Performance of a deep learning-based CT image denoising method: Generalizability over dose, reconstruction kernel, and slice thickness.

PURPOSE: Deep learning (DL) is rapidly finding applications in low-dose CT image denoising. While having the potential to improve the image quality (IQ) over the filtered back projection method (FBP) and produce images quickly, performance generalizability of the data-driven DL methods is not fully understood yet. The main purpose of this work is to investigate the performance generalizability of a low-dose CT image denoising neural network in data acquired under different scan conditions, particularly relating to these three parameters: reconstruction kernel, slice thickness, and dose (noise) level. A secondary goal is to identify any underlying data property associated with the CT scan settings that might help predict the generalizability of the denoising network. METHODS: We select the residual encoder-decoder convolutional neural network (REDCNN) as an example of a low-dose CT image denoising technique in this work. To study how the network generalizes on the three imaging parameters, we grouped the CT volumes in the Low-Dose Grand Challenge (LDGC) data into three pairs of training datasets according to their imaging parameters, changing only one parameter in each pair. We trained REDCNN with them to obtain six denoising models. We test each denoising model on datasets of matching and mismatching parameters with respect to its training sets regarding dose, reconstruction kernel, and slice thickness, respectively, to evaluate the denoising performance changes. Denoising performances are evaluated on patient scans, simulated phantom scans, and physical phantom scans using IQ metrics including mean-squared error (MSE), contrast-dependent modulation transfer function (MTF), pixel-level noise power spectrum (pNPS), and low-contrast lesion detectability (LCD). RESULTS: REDCNN had larger MSE when the testing data were different from the training data in reconstruction kernel, but no significant MSE difference when varying slice thickness in the testing data. REDCNN trained with quarter-dose data had slightly worse MSE in denoising higher-dose images than that trained with mixed-dose data (17%-80%). The MTF tests showed that REDCNN trained with the two reconstruction kernels and slice thicknesses yielded images of similar image resolution. However, REDCNN trained with mixed-dose data preserved the low-contrast resolution better compared to REDCNN trained with quarter-dose data. In the pNPS test, it was found that REDCNN trained with smooth-kernel data could not remove high-frequency noise in the test data of sharp kernel, possibly because the lack of high-frequency noise in the smooth-kernel data limited the ability of the trained model in removing high-frequency noise. Finally, in the LCD test, REDCNN improved the lesion detectability over the original FBP images regardless of whether the training and testing data had matching reconstruction kernels. CONCLUSIONS: REDCNN is observed to be poorly generalizable between reconstruction kernels, more robust in denoising data of arbitrary dose levels when trained with mixed-dose data, and not highly sensitive to slice thickness. It is known that reconstruction kernel affects the in-plane pNPS shape of a CT image, whereas slice thickness and dose level do not, so it is possible that the generalizability performance of this CT image denoising network highly correlates to the pNPS similarity between the testing and training data.

Efficient Regularized Field Map Estimation in 3D MRI.

Magnetic field inhomogeneity estimation is important in some types of magnetic resonance imaging (MRI), including field-corrected reconstruction for fast MRI with long readout times, and chemical shift based water-fat imaging. Regularized field map estimation methods that account for phase wrapping and noise involve nonconvex cost functions that require iterative algorithms. Most existing minimization techniques were computationally or memory intensive for 3D datasets, and are designed for single-coil MRI. This paper considers 3D MRI with optional consideration of coil sensitivity, and addresses the multi-echo field map estimation and water-fat imaging problem. Our efficient algorithm uses a preconditioned nonlinear conjugate gradient method based on an incomplete Cholesky factorization of the Hessian of the cost function, along with a monotonic line search. Numerical experiments show the computational advantage of the proposed algorithm over state-of-the-art methods with similar memory requirements.

Efficient Dynamic Parallel MRI Reconstruction for the Low-Rank Plus Sparse Model.

The low-rank plus sparse (L+S) decomposition model enables the reconstruction of under-sampled dynamic parallel magnetic resonance imaging (MRI) data. Solving for the low-rank and the sparse components involves non-smooth composite convex optimization, and algorithms for this problem can be categorized into proximal gradient methods and variable splitting methods. This paper investigates new efficient algorithms for both schemes. While current proximal gradient techniques for the L+S model involve the classical iterative soft thresholding algorithm (ISTA), this paper considers two accelerated alternatives, one based on the fast iterative shrinkage-thresholding algorithm (FISTA), and the other with the recent proximal optimized gradient method (POGM). In the augmented Lagrangian (AL) framework, we propose an efficient variable splitting scheme based on the form of the data acquisition operator, leading to simpler computation than the conjugate gradient (CG) approach required by existing AL methods. Numerical results suggest faster convergence of the efficient implementations for both frameworks, with POGM providing the fastest convergence overall and the practical benefit of being free of algorithm tuning parameters.

Numerical methods for polyline-to-point-cloud registration with applications to patient-specific stent reconstruction.

We present novel numerical methods for polyline-to-point-cloud registration and their application to patient-specific modeling of deployed coronary artery stents from image data. Patient-specific coronary stent reconstruction is an important challenge in computational hemodynamics and relevant to the design and improvement of the prostheses. It is an invaluable tool in large-scale clinical trials that computationally investigate the effect of new generations of stents on hemodynamics and eventually tissue remodeling. Given a point cloud of strut positions, which can be extracted from images, our stent reconstruction method aims at finding a geometrical transformation that aligns a model of the undeployed stent to the point cloud. Mathematically, we describe the undeployed stent as a polyline, which is a piecewise linear object defined by its vertices and edges. We formulate the nonlinear registration as an optimization problem whose objective function consists of a similarity measure, quantifying the distance between the polyline and the point cloud, and a regularization functional, penalizing undesired transformations. Using projections of points onto the polyline structure, we derive novel distance measures. Our formulation supports most commonly used transformation models including very flexible nonlinear deformations. We also propose 2 regularization approaches ensuring the smoothness of the estimated nonlinear transformation. We demonstrate the potential of our methods using an academic 2D example and a real-life 3D bioabsorbable stent reconstruction problem. Our results show that the registration problem can be solved to sufficient accuracy within seconds using only a few number of Gauss-Newton iterations.