Premium
ReUINet: A fast GNL distortion correction approach on a 1.0 T MRI‐Linac scanner
Author(s) -
Shan Shanshan,
Li Mao,
Li Mingyan,
Tang Fangfang,
Crozier Stuart,
Liu Feng
Publication year - 2021
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1002/mp.14861
Subject(s) - imaging phantom , computer science , linear particle accelerator , interpolation (computer graphics) , convolutional neural network , scanner , artificial intelligence , magnetic resonance imaging , distortion (music) , linear interpolation , encoding (memory) , computer vision , nuclear medicine , image (mathematics) , physics , pattern recognition (psychology) , optics , medicine , radiology , beam (structure) , amplifier , computer network , bandwidth (computing)
Purpose The hybrid system combining a magnetic resonance imaging (MRI) scanner with a linear accelerator (Linac) has become increasingly desirable for tumor treatment because of excellent soft tissue contrast and nonionizing radiation. However, image distortions caused by gradient nonlinearity (GNL) can have detrimental impacts on real‐time radiotherapy using MRI‐Linac systems, where accurate geometric information of tumors is essential. Methods In this work, we proposed a deep convolutional neural network‐based method to efficiently re cover u ndistorted i mages ( ReUINet ) for real‐time image guidance. The ReUINet , based on the encoder‐decoder structure, was created to learn the relationship between the undistorted images and distorted images. The ReUINet was pretrained and tested on a publically available brain MR image dataset acquired from 23 volunteers. Then, transfer learning was adopted to implement the pretrained model (i.e., network with optimal weights) on the experimental three‐dimensional (3D) grid phantom and in‐vivo pelvis image datasets acquired from the 1.0 T Australian MRI‐Linac system. Results Evaluations on the phantom (768 slices) and pelvis data (88 slices) showed that the ReUINet achieved improvement over 15 times and 45 times on computational efficiency in comparison with standard interpolation and GNL‐encoding methods, respectively. Moreover, qualitative and quantitative results demonstrated that the ReUINet provided better correction results than the standard interpolation method, and comparable performance compared to the GNL‐encoding approach. Conclusions Validated by simulation and experimental results, the proposed ReUINet showed promise in obtaining accurate MR images for the implementation of real‐time MRI‐guided radiotherapy.