This study investigates the use of the unsupervised deep learning framework VoxelMorph for deformable registration of longitudinal abdominopelvic CT images acquired in patients with bone metastases from breast cancer. The CT images were refined prior to registration by automatically removing the CT table and all other extra-corporeal components. To improve the learning capabilities of VoxelMorph when only a limited amount of training data is available, a novel incremental training strategy is proposed based on simulated deformations of consecutive CT images. In a 4-fold cross-validation scheme, the incremental training strategy achieved significantly better registration performance compared to training on a single volume. Although our deformable image registration method did not outperform iterative registration using NiftyReg (considered as a benchmark) in terms of registration quality, the registrations were approximately 300 times faster. This study showed the feasibility of deep learning based deformable registration of longitudinal abdominopelvic CT images via a novel incremental training strategy based on simulated deformations.
@article{arxiv.2005.07545,
title = {3D deformable registration of longitudinal abdominopelvic CT images using unsupervised deep learning},
author = {Maureen van Eijnatten and Leonardo Rundo and K. Joost Batenburg and Felix Lucka and Emma Beddowes and Carlos Caldas and Ferdia A. Gallagher and Evis Sala and Carola-Bibiane Schönlieb and Ramona Woitek},
journal= {arXiv preprint arXiv:2005.07545},
year = {2020}
}