Denoised and texture enhanced MVCT to improve soft tissue conspicuity

Self-supervised denoising of grating-based phase-contrast computed tomography

In the last decade, grating-based phase-contrast computed tomography (gbPC-CT) has received growing interest. It provides additional information about the refractive index decrement in the sample. This signal shows an increased soft-tissue contrast. However, the resolution dependence of the signal poses a challenge: its contrast enhancement is overcompensated by the low resolution in low-dose applications such as clinical computed tomography. As a result, the implementation of gbPC-CT is currently tied to a higher dose. To reduce the dose, we introduce the self-supervised deep learning network Noise2Inverse into the field of gbPC-CT. We evaluate the behavior of the Noise2Inverse parameters on the phase-contrast results. Afterward, we compare its results with other denoising methods, namely the Statistical Iterative Reconstruction, Block Matching 3D, and Patchwise Phase Retrieval. In the example of Noise2Inverse, we show that deep learning networks can deliver superior denoising results with respect to the investigated image quality metrics. Their application allows to increase the resolution while maintaining the dose. At higher resolutions, gbPC-CT can naturally deliver higher contrast than conventional absorption-based CT. Therefore, the application of machine learning-based denoisers shifts the dose-normalized image quality in favor of gbPC-CT, bringing it one step closer to medical application.

A systematic review of deep learning-based denoising for low-dose computed tomography from a perceptual quality perspective

Low-dose computed tomography (LDCT) scans are essential in reducing radiation exposure but often suffer from significant image noise that can impair diagnostic accuracy. While deep learning approaches have enhanced LDCT denoising capabilities, the predominant reliance on objective metrics like PSNR and SSIM has resulted in over-smoothed images that lack critical detail. This paper explores advanced deep learning methods tailored specifically to improve perceptual quality in LDCT images, focusing on generating diagnostic-quality images preferred in clinical practice. We review and compare current methodologies, including perceptual loss functions and generative adversarial networks, addressing the significant limitations of current benchmarks and the subjective nature of perceptual quality evaluation. Through a systematic analysis, this study underscores the urgent need for developing methods that balance both perceptual and diagnostic quality, proposing new directions for future research in the field.

Technical note: Preprocessing of portal images to improve image quality of VMAT-CT

BACKGROUND

The concept of volumetric modulated arc therapy-computed tomography (VMAT-CT) was proposed more than a decade ago. However, its application has been very limited mainly due to the poor image quality. More specifically, the blurred areas in electronic portal imaging device (EPID) images collected during VMAT heavily degrade the image quality of VMAT-CT.

PURPOSE

The goal of this study was to propose systematic methods to preprocess EPID images and improve the image quality of VMAT-CT.

METHODS

Online region-based active contour method was introduced to binarize portal images. Multi-leaf collimator (MLC) motion modeling was developed to remove the MLC motion blur. Outlier filtering was then applied to replace the remaining artifacts with plausible data. To assess the impact of these preprocessing methods on the image quality of VMAT-CT, 44 clinical VMAT plans for several treatment sites (lung, esophagus, and head & neck) were delivered to a Rando phantom, and several real-patient cases were also acquired. VMAT-CT reconstruction was attempted for all the cases, and image quality was evaluated.

RESULTS

All three preprocessing methods could effectively remove the blurred edges of EPID images. The combined preprocessing methods not only saved VMAT-CT from distortions and artifacts, but also increased the percentage of VMAT plans that can be reconstructed.

CONCLUSIONS

The systematic preprocessing of portal images improves the image quality of VMAT-CT significantly, and facilitates the application of VMAT-CT as an effective image guidance tool.

CoreDiff: Contextual Error-Modulated Generalized Diffusion Model for Low-Dose CT Denoising and Generalization

Low-dose computed tomography (CT) images suffer from noise and artifacts due to photon starvation and electronic noise. Recently, some works have attempted to use diffusion models to address the over-smoothness and training instability encountered by previous deep-learning-based denoising models. However, diffusion models suffer from long inference time due to a large number of sampling steps involved. Very recently, cold diffusion model generalizes classical diffusion models and has greater flexibility. Inspired by cold diffusion, this paper presents a novel COntextual eRror-modulated gEneralized Diffusion model for low-dose CT (LDCT) denoising, termed CoreDiff. First, CoreDiff utilizes LDCT images to displace the random Gaussian noise and employs a novel mean-preserving degradation operator to mimic the physical process of CT degradation, significantly reducing sampling steps thanks to the informative LDCT images as the starting point of the sampling process. Second, to alleviate the error accumulation problem caused by the imperfect restoration operator in the sampling process, we propose a novel ContextuaL Error-modulAted Restoration Network (CLEAR-Net), which can leverage contextual information to constrain the sampling process from structural distortion and modulate time step embedding features for better alignment with the input at the next time step. Third, to rapidly generalize the trained model to a new, unseen dose level with as few resources as possible, we devise a one-shot learning framework to make CoreDiff generalize faster and better using only one single LDCT image (un)paired with normal-dose CT (NDCT). Extensive experimental results on four datasets demonstrate that our CoreDiff outperforms competing methods in denoising and generalization performance, with clinically acceptable inference time. Source code is made available at https://github.com/qgao21/CoreDiff.

A qualitative study of improving megavoltage computed tomography image quality and maintaining dose accuracy using cycleGAN-based image synthesis

BACKGROUND

Due to inconsistent positioning, tumor shrinking, and weight loss during fractionated treatment, the initial plan was no longer appropriate after a few fractional treatments, and the patient will require adaptive helical tomotherapy (HT) to overcome the issue. Patients are scanned with megavoltage computed tomography (MVCT) before each fractional treatment, which is utilized for patient setup and provides information for dose reconstruction. However, the low contrast and high noise of MVCT make it challenging to delineate treatment targets and organs at risk (OAR).

PURPOSE

This study developed a deep-learning-based approach to generate high-quality synthetic kilovoltage computed tomography (skVCT) from MVCT and meet clinical dose requirements.

METHODS

Data from 41 head and neck cancer patients were collected; 25 (2995 slices) were used for training, and 16 (1898 slices) for testing. A cycle generative adversarial network (cycleGAN) based on attention gate and residual blocks was used to generate MVCT-based skVCT. For the 16 patients, kVCT-based plans were transferred to skVCT images and electron density profile-corrected MVCT images to recalculate the dose. The quantitative indices and clinically relevant dosimetric metrics, including the mean absolute error (MAE), structural similarity index measure (SSIM), peak signal-to-noise ratio (PSNR), gamma passing rates, and dose-volume-histogram (DVH) parameters (Dmax , Dmean , Dmin ), were used to assess the skVCT images.

RESULTS

The MAE, PSNR, and SSIM of MVCT were 109.6 ± 12.3 HU, 27.5 ± 1.1 dB, and 91.9% ± 1.7%, respectively, while those of skVCT were 60.6 ± 9.0 HU, 34.0 ± 1.9 dB, and 96.5% ± 1.1%. The image quality and contrast were enhanced, and the noise was reduced. The gamma passing rates improved from 98.31% ± 1.11% to 99.71% ± 0.20% (2 mm/2%) and 99.77% ± 0.18% to 99.98% ± 0.02% (3 mm/3%). No significant differences (p > 0.05) were observed in DVH parameters between kVCT and skVCT.

CONCLUSION

With training on a small data set (2995 slices), the model successfully generated skVCT with improved image quality, and the dose calculation accuracy was similar to that of MVCT. MVCT-based skVCT can increase treatment accuracy and offer the possibility of implementing adaptive radiotherapy.

StruNet: Perceptual and low-rank regularized transformer for medical image denoising

BACKGROUND

Various types of noise artifacts inevitably exist in some medical imaging modalities due to limitations of imaging techniques, which impair either clinical diagnosis or subsequent analysis. Recently, deep learning approaches have been rapidly developed and applied on medical images for noise removal or image quality enhancement. Nevertheless, due to complexity and diversity of noise distribution representations in different medical imaging modalities, most of the existing deep learning frameworks are incapable to flexibly remove noise artifacts while retaining detailed information. As a result, it remains challenging to design an effective and unified medical image denoising method that will work across a variety of noise artifacts for different imaging modalities without requiring specialized knowledge in performing the task.

PURPOSE

In this paper, we propose a novel encoder-decoder architecture called Swin transformer-based residual u-shape Network (StruNet), for medical image denoising.

METHODS

Our StruNet adopts a well-designed block as the backbone of the encoder-decoder architecture, which integrates Swin Transformer modules with residual block in parallel connection. Swin Transformer modules could effectively learn hierarchical representations of noise artifacts via self-attention mechanism in non-overlapping shifted windows and cross-window connection, while residual block is advantageous to compensate loss of detailed information via shortcut connection. Furthermore, perceptual loss and low-rank regularization are incorporated into loss function respectively in order to constrain the denoising results on feature-level consistency and low-rank characteristics.

RESULTS

To evaluate the performance of the proposed method, we have conducted experiments on three medical imaging modalities including computed tomography (CT), optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA).

CONCLUSIONS

The results demonstrate that the proposed architecture yields a promising performance of suppressing multiform noise artifacts existing in different imaging modalities.

Transformer With Double Enhancement for Low-Dose CT Denoising

Increasingly serious health problems have made the usage of computed tomography surge. Therefore, algorithms for processing CT images are becoming more and more abundant. These algorithms can lessen the harm of cumulative radiation in CT technology for the patient while eliminating the noise of image caused by dose reduction. However, the mainstream CNN-based algorithms are inefficient when dealing with features in broad regions. Inspired by the large receptive field of transformer framework, this paper designs an end-to-end low-dose CT (LDCT) denoising network based on the transformer. The overall network contains a main branch and dual side branches. Specifically, the overlapping-free window-based self-attention transformer block is adopted on the main branch to realize image denoising. On the dual side branches, we propose double enhancement module to enrich edge, texture, and context information of LDCT images. Meanwhile, the receptive field of network is further enlarged after processing, which is helpful for building model's long-range dependencies. The outputs of the side branches are concatenated for enhancing information and generating high-quality CT images. In addition, to better train the network, we introduce a compound loss function including mean squared error (MSE), multi-scale perceptual (MSP), and Sobel-L1 (SL) to make the denoised image closer to the targeted norm-dose CT (NDCT) image. Lastly, we conducted experiments on two clinical datasets including abdomen, head, and chest LDCT images with 25%, 25%, and 10% of the full dose, respectively. The experimental results demonstrated that the proposed DEformer achieved better denoising performance than the existing algorithms.

Generative adversarial network with radiomic feature reproducibility analysis for computed tomography denoising

BACKGROUND

Most computed tomography (CT) denoising algorithms have been evaluated using image quality analysis (IQA) methods developed for natural image, which do not adequately capture the texture details in medical imaging. Radiomics is an emerging image analysis technique that extracts texture information to provide a more objective basis for medical imaging diagnostics, overcoming the subjective nature of traditional methods. By utilizing the difficulty of reproducing radiomics features under different imaging protocols, we can more accurately evaluate the performance of CT denoising algorithms.

METHOD

We introduced radiomic feature reproducibility analysis as an evaluation metric for a denoising algorithm. Also, we proposed a low-dose CT denoising method based on a generative adversarial network (GAN), which outperformed well-known CT denoising methods.

RESULTS

Although the proposed model produced excellent results visually, the traditional image assessment metrics such as peak signal-to-noise ratio and structural similarity failed to show distinctive performance differences between the proposed method and the conventional ones. However, radiomic feature reproducibility analysis provided a distinctive assessment of the CT denoising performance. Furthermore, radiomic feature reproducibility analysis allowed fine-tuning of the hyper-parameters of the GAN.

CONCLUSION

We demonstrated that the well-tuned GAN architecture outperforms the well-known CT denoising methods. Our study is the first to introduce radiomics reproducibility analysis as an evaluation metric for CT denoising. We look forward that the study may bridge the gap between traditional objective and subjective evaluations in the clinical medical imaging field.

M3NAS: Multi-Scale and Multi-Level Memory-Efficient Neural Architecture Search for Low-Dose CT Denoising

Lowering the radiation dose in computed tomography (CT) can greatly reduce the potential risk to public health. However, the reconstructed images from dose-reduced CT or low-dose CT (LDCT) suffer from severe noise which compromises the subsequent diagnosis and analysis. Recently, convolutional neural networks have achieved promising results in removing noise from LDCT images. The network architectures that are used are either handcrafted or built on top of conventional networks such as ResNet and U-Net. Recent advances in neural network architecture search (NAS) have shown that the network architecture has a dramatic effect on the model performance. This indicates that current network architectures for LDCT may be suboptimal. Therefore, in this paper, we make the first attempt to apply NAS to LDCT and propose a multi-scale and multi-level memory-efficient NAS for LDCT denoising, termed M3NAS. On the one hand, the proposed M3NAS fuses features extracted by different scale cells to capture multi-scale image structural details. On the other hand, the proposed M3NAS can search a hybrid cell- and network-level structure for better performance. In addition, M3NAS can effectively reduce the number of model parameters and increase the speed of inference. Extensive experimental results on two different datasets demonstrate that the proposed M3NAS can achieve better performance and fewer parameters than several state-of-the-art methods. In addition, we also validate the effectiveness of the multi-scale and multi-level architecture for LDCT denoising, and present further analysis for different configurations of super-net.

Compound feature attention network with edge enhancement for low-dose CT denoising

BACKGROUND

Low-dose CT (LDCT) images usually contain serious noise and artifacts, which weaken the readability of the image.

OBJECTIVE

To solve this problem, we propose a compound feature attention network with edge enhancement for LDCT denoising (CFAN-Net), which consists of an edge-enhanced module and a proposed compound feature attention block (CFAB).

METHODS

The edge enhancement module extracts edge details with the trainable Sobel convolution. CFAB consists of an interactive feature learning module (IFLM), a multi-scale feature fusion module (MFFM), and a joint attention module (JAB), which removes noise from LDCT images in a coarse-to-fine manner. First, in IFLM, the noise is initially removed by cross-latitude interactive judgment learning. Second, in MFFM, multi-scale and pixel attention are integrated to explore fine noise removal. Finally, in JAB, we focus on key information, extract useful features, and improve the efficiency of network learning. To construct a high-quality image, we repeat the above operation by cascading CFAB.

RESULTS

By applying CFAN-Net to process the 2016 NIH AAPM-Mayo LDCT challenge test dataset, experiments show that the peak signal-to-noise ratio value is 33.9692 and the structural similarity value is 0.9198.

CONCLUSIONS

Compared with several existing LDCT denoising algorithms, CFAN-Net effectively preserves the texture of CT images while removing noise and artifacts.

Multislice input for 2D and 3D residual convolutional neural network noise reduction in CT

Purpose

Deep convolutional neural network (CNN)-based methods are increasingly used for reducing image noise in computed tomography (CT). Current attempts at CNN denoising are based on 2D or 3D CNN models with a single- or multiple-slice input. Our work aims to investigate if the multiple-slice input improves the denoising performance compared with the single-slice input and if a 3D network architecture is better than a 2D version at utilizing the multislice input.

Approach

Two categories of network architectures can be used for the multislice input. First, multislice images can be stacked channel-wise as the multichannel input to a 2D CNN model. Second, multislice images can be employed as the 3D volumetric input to a 3D CNN model, in which the 3D convolution layers are adopted. We make performance comparisons among 2D CNN models with one, three, and seven input slices and two versions of 3D CNN models with seven input slices and one or three output slices. Evaluation was performed on liver CT images using three quantitative metrics with full-dose images as reference. Visual assessment was made by an experienced radiologist.

Results

When the input channels of the 2D CNN model increases from one to three to seven, a trend of improved performance was observed. Comparing the three models with the seven-slice input, the 3D CNN model with a one-slice output outperforms the other models in terms of noise texture and homogeneity in liver parenchyma as well as subjective visualization of vessels.

Conclusions

We conclude the that multislice input is an effective strategy for improving performance for 2D deep CNN denoising models. The pure 3D CNN model tends to have a better performance than the other models in terms of continuity across axial slices, but the difference was not significant compared with the 2D CNN model with the same number of slices as the input.

Learning low-dose CT degradation from unpaired data with flow-based model

BACKGROUND

There has been growing interest in low-dose computed tomography (LDCT) for reducing the X-ray radiation to patients. However, LDCT always suffers from complex noise in reconstructed images. Although deep learning-based methods have shown their strong performance in LDCT denoising, most of them require a large number of paired training data of normal-dose CT (NDCT) images and LDCT images, which are hard to acquire in the clinic. Lack of paired training data significantly undermines the practicability of supervised deep learning-based methods. To alleviate this problem, unsupervised or weakly supervised deep learning-based methods are required.

PURPOSE

We aimed to propose a method that achieves LDCT denoising without training pairs. Specifically, we first trained a neural network in a weakly supervised manner to simulate LDCT images from NDCT images. Then, simulated training pairs could be used for supervised deep denoising networks.

METHODS

We proposed a weakly supervised method to learn the degradation of LDCT from unpaired LDCT and NDCT images. Concretely, LDCT and normal-dose images were fed into one shared flow-based model and projected to the latent space. Then, the degradation between low-dose and normal-dose images was modeled in the latent space. Finally, the model was trained by minimizing the negative log-likelihood loss with no requirement of paired training data. After training, an NDCT image can be input to the trained flow-based model to generate the corresponding LDCT image. The simulated image pairs of NDCT and LDCT can be further used to train supervised denoising neural networks for test.

RESULTS

Our method achieved much better performance on LDCT image simulation compared with the most widely used image-to-image translation method, CycleGAN, according to the radial noise power spectrum. The simulated image pairs could be used for any supervised LDCT denoising neural networks. We validated the effectiveness of our generated image pairs on a classic convolutional neural network, REDCNN, and a novel transformer-based model, TransCT. Our method achieved mean peak signal-to-noise ratio (PSNR) of 24.43dB, mean structural similarity (SSIM) of 0.785 on an abdomen CT dataset, mean PSNR of 33.88dB, mean SSIM of 0.797 on a chest CT dataset, which outperformed several traditional CT denoising methods, the same network trained by CycleGAN-generated data, and a novel transfer learning method. Besides, our method was on par with the supervised networks in terms of visual effects.

CONCLUSION

We proposed a flow-based method to learn LDCT degradation from only unpaired training data. It achieved impressive performance on LDCT synthesis. Next, we could train neural networks with the generated paired data for LDCT denoising. The denoising results are better than traditional and weakly supervised methods, comparable to supervised deep learning methods.

ERA-WGAT: Edge-enhanced residual autoencoder with a window-based graph attention convolutional network for low-dose CT denoising

Computed tomography (CT) has become a powerful tool for medical diagnosis. However, minimizing X-ray radiation risk for the patient poses significant challenges to obtain suitable low dose CT images. Although various low-dose CT methods using deep learning techniques have produced impressive results, convolutional neural network based methods focus more on local information and hence are very limited for non-local information extraction. This paper proposes ERA-WGAT, a residual autoencoder incorporating an edge enhancement module that performs convolution with eight types of learnable operators providing rich edge information and a window-based graph attention convolutional network that combines static and dynamic attention modules to explore non-local self-similarity. We use the compound loss function that combines MSE loss and multi-scale perceptual loss to mitigate the over-smoothing problem. Compared with current low-dose CT denoising methods, ERA-WGAT confirmed superior noise suppression and perceived image quality.

Generation of synthetic megavoltage CT for MRI-only radiotherapy treatment planning using a 3D deep convolutional neural network

BACKGROUND

Megavoltage computed tomography (MVCT) has been implemented on many radiotherapy treatment machines for on-board anatomical visualization, localization, and adaptive dose calculation. Implementing an MR-only workflow by synthesizing MVCT from magnetic resonance imaging (MRI) would offer numerous advantages for treatment planning and online adaptation.

PURPOSE

In this work, we sought to synthesize MVCT (sMVCT) datasets from MRI using deep learning to demonstrate the feasibility of MRI-MVCT only treatment planning.

METHODS

MVCTs and T1-weighted MRIs for 120 patients treated for head-and-neck cancer were retrospectively acquired and co-registered. A deep neural network based on a fully-convolutional 3D U-Net architecture was implemented to map MRI intensity to MVCT HU. Input to the model were volumetric patches generated from paired MRI and MVCT datasets. The U-Net was initialized with random parameters and trained on a mean absolute error (MAE) objective function. Model accuracy was evaluated on 18 withheld test exams. sMVCTs were compared to respective MVCTs. Intensity-modulated volumetric radiotherapy (IMRT) plans were generated on MVCTs of four different disease sites and compared to plans calculated onto corresponding sMVCTs using the gamma metric and dose-volume-histograms (DVHs).

RESULTS

MAE values between sMVCT and MVCT datasets were 93.3 ± 27.5, 78.2 ± 27.5, and 138.0 ± 43.4 HU for whole body, soft tissue, and bone volumes, respectively. Overall, there was good agreement between sMVCT and MVCT, with bone and air posing the greatest challenges. The retrospective dataset introduced additional deviations due to sinus filling or tumor growth/shrinkage between scans, differences in external contours due to variability in patient positioning, or when immobilization devices were absent from diagnostic MRIs. Dose distributions of IMRT plans evaluated for four test cases showed close agreement between sMVCT and MVCT images when evaluated using DVHs and gamma dose metrics, which averaged to 98.9 ± 1.0% and 96.8 ± 2.6% analyzed at 3%/3 mm and 2%/2 mm, respectively.

CONCLUSIONS

MVCT datasets can be generated from T1-weighted MRI using a 3D deep convolutional neural network with dose calculation on a sample sMVCT in close agreement with the MVCT. These results demonstrate the feasibility of using MRI-derived sMVCT in an MR-only treatment planning workflow.

Low-dose computed tomography image reconstruction via a multistage convolutional neural network with autoencoder perceptual loss network

BACKGROUND

Computed tomography (CT) is widely used in medical diagnoses due to its ability to non-invasively detect the internal structures of the human body. However, CT scans with normal radiation doses can cause irreversible damage to patients. The radiation exposure is reduced with low-dose CT (LDCT), although considerable speckle noise and streak artifacts in CT images and even structural deformation may result, significantly undermining its diagnostic capability.

METHODS

This paper proposes a multistage network framework which gradually divides the entire process into 2-staged sub-networks to complete the task of image reconstruction. Specifically, a dilated residual convolutional neural network (DRCNN) was used to denoise the LDCT image. Then, the learned context information was combined with the channel attention subnet, which retains local information, to preserve the structural details and features of the image and textural information. To obtain recognizable characteristic details, we introduced a novel self-calibration module (SCM) between the 2 stages to reweight the local features, which realizes the complementation of information at different stages while refining feature information. In addition, we also designed an autoencoder neural network, using a self-supervised learning scheme to train a perceptual loss neural network specifically for CT images.

RESULTS

We evaluated the diagnostic quality of the results and performed ablation experiments on the loss function and network structure modules to verify each module's effectiveness in the network. Our proposed network architecture obtained high peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and visual information fidelity (VIF) values in terms of quantitative evaluation. In the analysis of qualitative results, our network structure maintained a better balance between eliminating image noise and preserving image details. Experimental results showed that our proposed network structure obtained better metrics and visual evaluation.

CONCLUSIONS

This study proposed a new LDCT image reconstruction method by combining autoencoder perceptual loss networks with multistage convolutional neural networks (MSCNN). Experimental results showed that the newly proposed method has performance than other methods.

Improvement of megavoltage computed tomography image quality for adaptive helical tomotherapy using cycleGAN-based image synthesis with small datasets

PURPOSE

Megavoltage computed tomography (MVCT) offers an opportunity for adaptive helical tomotherapy. However, high noise and reduced contrast in the MVCT images due to a decrease in the imaging dose to patients limits its usability. Therefore, we propose an algorithm to improve the image quality of MVCT.

METHODS

The proposed algorithm generates kilovoltage CT (kVCT)-like images from MVCT images using a cycle-consistency generative adversarial network (cycleGAN)-based image synthesis model. Data augmentation using an affine transformation was applied to the training data to overcome the lack of data diversity in the network training. The mean absolute error (MAE), root-mean-square error (RMSE), peak signal-to-noise ratio (PSNR), and structural similarity index measure (SSIM) were used to quantify the correction accuracy of the images generated by the proposed algorithm. The proposed method was validated by comparing the images generated with those obtained from conventional and deep learning-based image processing method through non-augmented datasets.

RESULTS

The average MAE, RMSE, PSNR, and SSIM values were 18.91 HU, 69.35 HU, 32.73 dB, and 95.48 using the proposed method, respectively, whereas cycleGAN with non-augmented data showed inferior results (19.88 HU, 70.55 HU, 32.62 dB, 95.19, respectively). The voxel values of the image obtained by the proposed method also indicated similar distributions to those of the kVCT image. The dose-volume histogram of the proposed method was also similar to that of electron density corrected MVCT.

CONCLUSIONS

The proposed algorithm generates synthetic kVCT images from MVCT images using cycleGAN with small patient datasets. The image quality achieved by the proposed method was correspondingly improved to the level of a kVCT image while maintaining the anatomical structure of an MVCT image. The evaluation of dosimetric effectiveness of the proposed method indicates the applicability of accurate treatment planning in adaptive radiation therapy.

Improved contrast and noise of megavoltage computed tomography (MVCT) through cycle-consistent generative machine learning

PURPOSE

Megavoltage computed tomography (MVCT) has been implemented on many radiation therapy treatment machines as a tomographic imaging modality that allows for three-dimensional visualization and localization of patient anatomy. Yet MVCT images exhibit lower contrast and greater noise than its kilovoltage CT (kVCT) counterpart. In this work, we sought to improve these disadvantages of MVCT images through an image-to-image-based machine learning transformation of MVCT and kVCT images. We demonstrated that by learning the style of kVCT images, MVCT images can be converted into high-quality synthetic kVCT (skVCT) images with higher contrast and lower noise, when compared to the original MVCT.

METHODS

Kilovoltage CT and MVCT images of 120 head and neck (H&N) cancer patients treated on an Accuray TomoHD system were retrospectively analyzed in this study. A cycle-consistent generative adversarial network (CycleGAN) machine learning, a variant of the generative adversarial network (GAN), was used to learn Hounsfield Unit (HU) transformations from MVCT to kVCT images, creating skVCT images. A formal mathematical proof is given describing the interplay between function sensitivity and input noise and how it applies to the error variance of a high-capacity function trained with noisy input data. Finally, we show how skVCT shares distributional similarity to kVCT for various macro-structures found in the body.

RESULTS

Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were improved in skVCT images relative to the original MVCT images and were consistent with kVCT images. Specifically, skVCT CNR for muscle-fat, bone-fat, and bone-muscle improved to 14.8 ± 0.4, 122.7 ± 22.6, and 107.9 ± 22.4 compared with 1.6 ± 0.3, 7.6 ± 1.9, and 6.0 ± 1.7, respectively, in the original MVCT images and was more consistent with kVCT CNR values of 15.2 ± 0.8, 124.9 ± 27.0, and 109.7 ± 26.5, respectively. Noise was significantly reduced in skVCT images with SNR values improving by roughly an order of magnitude and consistent with kVCT SNR values. Axial slice mean (S-ME) and mean absolute error (S-MAE) agreement between kVCT and MVCT/skVCT improved, on average, from -16.0 and 109.1 HU to 8.4 and 76.9 HU, respectively.

CONCLUSIONS

A kVCT-like qualitative aid was generated from input MVCT data through a CycleGAN instance. This qualitative aid, skVCT, was robust toward embedded metallic material, dramatically improves HU alignment from MVCT, and appears perceptually similar to kVCT with SNR and CNR values equivalent to that of kVCT images.

DaNet: dose-aware network embedded with dose-level estimation for low-dose CT imaging

Many deep learning (DL)-based image restoration methods for low-dose CT (LDCT) problems directly employ the end-to-end networks on low-dose training data without considering dose differences. However, the radiation dose difference has a great impact on the ultimate results, and lower doses increase the difficulty of restoration. Moreover, there is increasing demand to design and estimate acceptable scanning doses for patients in clinical practice, necessitating dose-aware networks embedded with adaptive dose estimation. In this paper, we consider these dose differences of input LDCT images and propose an adaptive dose-aware network. First, considering a large dose distribution range for simulation convenience, we coarsely define five dose levels in advance as lowest, lower, mild, higher and highest radiation dose levels. Instead of directly building the end-to-end mapping function between LDCT images and high-dose CT counterparts, the dose level is primarily estimated in the first stage. In the second stage, the adaptively learned low-dose level is used to guide the image restoration process as the pattern of prior information through the channel feature transform. We conduct experiments on a simulated dataset based on original high dose parts of American Association of Physicists in Medicine challenge datasets from the Mayo Clinic. Ablation studies validate the effectiveness of the dose-level estimation, and the experimental results show that our method is superior to several other DL-based methods. Specifically, our method provides obviously better performance in terms of the peak signal-to-noise ratio and visual quality reflected in subjective scores. Due to the dual-stage process, our method may suffer limitations under more parameters and coarse dose-level definitions, and thus, further improvements in clinical practical applications with different CT equipment vendors are planned in future work.

SACNN: Self-Attention Convolutional Neural Network for Low-Dose CT Denoising With Self-Supervised Perceptual Loss Network

Computed tomography (CT) is a widely used screening and diagnostic tool that allows clinicians to obtain a high-resolution, volumetric image of internal structures in a non-invasive manner. Increasingly, efforts have been made to improve the image quality of low-dose CT (LDCT) to reduce the cumulative radiation exposure of patients undergoing routine screening exams. The resurgence of deep learning has yielded a new approach for noise reduction by training a deep multi-layer convolutional neural networks (CNN) to map the low-dose to normal-dose CT images. However, CNN-based methods heavily rely on convolutional kernels, which use fixed-size filters to process one local neighborhood within the receptive field at a time. As a result, they are not efficient at retrieving structural information across large regions. In this paper, we propose a novel 3D self-attention convolutional neural network for the LDCT denoising problem. Our 3D self-attention module leverages the 3D volume of CT images to capture a wide range of spatial information both within CT slices and between CT slices. With the help of the 3D self-attention module, CNNs are able to leverage pixels with stronger relationships regardless of their distance and achieve better denoising results. In addition, we propose a self-supervised learning scheme to train a domain-specific autoencoder as the perceptual loss function. We combine these two methods and demonstrate their effectiveness on both CNN-based neural networks and WGAN-based neural networks with comprehensive experiments. Tested on the AAPM-Mayo Clinic Low Dose CT Grand Challenge data set, our experiments demonstrate that self-attention (SA) module and autoencoder (AE) perceptual loss function can efficiently enhance traditional CNNs and can achieve comparable or better results than the state-of-the-art methods.

Gradient regularized convolutional neural networks for low-dose CT image enhancement

The potential risks of x-ray to patients have transferred the public's attention from normal dose CT (NDCT) to low-dose CT (LDCT). However, simply lowering the radiation dose of the CT system will significantly degrade the quality of CT images such as noise and artifacts, which compromises the diagnostic performance. Hence, various methods have been proposed to solve this problem over the past decades. Although these methods have achieved impressive results, they also suffer from a drawback of smoothing image details after denoising, which makes it difficult for clinical diagnosis and treatment. To address this issue, this paper introduces a novel gradient regularization method for LDCT enhancement. Rather than common methods which only consider the pixel-wise gray value loss in the reconstruction procedure, we also take the image gradient loss into consideration to preserve image details. By combining the gradient regularization method and the convolutional neural network (CNN) framework, a gradient regularized convolutional neural network (GRCNN) is proposed to enhance LDCT images which has achieved promising performance in our experiments both visually and quantitatively.

A Megavoltage CT Image Enhancement Method for Image-Guided and Adaptive Helical TomoTherapy

Purpose: To propose a novel method to improve the mega-voltage CT (MVCT) image quality for helical TomoTherapy while maintaining the stability on dose calculation.

Materials and Methods: The Block-Matching 3D-transform (BM3D) and Discriminative Feature Representation (DFR) methods were combined into a novel BM3D + DFR method for their respective advantages. A phantom (Catphan504) and three serials of clinical (head & neck, chest, and pelvis) MVCT images from 30 patients were acquired using the helical TomoTherapy system. The contrast-to-noise ratio (CNR) and edge detection algorithm (canny) was employed for image quality comparisons between the original and BM3D + DFR enhanced MVCT. A simulated rectangular field of 6 MV X-ray beams were vertically delivered on the original and post-processed MVCT serials of the same CT density phantom, and the dose curves on both serials were compared to test the effects of image enhancement on dose calculation accuracy.

Results: In total, 466 transversal MVCT slices were acquired and processed by both BM3D and the proposed BM3D + DFR methods. Compared to the original MVCT image, the BM3D + DFR method presented a remarkable improvement in terms of the soft tissue contrast and noise reduction. For the phantom image, the CNR of the region of interest (ROI) was improved from 1.70 to 4.03. The average CNR of ROIs for 10 patients from each anatomical group, were increased significantly from 1.45 ± 1.51 to 2.09 ± 1.68 for the head & neck (p < 0.001), from 0.92 ± 0.78 to 1.36 ± 0.85 for the chest (p < 0.001), and from 1.12 ± 1.22 to 1.76 ± 1.31 for the pelvis (p < 0.001), respectively. The canny edge detection operator showed that BM3D + DFR provided clearer organ boundaries with less chaos. The root-mean-square of the dosimetry difference on the iso-center passed horizontal dose profile curves and vertical percentage depth dose curves were only 0.09% and 0.06%, respectively.

Conclusions: The proposed BM3D + DFR method is feasible to improve the soft tissue contrast for the original MVCT images with coincidence in dose calculation and without compromising resolution. After integration in clinical workflow, the post-processed MVCT may be better applied on image-guided and adaptive helical TomoTherapy.

A post-processing method based on interphase motion correction and averaging to improve image quality of 4D magnetic resonance imaging: a clinical feasibility study

METHODS:

Nine patients (seven pancreas, one liver, and one lung) were recruited. 4D-MRI was performed using two prototype k-space sorted techniques, stack-of-stars (SOS) and koosh-ball (KB) acquisitions. Post-processing using MoCoAve was implemented for both methods. Image quality score, apparent SNR (aSNR), sharpness, motion trajectory and standard deviation (σ_GTV) of the gross tumor volumes were compared between original and MoCoAve image sets.

RESULTS:

All subjects successfully underwent 4D-MRI scans and MoCoAve was performed on all data sets. Significantly higher image quality scores (2.64 ± 0.39 vs 1.18 ± 0.34, p = 0.001) and aSNR (37.6 ± 15.3 vs 18.1 ± 5.7, p = 0.001) was observed in the MoCoAve images when compared to the original images. High correlation in tumor motion trajectories in the superoinferior direction (SI: 0.91 ± 0.08) and weaker in the anteroposterior (AP: 0.51 ± 0.44) and mediolateral (ML: 0.37 ± 0.23) directions, similar image sharpness (0.367 ± 0.068 vs 0.369 ± 0.072, p = 0.805), and minimal average absolute difference (0.47 ± 0.34  mm) of the motion trajectory profiles was found between the two image sets. The σ_GTV in pancreas patients was significantly (p = 0.039) lower in MoCoAve images (1.48 ± 1.35  cm3) than in the original images (2.17 ± 1.31  cm3).

CONCLUSION:

MoCoAve using interphase motion correction and averaging has shown promise as a post-processing method for improving k-space sorted (SOS and KB) 4D-MRI image quality in thoracic and abdominal cancer patients.

ADVANCES IN KNOWLEDGE:

The proposed method is an image based post-processing method that could be applied to many k-space sorted 4D-MRI methods for improved image quality and signal-to-noise ratio while preserving image sharpness and respiratory motion fidelity. It is a useful technique for the radiotherapy planning community who are interested in using 4D-MRI but aren't satisfied with their current MR image quality.

Technical Note: Iterative megavoltage CT (MVCT) reconstruction using block-matching 3D-transform (BM3D) regularization

PURPOSE

Megavoltage CT (MVCT) images are noisier than kilovoltage CT (KVCT) due to low detector efficiency to high-energy x rays. Conventional denoising methods compromise edge resolution and low-contrast object visibility. In this work, we incorporated block-matching 3D-transform shrinkage (BM3D) transformation into MVCT iterative reconstruction as nonlocal patch-wise regularization.

METHODS

The iterative reconstruction was achieved by adding to the existing least square data fidelity objective a regularization term, formulated as the L1 norm of the BM3D transformed image. A Fast Iterative Shrinkage-Thresholding Algorithm (FISTA) was adopted to accelerate CT reconstruction. The proposed method was compared against total variation (TV) regularization, BM3D postprocess method, and filtered back projection (FBP).

RESULTS

In the Catphan phantom study, BM3D regularization better enhances low-contrast objects compared with TV regularization and BM3D postprocess method at the same noise level. The spatial resolution using BM3D regularization is 2.79 and 2.55 times higher than that using the TV regularization at 50% of the modulation transfer function (MTF) magnitude, for the fully sampled reconstruction and down-sampled reconstruction, respectively. The BM3D regularization images show better bony details and low-contrast soft tissues, on the head and neck (H&N) and prostate patient images.

CONCLUSIONS

The proposed iterative BM3D regularization CT reconstruction method takes advantage of both the BM3D denoising capability and iterative reconstruction data fidelity consistency. This novel approach is superior to TV regularized iterative reconstruction or BM3D postprocess for improving noisy MVCT image quality.

Stacked competitive networks for noise reduction in low-dose CT

Since absorption of X-ray radiation has the possibility of inducing cancerous, genetic and other diseases to patients, researches usually attempt to reduce the radiation dose. However, reduction of the radiation dose associated with CT scans will unavoidably increase the severity of noise and artifacts, which can seriously affect diagnostic confidence. Due to the outstanding performance of deep neural networks in image processing, in this paper, we proposed a Stacked Competitive Network (SCN) approach to noise reduction, which stacks several successive Competitive Blocks (CB). The carefully handcrafted design of the competitive blocks was inspired by the idea of multi-scale processing and improvement the network's capacity. Qualitative and quantitative evaluations demonstrate the competitive performance of the proposed method in noise suppression, structural preservation, and lesion detection.

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.

4D-CT scans reveal reduced magnitude of respiratory liver motion achieved by different abdominal compression plate positions in patients with intrahepatic tumors undergoing helical tomotherapy

PURPOSE

While abdominal compression (AC) can be used to reduce respiratory liver motion in patients receiving helical tomotherapy for hepatocellular carcinoma, the nature and extent of this effect is not well described. The purpose of this study was to evaluate the changes in magnitude of three-dimensional liver motion with abdominal compression using four-dimensional (4D) computed tomography (CT) images of several plate positions.

METHODS

From January 2012 to October 2015, 72 patients with intrahepatic carcinoma and divided into four groups underwent 4D-CT scans to assess respiratory liver motion. Of the 72 patients, 19 underwent abdominal compression of the cephalic area between the subxiphoid and umbilicus (group A), 16 underwent abdominal compression of the caudal region between the subxiphoid area and the umbilicus (group B), 11 patients underwent abdominal compression of the caudal umbilicus (group C), and 26 patients remained free breathing (group D). 4D-CT images were sorted into ten-image series, according to the respiratory phase from the end inspiration to the end expiration, and then transferred to treatment planning software. All liver contours were drawn by a single physician and confirmed by a second physician. Liver relative coordinates were automatically generated to calculate the liver respiratory motion in different axial directions to compile the 10 ten contours into a single composite image. Differences in respiratory liver motion were assessed with a one-way analysis of variance test of significance.

RESULTS

The average respiratory liver motion in the Y axial direction was 4.53 ± 1.16, 7.56 ± 1.30, 9.95 ± 2.32, and 9.53 ± 2.62 mm in groups A, B, C, and D, respectively, with a significant change among the four groups (p < 0.001). Abdominal compression was most effective in group A (compression plate on the subxiphoid area), with liver displacement being 2.53 ± 0.93, 4.53 ± 1.16, and 2.14 ± 0.92 mm on the X-, Y-, and Z-axes, respectively. There was no significant difference in respiratory liver motion between group C (displacement: 3.23 ± 1.47, 9.95 ± 2.32, and 2.92 ± 1.10 mm on the X-, Y-, and Z-axes, respectively) and group D (displacement: 3.35 ± 1.55, 9.53 ± 2.62, and 3.35 ± 1.73 mm on the X-, Y-, and Z-axes, respectively). Abdominal compression was least effective in group C (compression on caudal umbilicus), with liver motion in this group similar to that of free-breathing patients (group D).

CONCLUSIONS

4D-CT scans revealed significant liver motion control via abdominal compression of the subxiphoid area; however, this control of liver motion was not observed with compression of the caudal umbilicus. The authors, therefore, recommend compression of the subxiphoid area in patients undergoing external radiotherapy for intrahepatic carcinoma.

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.

Nonlocal Means Denoising of Self-Gated and k-Space Sorted 4-Dimensional Magnetic Resonance Imaging Using Block-Matching and 3-Dimensional Filtering: Implications for Pancreatic Tumor Registration and Segmentation

PURPOSE

To denoise self-gated k-space sorted 4-dimensional magnetic resonance imaging (SG-KS-4D-MRI) by applying a nonlocal means denoising filter, block-matching and 3-dimensional filtering (BM3D), to test its impact on the accuracy of 4D image deformable registration and automated tumor segmentation for pancreatic cancer patients.

METHODS AND MATERIALS

Nine patients with pancreatic cancer and abdominal SG-KS-4D-MRI were included in the study. Block-matching and 3D filtering was adapted to search in the axial slices/frames adjacent to the reference image patch in the spatial and temporal domains. The patches with high similarity to the reference patch were used to collectively denoise the 4D-MRI image. The pancreas tumor was manually contoured on the first end-of-exhalation phase for both the raw and the denoised 4D-MRI. B-spline deformable registration was applied to the subsequent phases for contour propagation. The consistency of tumor volume defined by the standard deviation of gross tumor volumes from 10 breathing phases (σ_GTV), tumor motion trajectories in 3 cardinal motion planes, 4D-MRI imaging noise, and image contrast-to-noise ratio were compared between the raw and denoised groups.

RESULTS

Block-matching and 3D filtering visually and quantitatively reduced image noise by 52% and improved image contrast-to-noise ratio by 56%, without compromising soft tissue edge definitions. Automatic tumor segmentation is statistically more consistent on the denoised 4D-MRI (σ_GTV = 0.6 cm(3)) than on the raw 4D-MRI (σ_GTV = 0.8 cm(3)). Tumor end-of-exhalation location is also more reproducible on the denoised 4D-MRI than on the raw 4D-MRI in all 3 cardinal motion planes.

CONCLUSIONS

Block-matching and 3D filtering can significantly reduce random image noise while maintaining structural features in the SG-KS-4D-MRI datasets. In this study of pancreatic tumor segmentation, automatic segmentation of GTV in the registered image sets is shown to be more consistent on the denoised 4D-MRI than on the raw 4D-MRI.

Radiation-induced CT number changes in GTV and parotid glands during the course of radiation therapy for nasopharyngeal cancer

OBJECTIVE

To investigate the changes in CT number (CTN) in gross tumour volume (GTV) and organs at risk (OARs) during the course of radiation therapy (RT) for nasopharyngeal cancer (NPC).

METHODS

Daily megavoltage CT (MVCT) data collected from 30 patients with NPC treated with a prescription dose of 70 Gy in 30-33 fractions using helical tomotherapy were retrospectively analyzed. The contours of GTV and OARs on daily MVCTs were obtained by populating the planning contours from planning CT to daily MVCTs with manual editing, if necessary. The changes of GTV and OAR volumes and the histograms of CTN in the GTV and OARs during the course of RT delivery were analyzed.

RESULTS

Volumes of GTV and parotid glands were reduced during the course of radiation treatment, with an average shrinkage rate of 0.23% per day (range, 0.02-0.8%) and 1.2% per day (range, 0.2-2.3%), respectively. The mean CTN changes in GTV and ipsilateral and contralateral parotid glands were reduced by 52 ± 35 HU, 18 ± 20 HU and 17 ± 22 HU, respectively. For GTV, the CTN and GTV volume decreases were found to be correlated with each other (p < 0.0001). No noticeable CTN change was found in the spinal cord and non-specified tissue irradiated with low doses.

CONCLUSION

The CTN changes in GTV and parotids are measurable during the delivery of fractionated radiotherapy for NPC, were associated with the doses received (the number of fractions delivered) and were patient specific.

ADVANCES IN KNOWLEDGE

The CTN change during radiotherapy is dose dependent and is measurable for NPC.

Four-Dimensional Magnetic Resonance Imaging With 3-Dimensional Radial Sampling and Self-Gating-Based K-Space Sorting: Early Clinical Experience on Pancreatic Cancer Patients

PURPOSE

To apply a novel self-gating k-space sorted 4-dimensional MRI (SG-KS-4D-MRI) method to overcome limitations due to anisotropic resolution and rebinning artifacts and to monitor pancreatic tumor motion.

METHODS AND MATERIALS

Ten patients were imaged using 4D-CT, cine 2-dimensional MRI (2D-MRI), and the SG-KS-4D-MRI, which is a spoiled gradient recalled echo sequence with 3-dimensional radial-sampling k-space projections and 1-dimensional projection-based self-gating. Tumor volumes were defined on all phases in both 4D-MRI and 4D-CT and then compared.

RESULTS

An isotropic resolution of 1.56 mm was achieved in the SG-KS-4D-MRI images, which showed superior soft-tissue contrast to 4D-CT and appeared to be free of stitching artifacts. The tumor motion trajectory cross-correlations (mean ± SD) between SG-KS-4D-MRI and cine 2D-MRI in superior-inferior, anterior-posterior, and medial-lateral directions were 0.93 ± 0.03, 0.83 ± 0.10, and 0.74 ± 0.18, respectively. The tumor motion trajectories cross-correlations between SG-KS-4D-MRI and 4D-CT in superior-inferior, anterior-posterior, and medial-lateral directions were 0.91 ± 0.06, 0.72 ± 0.16, and 0.44 ± 0.24, respectively. The average standard deviation of gross tumor volume calculated from the 10 breathing phases was 0.81 cm(3) and 1.02 cm(3) for SG-KS-4D-MRI and 4D-CT, respectively (P=.012).

CONCLUSIONS

A novel SG-KS-4D-MRI acquisition method capable of reconstructing rebinning artifact-free, high-resolution 4D-MRI images was used to quantify pancreas tumor motion. The resultant pancreatic tumor motion trajectories agreed well with 2D-cine-MRI and 4D-CT. The pancreatic tumor volumes shown in the different phases for the SG-KS-4D-MRI were statistically significantly more consistent than those in the 4D-CT.