Diagnosis of Benign or Malignant Mammography Using Radiomics with Generated Whole Breast Masks

In recent years, research in computer-aided medical image diagnosis has advanced, and studies on the benign and malignant diagnosis of mammography, which has become widespread in breast cancer screening, have also progressed. Deep learning is commonly used in computer-aided medical image diagnosis, but research utilizing Radiomics for tumor diagnosis is also advancing. However, Radiomics, which requires tumor masks for each image as input, incurs high labor costs from specialized medical professionals to create these masks. Therefore, when we input whole breast masks generated from mammography into Radiomics and compared the results of benign and malignant classification, it was suggested that although it does not surpass the use of tumor masks, it is still a valuable approach.