Multimodality image registration in the head‐and‐neck using a deep learning‐derived synthetic CT as a bridge

Purpose To develop and demonstrate the efficacy of a novel head‐and‐neck multimodality image registration technique using deep‐learning‐based cross‐modality synthesis. Methods and Materials Twenty‐five head‐and‐neck patients received magnetic resonance (MR) and computed tomography (CT) (CT aligned ) scans on the same day with the same immobilization. Fivefold cross validation was used with all of the MR‐CT pairs to train a neural network to generate synthetic CTs from MR images. Twenty‐four of 25 patients also had a separate CT without immobilization (CT non‐aligned ) and were used for testing. CT non‐aligned 's were deformed to the synthetic CT, and compared to CT non‐aligned registered to MR. The same registrations were performed from MR to CT non‐aligned and from synthetic CT to CT non‐aligned . All registrations used B‐splines for modeling the deformation, and mutual information for the objective. Results were evaluated using the 95% Hausdorff distance among spinal cord contours, landmark error, inverse consistency, and Jacobian determinant of the estimated deformation fields. Results When large initial rigid misalignment is present, registering CT to MRI‐derived synthetic CT aligns the cord better than a direct registration. The average landmark error decreased from 9.8 ± 3.1 mm in MR→CT non‐aligned to 6.0 ± 2.1 mm in CT synth →CT non‐aligned deformable registrations. In the CT to MR direction, the landmark error decreased from 10.0 ± 4.3 mm in CT non‐aligned →MR deformable registrations to 6.6 ± 2.0 mm in CT non‐aligned →CT synth deformable registrations. The Jacobian determinant had an average value of 0.98. The proposed method also demonstrated improved inverse consistency over the direct method. Conclusions We showed that using a deep learning‐derived synthetic CT in lieu of an MR for MR→CT and CT→MR deformable registration offers superior results to direct multimodal registration.