Deep learning-based group-wise registration for longitudinal MRI analysis in glioma
This work addresses the problem of efficient and accurate glioma growth analysis for medical researchers, though it is incremental as it builds on existing registration methods with a learning-based approach.
The paper tackled the challenge of longitudinal MRI registration for glioma, which is complicated by mass-effects and tissue changes, by proposing a deep learning-based group-wise registration method that achieved comparable Dice coefficients and more detailed registrations while reducing runtime to under a minute.
Glioma growth may be quantified with longitudinal image registration. However, the large mass-effects and tissue changes across images pose an added challenge. Here, we propose a longitudinal, learning-based, and groupwise registration method for the accurate and unbiased registration of glioma MRI. We evaluate on a dataset from the Glioma Longitudinal AnalySiS consortium and compare it to classical registration methods. We achieve comparable Dice coefficients, with more detailed registrations, while significantly reducing the runtime to under a minute. The proposed methods may serve as an alternative to classical toolboxes, to provide further insight into glioma growth.