CT Noise Research Articles

Comprehensive Approach to TB Diagnosis Powered by Machine Learning in Endemic Countries

Abstract Introduction/Objective Tuberculosis (TB), a leading cause of mortality worldwide, is primarily diagnosed through methods that face numerous challenges, including lengthy result times and diagnostic inaccuracy, particularly in the context of drug-resistant strains and latent infections. This study proposes a novel approach that integrates multiple diagnostic modalities with machine learning to enhance TB diagnosis accuracy. Methods/Case Report The Machine learning model was tailored for the classification of radiological images, specifically CT scans, using original DICOM images to preserve valuable voxel information. The study involved a curated dataset of CT scans from Pakistan, TB-positive (smear+/Culture+ n=72; smear+/culture- n= 3; smear-/Culture+ n=16; smear-/Culture- n=14; clinical and radiological diagnosis only n= 12) and healthy (n=103) scans from COVID-CTSet were included in the training to assess the specificity of the model. Data preprocessing included noise reduction, histogram matching, and data augmentation techniques to ensure a robust training dataset. The methodology also incorporated using k-fold cross validation and a custom learning rate scheduling strategy to optimize the model’s training process. To further test its robustness, the model was evaluated by swapping random patches and analyzing 60 mm patches from the entire CT scan for local feature categorization. Results (if a Case Study enter NA) The model has demonstrated outstanding performance, with sensitivity and specificity exceeding 95% on our internal dataset (n=117) and aligning well with clinical, smear, culture, and anti-TB antibody findings. It also showed promising results on the external NIH TB Portal’s dataset (n= 1,460). The model is robust to variations in CT scanner image quality and noise profiles, maintaining strong generalization across datasets. Even when analyzing random 60mm patches, effectively classifying all local features Conclusion Our study enhances tuberculosis diagnostic methods in endemic regions by blending machine learning with conventional diagnostic techniques. This approach improves TB detection, particularly the early and accurate identification in areas impacted by drug-resistant strains and asymptomatic latent infections. By integrating radiological and clinico-pathological data, we aim to develop a comprehensive diagnostic score. Moving forward, we plan to further integrate this method into clinical decision support systems and adapt the approach to other infectious diseases.

Head and neck tumor segmentation from [18F]F-FDG PET/CT images based on 3D diffusion model

Objective. Head and neck (H&N) cancers are among the most prevalent types of cancer worldwide, and [18F]F-FDG PET/CT is widely used for H&N cancer management. Recently, the diffusion model has demonstrated remarkable performance in various image-generation tasks. In this work, we proposed a 3D diffusion model to accurately perform H&N tumor segmentation from 3D PET and CT volumes. Approach. The 3D diffusion model was developed considering the 3D nature of PET and CT images acquired. During the reverse process, the model utilized a 3D UNet structure and took the concatenation of 3D PET, CT, and Gaussian noise volumes as the network input to generate the tumor mask. Experiments based on the HECKTOR challenge dataset were conducted to evaluate the effectiveness of the proposed diffusion model. Several state-of-the-art techniques based on U-Net and Transformer structures were adopted as the reference methods. Benefits of employing both PET and CT as the network input, as well as further extending the diffusion model from 2D to 3D, were investigated based on various quantitative metrics and qualitative results. Main results. Results showed that the proposed 3D diffusion model could generate more accurate segmentation results compared with other methods (mean Dice of 0.739 compared to less than 0.726 for other methods). Compared to the diffusion model in 2D form, the proposed 3D model yielded superior results (mean Dice of 0.739 compared to 0.669). Our experiments also highlighted the advantage of utilizing dual-modality PET and CT data over only single-modality data for H&N tumor segmentation (with mean Dice less than 0.570). Significance. This work demonstrated the effectiveness of the proposed 3D diffusion model in generating more accurate H&N tumor segmentation masks compared to the other reference methods.

Perceptual thresholds for differences in CT noise texture.

The average () or peak () noise power spectrum (NPS) frequency is often used as a one-parameter descriptor of the CT noise texture. Our study develops a more complete two-parameter model of the CT NPS and investigates the sensitivity of human observers to changes in it. A model of CT NPS was created based on its and a half-Gaussian fit () to the downslope. Two-alternative forced-choice staircase studies were used to determine perceptual thresholds for noise texture, defined as parameter differences with a predetermined level of discrimination performance (80% correct). Five imaging scientist observers performed the forced-choice studies for eight directions in the -space, for two reference NPSs (corresponding to body and lung kernels). The experiment was repeated with 32 radiologists, each evaluating a single direction in the -space. NPS differences were quantified by the noise texture contrast (), the integral of the absolute NPS difference. The two-parameter NPS model was found to be a good representation of various clinical CT reconstructions. Perception thresholds for alone are for body and for lung NPSs. For , these values are 0.15 and , respectively. Thresholds change if the other parameter also changes. Different NPSs with the same or can be discriminated. Nonradiologist observers did not need more than radiologists. or is insufficient to describe noise texture completely. The discrimination of noise texture changes depending on its frequency content. Radiologists do not discriminate noise texture changes better than nonradiologists.

Learned high-resolution cardiac CT imaging from ultra-high-resolution PCD-CT.

Coronary computed tomography angiography (cCTA) is a widely used non-invasive diagnostic exam for patients with coronary artery disease (CAD). However, most clinical CT scanners are limited in spatial resolution from use of energy-integrating detectors (EIDs). Radiological evaluation of CAD is challenging, as coronary arteries are small (3-4 mm diameter) and calcifications within them are highly attenuating, leading to blooming artifacts. As such, this is a task well suited for high spatial resolution. Recently, photon-counting-detector (PCD) CT became commercially available, allowing for ultra-high resolution (UHR) data acquisition. However, PCD-CTs are costly, restricting widespread accessibility. To address this problem, we propose a super resolution convolutional neural network (CNN): ILUMENATE (Improved LUMEN visualization through Artificial super-resoluTion imagEs), creating a high resolution (HR) image simulating UHR PCD-CT. The network was trained and validated using patches extracted from 8 patients with a modified U-Net architecture. Training input and labels consisted of UHR PCD-CT images reconstructed with a smooth kernel degrading resolution (LR input) and sharp kernel (HR label). The network learned the resolution difference and was tested on 5 unseen LR patients. We evaluated network performance quantitatively and qualitatively through visual inspection, line profiles to assess spatial resolution improvements, ROIs for CT number stability and noise assessment, structural similarity index (SSIM), and percent diameter luminal stenosis. Overall, ILUMENATE improved images quantitatively and qualitatively, creating sharper edges more closely resembling reconstructed HR reference images, maintained stable CT numbers with less than 4% difference, reduced noise by 28%, maintained structural similarity (average SSIM = 0.70), and reduced percent diameter stenosis with respect to input images. ILUMENATE demonstrates potential impact for CAD patient management, improving the quality of LR CT images bringing them closer to UHR PCD-CT images.

Evaluation of data uncertainty for deep-learning-based CT noise reduction using ensemble patient data and a virtual imaging trial framework.

Deep learning-based image reconstruction and noise reduction (DLIR) methods have been increasingly deployed in clinical CT. Accurate assessment of their data uncertainty properties is essential to understand the stability of DLIR in response to noise. In this work, we aim to evaluate the data uncertainty of a DLIR method using real patient data and a virtual imaging trial framework and compare it with filtered-backprojection (FBP) and iterative reconstruction (IR). The ensemble of noise realizations was generated by using a realistic projection domain noise insertion technique. The impact of varying dose levels and denoising strengths were investigated for a ResNet-based deep convolutional neural network (DCNN) model trained using patient images. On the uncertainty maps, DCNN shows more detailed structures than IR although its bias map has less structural dependency, which implies that DCNN is more sensitive to small changes in the input. Both visual examples and histogram analysis demonstrated that hotspots of uncertainty in DCNN may be associated with a higher chance of distortion from the truth than IR, but it may also correspond to a better detection performance for some of the small structures.

Image quality evaluation in deep-learning-based CT noise reduction using virtual imaging trial methods: Contrast-dependent spatial resolution.

Deep-learning-based image reconstruction and noise reduction methods (DLIR) have been increasingly deployed in clinical CT. Accurate image quality assessment of these methods is challenging as the performance measured using physical phantoms may not represent the true performance of DLIR in patients since DLIR is trained mostly on patient images. In this work, we aim to develop a patient-data-based virtual imaging trial framework and, as a first application, use it to measure the spatial resolution properties of a DLIR method. The patient-data-based virtual imaging trial framework consists of five steps: (1) insertion of lesions into projection domain data using the acquisition geometry of the patient exam to simulate different lesion characteristics; (2) insertion of noise into projection domain data using a realistic photon statistical model of the CT system to simulate different dose levels; (3) creation of DLIR-processed images from projection or image data; (4) creation of ensembles of DLIR-processed patient images from a large number of noise and lesion realizations; and (5) evaluation of image quality using ensemble DLIR images. This framework was applied to measure the spatial resolution of a ResNet based deep convolutional neural network (DCNN) trained on patient images. Lesions in a cylindrical shape and different contrast levels (-500, -100, -50, -20, -10 HU) were inserted to the lower right lobe of the liver in a patient case. Multiple dose levels were simulated (50%, 25%, 12.5%). Each lesion and dose condition had 600 noise realizations. Multiple reconstruction and denoising methods were used on all the noise realizations, including the original filtered-backprojection (FBP), iterative reconstruction (IR), and the DCNN method with three different strength setting (DCNN-weak, DCNN-medium, and DCNN-strong). Mean lesion signal was calculated by performing ensemble averaging of all the noise realizations for each lesion and dose condition and then subtracting the lesion-present images from the lesion absent images. Modulation transfer functions (MTFs) both in-plane and along the z-axis were calculated based on the mean lesion signals. The standard deviations of MTFs at each condition were estimated with bootstrapping: randomly sampling (with replacement) all the DLIR/FBP/IR images from the ensemble data (600 samples) at each condition. The impact of varying lesion contrast, dose levels, and denoising strengths were evaluated. Statistical analysis with paired t-test was used to compare the z-axis and in-plane spatial resolution of five algorithms for five different contrasts and three dose levels. The in-plane and z-axis spatial resolution degradation of DCNN becomes more severe as the contrast or radiation dose decreased, or DCNN denoising strength increased. In comparison with FBP, a 59.5% and 4.1% reduction of in-plane and z-axis MTF (in terms of spatial frequencies at 50% MTF), respectively, was observed at low contrast (-10 HU) for DCNN with the highest denoising strength at 25% routine dose level. When the dose level reduces from 50% to 12.5% of routine dose, the in-plane and z-axis MTFs reduces from 92.1% to 76.3%, and from 98.9% to 95.5%, respectively, at contrast of -100 HU, using FBP as the reference. For most conditions of contrasts and dose levels, significant differences were found among the five algorithms, with the following relationship in both in-plane and cross-plane spatial resolution: FBP>DCNN-Weak>IR>DCNN-Medium>DCNN-Strong. The spatial resolution difference among algorithms decreases at higher contrast or dose levels. A patient-data-based virtual imaging trial framework was developed and applied to measuring the spatial resolution properties of a DCNN noise reduction method at different contrast and dose levels using real patient data. As with other non-linear image reconstruction and post-processing techniques, the evaluated DCNN method degraded the in-plane and z-axis spatial resolution at lower contrast levels, lower radiation dose, and higher denoising strength.

Improving the detection of hypo-vascular liver metastases in multiphase contrast-enhanced CT with slice thickness less than 5 mm using DenseNet

IntroductionThinner slices are more susceptible in detecting small lesions but suffer from higher statistical fluctuation. This work aimed to reduce image noise in multiphase contrast-enhanced CT reconstructed with slice thickness thinner than the clinical setting (i.e., 5 mm) using convolutional neural network (CNN) for enabling better detection of hypo-vascular liver metastasis. MethodsA DenseNet model was used to generate noise map for multiphase CT reconstructed with slice thickness of 2.5 mm and 1.25 mm. Image denoising was conducted by subtracting the CNN-generated noise map from CT images with reduced photon flux due to thinner slice thickness. The performance of DenseNet was evaluated on CT scans of electron density phantoms and patients with hypovascular liver metastases less than 1.5 cm in terms of Hounsfield Unit (HU) variation, statistical fluctuation, and contrast-to-noise ratio (CNR). ResultsThe phantom study demonstrated that the CNN-based denoising method was able to reduce statistical fluctuation in CT images reconstructed with slice thickness of 2.5 mm and 1.25 mm without causing significant edge blurring or variation in HU values. With regards to patient study, it was found that the denoised 2.5-mm and 1.25-mm slices had higher CNR than the conventional 5-mm slices for hypo-vascular liver metastases in all 4 phases of multiphase CT. ConclusionOur results demonstrated that the detection of hypo-vascular liver metastases in multiphase contrast-enhanced CT with slice thickness less than 5 mm could be improved by using the CNN-based denoising method. Implications for practiceReconstruction slice thickness has a strong influence on the image quality of CT imaging. A CNN-based denoising method was used in this work to reduce the image noise in multiphase contrast-enhanced CT reconstructed with slice thickness less than 5 mm.

Enhancing Soft Tissue Differentiation with Different Dual-Energy CT Systems: A Phantom Study

To quantitatively evaluate the possible advantages of quantifying and differentiating various soft tissues using virtual monochromatic images (VMI) derived from different dual-energy computed tomography (DECT) technologies. This study involved four DECT scanners with different technologies. CIRS phantom images were acquired in single-energy (SECT) and DECT modes with each scanner. The analysis focused on five equivalent soft-tissue inserts: adipose, breast, liver, muscle, and bone (200 mg). The signal-to-noise ratio (SNR) was calculated for each equivalent soft-tissue insert. Finally, the contrasts of tissue pairs between DECT and SECT images were compared using Wilcoxon signed-rank tests adjusted for multiple comparisons. Average CT numbers and noise showed a significant difference pattern between DECT with respect to SECT for each CT scanner. Generally, energy levels of 70 keV or higher led to improved SNR in VMI for most of the equivalent soft-tissue inserts. However, energy levels of 40–50 keV showed significantly higher contrasts in most of the equivalent soft-tissue insert pairs. DECT images at low energies, especially at 40–50 keV, outperform SECT images in discriminating soft tissues across all four DECT technologies. The combined use of DECT images reconstructed at different energy levels provides a more comprehensive set of information for diagnostic and/or radiotherapy evaluation compared to SECT. Some differences between scanners are evident, depending on the DECT acquisition technique and reconstruction method.

CT image restoration method via total variation and L 0 smoothing filter

Abstract In X-ray CT imaging, there are some cases where the obtained CT images have serious ring artifacts and noise, and these degraded CT images seriously affect the quality of clinical diagnosis. Thus, developing an effective method that can simultaneously suppress ring artifacts and noise is of great importance. Total variation (TV) is a famous prior regularization for image denoising in the image processing field, however, for degraded CT images, it can suppress the noise but fail to reduce the ring artifacts. To address this issue, the L 0 L_{0} smoothing filter is incorporated with TV prior for CT ring artifacts and noise removal problem where the problem is transformed into several optimization sub-problems which are iteratively solved. The experiments demonstrate that the ring artifacts and noise presented in the CT image can be effectively suppressed by the proposed method and meanwhile the detailed features such as edge structure can be well preserved. As the superiority of TV and L 0 L_{0} smoothing filters are fully utilized, the performance of the proposed method is better than the existing methods such as the TV-based method and L 0 L_{0} -based method.

Value of four-dimensional computed tomography angiography combined with stromal cell-derived factor-1 for differentiating ruptured intracranial aneurysms and assessing risk of ruptures

Abstract Background: To analyze the value of four-dimensional computed tomography angiography (4D-CTA) combined with stromal cell-derived factor-1 (SDF-1) for differentiating ruptured intracranial aneurysms and assessing the risk of ruptures. Methods: Fifty patients with unruptured intracranial aneurysms and 50 patients with ruptured ones were included in non-rupture group 1 and rupture group 1, respectively. All patients underwent 4D-CTA and 3D-CTA, and the serum SDF-1 level was detected. Non-rupture group 1 was followed up for 12 months. On this basis, the patients with ruptured aneurysms were included in rupture group 2 and those without ruptured aneurysms were assigned to non-rupture group 2. Results: The AUC values of Wn, AR, L, SR, SDF-1 and their combination for the diagnosis of ruptured intracranial aneurysms were all >0.70, especially their combination. Wn, AR, L, and SR were higher in rupture group 2 than those in non-rupture group 2 (P<0.05). The level of SDF-1 in rupture group 2 [(142.48±11.23) μg/L] was higher than that in non-rupture group 2 [(128.03±10.28) μg/L] (P<0.05). Wn, AR, L, SR, and SDF-1 and their combination all had the AUC values of >0.70 for the prediction of ruptured intracranial aneurysms, especially their combination. The CT values of the internal carotid artery and middle cerebral artery and the CT value and noise of brain parenchyma in 4D-CTA were higher than those in 3D-CTA (P<0.05). Conclusions: 4D-CTA combined with SDF-1 can effectively differentiate ruptured intracranial aneurysms and predict the risk of ruptures.

Invisible Shield: Unveiling an Efficient Watermarking Solution for Medical Imaging Security

Securing medical imaging poses a significant challenge in preserving the confidentiality of healthcare data. Numerous research efforts have focused on fortifying these images, with encryption emerging as a primary solution for maintaining data integrity without compromising confidentiality. However, applying conventional encryption techniques directly to e-health data encounters hurdles, including limitations in data size, redundancy, and capacity, particularly in open-channel patient data transmissions. As a result, the unique characteristics of images, marked by their risk of data loss and the need for confidentiality, make preserving the privacy of data contents a complex task. This underscores the pressing need for innovative approaches to ensure the security and confidentiality of sensitive healthcare information within medical images. The proposed algorithm outperforms referenced algorithms in both image fidelity and steganographic capacity across diverse medical imaging modalities. It consistently achieves higher Peak Signal-to-Noise Ratio (PSNR) values, indicating superior image fidelity, reduced noise, and preserved signal quality in CT, MRI, ultrasound, and X-ray modalities. The experimental results demonstrate a considerable improvement in both the Peak Signal-to-Noise Ratio (PSNR) and maximum embedding capacity. Specifically, the average PSNR value for the X-ray modality reached a notable 73 dB, signifying superior image quality. Moreover, the CT modality exhibited the highest maximum embedding capacity, measured at 0.52, showcasing its efficiency in accommodating data within the images. Moreover, the algorithm consistently offers increased steganographic data hiding capacity in these images without perceptibly degrading their quality or integrity.

A simplified low-cost phantom for image quality assessment of dental cone beam computed tomography unit.

A standardised testing protocol for evaluation of a wide range of dental cone beam computed tomography (CBCT) performance and image quality (IQ) parameters is still limited and commercially available testing tool is unaffordable by some centres. This study aims to assess the performance of a low-cost fabricated phantom for image quality assessment (IQA) of digital CBCT unit. A customised polymethyl methacrylate (PMMA) cylindrical phantom was developed for performance evaluation of Planmeca ProMax 3D Mid digital dental CBCT unit. The fabricated phantom consists of four different layers for testing specific IQ parameters such as CT number accuracy and uniformity, noise and CT number linearity. The phantom was scanned using common scanning protocols in clinical routine (90.0 kV, 8.0 mA and 13.6 s). In region-of-interest (ROI) analysis, the mean CT numbers (in Hounsfield unit, HU) and noise for water and air were determined and compared with the reference values (0 HU for water and -1000 HU for air). For linearity test, the correlation between the measured HU of different inserts with their density was studied. The average CT number were -994.1 HU and -2.4 HU, for air and water, respectively and the differences were within the recommended acceptable limit. The linearity test showed a strong positive correlation (R2 = 0.9693) between the measured HU and their densities. The fabricated IQ phantom serves as a simple and affordable testing tool for digital dental CBCT imaging.

Exploration of the pulse pileup effects in a clinical CdTe-based photon-counting computed tomography.

High tube current generates a high flux of x-rays to photon counting detectors (PCDs) that can potentially result in the piling up of pulses formed by concurrent photons, which can cause count loss and energy resolution degradation. To evaluate the performance of clinical photon-counting CT (PCCT) systems in high flux, potentially influenced by pulse pileup effects, in terms of task-generic image quality metrics. A clinical phantom was scanned on a commercial PCCT scanner (NAEOTOM Alpha, Siemens) at 120kV under fourteen different tube current levels (40-1000mA) with a rotation time of 0.25 s and a pitch of 1. The dose levels corresponded to CTDIvol (32cm phantom) of 0.79-19.8mGy. CT sinograms were reconstructed using QIR-off mode (noniterative reconstruction algorithm), Br44 kernel, and a voxel size of . imQuest, an open-source MATLAB-based software package was used to calculate noise power spectrum (NPS), task transfer function (TTF), contrast-to-noise ratio (CNR), and CT number according to AAPM Task Group 233 metrology. The 50% cut-off frequency of TTF (f50 ) remained mostly constant across all higher tube currents for all inserts, namely polyethylene, bone, air, and acrylic. Using the lowest two data points (40 and 80mA), the expected relationship between noise magnitude and tube current was determined to be noise mA-0.47 . The measured noise magnitude were up to 11.1% higher than the expected value at the highest tube current. The average frequency of NPS (fav ) decreased from 0.32 to 0.29mm-1 as tube current increased from 40 to 1000mA. No considerable effects were observed in CT number measurement of any insert; however, CT numbers for air and bone changed almost monotonically as tube current increased. Absolute CNR increased monotonically for all inserts; however, the difference between measured and expected CNRs were approximately -6% to 12% across all tube currents. Increasing tube currents did not affect the spatial resolution, but slightly affected the CT number and noise measurements of the clinical PCCT system. However, the effects were only considerable at clinically irrelevant tube currents used on a small 20-cm phantom. In general clinical practices, automatic exposure control techniques are used to decrease the variation of flux on the detector, which alleviates the chances of detector saturation due to high count rates. The observed effects could be due to pulse pileup, signal-dependent filtration of the system, or nonlinearities in the reconstruction algorithm. In conclusion, either the deadtime of the detector used in the photon-counting CT system is shorter such that count losses due to pulse pileup are negligible, or pulse pileup has inconsiderable effects on the image quality of clinical photon-counting CT systems in routine clinical practice due to possible corrections applied on the system.

A new method for fault identification of T-connection transmission line based on multi-scale traveling wave reactive power and random forest.

Though the traditional fault diagnosis method of T-connected transmission lines can identify the faults inside and outside the area, it can not identify the specific branches. To improve the accuracy and reliability of fault diagnosis of T-connection transmission lines, a new method is proposed to identify specific faulty branches of T-connection transmission lines based on multi-scale traveling wave reactive power and random forest. Based on the S-transform, the mean and sum ratios of the corresponding short-time series traveling wave reactive powers of each two traveling wave protection units at multiple frequencies are calculated respectively to form the fault feature vector sample set of the T-connection transmission line. A random forest fault branch identification model is established, and it is trained and tested by the fault feature sample set of T-connection transmission line to identify the fault branch. The simulation results show that the proposed algorithm can identify the branch where the fault is located inside and outside the protection zone of T-connection transmission line quickly and accurately under various working conditions. This method also shows good performance to identify faults even under the situation of CT saturation, noise influence and data loss.

Comprehensive evaluations of a prototype full field-of-view photon counting CT system through phantom studies.

Objective. Photon counting CT (PCCT) has been a research focus in the last two decades. Recent studies and advancements have demonstrated that systems using semiconductor-based photon counting detectors (PCDs) have the potential to provide better contrast, noise and spatial resolution performance compared to conventional scintillator-based systems. With multi-energy threshold detection, PCD can simultaneously provide the photon energy measurement and enable material decomposition for spectral imaging. In this work, we report a performance evaluation of our first CdZnTe-based prototype full-size PCCT system through various phantom imaging studies.Approach.This prototype system supports a 500 mm scan field-of-view and 10 mmz-coverage at isocenter. Phantom scans were acquired using 120 kVp from 50 to 400 mAs to assess the imaging performance on: CT number accuracy, uniformity, noise, spatial resolution, material differentiation and quantification.Main results.Both qualitative and quantitative evaluations show that PCCT, under the tested conditions, has superior imaging performance with lower noise and improved spatial resolution compared to conventional energy integrating detector (EID)-CT. Using projection domain material decomposition approach with multiple energy bin measurements, PCCT virtual monoenergetic images have lower noise, and good accuracy in quantifying iodine and calcium concentrations. These results lead to increased contrast-to-noise ratio (CNR) for both high and low contrast study objects compared to EID-CT at matched dose and spatial resolution. PCCT can also generate super-high resolution images using much smaller detector pixel size than EID-CT and greatly improve image spatial resolution.Significance.Improved spatial resolution and quantification accuracy with reduced image noise of the PCCT images can potentially lead to better diagnosis at reduced radiation dose compared to conventional EID-CT. Increased CNR achieved by PCCT suggests potential reduction in iodine contrast media load, resulting in better patient safety and reduced cost.

Dual-domain fusion deep convolutional neural network for low-dose CT denoising.

In view of the underlying health risks posed by X-ray radiation, the main goal of the present research is to achieve high-quality CT images at the same time as reducing x-ray radiation. In recent years, convolutional neural network (CNN) has shown excellent performance in removing low-dose CT noise. However, previous work mainly focused on deepening and feature extraction work on CNN without considering fusion of features from frequency domain and image domain. To address this issue, we propose to develop and test a new LDCT image denoising method based on a dual-domain fusion deep convolutional neural network (DFCNN). This method deals with two domains, namely, the DCT domain and the image domain. In the DCT domain, we design a new residual CBAM network to enhance the internal and external relations of different channels while reducing noise to promote richer image structure information. For the image domain, we propose a top-down multi-scale codec network as a denoising network to obtain more acceptable edges and textures while obtaining multi-scale information. Then, the feature images of the two domains are fused by a combination network. The proposed method was validated on the Mayo dataset and the Piglet dataset. The denoising algorithm is optimal in both subjective and objective evaluation indexes as compared to other state-of-the-art methods reported in previous studies. The study results demonstrate that by applying the new fusion model denoising, denoising results in both image domain and DCT domain are better than other models developed using features extracted in the single image domain.

The efficacy of low-dose CT with deep learning image reconstruction in the surveillance of incidentally detected pancreatic cystic lesions.

To evaluate the efficacy of low-dose CT (LDCT) with deep learning image reconstruction (DLIR) for the surveillance of pancreatic cystic lesions (PCLs) compared with standard-dose CT (SDCT) with adaptive statistical iterative reconstruction (ASIR-V). The study enrolled 103 patients who underwent pancreatic CT for follow-up of incidentally detected PCLs. The CT protocol included LDCT in the pancreatic phase with 40% ASIR-V, DLIR at medium (DLIR-M) and high levels (DLIR-H), and SDCT in the portal-venous phase with 40% ASIR-V. The overall image quality and conspicuity of PCLs were qualitatively assessed using five-point scales by two radiologists. The size of PCLs, presence of thickened/enhancing walls, enhancing mural nodules, and main pancreatic duct dilatation were reviewed. CT noise and cyst-to-pancreas contrast-to-noise ratio (CNR) were measured. Qualitative and quantitative parameters were analyzed using the chi-squared test, one-way ANOVA, and t-test. Additionally, interobserver agreement was analyzed using the kappa and weighted-kappa statistics. The volume CT dose-indexes in LDCT and SDCT were 3.0 ± 0.6mGy and 8.4 ± 2.9mGy, respectively. LDCT with DLIR-H showed the highest overall image quality, the lowest noise, and the highest CNR. The PCL conspicuity in LDCT with either DLIR-M or DLIR-H was not significantly different from that in SDCT with ASIR-V. Other findings depicting PCLs also revealed no significant differences between LDCT with DLIR and SDCT with ASIR-V. Moreover, the results revealed good or excellent interobserver agreement. LDCT with DLIR has a comparable performance with SDCT for the follow-up of incidentally detected PCLs.

Dual-scale similarity-guided cycle generative adversarial network for unsupervised low-dose CT denoising

Removing the noise in low-dose CT (LDCT) is crucial to improving the diagnostic quality. Previously, many supervised or unsupervised deep learning-based LDCT denoising algorithms have been proposed. Unsupervised LDCT denoising algorithms are more practical than supervised ones since they do not need paired samples. However, unsupervised LDCT denoising algorithms are rarely used clinically due to their unsatisfactory denoising ability. In unsupervised LDCT denoising, the lack of paired samples makes the direction of gradient descent full of uncertainty. On the contrary, paired samples used in supervised denoising allow the parameters of networks to have a clear direction of gradient descent. To bridge the gap in performance between unsupervised and supervised LDCT denoising, we propose dual-scale similarity-guided cycle generative adversarial network (DSC-GAN). DSC-GAN uses similarity-based pseudo-pairing to better accomplish unsupervised LDCT denoising. We design a Vision Transformer-based global similarity descriptor and a residual neural network-based local similarity descriptor for DSC-GAN to effectively describe the similarity between two samples. During training, pseudo-pairs, i.e., similar LDCT samples and normal-dose CT (NDCT) samples, dominate parameter updates. Thus, the training can achieve equivalent effect as training with paired samples. Experiments on two datasets demonstrate that DSC-GAN beats the state-of-the-art unsupervised algorithms and reaches a level close to supervised LDCT denoising algorithms.

Quantitative CT Evaluation of Emphysema Progression over 10 Years in the COPDGene Study.

Background Long-term studies of chronic obstructive pulmonary disease (COPD) can evaluate emphysema progression. Adjustment for differences in equipment and scanning protocols of individual CT examinations have not been studied extensively. Purpose To evaluate emphysema progression in current and former smokers in the COPDGene cohort over three imaging points obtained at 5-year intervals accounting for individual CT parameters. Materials and Methods Current and former cigarette smokers enrolled between 2008 and 2011 from the COPDGene study were prospectively followed for 10 years between 2008 and 2020. Extent of emphysema as adjusted lung density (ALD) from quantitative CT was measured at baseline and at 5- and 10-year follow-up. Linear mixed models adjusted for CT technical characteristics were constructed to evaluate emphysema progression. Mean annual changes in ALD over consecutive 5-year study periods were estimated by smoking status and baseline emphysema. Results Of 8431 participants at baseline (mean age, 60 years ± 9 [SD]; 3905 female participants), 4913 were at 5-year follow-up and 1544 participants were at 10-year follow-up. There were 4134 (49%) participants who were current smokers, and 4449 (53%) participants had more than trace emphysema at baseline. Current smokers with more than trace emphysema showed the largest decline in ALD, with mean annual decreases of 1.4 g/L (95% CI: 1.2, 1.5) in the first 5 years and 0.9 g/L (95% CI: 0.7, 1.2) in the second 5 years. Accounting for CT noise, field of view, and scanner model improved model fit for estimation of emphysema progression (P < .001 by likelihood ratio test). Conclusion Evaluation at CT of emphysema progression in the COPDGene study showed that, during the span of 10 years, participants with pre-existing emphysema who continued smoking had the largest decline in ALD. Adjusting for CT equipment and protocol factors improved these longitudinal estimates. Clinical trial registration no. NCT00608764 © RSNA, 2023 Supplemental material is available for this article. See the editorial by Parraga and Kirby in this issue.

Dose reduction and image enhancement in micro-CT using deep learning.

In preclinical settings, micro-computed tomography (CT) provides a powerful tool to acquire high resolution anatomical images of rodents and offers the advantage to in vivo non-invasively assess disease progression and therapy efficacy. Much higher resolutions are needed to achieve scale-equivalent discriminatory capabilities in rodents as those in humans. High resolution imaging however comes at the expense of increased scan times and higher doses. Specifically, with preclinical longitudinal imaging, there are concerns that dose accumulation may affect experimental outcomes of animal models. Dose reduction efforts under the ALARA (as low as reasonably achievable) principles are thus a key point of attention. However, low dose CT acquisitions inherently induce higher noise levels which deteriorate image quality and negatively impact diagnostic performance. Many denoising techniques already exist, and deep learning (DL) has become increasingly popular for image denoising, but research has mostly focused on clinical CT with limited studies conducted on preclinical CT imaging. We investigate the potential of convolutional neural networks (CNN) for restoring high quality micro-CT images from low dose (noisy) images. The novelty of the CNN denoising frameworks presented in this work consists of utilizingimage pairs with realistic CT noise present in the input as well as thetarget image used for the model training; a noisier image acquired with a low dose protocol is matched to a less noisy image acquired with a higher dose scan of the same mouse. Low and high dose ex vivo micro-CT scans of 38 mice were acquired. Two CNN models, based on a 2D and 3D four-layer U-Net, were trained with mean absolute error (30 training, 4 validation and 4 test sets). To assess denoising performance, ex vivo mice and phantom data were used. Both CNN approaches were compared to existing methods, like spatial filtering (Gaussian, Median, Wiener) and iterative total variation image reconstruction algorithm. Image quality metrics were derived from the phantom images. A first observer study (n=23) was set-up to rank overall quality of differently denoised images. A second observer study (n=18) estimated the dose reduction factor of the investigated 2D CNN method. Visual and quantitative results show that both CNN algorithms exhibit superior performance in terms of noise suppression, structural preservation and contrast enhancement over comparator methods. The quality scoring by 23 medical imaging experts also indicates that the investigated 2D CNN approach is consistently evaluated as the best performing denoising method. Results from the second observer study and quantitative measurements suggest that CNN-based denoising could offer a 2-4× dose reduction, with an estimated dose reduction factor of about 3.2 for the considered 2D network. Our results demonstrate the potential of DL in micro-CT for higher quality imaging at low dose acquisition settings. In the context of preclinical research, this offers promising future prospects for managing the cumulative severity effects of radiation in longitudinal studies.