dtiRIM: A generalisable deep learning method for diffusion tensor imaging

Diffusion weighted MRI is an indispensable tool for routine patient screening and diagnostics of pathology. Re-cently, several deep learning methods have been proposed to quantify diffusion parameters, but poor generalisa-tion to new data prevents broader use of these methods, as they require retraining of the neural network for each new scan protocol. In this work, we present the dtiRIM, a new deep learning method for Diffusion Tensor Imaging (DTI) based on the Recurrent Inference Machines. Thanks to its ability to learn how to solve inverse problems and to use the diffusion tensor model to promote data consistency, the dtiRIM can generalise to variations in the acquisition settings. This enables a single trained network to produce high quality tensor estimates for a variety of cases. We performed extensive validation of our method using simulation and in vivo data, and compared it to the Iterated Weighted Linear Least Squares (IWLLS), the approach of the state-of-the-art MRTrix3 software, and to an implementation of the Maximum Likelihood Estimator (MLE). Our results show that dtiRIM predictions present low dependency on tissue properties, anatomy and scanning parameters, with results comparable to or better than both IWLLS and MLE. Further, we demonstrate that a single dtiRIM model can be used for a diver-sity of data sets without significant loss in quality, representing, to our knowledge, the first generalisable deep learning based solver for DTI.

Automated summary

Diffusion weighted MRI is essential for patient screening and diagnosis. However, current deep learning methods for quantifying diffusion parameters lack generalization, requiring retraining for each new scan. In this study, we propose dtiRIM, a deep learning method for Diffusion Tensor Imaging with superior generalization due to its ability to solve inverse problems and promote data consistency using the diffusion tensor model. Results show that dtiRIM produces high-quality tensor estimates, comparable or better than existing methods, with low dependency on tissue properties and scanning parameters. Furthermore, a single dtiRIM model can be used for diverse datasets without significant loss in quality, making it a groundbreaking solution for DTI analysis.