A Comparison of Breast Tumors Classifications Based on Mammography and Digital Breast Tomosynthesis: A Fractal Analysis Study

摘要

In this study, we present a radiomics approach to classify the benign and malignant breast tumors based on the fractal features: fractal dimension and lacunarity. 91 cases samples were studied and the features extracted from both mammography(MAMMO) and digital breast tomosynthesis(DBT) were involved to perform the prediction. With three classification algorithms applied, the classification models gave accuracies from 0.61 to 0.85 in the prediction based on MAMMO, and accuracies from 0.85 to 0.91 in the prediction based on DBT.