Abstract:BackgroundAn automated, accurate, and efficient lung four-dimensional computed tomography (4DCT) image registration method is clinically important to quantify respiratory motion for optimal motion management.PurposeThe purpose of this work is to develop a weakly supervised deep learning method for 4DCT lung deformable image registration (DIR).MethodsThe landmark-driven cycle network is proposed as a deep learning platform that performs DIR of individual phase datasets in a simulation 4DCT. This proposed network comprises a generator and a discriminator. The generator accepts moving and target CTs as input and outputs the deformation vector fields (DVFs) to match the two CTs. It is optimized during both forward and backward paths to enhance the bi-directionality of DVF generation. Further, the landmarks are used to weakly supervise the generator network. Landmark-driven loss is used to guide the generator's training. The discriminator then judges the realism of the deformed CT to provide extra DVF regularization.ResultsWe performed four-fold cross-validation on 10 4DCT datasets from the public DIR-Lab dataset and a hold-out test on our clinic dataset, which included 50 4DCT datasets. The DIR-Lab dataset was used to evaluate the performance of the proposed method against other methods in the literature by calculating the DIR-Lab Target Registration Error (TRE). The proposed method outperformed other deep learning-based methods on the DIR-Lab datasets in terms of TRE. Bi-directional and landmark-driven loss were shown to be effective for obtaining high registration accuracy. The mean and standard deviation of TRE for the DIR-Lab datasets was 1.20 & PLUSMN; 0.72 mm and the mean absolute error (MAE) and structural similarity index (SSIM) for our datasets were 32.1 & PLUSMN; 11.6 HU and 0.979 & PLUSMN; 0.011, respectively.ConclusionThe landmark-driven cycle network has been validated and tested for automatic deformable image registration of patients' lung 4DCTs with results comparable to or better than competing methods.

GPU-accelerated Block Matching Algorithm for Deformable Registration of Lung CT Images

Rapid 2D-3D Medical Image Registration Based on CUDA

Faster dense deformable image registration by utilizing both CPU and GPU

An Unsupervised Deep Learning Framework for Large-Scale Lung CT Deformable Image Registration

Implementation and evaluation of various demons deformable image registration algorithms on GPU

2D-3D Medical Image Registration Based on GPU

[Accurate 3D free-form registration between fan-beam CT and cone-beam CT].

CNN-based Lung CT Registration with Multiple Anatomical Constraints

Accelerated Gradient-Based Free Form Deformable Registration for Online Adaptive Radiotherapy

Learning Deformable Point Set Registration with Regularized Dynamic Graph CNNs for Large Lung Motion in COPD Patients

Lung 4D CT Image Registration Based on High-Order Markov Random Field.

Deformable Lung 4DCT Image Registration Via Landmark-Driven Cycle Network

Estimation of Large Motion in Lung CT by Integrating Regularized Keypoint Correspondences into Dense Deformable Registration

Deep-Learning-based Fast and Accurate 3D CT Deformable Image Registration in Lung Cancer

Residual Aligner-based Network (RAN): Motion-separable structure for coarse-to-fine discontinuous deformable registration

A fast CT and CT-fluoroscopy registration algorithm with respiratory motion compensation for image-guided lung intervention.

Estimation of Lung Motion Using Deformable Image Registration Based on Compressible Flow

Residual Fully Convolutional Deformable Registration Network for Deformable Lung CT Images

Unsupervised Computed Tomography and Cone-Beam Computed Tomography Image Registration Using a Dual Attention Network

An accelerative method for multimodality medical image registration based on CUDA

Respiratory Deformation Registration in 4D-Ct/cone Beam CT Using Deep Learning.