Significant Reduction Of Artifacts Research Articles

Correction of scatter in CBCT with a new grid design

Objective. This study aims at developing a simple and rapid Compton scatter correction approach for cone-beam CT (CBCT) imaging. Approach. In this work, a new Compton scatter estimation model is established based on two distinct CBCT scans: one measures the full primary and scatter signals without anti-scatter grid (ASG), and the other measures a portion of primary and scatter signals with ASG. To accelerate the entire data acquisition speed, a half anti-scatter grid (h-ASG) that covers half of the full detector surface is proposed. As a result, the distribution of scattered x-ray photons could be estimated from a single CBCT scan. Physical phantom experiments are conducted to validate the performance of the newly proposed scatter correction approach. Main results. Results demonstrate that the proposed half grid approach can quickly and precisely estimate the distribution of scattered x-ray photons from only one single CBCT scan, resulting in a significant reduction of shading artifacts. In addition, it is found that the h-ASG approach is less sensitive to the grid transmission fractions, grid ratio and object size, indicating a robust performance of the new method. Significance. In the future, the Compton scatter artifacts can be quickly corrected using a half grid in CBCT imaging.

Deep Learning Reconstruction of Prospectively Accelerated MRI of the Pancreas: Clinical Evaluation of Shortened Breath-Hold Examinations With Dixon Fat Suppression.

Deep learning (DL)-enabled magnetic resonance imaging (MRI) reconstructions can enable shortening of breath-hold examinations and improve image quality by reducing motion artifacts. Prospective studies with DL reconstructions of accelerated MRI of the upper abdomen in the context of pancreatic pathologies are lacking. In a clinical setting, the purpose of this study is to investigate the performance of a novel DL-based reconstruction algorithm in T1-weighted volumetric interpolated breath-hold examinations with partial Fourier sampling and Dixon fat suppression (hereafter, VIBE-DixonDL). The objective is to analyze its impact on acquisition time, image sharpness and quality, diagnostic confidence, pancreatic lesion conspicuity, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR). This prospective single-center study included participants with various pancreatic pathologies who gave written consent from January 2023 to September 2023. During the same session, each participant underwent 2 MRI acquisitions using a 1.5 T scanner: conventional precontrast and postcontrast T1-weighted VIBE acquisitions with Dixon fat suppression (VIBE-Dixon, reference standard) using 4-fold parallel imaging acceleration and 6-fold accelerated VIBE-Dixon acquisitions with partial Fourier sampling utilizing a novel DL reconstruction tailored to the acquisition. A qualitative image analysis was performed by 4 readers. Acquisition time, image sharpness, overall image quality, image noise and artifacts, diagnostic confidence, as well as pancreatic lesion conspicuity and size were compared. Furthermore, a quantitative analysis of SNR and CNR was performed. Thirty-two participants were evaluated (mean age ± SD, 62 ± 19 years; 20 men). The VIBE-DixonDL method enabled up to 52% reduction in average breath-hold time (7 seconds for VIBE-DixonDL vs 15 seconds for VIBE-Dixon, P < 0.001). A significant improvement of image sharpness, overall image quality, diagnostic confidence, and pancreatic lesion conspicuity was observed in the images recorded using VIBE-DixonDL (P < 0.001). Furthermore, a significant reduction of image noise and motion artifacts was noted in the images recorded using the VIBE-DixonDL technique (P < 0.001). In addition, for all readers, there was no evidence of a difference in lesion size measurement between VIBE-Dixon and VIBE-DixonDL. Interreader agreement between VIBE-Dixon and VIBE-DixonDL regarding lesion size was excellent (intraclass correlation coefficient, >90). Finally, a statistically significant increase of pancreatic SNR in VIBE-DIXONDL was observed in both the precontrast (P = 0.025) and postcontrast images (P < 0.001). Also, an increase of splenic SNR in VIBE-DIXONDL was observed in both the precontrast and postcontrast images, but only reaching statistical significance in the postcontrast images (P = 0.34 and P = 0.003, respectively). Similarly, an increase of pancreas CNR in VIBE-DIXONDL was observed in both the precontrast and postcontrast images, but only reaching statistical significance in the postcontrast images (P = 0.557 and P = 0.026, respectively). The prospectively accelerated, DL-enhanced VIBE with Dixon fat suppression was clinically feasible. It enabled a 52% reduction in breath-hold time and provided superior image quality, diagnostic confidence, and pancreatic lesion conspicuity. This technique might be especially useful for patients with limited breath-hold capacity.

Metal Artifact Correction in Industrial CT Images Based on a Dual-Domain Joint Deep Learning Framework

Industrial computed tomography (CT) images reconstructed directly from projection data using the filtered back projection (FBP) method exhibit strong metal artifacts due to factors such as beam hardening, scatter, statistical noise, and deficiencies in the reconstruction algorithms. Traditional correction approaches, confined to either the projection domain or the image domain, fail to fully utilize the rich information embedded in the data. To leverage information from both domains, we propose a joint deep learning framework that integrates UNet and ResNet architectures for the correction of metal artifacts in CT images. Initially, the UNet network is employed to correct the imperfect projection data (sinograms), the output of which serves as the input for the CT image reconstruction unit. Subsequently, the reconstructed CT images are fed into the ResNet, with both networks undergoing a joint training process to optimize image quality. We take the projection data obtained by analytical simulation as the data set. The resulting optimized industrial CT images show a significant reduction in metal artifacts, with the average Peak Signal-to-Noise Ratio (PSNR) reaching 36.13 and the average Structural Similarity Index (SSIM) achieving 0.953. By conducting simultaneous correction in both the projection and image domains, our method effectively harnesses the complementary information from both, exhibiting a marked improvement in correction results over the deep learning-based single-domain corrections. The generalization capability of our proposed method is further verified in ablation experiments and multi-material phantom CT artifact correction.

Estimate and compensate head motion in non-contrast head CT scans using partial angle reconstruction and deep learning.

Patient head motion is a common source of image artifacts in computed tomography (CT) of the head, leading to degraded image quality and potentially incorrect diagnoses. The partial angle reconstruction (PAR) means dividing the CT projection into several consecutive angular segments and reconstructing each segment individually. Although motion estimation and compensation using PAR has been developed and investigated in cardiac CT scans, its potential for reducing motion artifacts in head CT scans remains unexplored. To develop a deep learning (DL) model capable of directly estimating head motion from PAR images of head CT scans and to integrate the estimated motion into an iterative reconstruction process to compensate for the motion. Head motion is considered as a rigid transformation described by six time-variant variables, including the three variables for translation and three variables for rotation. Each motion variable is modeled using a B-spline defined by five control points (CP) along time. We split the full projections from 360° into 25 consecutive PARs and subsequently input them into a convolutional neural network (CNN) that outputs the estimated CPs for each motion variable. The estimated CPs are used to calculate the object motion in each projection, which are incorporated into the forward and backprojection of an iterative reconstruction algorithm to reconstruct the motion-compensated image. The performance of our DL model is evaluated through both simulation and phantom studies. The DL model achieved high accuracy in estimating head motion, as demonstrated in both the simulation study (mean absolute error (MAE) ranging from 0.28 to 0.45mm or degree across different motion variables) and the phantom study (MAE ranging from 0.40 to 0.48mm or degree). The resulting motion-corrected image, , exhibited a significant reduction in motion artifacts when compared to the traditional filtered back-projection reconstructions, which is evidenced both in the simulation study (image MAE drops from 178 33HU to 37 9HU, structural similarity index (SSIM) increases from 0.60 0.06 to 0.98 0.01) and the phantom study (image MAE drops from 117 17HU to 42 19HU, SSIM increases from 0.83 0.04 to 0.98 0.02). We demonstrate that using PAR and our proposed deep learning model enables accurate estimation of patient head motion and effectively reduces motion artifacts in the resulting head CT images.

Fast and accessible T2 mapping using off-resonance corrected DESPOT2 with application to 3D prostate

PurposeMost T1 and T2 mapping take long acquisitions or needs specialized sequences not widely accessible on clinical scanners. An available solution is DESPOT1/T2 (Driven equilibrium single pulse observation of T1/T2). DESPOT1/T2 uses Spoiled gradient-echo (SPGR) and balanced Steady-State Free Precession (bSSFP) sequences, offering an accessible and reliable way for 3D accelerated T1/T2 mapping. However, bSSFP is prone to off-resonance artifacts, limiting the application of DESPOT2 in regions with high susceptibility contrasts, like the prostate. Our proposal, DESPO+, employs the full bSSFP and SPGR models with a dictionary-based method to reconstruct 3D T1/T2 maps in the prostate region without off-resonance banding. MethodsDESPO+ modifies the bSSFP acquisition of the original variable flip angle DESPOT2. DESPO+ uses variable repetition and echo times, employing a dictionary-based method of the full bSSFP and SPGR models to reconstruct T1, T2, and Proton Density (PD) simultaneously. The proposed DESPO+ method underwent testing through simulations, T1/T2 phantoms, and on fourteen healthy subjects. ResultsThe results reveal a significant reduction in T2 map banding artifacts compared to the original DESPOT2 method. DESPO+ approach reduced T2 errors by up to seven times compared to DESPOT2 in simulations and phantom experiments. We also synthesized in-vivo T1-weighted/T2-weighted images from the acquired maps using a spin-echo model to verify the map's quality when lacking a reference. For in-vivo imaging, the synthesized images closely resemble those from the clinical MRI protocol, reducing scan time by around 50% compared to traditional spin-echo T1-weighted/T2-weighted acquisitions. ConclusionDESPO+ provides an off-resonance insensitive and clinically available solution, enabling high-resolution 3D T1/T2 mapping and synthesized T1-weighted/T2-weighted images for the entire prostate, all achieved within a short scan time of 3.6 min, similar to DESPOT1/T2.

Clinical Evaluation of a 2-Minute Ultrafast Brain MR Protocol for Evaluation of Acute Pathology in the Emergency and Inpatient Settings.

The use of MR imaging in emergency settings has been limited by availability, long scan times, and sensitivity to motion. This study assessed the diagnostic performance of an ultrafast brain MR imaging protocol for evaluation of acute intracranial pathology in the emergency department and inpatient settings. Sixty-six adult patients who underwent brain MR imaging in the emergency department and inpatient settings were included in the study. All patients underwent both the reference and the ultrafast brain MR protocols. Both brain MR imaging protocols consisted of T1-weighted, T2/T2*-weighted, FLAIR, and DWI sequences. The ultrafast MR images were reconstructed by using a machine-learning assisted framework. All images were reviewed by 2 blinded neuroradiologists. The average acquisition time was 2.1 minutes for the ultrafast brain MR protocol and 10 minutes for the reference brain MR protocol. There was 98.5% agreement on the main clinical diagnosis between the 2 protocols. In head-to-head comparison, the reference protocol was preferred in terms of image noise and geometric distortion (P < .05 for both). The ultrafast ms-EPI protocol was preferred over the reference protocol in terms of reduced motion artifacts (P < .01). Overall diagnostic quality was not significantly different between the 2 protocols (P > .05). The ultrafast brain MR imaging protocol provides high accuracy for evaluating acute pathology while only requiring a fraction of the scan time. Although there was greater image noise and geometric distortion on the ultrafast brain MR protocol images, there was significant reduction in motion artifacts with similar overall diagnostic quality between the 2 protocols.

A Novel Reconstruction Technique to Reduce Stair-Step Artifacts in Sequential Mode Coronary CT Angiography.

Prospective electrocardiography-triggering is one of the most commonly used cardiac computed tomography (CT) scan modes but can be susceptible to stair-step artifacts in the transition areas of an acquisition over multiple cardiac cycles. We evaluated a novel reconstruction algorithm to reduce the occurrence and severity of such artifacts in sequential coronary CT angiography. In this institutional review board-approved, retrospective study, 50 consecutive patients (16 females; mean age, 58.9 ± 15.2) were included who underwent coronary CT angiography on a dual-source photon-counting detector CT in the sequential ultra-high-resolution mode with a detector collimation of 120 × 0.2 mm. Each scan was reconstructed without (hereafter called standard reconstruction) and with the novel ZeeFree reconstruction algorithm, which aims to minimize stair-step artifacts. The presence and extent of stair-step artifacts were rated by 2 independent, blinded readers on a 4-point discrete visual scale. The relationship between the occurrences of artifacts was correlated with the average and variability of heart rate and with patient characteristics. A total of 504 coronary segments were included into the analyses. In standard reconstructions, reader 1 reported stair-step artifacts in 40/504 (7.9%) segments, from which 12/504 led to nondiagnostic image quality (2.4% of all segments). Reader 2 reported 56/504 (11.1%) stair-step artifacts, from which 11/504 lead to nondiagnostic image quality (2.2% of all segments). With the ZeeFree algorithm, 9/12 (75%) and 8/11 (73%) of the nondiagnostic segments improved to a diagnostic quality for readers 1 and 2, respectively. The ZeeFree reconstruction algorithm significantly reduced the frequency and extent of stair-step artifacts compared with standard reconstructions for both readers (P < 0.001, each). Heart rate variability and body mass index were significantly related to the occurrence of stair-step artifacts (P < 0.05). Our study demonstrates the feasibility and effectiveness of a novel reconstruction algorithm leading to a significant reduction of stair-step artifacts and, hence, a reduction of coronary segments with a nondiagnostic image quality in sequential ultra-high-resolution coronary photon-counting detector CT angiography.

Photon-Counting Detector CT: Clinical Utility of Virtual Monoenergetic Imaging Combined With Tin Prefiltration to Reduce Metal Artifacts in the Postoperative Ankle.

The aim of this study was to compare the effectiveness and clinical utility of virtual monoenergetic image (VMI) reconstructions in computed tomography (CT) scans with and without tin prefiltration on a photon-counting detector (PCD) CT system to reduce metal implant artifacts in the postoperative ankle. This retrospective study included patients with internal fixation of the ankle scanned with and without tin prefiltration (Sn) on a PCD CT scanner between March and October 2023. Virtual monoenergetic images between 60 and 190 keV were reconstructed with a 10-keV increment in a bone kernel for both acquisitions (VMISn and VMIStd, respectively). Noise measurements assessed artifact reduction in the most prominent near-metal image distortions and were compared between acquisitions modes as well as between polychromatic images and VMIs. Three readers assessed the visibility of osseous healing along with interpretability and artifact extent for 5 reconstruction levels. A total of 48 patients (21 females, 27 males; mean age, 55.1 ± 19.4 years) were included in this study. Tin-prefiltered acquisitions (n = 30) had a lower artifact level for polychromatic images and VMIs compared with non-tin-prefiltered acquisitions (n = 18; P ≤ 0.043). A significant reduction of metal artifacts was observed for VMISn ≥120 keV compared with polychromatic images (hyperdense artifacts: 40.2 HU [interquartile range (IQR) 39.8] vs 14.0 HU [IQR 11.1]; P ≤ 0.01 and hypodense artifacts: 91.2 HU [IQR 82.4] vs 29.7 HU [IQR 39.6]; P ≤ 0.001). For VMIStd, this applied to reconstructions ≥100 keV (hyperdense artifacts: 57.7 HU [IQR 33.4] vs 19.4 HU [IQR 27.6]; P ≤ 0.001 and hypodense artifacts: 106.9 HU [IQR 76.1] vs 57.4 HU [IQR 55.7]; P ≤ 0.021). For visibility of osseous healing, VMISn at 120 keV yielded higher ratings compared with polychromatic images (P ≤ 0.001), whereas image interpretability was rated better (P = 0.023), and artifact extent was rated lower (P ≤ 0.001) compared with polychromatic images. Tin-prefiltered VMI at 120 keV showed a significant reduction in metal artifacts compared with polychromatic images, whereas visibility of osseous healing and image interpretability was improved. Therefore, tin-prefiltration PCD CT with VMI reconstructions may be a helpful complement to postsurgical CT imaging of the ankle in patients with metal implants.

Convolutional network denoising for acceleration of multi-shot diffusion MRI

Multi-shot echo planar imaging is a promising technique to reduce geometric distortions and increase spatial resolution in diffusion-weighted MRI (DWI), at the expense of increased scan time. Moreover, performing DWI in the body requires multiple repetitions to obtain sufficient signal-to-noise ratio, which further increases the scan time. This work proposes to reduce the number of repetitions and perform denoising of high b-value images using a convolutional network denoising trained on single-shot DWI to accelerate the acquisition of multi-shot DWI. Convolutional network denoising is demonstrated to accelerate the acquisition of 2-shot DWI by a factor of 4 compared to the clinical standard on patients with rectal cancer. Image quality was evaluated using qualitative scores from expert body radiologists between accelerated and non-accelerated acquisition. Additionally, the effect of convolutional network denoising on each image quality score was analyzed using a Wilcoxon signed-rank test. Convolutional network denoising would enable to increase the number of shots without increasing scan time for significant geometric artifact reduction and spatial resolution increase.

Artifact Reduction Proton Density Magnetic Resonance Imaging Can Better Visualize Unicompartmental Knee Arthroplasty Components but Does Not Improve Measurement Accuracy at 3T: An In Vitro Phantom Study.

Background There are no studies of the efficacy of slice encoding for metal artifact correction (SEMAC) magnetic resonance imaging (MRI) at 3T for patients following unicompartmental knee arthroplasty (UKA), although the artifact is expected to increase compared with 1.5T. Purpose To clarify whether SEMAC MRI can better visualize UKA components and improve measurement accuracy at 3T MRI. Materials and methods The phantom consisted of femoral and tibial standard UKA components embedded in agarose gel. The MR images were scanned on a 3T MR system including proton density (PD) MR images. Six orthopedic surgeons blinded to the size and details of the components independently scored the diagnostic value for measurement and measured the lengths of the femoral posterior condyle, femoral peg, anterior-posterior (AP) tibial component, medial-lateral (ML) tibial component, and tibial keel, with and without SEMAC. Visualization scores were stratified as 0 = definitely nondiagnostic, 1 = probably nondiagnostic, 2 = possibly diagnostic, 3 = probably diagnostic, and 4 = definitely diagnostic. In addition, the differences between actual length and 95% confidence intervals of five measurement points were analyzed. Results The diagnostic values of the posterior condyle (2.0; 1.5 vs. 0; 0) and femoral peg (1.5; 1.0 vs. 0; 0) were significantly better in SEMAC-PD MRI than in non-SEMAC-PD MRI (P<0.05). On the other hand, there were no significant differences in the visualizations of AP, ML, and keel of the tibial components. Measurements of the femoral posterior condyle and tibial keel approached the actual length, but were not involved within the 95% confidence interval(actual length, 19.4 mm vs. 95% CI, 15.7-19.1 mm). Conclusion A significant reduction of metal artifacts was observed only around the femoral component in SEMAC-PD MRI. Despite artifact reduction, this sequence did not result in better visualization for measurement.

A correction algorithm for improved magnetic field monitoring with distal field probes.

A significant source of artifacts in MRI are field fluctuations. Field monitoring is a new technology that allows measurement of field dynamics during a scan via "field probes," which can be used to improve image reconstruction. Ideally, probes are located within the volume where gradients produce nominally linear field patterns. However, in some situations probes must be located far from isocenter where rapid field variation can arise, leading to erroneous field-monitoring characterizations and images. This work aimed to develop an algorithm that improves the robustness of field dynamics in these situations. The algorithm is split into three components. Component 1 calculates field dynamics one spatial order at a time, whereas the second implements a weighted least squares solution based on probe distance. Component 3 then calculates phase residuals and removes the residual phase for distant probes before recalculation. Two volunteers and a phantom were scanned on a 7T MRI using diffusion-weighted sequences, and field monitoring was performed. Image reconstructions were informed with field dynamics calculated conventionally, and with the correction algorithm, after which in vivo images were compared qualitatively and phantom image error was quantitatively assessed. The algorithm was able to correct corrupted field dynamics, resulting in image-quality improvements. Significant artifact reduction was observed when correcting higher-order fits. Stepwise fitting provided the most correction benefit, which was marginally improved when adding the other correction strategies. The proposed algorithm can mitigate effects of phase errors in field monitoring, providing improved characterization of field dynamics.

Metal artifact reduction in patients with total hip replacements: evaluation of clinical photon counting CT using virtual monoenergetic images

ObjectivesTo investigate photon-counting CT (PCCT)–derived virtual monoenergetic images (VMI) for artifact reduction in patients with unilateral total hip replacements (THR).MethodsForty-two patients with THR and portal-venous phase PCCT of the abdomen and pelvis were retrospectively included. For the quantitative analysis, region of interest (ROI)–based measurements of hypodense and hyperdense artifacts, as well as of artifact-impaired bone and the urinary bladder, were conducted, and corrected attenuation and image noise were calculated as the difference of attenuation and noise between artifact-impaired and normal tissue. Two radiologists qualitatively evaluated artifact extent, bone assessment, organ assessment, and iliac vessel assessment using 5-point Likert scales.ResultsVMI110keV yielded a significant reduction of hypo- and hyperdense artifacts compared to conventional polyenergetic images (CI) and the corrected attenuation closest to 0, indicating best possible artifact reduction (hypodense artifacts: CI: 237.8 ± 71.4 HU, VMI110keV: 8.5 ± 122.5 HU; p < 0.05; hyperdense artifacts: CI: 240.6 ± 40.8 HU vs. VMI110keV: 13.0 ± 110.4 HU; p < 0.05). VMI110keV concordantly provided best artifact reduction in the bone and bladder as well as the lowest corrected image noise. In the qualitative assessment, VMI110keV received the best ratings for artifact extent (CI: 2 (1–3), VMI110keV: 3 (2–4); p < 0.05) and bone assessment (CI: 3 (1–4), VMI110keV: 4 (2–5); p < 0.05), whereas organ and iliac vessel assessments were rated highest in CI and VMI70keV.ConclusionsPCCT-derived VMI effectively reduce artifacts from THR and thereby improve assessability of circumjacent bone tissue. VMI110keV yielded optimal artifact reduction without overcorrection, yet organ and vessel assessments at that energy level and higher were impaired by loss of contrast.Clinical relevance statementPCCT-enabled artifact reduction is a feasible method for improving assessability of the pelvis in patients with total hip replacements at clinical routine imaging.Key Points• Photon-counting CT-derived virtual monoenergetic images at 110 keV yielded best reduction of hyper- and hypodense artifacts, whereas higher energy levels resulted in artifact overcorrection.• The qualitative artifact extent was reduced best in virtual monoenergetic images at 110 keV, facilitating an improved assessment of the circumjacent bone.• Despite significant artifact reduction, assessment of pelvic organs as well as vessels did not profit from energy levels higher than 70 keV, due to the decline in image contrast.

Single-Fiber Transducer Arrays for 3-D Ultrasound Computed Tomography: From Requirements to Results

A potential method for future breast cancer screening is 3D ultrasound computed tomography (USCT). The utilized image reconstruction algorithms require transducer characteristics fundamentally different from conventional transducer arrays, leading to the necessity of a custom design. This design has to provide random transducer positioning, isotropic sound emission as well as a large bandwidth and wide opening angle. In this paper, we present a new transducer array design to be utilized in a third generation 3D USCT system. Each system requires 128 cylindrical arrays, mounted into the shell of a hemispherical measurement vessel. Each new array contains a 0.6 mm thick disk with 18 single PZT fibers (0.46 mm diameter) embedded in a polymer matrix. Randomized positioning of the fibers is achieved with an arrange-and-fill process. The single-fiber disks are connected on both ends with a matching and backing disk using simple stacking and adhesives. This enables fast and scalable production. We characterized the acoustic field of 54 transducers with a hydrophone. Measurements in 2D showed isotropic acoustic fields. The mean bandwidth and opening angle are 131% and 42°, respectively (both -10 dB). The large bandwidth arises from two resonances within the utilized frequency range. Parameter studies using different models showed that the realized design is already close to the achievable optimum for the transducer technology used. Two 3D USCT systems were equipped with the new arrays. First images show promising results, with an increase in image contrast and a significant reduction of artifacts.

Micro-CT imaging of multiple K-edge elements using GaAs and CdTe photon counting detectors

Objective. To evaluate the performance of two photon-counting (PC) detectors based on different detector materials, gallium arsenide (GaAs) and cadmium telluride (CdTe), for PC micro-CT imaging of phantoms with multiple contrast materials. Another objective is to determine if combining these two detectors in the same micro-CT system can offer higher spectral performance and significant artifact reduction compared to a single detector system. Approach. We have constructed a dual-detector, micro-CT system equipped with two PCDs based on different detector materials: gallium arsenide (GaAs) and cadmium telluride (CdTe). We demonstrate the performance of these detectors for PC micro-CT imaging of phantoms with up to 5 contrast materials with K-edges spread across the x-ray spectrum ranging from iodine with a K-edge at 33.2 keV to bismuth with a K-edge at 90.5 keV. We also demonstrate the use of our system to image a mouse prepared with both iodine and bismuth contrast agents to target different biological systems. Main results. When using the same dose and scan parameters, GaAs shows increased low energy (<50 keV) spectral sensitivity and specificity compared to CdTe. However, GaAs performance at high energies suffers from spectral artifacts and has comparatively low photon counts indicating wasted radiation dose. We demonstrate that combining a GaAs-based and a CdTe-based PC detector in the same micro-CT system offers higher spectral performance and significant artifact reduction compared to a single detector system. Significance. More accurate PC micro-CT using a GaAs PCD alone or in combination with a CdTe PCD could serve for developing new contrast agents such as nanoparticles that show promise in the developing field of theranostics (therapy and diagnostics).

Artifacts reduction in high-acutance phase images for X-ray grating interferometry.

X-ray grating-based techniques often lead to artifacts in the phase retrieval process of phase objects presenting very fast spatial transitions or sudden jumps, especially in the field of non-destructive testing and evaluation. In this paper, we present a method that prevents the emergence of artifacts by building an interferogram corrected from any variations of the object intensity and given as input in the phase retrieval process. For illustration, this method is applied to a carbon fiber specimen imaged by a microfocus X-ray tube and a single 2D grating. A significant reduction of artifacts has been obtained, by a factor higher than 10. This evaluation has been performed experimentally thanks to the Confidence Map tool, a recently developed method that estimates the error distribution from the phase gradient information.

Effects of Botulinum Toxin Injection on Reducing Myogenic Artifacts during Video-EEG Monitoring: A Longitudinal Study

ABSTRACT Medically refractory seizures affect one-third of patients with epilepsy (PwE), for whom epilepsy surgery is considered. Video electroencephalography (vEEG) monitoring is a fundamental tool for pre-operative seizure localization. Facial and cranial myogenic artifacts can obscure vEEG findings, thus interfering with seizure localization. Studies have shown the beneficial effects of botulinum toxin type A (BTX-A) injection into cranial muscles for reducing myogenic artifacts. This longitudinal study aimed to assess the effects of BTX-A injection on these artifacts. Twenty-two patients with medically refractory hypermotor seizures with daily seizure frequency and undetermined epilepsy localization were included in this study and underwent Dysport® injection (200 units) into the frontotemporal region. vEEG recordings were performed at baseline (one week before the injection), and at three days and six days post-injection. Before and after the injection, the amplitudes of myogenic artifacts were compared during various states (ictal, blinking, chewing, bruxism, head lateralization, scowling, talking, and yawning). BTX-A injection significantly reduced the amplitudes of EEG myogenic artifacts, except during blinking (day three) and talking (days three and six). On day six, significant reduction in EEG myogenic artifacts were noted during blinking, chewing, and bruxism for the greatest number of patients (95.5%, 90.9%, 81.8%), while significant reductions in EEG myogenic artifacts during talking, head lateralization, and ictal phase were associated with the least number of patients (22.7%, 36.3%, and 40.9%). Therefore, BTX-A injection could be a convenient method for filtering myogenic contamination, improving EEG interpretation, and facilitating seizure localization in patients with medically refractory seizures.

Deep learning based correction of low performing pixel in computed tomography

Low Performing Pixel (LPP)/bad pixel in CT detectors cause ring and streaks artifacts, structured non-uniformities and deterioration of the image quality. These artifacts make the image unusable for diagnostic purposes. A missing/defective detector pixel translates to a channel missing across all views in sinogram domain and its effect gets spill over entire image in reconstruction domain as artifacts. Most of the existing ring and streak removal algorithms perform correction only in the reconstructed image domain. In this work, we propose a supervised deep learning algorithm that operates in sinogram domain to remove distortions cause by the LPP. This method leverages CT scan geometry, including conjugate ray information to learn the interpolation in sinogram domain. While the experiments are designed to cover the entire detector space, we emphasize on LPPs near detector iso-center as these have most adverse impact on image quality specially if the LPPs fall on the high frequency region (bone-tissue interface). We demonstrated efficacy of the proposed method using data acquired on GE RevACT multi-slice CT system with flat-panel detector. Experimental results on head scans show significant reduction in ring artifacts regardless of LPP location in the detector geometry. We have simulated isolated LPPs accounting for 5% and 10% of total channels. Detailed statistical analysis illustrates approximately 5dB improvement in SNR in both sinogram and reconstruction domain as compared to classical bicubic and Lagrange interpolation methods. Also, with reduction in ring and streak artifacts, the perceptual image quality is improved across all the test images.

Influence of CT metal artifact reduction on SPECT/CT quantification of bone scintigraphy – Retrospective study for selected types of metal implants

Implanted metal prostheses can cause severe artifacts in reconstructed computed tomography (CT) images. To reduce the diagnostic impact of these artifacts and improve attenuation correction in single photon emission computed tomography (SPECT), an algorithm of iterative metal artifact reduction (iMAR) for SPECT/CT systems was developed. The aims of this study were (a) to assess the difference in visual image quality by comparing CT and SPECT images reconstructed with and without iMAR and (b) to determine the influence of iMAR on quantitative 99mTc-uptake in SPECT/CT. This retrospective study includes 21 patients with implanted metal prostheses who underwent SPECT/CT bone scintigraphy. CT data were reconstructed with iMAR and without (noMAR) and were used for attenuation correction of SPECT data for xSPECT Quant and xSPECT Bone reconstruction. The effect of iMAR on image quality was evaluated by visual analysis and the effect on quantitative SPECT/CT was assessed by measuring HU values and absolute uptake values (kBq/mL) in volumes of interest (VOIs). There was a significant reduction of visible metal artifacts with iMAR (p<0.01) in the CT images, but visual differences in the SPECT images were minor. The values of quantitative tracer uptake in VOIs near metal implants were lower for iMAR vs. noMAR xSPECT Quant (p<0.01). Only VOIs near metal showed significant differences in HU values, which were 14.6% lower for iMAR CT (p<0.01). The use of iMAR reduces metal artifacts in CT and improves the perceived image quality. Although in some cases a significant difference in the quantitative evaluation of SPECT/CT was observed, the influence of iMAR can be considered small in relation to other factors in the clinical setting.

Learning-based motion artifact removal networks for quantitative mapping.

To introduce two novel learning-based motion artifact removal networks (LEARN) for the estimation of quantitative motion- and -inhomogeneity-corrected maps from motion-corrupted multi-Gradient-Recalled Echo (mGRE) MRI data. We train two convolutional neural networks (CNNs) to correct motion artifacts for high-quality estimation of quantitative -inhomogeneity-corrected maps from mGRE sequences. The first CNN, LEARN-IMG, performs motion correction on complex mGRE images, to enable the subsequent computation of high-quality motion-free quantitative (and any other mGRE-enabled) maps using the standard voxel-wise analysis or machine learning-based analysis. The second CNN, LEARN-BIO, is trained to directly generate motion- and -inhomogeneity-corrected quantitative maps from motion-corrupted magnitude-only mGRE images by taking advantage of the biophysical model describing the mGRE signal decay. We show that both CNNs trained on synthetic MR images are capable of suppressing motion artifacts while preserving details in the predicted quantitative maps. Significant reduction of motion artifacts on experimental in vivo motion-corrupted data has also been achieved by using our trained models. Both LEARN-IMG and LEARN-BIO can enable the computation of high-quality motion- and -inhomogeneity-corrected maps. LEARN-IMG performs motion correction on mGRE images and relies on the subsequent analysis for the estimation of maps, while LEARN-BIO directly performs motion- and -inhomogeneity-corrected estimation. Both LEARN-IMG and LEARN-BIO jointly process all the available gradient echoes, which enables them to exploit spatial patterns available in the data. The high computational speed of LEARN-BIO is an advantage that can lead to a broader clinical application.

Accelerated Three-dimensional T2-Weighted Turbo-Spin-Echo Sequences with Inner-Volume Excitation and Iterative Denoising in the Setting of Pelvis MRI at 1.5T: Impact on Image Quality and Lesion Detection

To investigate image quality and rate of lesion detection in a novel three-dimensional T2-weighted turbo-spin-echo sequence with inner-volume excitation (zoomed imaging) and iterative denoising processing in pelvic MRI at 1.5T. Two-dimensional T2-weighted turbo-spin-echo sequences were used as the clinical reference standard (2D-T2-TSE). This is a prospective study of patients with various pelvic pathologies. Each patient underwent standard 2D-T2-TSE in three planes with two-fold acceleration as well as a single three-dimensional T2-TSE in the sagittal plane with four-fold acceleration known as Sampling-Perfection-with-Application-optimized-Contrast-using-different-flip-angle-Evolutions (3D-T2-SPACE). The 3D-T2-SPACE images were reconstructed in three orthogonal planes at a slice thickness of 2 mm (vs. 2D-T2-TSE at 4 mm). Two radiologists conducted a qualitative image analysis on standard 2D-T2-TSE and multiplanar reconstructed 3D-T2-SPACE images. These parameters were compared and inter-reader agreement was computed. Furthermore, each reader documented the observed lesions of various pelvic organs. The rate of lesion detection was compared between readers and sequences. Inter-reader and inter-sequence agreement were computed. Forty patients (25 females) were included. Mean patient age was 58 ± 13 years. 3D-T2-SPACE enabled an approximate 22% reduction of acquisition time and 50% of reconstructed slice thickness. 3D-T2-SPACE showed fewer artifacts than 2D-T2-TSE (p < 0.001). However, 2D-T2-TSE was rated to have significantly higher signal intensity than 3D-T2-SPACE (p < 0.001). There were no significant differences between the two sequences regarding all other parameters. Inter-reader agreement regarding image quality parameters was substantial (Kappa=0.772). For all analyzed pelvic anatomic structures, inter-reader and inter-sequence agreement for lesion detection was excellent (Kappa > 0.80). 3D-T2-SPACE with the inner-volume excitation and iterative denoising is clinically feasible at 1.5 T, enabling faster imaging, thinner slices, and significant reduction of artifacts. Despite that signal intensity was inferior in the SPACE images, overall image quality, diagnostic confidence and lesion detection were not compromised. This prospective study sets the stage for further clinical implementation and future investigations tailored to specific indications in pelvis MRI.