CAD scheme using a dataset of digital mammograms.

Main Methods:

• An 8-layer deep learning network with convolution-max-pooling layers was designed for automatic feature extraction.
• A multilayer perceptron (MLP) classifier was employed for feature categorization.
• A four-fold cross-validation method was used to train and test the deep learning network on 560 regions of interest (ROIs) from digital mammograms.

Main Results:

• The CAD scheme achieved an area under the receiver operating characteristic curve (AUC) ranging from 0.696±0.044 to 0.836±0.036 across four testing folds.
• The overall AUC for the entire dataset was 0.790±0.019, indicating good classification performance.
• The system successfully classified malignant and benign masses without requiring lesion segmentation or manual feature engineering.

Conclusions:

• Deep learning-based CAD schemes are feasible for classifying breast masses in mammograms.
• The proposed method offers an automated approach, eliminating the need for manual feature extraction and selection.
• This technology has the potential to improve the accuracy and efficiency of mammogram interpretation.