Classification in Dynamic Contrast-enhanced Breast MRI: Comparison of Cluster-based approach and BI-RADS Criteria

This study introduces an automatic classification of kinetic patterns on dynamic contrast-enhanced (DCE) breast MR images. We performed k-means cluster analysis technique from real world tumor cases into a reasonable number of reference pattern set that is representative of each cluster. This technique, then, classifies the tumor specific patterns from the given MRI data by measuring vector distances from the reference pattern set. The detections are also compared with color-highlighted texture patterns by reference pattern of BI-RAIDS kinetic criteria. This method will enhance the radiologists’ capability to efficiently identify and characterize the multi-dimensional patterns from tumors on DCE breast MRI.