Analyzing breast tumor heterogeneity to predict the response to chemotherapy using 3D MR images registration

Abstract

Breast tumor structure is extremely heterogeneous. This heterogeneity changes spatially during chemotherapy treatment. This was correlated with how the tumor reacts to Neoadjuvant chemotherapy (treatment presiding surgery). A significant number of studies adopting Magnetic Resonance Imaging (MRI) have looked into the quantification of intratumor heterogeneity of breast cancer. Nevertheless, a limited number of them are interested chiefly in evaluating breast cancer heterogeneity as index of response to treatment. In this paper, we present a new approach that compares breast tumor heterogeneity degree, using acquired images before and after chemotherapy. The purpose of our study is to help radiologists to predict the effectiveness of Neoadjuvant chemotherapy as soon as possible. Indeed, many patients with breast cancer have a non-responding tumor type to chemotherapy. However, these patients still requiring avoidable chemotherapies during a long time, which causes several undesirable effects. We evaluate our study using a data set constructed by Dynamic Contrast Enhanced (DCE-MRI) and Diffusion Weighed sequences of MRI (DW-MRI), consisting of 64 adult patients. Our approach consists to apply a volumetric registration between images acquired before and after chemotherapy to quantify induced changes on the tumor by the first chemotherapy session. The first step of our approach is to segment the volume of interest (VOI). Then, a coherent volumetric registration will provide to make a voxel by voxel comparison of breast tumor volume. Therefore, The breast tumor response rate to the first chemotherapy session was obtained, by comparing each voxel intensity in DCE-MR and DW-MRI sequences before and after the chemotherapy. This approach will not only provide the breast tumor response degree to chemotherapy, but also monitoring tumor regions that have responded, not responded and tumor regions that have recognized disease progression during chemotherapy session. This will allow radiologists and oncologists to decide if the patient will continue to require chemotherapy, or applying other alternative solutions, without wasting time in unnecessary chemotherapy sessions.