Abstract #0991

Support vector machine for breast cancer classification using DWI histogram features: preliminary study

Igor Vidić¹, Liv Egnell¹, Jose R. Teruel², Torill E. Sjøbakk³, Neil P. Jerome³, Agnes Østlie⁴, Hans E. Fjøsne^5,6, Tone F. Bathen³, and Pål Erik Goa¹

¹Department of Physics, Norwegian University of Science and Technology (NTNU), Trondheim, Norway, ²Department of Radiology, University of California, La Jolla, CA, United States, ³Department of Circulation and Medical Imaging, Norwegian University of Science and Technology (NTNU), Trondheim, Norway, ⁴Clinic of Radiology and Nuclear Medicine, St. Olavs University Hospital, Trondheim, Norway, ⁵Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology (NTNU), Trondheim, Norway, ⁶Department of Surgery, St. Olavs University Hospital, Trondheim, Norway

In this work we use the machine learning method support vector machine (SVM) to classify malignant and benign tumors, as well as ER+HER2- and ER+HER2+. As feature we use histogram properties of DWI-models (RED, ADC, IVIM) parameters as features. Our study showed that SVM classifiers using combinations of features from different models have predictive power in both analyses, also it performed better than SVM using combination of parameters obtained only from one of the models. The results are encouraging because SVM with DWI parameters can potentialy hinder unnecessary biopsies.

This abstract and the presentation materials are available to members only; a login is required.

Join Here