Meeting Banner
Abstract #4336

Complementary value of End-to-end Deep Learning and Radiomics in Breast Cancer Classification on Diffusion-Weighted MRI

Paul F. Jaeger1, Sebastian Bickelhaupt2, Frederik Bernd Laun3,4, Wolfgang Lederer5, Heidi Daniel6, Tristan Anselm Kuder4, Stefan Delorme2, Heinz-Peter Schlemmer2, Franziska Steudle2, and Klaus Hermann Maier-Hein1

1Division of Medical Image Computing, German Cancer Research Center, Heidelberg, Germany, 2Department of Radiology, German Cancer Research Center, Heidelberg, Germany, 3Institute of Radiology, University Hospital Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany, 4Medical Physics in Radiology, German Cancer Research Center, Heidelberg, Germany, 5Radiological Practice, ATOS Clinic, Heidelberg, Germany, 6Radiology Center Mannheim, Mannheim, Germany

Two fundamentally different approaches have been proposed recently for the classification of breast lesions on diffusion-weighted MRI Images: “Radiomics” extracts quantitative parameters by fitting a biophysical model to the q-space signal and subsequently computes handcrafted features to feed a classifier. Convolutional neural networks on the other hand autonomously learn all processing components in an end-to-end training. To date it is unclear how the two methods compare with respect to overall performance, complementary value of features and combinability. We address these open research questions and propose a combined model that significantly outperforms the two standalone approaches.

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

Join Here