Image Quality Of Low-dose CT Research Articles

StatNet: Statistical Image Restoration for Low-Dose CT using Deep Learning

Deep learning has recently attracted widespread interest as a means of reducing noise in low-dose CT (LDCT) images. Deep convolutional neural networks (CNNs) are typically trained to transfer high-quality image features of normal-dose CT (NDCT) images to LDCT images. However, existing deep learning approaches for denoising LDCT images often overlook the statistical property of CT images. In this paper, we propose an approach to statistical image restoration for LDCT using deep learning. We introduce a loss function to incorporate the noise property in image domain derived from the noise statistics in sinogram domain. In order to capture the spatially-varying statistics of CT images, we increase the receptive fields of the neural network to cover full-size CT slices. In addition, the proposed network utilizes $z$-directional correlation by taking multiple consecutive CT slices as input. For performance evaluation, the proposed networks are trained and validated with a public dataset consisting of LDCT-NDCT image pairs. We also perform a retrospective study by testing the networks with clinical LDCT images. The experimental results show that the denoising networks successfully reduce the noise level and restore the image details without adding artifacts. This study demonstrates that the statistical deep learning approach can restore the image quality of LDCT without loss of anatomical information.

Image Quality and Lesion Detection on Deep Learning Reconstruction and Iterative Reconstruction of Submillisievert Chest and Abdominal CT.

OBJECTIVE. The objective of this study was to compare image quality and clinically significant lesion detection on deep learning reconstruction (DLR) and iterative reconstruction (IR) images of submillisievert chest and abdominopelvic CT. MATERIALS AND METHODS. Our prospective multiinstitutional study included 59 adult patients (33 women, 26 men; mean age ± SD, 65 ± 12 years old; mean body mass index [weight in kilograms divided by the square of height in meters] = 27 ± 5) who underwent routine chest (n = 22; 16 women, six men) and abdominopelvic (n = 37; 17 women, 20 men) CT on a 640-MDCT scanner (Aquilion ONE, Canon Medical Systems). All patients gave written informed consent for the acquisition of low-dose (LD) CT (LDCT) after a clinically indicated standard-dose (SD) CT (SDCT). The SDCT series (120 kVp, 164-644 mA) were reconstructed with interactive reconstruction (IR) (adaptive iterative dose reduction [AIDR] 3D, Canon Medical Systems), and the LDCT (100 kVp, 120 kVp; 30-50 mA) were reconstructed with filtered back-projection (FBP), IR (AIDR 3D and forward-projected model-based iterative reconstruction solution [FIRST], Canon Medical Systems), and deep learning reconstruction (DLR) (Advanced Intelligent Clear-IQ Engine [AiCE], Canon Medical Systems). Four subspecialty-trained radiologists first read all LD image sets and then compared them side-by-side with SD AIDR 3D images in an independent, randomized, and blinded fashion. Subspecialty radiologists assessed image quality of LDCT images on a 3-point scale (1 = unacceptable, 2 = suboptimal, 3 = optimal). Descriptive statistics were obtained, and the Wilcoxon sign rank test was performed. RESULTS. Mean volume CT dose index and dose-length product for LDCT (2.1 ± 0.8 mGy, 49 ± 13mGy·cm) were lower than those for SDCT (13 ± 4.4 mGy, 567 ± 249 mGy·cm) (p < 0.0001). All 31 clinically significant abdominal lesions were seen on SD AIDR 3D and LD DLR images. Twenty-five, 18, and seven lesions were detected on LD AIDR 3D, LD FIRST, and LD FBP images, respectively. All 39 pulmonary nodules detected on SD AIDR 3D images were also noted on LD DLR images. LD DLR images were deemed acceptable for interpretation in 97% (35/37) of abdominal and 95-100% (21-22/22) of chest LDCT studies (p = 0.2-0.99). The LD FIRST, LD AIDR 3D, and LD FBP images had inferior image quality compared with SD AIDR 3D images (p < 0.0001). CONCLUSION. At submillisievert chest and abdominopelvic CT doses, DLR enables image quality and lesion detection superior to commercial IR and FBP images.

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.

Deep learning-based image quality improvement for low-dose computed tomography simulation in radiation therapy.

Low-dose computed tomography (CT) is desirable for treatment planning and simulation in radiation therapy. Multiple rescanning and replanning during the treatment course with a smaller amount of dose than a single conventional full-dose CT simulation is a crucial step in adaptive radiation therapy. We developed a machine learning-based method to improve image quality of low-dose CT for radiation therapy treatment simulation. We used a residual block concept and a self-attention strategy with a cycle-consistent adversarial network framework. A fully convolution neural network with residual blocks and attention gates (AGs) was used in the generator to enable end-to-end transformation. We have collected CT images from 30 patients treated with frameless brain stereotactic radiosurgery (SRS) for this study. These full-dose images were used to generate projection data, which were then added with noise to simulate the low-mAs scanning scenario. Low-dose CT images were reconstructed from this noise-contaminated projection data and were fed into our network along with the original full-dose CT images for training. The performance of our network was evaluated by quantitatively comparing the high-quality CT images generated by our method with the original full-dose images. When mAs is reduced to 0.5% of the original CT scan, the mean square error of the CT images obtained by our method is , with respect to the original full-dose images. The proposed method successfully improved the noise, contract-to-noise ratio, and nonuniformity level to be close to those of full-dose CT images and outperforms a state-of-the-art iterative reconstruction method. Dosimetric studies show that the average differences of dose-volume histogram metrics are ( ). These quantitative results strongly indicate that the denoised low-dose CT images using our method maintains image accuracy and quality and are accurate enough for dose calculation in current CT simulation of brain SRS treatment. We also demonstrate the great potential for low-dose CT in the process of simulation and treatment planning.

Deep Learning for Low-Dose CT Denoising Using Perceptual Loss and Edge Detection Layer

Low-dose CT denoising is a challenging task that has been studied by many researchers. Some studies have used deep neural networks to improve the quality of low-dose CT images and achieved fruitful results. In this paper, we propose a deep neural network that uses dilated convolutions with different dilation rates instead of standard convolution helping to capture more contextual information in fewer layers. Also, we have employed residual learning by creating shortcut connections to transmit image information from the early layers to later ones. To further improve the performance of the network, we have introduced a non-trainable edge detection layer that extracts edges in horizontal, vertical, and diagonal directions. Finally, we demonstrate that optimizing the network by a combination of mean-square error loss and perceptual loss preserves many structural details in the CT image. This objective function does not suffer from over smoothing and blurring effects causing by per-pixel loss and grid-like artifacts resulting from perceptual loss. The experiments show that each modification to the network improves the outcome while changing the complexity of the network, minimally.

Low-Dose CT Image Restoration Based on Adaptive Prior Feature Matching and Nonlocal Means

Although the low-dose CT (LDCT) technique can reduce the radiation damage to patients, it will be highly detrimental to the reconstructed image quality. The normal-dose scan assisted algorithms have shown their potential in improving LDCT image quality by using a registered previously scanned normal-dose CT (NDCT) reference to regularize the corresponding LDCT target. The major drawback of such methods is the requirement of a previous patient-specific NDCT scan, which limits their clinical application. To address these problems, this paper proposed adaptive prior feature matching method for better restoration of the LDCT image. The innovation lies in construction of offline texture feature database and online adaptive prior feature matching integrated with the NLM regularization. Specifically, the prior features were extracted by the gray level co-occurrence matrix (GLCM) from regions of interest (ROIs) in existing NDCT scans of population patients. For online adaptive prior feature matching, ROIs with their texture features being similar to those of the current noisy target ROI are selected from the database as the references for the NLM regularization. The effectiveness of the proposed algorithm is validated by clinical lung cancer studies, the gain over traditional methods is noticeable in terms of both noise suppression and textures preservation.

A phantom study of the performance of model-based iterative reconstruction in low-dose chest and abdominal CT: When are benefits maximized?

IntroductionThe aim of this study was to assess and compare the effects of CT image reconstruction techniques on low-dose CT image quality using phantoms. MethodsAnthropomorphic torso and spatial/contrast-resolution phantoms were scanned at decreasing tube currents between 400 and 10 mA. CT thorax and abdomen/pelvis series were reconstructed with filtered back projection (FBP) alone, combined 40% adaptive statistical iterative reconstruction & FBP (ASIR40), and model-based iterative reconstruction (MBIR) [(resolution-preference 05 (RP05) and RP20 in the thorax and RP05 and noise-reduction 05 (NR05) in the abdomen)]. Two readers rated image quality quantitatively and qualitatively. ResultsIn thoracic CT, objective image noise on MBIR RP05 data sets outperformed FBP at 200, 100, 50 and 10 mA and outperformed ASIR40 at 50 and 10 mA (p < 0.001). MBIR RP20 outperformed FBP at 50 and 10 mA and outperformed ASIR40 at 10 mA (p < 0.001). Compared with both FBP and ASIR40, MBIR RP05 demonstrated significantly better signal-to-noise ratio (SNR) at 10 mA. In abdomino-pelvic CT, MBIR RP05 and NR05 outperformed FBP and ASIR at all tube current levels for objective image noise. NR05 demonstrated greater SNR at 200, 100, 50 and 10 mA and RP05 demonstrated greater SNR at 50 and 10 mA compared with both FBP and ASIR. MBIR images demonstrated better subjective image quality scores. Spatial resolution, low-contrast detectability and contrast-to-noise ratio (CNR) were comparable between image reconstruction techniques. ConclusionCTs reconstructed with MBIR have lower image noise and improved image quality compared with FBP and ASIR. These effects increase with reduced radiation exposure confirming optimal use for low-dose CT imaging.

Detection of Relapse by Low-dose Computed Tomography During Surveillance in Stage I Testicular Germ Cell Tumours

BackgroundStandard-dose computed tomography (SDCT) scans are associated with radiation exposure during stage I testicular cancer surveillance. ObjectiveTo evaluate low-dose CT (LDCT) for clinical use. Design, setting, and participantsIn this single-arm prospective study, patients on surveillance for stage I testicular germ cell tumour underwent SDCT and LDCT scans on their first visit after enrolment. The adequacy of LDCT image quality was assessed for subsequent use. Patients were followed with LDCT only and suspected relapse was confirmed by SDCT. Outcome measures and statistical analysisWe assessed whether initial LDCT scans were of sufficient quality for routine clinical use. We compared mean differences in nodal size at relapse between LDCT and SDCT using a one-sample paired t test. The relapse free-rate was calculated using the Kaplan-Meier method. Results and limitationsOf 257 patients, one was excluded because of inadequate image quality. At median follow-up of 5.25 yr, 35 patients had relapsed, 33 with retroperitoneal lymphadenopathy. The 2- and 5-yr relapse-free rates were 89.5% and 85.3%, respectively. The mean size of retroperitoneal nodal relapse was 17.3 and 17.5mm on the short axis, 23.2 and 22.7mm on the long axis, and 26.1 and 26.7mm on craniocaudal length for LDCT and SDCT, respectively. The mean difference between LDCT and SDCT was 0.14mm (p=0.55) short axis, -0.54mm (p=0.092) long axis, and -0.51mm (p=0.086) length. A limitation was the lack of a control arm. ConclusionsLDCT image quality was adequate for clinical use, and retroperitoneal nodal relapse was detected with minimal differences seen between LD and SDCT. LDCT can be safely adopted and will decrease overall radiation exposure in stage I germ cell tumour surveillance. Patient summaryWe studied the use of low-dose computed tomography scans for detecting testicular cancer recurrence in lymph nodes of the abdomen and pelvis and found that they were safe, effective and would potentially reduce overall X-ray exposure.This trial is registered at ClinicalTrials.gov as NCT03142802.

Improved image quality of low-dose CT combining with iterative model reconstruction algorithm for response assessment in patients after treatment of malignant tumor.

To evaluate the image quality and radiation dose of low-dose (LD) computed tomography (LD-CT) combining with iterative model reconstruction (IMR) algorithm for response assessment in patients after treatment of malignant tumor compared with routine-dose CT (RD-CT). Forty-seven patients [mean age 57.8±10.9 years, 30 males, body mass index (BMI) 22.09±2.35 kg/m2] after treatment of malignant tumor underwent contrast-enhanced chest and abdomen CT twice for response assessment with an interval of 6 months according to clinical routine. The first CT scans were performed with RD protocol at 120 kVp and images were reconstructed with filtered back projection (FBP) algorithm; while the second scans were performed with LD protocol at 100 kVp and images were reconstructed with FBP and IMR algorithm respectively. All scans were performed using an automatic tube current modulation technique with 150 mAs as reference. Objective image quality including CT attenuation, image noise, and contrast to noise ratio (CNR), and subjective image quality including artifacts, noise, visualization of small structures and confidence of targeted lesions, as well as lesion detection were assessed and compared. Effective radiation dose of LD-CT scans was reduced 54.8% compared to RD-CT scans (26.89±3.35 vs. 12.14±2.09 mSv). Higher CT attenuation was found in both LD-IMR and LD-FBP images compared to RD-FBP images. Better subjective image quality and CNR as well as lower objective noise were found in LD-IMR images (all, P<0.05). Two small lesions with the diameter less than 1 cm were missed in LD-FBP images, which were able to be observed in LD-IMR images. IMR is able to help more than half of reduction of radiation dose without compromising the quality of diagnostic information in patients after treatment of malignant tumors to chest and abdomen CT for response assessment.

Denoising Low-Dose CT Images Using Multiframe Blind Source Separation and Block Matching Filter

In order to reduce radiation effect during CT scans, low-dose techniques are employed in different medical imaging applications. But images in the low-dose CT tend to be rather noisy to be diagnostically useful. One way to improve the quality of low-dose CT images is to use a multiframe imaging technique. Here, we proposed a blind source separation (BSS) based CT image method using a multiframe low-dose image sequence. Because we found that BSS alone cannot denoise the image completely, we incorporated a nonlocal GroupWise block matching 3-D filter with BSS using the noise statistics, extracted from the noise components. With this technique, we produced a better quality image than that produced with a single frame half dose CT image and other multiframe imaging techniques, such as, frame averaging and applying the Wiener filter after BSS. Denoising performance, spatial resolution, and noise characteristics were measured by evaluating the peak signal to noise ratio, structural similarity index, modulation transfer function, and Bland–Altman analysis. This hybrid model shows a better denoising performance with less compromise in image details as more frames are included in an image sequence.

3D Feature Constrained Reconstruction for Low-Dose CT Imaging

Low-dose computed tomography (LDCT) images are often highly degraded by amplified mottle noise and streak artifacts. Maintaining image quality under low-dose scan protocols is a well-known challenge. Recently, sparse representation-based techniques have been shown to be efficient in improving such CT images. In this paper, we propose a 3D feature constrained reconstruction (3D-FCR) algorithm for LDCT image reconstruction. The feature information used in the 3D-FCR algorithm relies on a 3D feature dictionary constructed from available high quality standard-dose CT sample. The CT voxels and the sparse coefficients are sequentially updated using an alternating minimization scheme. The performance of the 3D-FCR algorithm was assessed through experiments conducted on phantom simulation data and clinical data. A comparison with previously reported solutions was also performed. Qualitative and quantitative results show that the proposed method can lead to a promising improvement of LDCT image quality.

Structurally-sensitive Multi-scale Deep Neural Network for Low-Dose CT Denoising.

Computed tomography (CT) is a popular medical imaging modality and enjoys wide clinical applications. At the same time, the x-ray radiation dose associated with CT scannings raises a public concern due to its potential risks to the patients. Over the past years, major efforts have been dedicated to the development of Low-Dose CT (LDCT) methods. However, the radiation dose reduction compromises the signal-to-noise ratio (SNR), leading to strong noise and artifacts that downgrade CT image quality. In this paper, we propose a novel 3D noise reduction method, called Structurally-sensitive Multi-scale Generative Adversarial Net (SMGAN), to improve the LDCT image quality. Specifically, we incorporate three-dimensional (3D) volumetric information to improve the image quality. Also, different loss functions for training denoising models are investigated. Experiments show that the proposed method can effectively preserve structural and textural information in reference to normal-dose CT (NDCT) images, and significantly suppress noise and artifacts. Qualitative visual assessments by three experienced radiologists demonstrate that the proposed method retrieves more information, and outperforms competing methods.

Adaptive non-local means on local principle neighborhood for noise/artifacts reduction in low-dose CT images.

Low-dose CT (LDCT) technique can reduce the x-ray radiation exposure to patients at the cost of degraded images with severe noise and artifacts. Non-local means (NLM) filtering has shown its potential in improving LDCT image quality. However, currently most NLM-based approaches employ a weighted average operation directly on all neighbor pixels with a fixed filtering parameter throughout the NLM filtering process, ignoring the non-stationary noise nature of LDCT images. In this paper, an adaptive NLM filtering scheme on local principle neighborhoods (PC-NLM) is proposed for structure-preserving noise/artifacts reduction in LDCT images. Instead of using neighboring patches directly, in the PC-NLM scheme, the principle component analysis (PCA) is first applied on local neighboring patches of the target patch to decompose the local patches into uncorrelated principle components (PCs), then a NLM filtering is used to regularize each PC of the target patch and finally the regularized components is transformed to get the target patch in image domain. Especially, in the NLM scheme, the filtering parameter is estimated adaptively from local noise level of the neighborhood as well as the signal-to-noise ratio (SNR) of the corresponding PC, which guarantees a "weaker" NLM filtering on PCs with higher SNR and a "stronger" filtering on PCs with lower SNR. The PC-NLM procedure is iteratively performed several times for better removal of the noise and artifacts, and an adaptive iteration strategy is developed to reduce the computational load by determining whether a patch should be processed or not in next round of the PC-NLM filtering. The effectiveness of the presented PC-NLM algorithm is validated by experimental phantom studies and clinical studies. The results show that it can achieve promising gain over some state-of-the-art methods in terms of artifact suppression and structure preservation. With the use of PCA on local neighborhoods to extract principal structural components, as well as adaptive NLM filtering on PCs of the target patch using filtering parameter estimated based on the local noise level and corresponding SNR, the proposed PC-NLM method shows its efficacy in preserving fine anatomical structures and suppressing noise/artifacts in LDCT images.

Learning-Based Artifact Removal via Image Decomposition for Low-Dose CT Image Processing

Streak artifacts and mottle noise often appear in low-dose CT (LDCT) images due to excessive quantum noise in low-dose X-ray imaging process, thus degrading CT image quality. This research is aimed at improving the quality of LDCT images via image decomposition and dictionary learning. The proposed method first decomposes a LDCT image into the low-frequency (LF) and high-frequency (HF) parts by a bilateral filter. The HF part is then decomposed into an artifact component and a tissue component by performing dictionary learning (DL) and sparse coding. The tissue component is combined with the LF part to obtain the artifact-suppressed image. At last, a DL method is applied to further reduce the residual artifacts and noise. Different from previous research works with sparse representation, the proposed method does not need to collect training images in advance. The results of numerical simulation and clinical data experiments indicate the effectiveness of the proposed approach.

Improving Low-dose Cardiac CT Images based on 3D Sparse Representation.

Cardiac computed tomography (CCT) is a reliable and accurate tool for diagnosis of coronary artery diseases and is also frequently used in surgery guidance. Low-dose scans should be considered in order to alleviate the harm to patients caused by X-ray radiation. However, low dose CT (LDCT) images tend to be degraded by quantum noise and streak artifacts. In order to improve the cardiac LDCT image quality, a 3D sparse representation-based processing (3D SR) is proposed by exploiting the sparsity and regularity of 3D anatomical features in CCT. The proposed method was evaluated by a clinical study of 14 patients. The performance of the proposed method was compared to the 2D spares representation-based processing (2D SR) and the state-of-the-art noise reduction algorithm BM4D. The visual assessment, quantitative assessment and qualitative assessment results show that the proposed approach can lead to effective noise/artifact suppression and detail preservation. Compared to the other two tested methods, 3D SR method can obtain results with image quality most close to the reference standard dose CT (SDCT) images.

CT pulmonary angiogram with 60% dose reduction: Influence of iterative reconstructions on image quality.

To compare the quality of low-dose CT images with sinogram affirmed iterative reconstruction (SAFIRE), and full-dose CT with filtered back projection reconstructions (FBP). Fifty pulmonary CT performed by a dual-source technique (120kVp; 110mAs) with (a) the same energy in both tubes, and (b) the distribution of reference mAs with 40% in tube A (44mAs) and 60% in tube B (66mAs). Each acquisition allowed reconstruction of: (a) full-dose images (with both tubes) with FBP reconstructions (group 1); and (b) low-dose images (from tube A) reconstructed with SAFIRE (group 2). Group 2 images presented: (a) a significant objective reduction in noise measured in the trachea on mediastinal (16.04±5.66 vs 17.66±5.84) (P=0.0284) and pulmonary (29.77±6.79 vs 37.96±9.03) (P<0.0001) images; (b) a similar subjective perception of noise and overall image quality (P=1), which was considered to be excellent in 66% (33/50) of the cases, with no influence on the detection of elementary pulmonary lesions of infiltration (98.4%; 95% CI=[96.9%-99.9%]). Despite a 60% reduction in radiation dose, the image quality with iterative reconstruction is objectively better and subjectively similar to full-dose FBP images.