An anatomically-informed correspondence initialisation method to improve learning-based registration for radiotherapy
This work addresses registration accuracy and efficiency in radiotherapy planning, but it is incremental as it builds on existing methods with a novel initialisation step.
The paper tackled the problem of improving interpatient CT non-rigid registration for radiotherapy by proposing an anatomically-informed initialisation method using learning-based correspondences, which reduced the mean distance-to-agreement by 1.8mm for included structures and 0.6mm for others while speeding up registration from 72 to 5 seconds.
We propose an anatomically-informed initialisation method for interpatient CT non-rigid registration (NRR), using a learning-based model to estimate correspondences between organ structures. A thin plate spline (TPS) deformation, set up using the correspondence predictions, is used to initialise the scans before a second NRR step. We compare two established NRR methods for the second step: a B-spline iterative optimisation-based algorithm and a deep learning-based approach. Registration performance is evaluated with and without the initialisation by assessing the similarity of propagated structures. Our proposed initialisation improved the registration performance of the learning-based method to more closely match the traditional iterative algorithm, with the mean distance-to-agreement reduced by 1.8mm for structures included in the TPS and 0.6mm for structures not included, while maintaining a substantial speed advantage (5 vs. 72 seconds).