Cine cone beam CT reconstruction using low-rank matrix factorization: algorithm and a proof-of-principle study

LEARN: Learned Experts' Assessment-Based Reconstruction Network for Sparse-Data CT

Compressive sensing (CS) has proved effective for tomographic reconstruction from sparsely collected data or under-sampled measurements, which are practically important for few-view computed tomography (CT), tomosynthesis, interior tomography, and so on. To perform sparse-data CT, the iterative reconstruction commonly uses regularizers in the CS framework. Currently, how to choose the parameters adaptively for regularization is a major open problem. In this paper, inspired by the idea of machine learning especially deep learning, we unfold the state-of-the-art "fields of experts"-based iterative reconstruction scheme up to a number of iterations for data-driven training, construct a learned experts' assessment-based reconstruction network (LEARN) for sparse-data CT, and demonstrate the feasibility and merits of our LEARN network. The experimental results with our proposed LEARN network produces a superior performance with the well-known Mayo Clinic low-dose challenge data set relative to the several state-of-the-art methods, in terms of artifact reduction, feature preservation, and computational speed. This is consistent to our insight that because all the regularization terms and parameters used in the iterative reconstruction are now learned from the training data, our LEARN network utilizes application-oriented knowledge more effectively and recovers underlying images more favorably than competing algorithms. Also, the number of layers in the LEARN network is only 50, reducing the computational complexity of typical iterative algorithms by orders of magnitude.

Statistical Iterative CBCT Reconstruction Based on Neural Network

Cone-beam computed tomography (CBCT) plays an important role in radiation therapy. Statistical iterative reconstruction (SIR) algorithms with specially designed penalty terms provide good performance for low-dose CBCT imaging. Among others, the total variation (TV) penalty is the current state-of-the-art in removing noises and preserving edges, but one of its well-known limitations is its staircase effect. Recently, various penalty terms with higher order differential operators were proposed to replace the TV penalty to avoid the staircase effect, at the cost of slightly blurring object edges. We developed a novel SIR algorithm using a neural network for CBCT reconstruction. We used a data-driven method to learn the "potential regularization term" rather than design a penalty term manually. This approach converts the problem of designing a penalty term in the traditional statistical iterative framework to designing and training a suitable neural network for CBCT reconstruction. We proposed using transfer learning to overcome the data deficiency problem and an iterative deblurring approach specially designed for the CBCT iterative reconstruction process during which the noise level and resolution of the reconstructed images may change. Through experiments conducted on two physical phantoms, two simulation digital phantoms, and patient data, we demonstrated the excellent performance of the proposed network-based SIR for CBCT reconstruction, both visually and quantitatively. Our proposed method can overcome the staircase effect, preserve both edges and regions with smooth intensity transition, and provide reconstruction results at high resolution and low noise level.

Motion guided Spatiotemporal Sparsity for high quality 4D-CBCT reconstruction

Conventional cone-beam computed tomography is often deteriorated by respiratory motion blur, which negatively affects target delineation. On the other side, the four dimensional cone-beam computed tomography (4D-CBCT) can be considered to describe tumor and organ motion. But for current on-board CBCT imaging system, the slow rotation speed limits the projection number at each phase, and the associated reconstructions are contaminated by noise and streak artifacts using the conventional algorithm. To address the problem, we propose a novel framework to reconstruct 4D-CBCT from the under-sampled measurements—Motion guided Spatiotemporal Sparsity (MgSS). In this algorithm, we try to divide the CBCT images at each phase into cubes (3D blocks) and track the cubes with estimated motion field vectors through phase, then apply regional spatiotemporal sparsity on the tracked cubes. Specifically, we recast the tracked cubes into four-dimensional matrix, and use the higher order singular value decomposition (HOSVD) technique to analyze the regional spatiotemporal sparsity. Subsequently, the blocky spatiotemporal sparsity is incorporated into a cost function for the image reconstruction. The phantom simulation and real patient data are used to evaluate this algorithm. Results show that the MgSS algorithm achieved improved 4D-CBCT image quality with less noise and artifacts compared to the conventional algorithms.

Principal component reconstruction (PCR) for cine CBCT with motion learning from 2D fluoroscopy

PURPOSE

This work aims to generate cine CT images (i.e., 4D images with high-temporal resolution) based on a novel principal component reconstruction (PCR) technique with motion learning from 2D fluoroscopic training images.

METHODS

In the proposed PCR method, the matrix factorization is utilized as an explicit low-rank regularization of 4D images that are represented as a product of spatial principal components and temporal motion coefficients. The key hypothesis of PCR is that temporal coefficients from 4D images can be reasonably approximated by temporal coefficients learned from 2D fluoroscopic training projections. For this purpose, we can acquire fluoroscopic training projections for a few breathing periods at fixed gantry angles that are free from geometric distortion due to gantry rotation, that is, fluoroscopy-based motion learning. Such training projections can provide an effective characterization of the breathing motion. The temporal coefficients can be extracted from these training projections and used as priors for PCR, even though principal components from training projections are certainly not the same for these 4D images to be reconstructed. For this purpose, training data are synchronized with reconstruction data using identical real-time breathing position intervals for projection binning. In terms of image reconstruction, with a priori temporal coefficients, the data fidelity for PCR changes from nonlinear to linear, and consequently, the PCR method is robust and can be solved efficiently. PCR is formulated as a convex optimization problem with the sum of linear data fidelity with respect to spatial principal components and spatiotemporal total variation regularization imposed on 4D image phases. The solution algorithm of PCR is developed based on alternating direction method of multipliers.

RESULTS

The implementation is fully parallelized on GPU with NVIDIA CUDA toolbox and each reconstruction takes about a few minutes. The proposed PCR method is validated and compared with a state-of-art method, that is, PICCS, using both simulation and experimental data with the on-board cone-beam CT setting. The results demonstrated the feasibility of PCR for cine CBCT and significantly improved reconstruction quality of PCR from PICCS for cine CBCT.

CONCLUSION

With a priori estimated temporal motion coefficients using fluoroscopic training projections, the PCR method can accurately reconstruct spatial principal components, and then generate cine CT images as a product of temporal motion coefficients and spatial principal components.

Low-Dose CT With a Residual Encoder-Decoder Convolutional Neural Network

Given the potential risk of X-ray radiation to the patient, low-dose CT has attracted a considerable interest in the medical imaging field. Currently, the main stream low-dose CT methods include vendor-specific sinogram domain filtration and iterative reconstruction algorithms, but they need to access raw data, whose formats are not transparent to most users. Due to the difficulty of modeling the statistical characteristics in the image domain, the existing methods for directly processing reconstructed images cannot eliminate image noise very well while keeping structural details. Inspired by the idea of deep learning, here we combine the autoencoder, deconvolution network, and shortcut connections into the residual encoder-decoder convolutional neural network (RED-CNN) for low-dose CT imaging. After patch-based training, the proposed RED-CNN achieves a competitive performance relative to the-state-of-art methods in both simulated and clinical cases. Especially, our method has been favorably evaluated in terms of noise suppression, structural preservation, and lesion detection.

Sinogram Blurring Matrix Estimation From Point Sources Measurements With Rank-One Approximation for Fully 3-D PET

An accurate system matrix is essential in positron emission tomography (PET) for reconstructing high quality images. To reduce storage size and image reconstruction time, we factor the system matrix into a product of a geometry projection matrix and a sinogram blurring matrix. The geometric projection matrix is computed analytically and the sinogram blurring matrix is estimated from point source measurements. Previously, we have estimated a 2-D blurring matrix for a preclinical PET scanner. The 2-D blurring matrix only considers blurring effects within a transaxial sinogram and does not compensate for inter-sinogram blurring effects. For PET scanners with a long axial field of view, inter-sinogram blurring can be a major problem influencing the image quality in the axial direction. Hence, the estimation of a 4-D blurring matrix is desirable to further improve the image quality. The 4-D blurring matrix estimation is an ill-conditioned problem due to the large number of unknowns. Here, we propose a rank-one approximation for each blurring kernel image formed by a row vector of the sinogram blurring matrix to improve the stability of the 4-D blurring matrix estimation. The proposed method is applied to the simulated data as well as the real data obtained from an Inveon microPET scanner. The results show that the newly estimated 4-D blurring matrix can improve the image quality over those obtained with a 2-D blurring matrix and requires less point source scans to achieve similar image quality compared with an unconstrained 4-D blurring matrix estimation.

Low-dose 4D cone-beam CT via joint spatiotemporal regularization of tensor framelet and nonlocal total variation

Excessive radiation exposure is still a major concern in 4D cone-beam computed tomography (4D-CBCT) due to its prolonged scanning duration. Radiation dose can be effectively reduced by either under-sampling the x-ray projections or reducing the x-ray flux. However, 4D-CBCT reconstruction under such low-dose protocols is prone to image artifacts and noise. In this work, we propose a novel joint regularization-based iterative reconstruction method for low-dose 4D-CBCT. To tackle the under-sampling problem, we employ spatiotemporal tensor framelet (STF) regularization to take advantage of the spatiotemporal coherence of the patient anatomy in 4D images. To simultaneously suppress the image noise caused by photon starvation, we also incorporate spatiotemporal nonlocal total variation (SNTV) regularization to make use of the nonlocal self-recursiveness of anatomical structures in the spatial and temporal domains. Under the joint STF-SNTV regularization, the proposed iterative reconstruction approach is evaluated first using two digital phantoms and then using physical experiment data in the low-dose context of both under-sampled and noisy projections. Compared with existing approaches via either STF or SNTV regularization alone, the presented hybrid approach achieves improved image quality, and is particularly effective for the reconstruction of low-dose 4D-CBCT data that are not only sparse but noisy.

Low-dose CT via convolutional neural network

In order to reduce the potential radiation risk, low-dose CT has attracted an increasing attention. However, simply lowering the radiation dose will significantly degrade the image quality. In this paper, we propose a new noise reduction method for low-dose CT via deep learning without accessing original projection data. A deep convolutional neural network is here used to map low-dose CT images towards its corresponding normal-dose counterparts in a patch-by-patch fashion. Qualitative results demonstrate a great potential of the proposed method on artifact reduction and structure preservation. In terms of the quantitative metrics, the proposed method has showed a substantial improvement on PSNR, RMSE and SSIM than the competing state-of-art methods. Furthermore, the speed of our method is one order of magnitude faster than the iterative reconstruction and patch-based image denoising methods.

TICMR: Total Image Constrained Material Reconstruction via Nonlocal Total Variation Regularization for Spectral CT

This work develops a material reconstruction method for spectral CT, namely Total Image Constrained Material Reconstruction (TICMR), to maximize the utility of projection data in terms of both spectral information and high signal-to-noise ratio (SNR). This is motivated by the following fact: when viewed as a spectrally-integrated measurement, the projection data can be used to reconstruct a total image without spectral information, which however has a relatively high SNR; when viewed as a spectrally-resolved measurement, the projection data can be utilized to reconstruct the material composition, which however has a relatively low SNR. The material reconstruction synergizes material decomposition and image reconstruction, i.e., the direct reconstruction of material compositions instead of a two-step procedure that first reconstructs images and then decomposes images. For material reconstruction with high SNR, we propose TICMR with nonlocal total variation (NLTV) regularization. That is, first we reconstruct a total image using spectrally-integrated measurement without spectral binning, and build the NLTV weights from this image that characterize nonlocal image features; then the NLTV weights are incorporated into a NLTV-based iterative material reconstruction scheme using spectrally-binned projection data, so that these weights serve as a high-SNR reference to regularize material reconstruction. Note that the nonlocal property of NLTV is essential for material reconstruction, since material compositions may have significant local intensity variations although their structural information is often similar. In terms of solution algorithm, TICMR is formulated as an iterative reconstruction method with the NLTV regularization, in which the nonlocal divergence is utilized based on the adjoint relationship. The alternating direction method of multipliers is developed to solve this sparsity optimization problem. The proposed TICMR method was validated using both simulated and experimental data. In comparison with FBP and total-variation-based iterative method, TICMR had improved image quality, e.g., contrast-to-noise ratio and spatial resolution.

Robust fluence map optimization via alternating direction method of multipliers with empirical parameter optimization

For the treatment planning during intensity modulated radiation therapy (IMRT) or volumetric modulated arc therapy (VMAT), beam fluence maps can be first optimized via fluence map optimization (FMO) under the given dose prescriptions and constraints to conformally deliver the radiation dose to the targets while sparing the organs-at-risk, and then segmented into deliverable MLC apertures via leaf or arc sequencing algorithms. This work is to develop an efficient algorithm for FMO based on alternating direction method of multipliers (ADMM). Here we consider FMO with the least-square cost function and non-negative fluence constraints, and its solution algorithm is based on ADMM, which is efficient and simple-to-implement. In addition, an empirical method for optimizing the ADMM parameter is developed to improve the robustness of the ADMM algorithm. The ADMM based FMO solver was benchmarked with the quadratic programming method based on the interior-point (IP) method using the CORT dataset. The comparison results suggested the ADMM solver had a similar plan quality with slightly smaller total objective function value than IP. A simple-to-implement ADMM based FMO solver with empirical parameter optimization is proposed for IMRT or VMAT.

A method for volumetric imaging in radiotherapy using single x-ray projection

PURPOSE

It is an intriguing problem to generate an instantaneous volumetric image based on the corresponding x-ray projection. The purpose of this study is to develop a new method to achieve this goal via a sparse learning approach.

METHODS

To extract motion information hidden in projection images, the authors partitioned a projection image into small rectangular patches. The authors utilized a sparse learning method to automatically select patches that have a high correlation with principal component analysis (PCA) coefficients of a lung motion model. A model that maps the patch intensity to the PCA coefficients was built along with the patch selection process. Based on this model, a measured projection can be used to predict the PCA coefficients, which are then further used to generate a motion vector field and hence a volumetric image. The authors have also proposed an intensity baseline correction method based on the partitioned projection, in which the first and the second moments of pixel intensities at a patch in a simulated projection image are matched with those in a measured one via a linear transformation. The proposed method has been validated in both simulated data and real phantom data.

RESULTS

The algorithm is able to identify patches that contain relevant motion information such as the diaphragm region. It is found that an intensity baseline correction step is important to remove the systematic error in the motion prediction. For the simulation case, the sparse learning model reduced the prediction error for the first PCA coefficient to 5%, compared to the 10% error when sparse learning was not used, and the 95th percentile error for the predicted motion vector was reduced from 2.40 to 0.92 mm. In the phantom case with a regular tumor motion, the predicted tumor trajectory was successfully reconstructed with a 0.82 mm error for tumor center localization compared to a 1.66 mm error without using the sparse learning method. When the tumor motion was driven by a real patient breathing signal with irregular periods and amplitudes, the average tumor center error was 0.6 mm. The algorithm robustness with respect to sparsity level, patch size, and presence or absence of diaphragm, as well as computation time, has also been studied.

CONCLUSIONS

The authors have developed a new method that automatically identifies motion information from an x-ray projection, based on which a volumetric image is generated.

Tensor-based dictionary learning for dynamic tomographic reconstruction

In dynamic computed tomography (CT) reconstruction, the data acquisition speed limits the spatio-temporal resolution. Recently, compressed sensing theory has been instrumental in improving CT reconstruction from far few-view projections. In this paper, we present an adaptive method to train a tensor-based spatio-temporal dictionary for sparse representation of an image sequence during the reconstruction process. The correlations among atoms and across phases are considered to capture the characteristics of an object. The reconstruction problem is solved by the alternating direction method of multipliers. To recover fine or sharp structures such as edges, the nonlocal total variation is incorporated into the algorithmic framework. Preclinical examples including a sheep lung perfusion study and a dynamic mouse cardiac imaging demonstrate that the proposed approach outperforms the vectorized dictionary-based CT reconstruction in the case of few-view reconstruction.

Accelerated NMR spectroscopy with low-rank reconstruction

Accelerated multi-dimensional NMR spectroscopy is a prerequisite for high-throughput applications, studying short-lived molecular systems and monitoring chemical reactions in real time. Non-uniform sampling is a common approach to reduce the measurement time. Here, a new method for high-quality spectra reconstruction from non-uniformly sampled data is introduced, which is based on recent developments in the field of signal processing theory and uses the so far unexploited general property of the NMR signal, its low rank. Using experimental and simulated data, we demonstrate that the low-rank reconstruction is a viable alternative to the current state-of-the-art technique compressed sensing. In particular, the low-rank approach is good in preserving of low-intensity broad peaks, and thus increases the effective sensitivity in the reconstructed spectra.

Accelerating dynamic magnetic resonance imaging (MRI) for lung tumor tracking based on low-rank decomposition in the spatial-temporal domain: a feasibility study based on simulation and preliminary prospective undersampled MRI

PURPOSE

To evaluate a low-rank decomposition method to reconstruct down-sampled k-space data for the purpose of tumor tracking.

METHODS AND MATERIALS

Seven retrospective lung cancer patients were included in the simulation study. The fully-sampled k-space data were first generated from existing 2-dimensional dynamic MR images and then down-sampled by 5 × -20 × before reconstruction using a Cartesian undersampling mask. Two methods, a low-rank decomposition method using combined dynamic MR images (k-t SLR based on sparsity and low-rank penalties) and a total variation (TV) method using individual dynamic MR frames, were used to reconstruct images. The tumor trajectories were derived on the basis of autosegmentation of the resultant images. To further test its feasibility, k-t SLR was used to reconstruct prospective data of a healthy subject. An undersampled balanced steady-state free precession sequence with the same undersampling mask was used to acquire the imaging data.

RESULTS

In the simulation study, higher imaging fidelity and low noise levels were achieved with the k-t SLR compared with TV. At 10 × undersampling, the k-t SLR method resulted in an average normalized mean square error <0.05, as opposed to 0.23 by using the TV reconstruction on individual frames. Less than 6% showed tracking errors >1 mm with 10 × down-sampling using k-t SLR, as opposed to 17% using TV. In the prospective study, k-t SLR substantially reduced reconstruction artifacts and retained anatomic details.

CONCLUSIONS

Magnetic resonance reconstruction using k-t SLR on highly undersampled dynamic MR imaging data results in high image quality useful for tumor tracking. The k-t SLR was superior to TV by better exploiting the intrinsic anatomic coherence of the same patient. The feasibility of k-t SLR was demonstrated by prospective imaging acquisition and reconstruction.