Methods for clinical evaluation of noise reduction techniques in abdominopelvic CT

Optimization of smoothing parameter for block matching and 3D filtering algorithm in low-dose chest and abdominal computed tomography images

Computed tomography (CT), known for its exceptionally high accuracy, is associated with a substantial dose of ionizing radiation. Low-dose protocols have been devised to address this issue; however, a reduction in the radiation dose can lead to a deficiency in the number of photons, resulting in quantum noise. Thus, the aim of this study was to optimize the smoothing parameter (σ-value) of the block matching and 3D filtering (BM3D) algorithm to effectively reduce noise in low-dose chest and abdominal CT images. Acquired images were subsequently analyze using quantitative evaluation metrics, including contrast to noise ratio (CNR), coefficient of variation (CV), and naturalness image quality evaluator (NIQE). Quantitative evaluation results demonstrated that the optimal σ-value for CNR, CV, and NIQE were 0.10, 0.11, and 0.09 in low-dose chest CT images respectively, whereas those in abdominal images were 0.12, 0.11, and 0.09, respectively. The average of the optimal σ-values, which produced the most improved results, was 0.10, considering both visual and quantitative evaluations. In conclusion, we demonstrated that the optimized BM3D algorithm with σ-value is effective for noise reduction in low-dose chest and abdominal CT images indicating its feasibility of in the clinical field.

Development and validation of the effective CNR analysis method for evaluating the contrast resolution of CT images

Contrast resolution is an important index for evaluating the signal detectability of computed tomographic (CT) images. Recently, various noise reduction algorithms, such as iterative reconstruction (IR) and deep learning reconstruction (DLR), have been proposed to reduce the image noise in CT images. However, these algorithms cause changes in the image noise texture and blurred image signals in CT images. Furthermore, the contrast-to-noise ratio (CNR) cannot be accurately evaluated in CT images reconstructed using noise reduction methods. Therefore, in this study, we devised a new method, namely, "effective CNR analysis," for evaluating the contrast resolution of CT images. We verified whether the proposed algorithm could evaluate the effective contrast resolution based on the signal detectability of CT images. The findings showed that the effective CNR values obtained using the proposed method correlated well with the subjective visual impressions of CT images. To investigate whether signal detectability was appropriately evaluated using effective CNR analysis, the conventional CNR analysis method was compared with the proposed method. The CNRs of the IR and DLR images calculated using conventional CNR analysis were 13.2 and 10.7, respectively. By contrast, those calculated using the effective CNR analysis were estimated to be 0.7 and 1.1, respectively. Considering that the signal visibility of DLR images was superior to that of IR images, our proposed effective CNR analysis was shown to be appropriate for evaluating the contrast resolution of CT images.

Patient-derived PixelPrint phantoms for evaluating clinical imaging performance of a deep learning CT reconstruction algorithm

Objective. Deep learning reconstruction (DLR) algorithms exhibit object-dependent resolution and noise performance. Thus, traditional geometric CT phantoms cannot fully capture the clinical imaging performance of DLR. This study uses a patient-derived 3D-printed PixelPrint lung phantom to evaluate a commercial DLR algorithm across a wide range of radiation dose levels.Method. The lung phantom used in this study is based on a patient chest CT scan containing ground glass opacities and was fabricated using PixelPrint 3D-printing technology. The phantom was placed inside two different size extension rings to mimic a small- and medium-sized patient and was scanned on a conventional CT scanner at exposures between 0.5 and 20 mGy. Each scan was reconstructed using filtered back projection (FBP), iterative reconstruction, and DLR at five levels of denoising. Image noise, contrast to noise ratio (CNR), root mean squared error, structural similarity index (SSIM), and multi-scale SSIM (MS SSIM) were calculated for each image.Results.DLR demonstrated superior performance compared to FBP and iterative reconstruction for all measured metrics in both phantom sizes, with better performance for more aggressive denoising levels. DLR was estimated to reduce dose by 25%-83% in the small phantom and by 50%-83% in the medium phantom without decreasing image quality for any of the metrics measured in this study. These dose reduction estimates are more conservative compared to the estimates obtained when only considering noise and CNR.Conclusion. DLR has the capability of producing diagnostic image quality at up to 83% lower radiation dose, which can improve the clinical utility and viability of lower dose CT scans. Furthermore, the PixelPrint phantom used in this study offers an improved testing environment with more realistic tissue structures compared to traditional CT phantoms, allowing for structure-based image quality evaluation beyond noise and contrast-based assessments.

Low-dose liver CT: image quality and diagnostic accuracy of deep learning image reconstruction algorithm

OBJECTIVES

To perform a comprehensive within-subject image quality analysis of abdominal CT examinations reconstructed with DLIR and to evaluate diagnostic accuracy compared to the routinely applied adaptive statistical iterative reconstruction (ASiR-V) algorithm.

MATERIALS AND METHODS

Oncologic patients were prospectively enrolled and underwent contrast-enhanced CT. Images were reconstructed with DLIR with three intensity levels of reconstruction (high, medium, and low) and ASiR-V at strength levels from 10 to 100% with a 10% interval. Three radiologists characterized the lesions and two readers assessed diagnostic accuracy and calculated signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), figure of merit (FOM), and subjective image quality, the latter with a 5-point Likert scale.

RESULTS

Fifty patients (mean age: 70 ± 10 years, 23 men) were enrolled and 130 liver lesions (105 benign lesions, 25 metastases) were identified. DLIR_H achieved the highest SNR and CNR, comparable to ASiR-V 100% (p ≥ .051). DLIR_M returned the highest subjective image quality (score: 5; IQR: 4-5; p ≤ .001) and significant median increase (29%) in FOM (p < .001). Differences in detection were identified only for lesions ≤ 0.5 cm: 32/33 lesions were detected with DLIR_M and 26 lesions were detected with ASiR-V 50% (p = .031). Lesion accuracy of was 93.8% (95% CI: 88.1, 97.3; 122 of 130 lesions) for DLIR and 87.7% (95% CI: 80.8, 92.8; 114 of 130 lesions) for ASiR-V 50%.

CONCLUSIONS

DLIR yields superior image quality and provides higher diagnostic accuracy compared to ASiR-V in the assessment of hypovascular liver lesions, in particular for lesions ≤ 0.5 cm.

CLINICAL RELEVANCE STATEMENT

Deep learning image reconstruction algorithm demonstrates higher diagnostic accuracy compared to iterative reconstruction in the identification of hypovascular liver lesions, especially for lesions ≤ 0.5 cm.

KEY POINTS

• Iterative reconstruction algorithm impacts image texture, with negative effects on diagnostic capabilities. • Medium-strength deep learning image reconstruction algorithm outperforms iterative reconstruction in the diagnostic accuracy of ≤ 0.5 cm hypovascular liver lesions (93.9% vs 78.8%), also granting higher objective and subjective image quality. • Deep learning image reconstruction algorithm can be safely implemented in routine abdominal CT protocols in place of iterative reconstruction.

Optimization Method to Predict Optimal Noise Reduction Parameters for the Non-Local Means Algorithm Based on the Scintillator Thickness in Radiography

The resulting image obtained from an X-ray imaging system depends significantly on the characteristics of the detector. In particular, when an X-ray image is acquired by thinning the detector, a relatively large amount of noise inevitably occurs. In addition, when a thick detector is used to reduce noise in X-ray images, blurring increases and the ability to distinguish target areas deteriorates. In this study, we aimed to derive the optimal X-ray image quality by deriving the optimal noise reduction parameters based on the non-local means (NLM) algorithm. The detectors used were of two thicknesses (96 and 140 μm), and images were acquired based on the IEC 62220-1-1:2015 RQA-5 protocol. The optimal parameters were derived by calculating the edge preservation index and signal-to-noise ratio according to the sigma value of the NLM algorithm. As a result, a sigma value of the optimized NLM algorithm (0.01) was derived, and this algorithm was applied to a relatively thin X-ray detector system to obtain appropriate noise level and spatial resolution data. The no-reference-based blind/referenceless image spatial quality evaluator value, which analyzes the overall image quality, was best when using the proposed method. In conclusion, we propose an optimized NLM algorithm based on a new method that can overcome the noise amplification problem in thin X-ray detector systems and is expected to be applied in various photon imaging fields in the future.

Evolution of CT radiation dose in pediatric patients undergoing hybrid 2-[18F]FDG PET/CT between 2007 and 2021

OBJECTIVES

To evaluate the evolution of CT radiation dose in pediatric patients undergoing hybrid 2-[18F]fluoro-2-deoxy-D-glucose (2-[18F]FDG) PET/CT between 2007 and 2021.

METHODS AND MATERIALS

Data from all pediatric patients aged 0-18 years who underwent hybrid 2-[18F]FDG PET/CT of the body between January 2007 and May 2021 were reviewed. Demographic and imaging parameters were collected. A board-certified radiologist reviewed all CT scans and measured image noise in the brain, liver, and adductor muscles.

RESULTS

294 scans from 167 children (72 females (43%); median age: 14 (IQR 10-15) years; BMI: median 17.5 (IQR 15-20.4) kg/m2) were included. CT dose index-volume (CTDIvol) and dose length product (DLP) both decreased significantly from 2007 to 2021 (both p < 0.001, Spearman's rho coefficients -0.46 and -0.35, respectively). Specifically, from 2007 to 2009 to 2019-2021 CTDIvol and DLP decreased from 2.94 (2.14-2.99) mGy and 309 (230-371) mGy*cm, respectively, to 0.855 (0.568-1.11) mGy and 108 (65.6-207) mGy*cm, respectively. From 2007 to 2021, image noise in the brain and liver remained constant (p = 0.26 and p = 0.06), while it decreased in the adductor muscles (p = 0.007). Peak tube voltage selection (in kilovolt, kV) of CT scans shifted from high kV imaging (140 or 120kVp) to low kV imaging (100 or 80kVp) (p < 0.001) from 2007 to 2021.

CONCLUSION

CT radiation dose in pediatric patients undergoing hybrid 2-[18F]FDG PET/CT has decreased in recent years equaling approximately one-third of the initial amount.

ADVANCES IN KNOWLEDGE

Over the past 15 years, CT radiation dose decreased considerably in pediatric patients undergoing hybrid imaging, while objective image quality may not have been compromised.

Application of Deep Learning-Based Denoising Technique for Radiation Dose Reduction in Dynamic Abdominal CT: Comparison with Standard-Dose CT Using Hybrid Iterative Reconstruction Method

The purpose is to evaluate whether deep learning-based denoising (DLD) algorithm provides sufficient image quality for abdominal computed tomography (CT) with a 30% reduction in radiation dose, compared to standard-dose CT reconstructed with conventional hybrid iterative reconstruction (IR). The subjects consisted of 50 patients who underwent abdominal CT with standard dose and reconstructed with hybrid IR (ASiR-V50%) and another 50 patients who underwent abdominal CT with approximately 30% less dose and reconstructed with ASiR-V50% and DLD at low-, medium- and high-strength (DLD-L, DLD-M and DLD-H, respectively). The standard deviation of attenuation in liver parenchyma was measured as image noise. Contrast-to-noise ratio (CNR) for portal vein on portal venous phase was calculated. Lesion conspicuity in 23 abdominal solid mass on the reduced-dose CT was rated on a 5-point scale: 0 (best) to -4 (markedly inferior). Compared with hybrid IR of standard-dose CT, DLD-H of reduced-dose CT provided significantly lower image noise (portal phase: 9.0 (interquartile range, 8.7-9.4) HU vs 12.0 (11.4-12.7) HU, P < 0.0001) and significantly higher CNR (median, 5.8 (4.4-7.4) vs 4.3 (3.3-5.3), P = 0.0019). As for DLD-M of reduced-dose CT, no significant difference was found in image noise and CNR compared to hybrid IR of standard-dose CT (P > 0.99). Lesion conspicuity scores for DLD-H and DLD-M were significantly better than hybrid IR (P < 0.05). Dynamic contrast-enhanced abdominal CT acquired with approximately 30% lower radiation dose and generated with the DLD algorithm exhibit lower image noise and higher CNR compared to standard-dose CT with hybrid IR.

Improvement of diagnostic performance of hyperacute ischemic stroke in head CT using an image-based noise reduction technique with non-black-boxed process

PURPOSE

This study aims to investigate whether an image-based noise reduction (INR) technique with a conventional rule-based algorithm involving no black-boxed processes can outperform an existing hybrid-type iterative reconstruction (HIR) technique, when applied to brain CT images for diagnosis of early CT signs, which generally exhibit low-contrast lesions that are difficult to detect.

METHODS

The subjects comprised 27 patients having infarctions within 4.5 h of onset and 27 patients with no change in brain parenchyma. Images with thicknesses of 5 mm and 0.625 mm were reconstructed by HIR. Images with a thickness of 0.625 mm reconstructed by filter back projection (FBP) were processed by INR. The contrast-to-noise ratios (CNRs) were calculated between gray and white matters; lentiform nucleus and internal capsule; infarcted and non-infarcted areas. Two radiologists subjectively evaluated the presence of hyperdense artery signs (HASs) and infarctions and visually scored three properties regarding image quality (0.625-mm HIR images were excluded because of their notably worse noise appearances).

RESULTS

The CNRs of INR were significantly better than those of HIR with P < 0.001 for all the indicators. INR yielded significantly higher areas under the curve for both infarction and HAS detections than HIR (P < 0.001). Also, INR significantly improved the visual scores of all the three indicators.

CONCLUSION

The INR incorporating a simple and reproducible algorithm was more effective than HIR in detecting early CT signs and can be potentially applied to CT images from a large variety of CT systems.

A Characterization of Deep Learning Reconstruction Applied to Dual-Energy Computed Tomography Monochromatic and Material Basis Images

OBJECTIVE

Advancements in computed tomography (CT) reconstruction have enabled image quality improvements and dose reductions. Previous advancements have included iterative and model-based reconstruction. The latest image reconstruction advancement uses deep learning, which has been evaluated for polychromatic imaging only. This article characterizes a commercially available deep learning imaging reconstruction applied to dual-energy CT.

METHODS

Monochromatic, iodine basis, and water basis images were reconstructed with filtered back projection (FBP), iterative (ASiR-V), and deep learning (DLIR) methods in a phantom experiment. Slice thickness, contrast-to-noise ratio, modulation transfer function, and noise power spectrum metrics were used to characterize ASiR-V and DLIR relative to FBP over a range of dose levels, phantom sizes, and iodine concentrations.

RESULTS

Slice thicknesses for ASiR-V and DLIR demonstrated no statistically significant difference relative to FBP for all measurement conditions. Contrast-to-noise ratio performance for DLIR-high and ASiR-V 40% at 2 mg I/mL on 40-keV images were 162% and 30% higher than FBP, respectively. Task-based modulation transfer function measurements demonstrated no clinically significant change between FBP and ASiR-V and DLIR on monochromatic or iodine basis images.

CONCLUSIONS

Deep learning image reconstruction enabled better image quality at lower monochromatic energies and on iodine basis images where image contrast is maximized relative to polychromatic or high-energy monochromatic images. Deep learning image reconstruction did not demonstrate thicker slices, decreased spatial resolution, or poor noise texture (ie, "plastic") relative to FBP.

Characteristics of the deep learning-based virtual monochromatic image with fast kilovolt-switching CT: a phantom study

PURPOSE

We assessed the physical properties of virtual monochromatic images (VMIs) obtained with different energy levels in various contrast settings and radiation doses using deep learning-based spectral computed tomography (DL-Spectral CT) and compared the results with those from single-energy CT (SECT) imaging.

MATERIALS AND METHODS

A Catphan^® 600 phantom was scanned by DL-Spectral CT at various radiation doses. We reconstructed the VMIs obtained at 50, 70, and 100 keV. SECT (120 kVp) images were acquired at the same radiation doses. The standard deviations of the CT number and noise power spectrum (NPS) were calculated for noise characterization. We evaluated the spatial resolution by determining the 10% task-based transfer function (TTF) level, and we assessed the task-based detectability index (d').

RESULTS

Regardless of the radiation dose, the noise was the lowest at 70 keV VMI. The NPS showed that the noise amplitude at all spatial frequencies was the lowest among other VMI and 120 kVp images. The spatial resolution was higher for 70 keV VMI compared to the other VMIs, except for high-contrast objects. The d' of 70 keV VMI was the highest among the VMI and 120 kVp images at all radiation doses and contrast settings. The d' of the 70 keV VMIs at the minimum dose was higher than that at the maximum dose in any other image.

CONCLUSION

The physical properties of the DL-Spectral CT VMIs varied with the energy level. The 70 keV VMI had the highest detectability by far among the VMI and 120-kVp images. DL-Spectral CT may be useful to reduce radiation doses.

Ultra-low-dose hepatic multiphase CT using deep learning-based image reconstruction algorithm focused on arterial phase in chronic liver disease: A non-inferiority study

PURPOSE

This study determined whether image quality and detectability of ultralow-dose hepatic multiphase CT (ULDCT, 33.3% dose) using a vendor-agnostic deep learning model(DLM) are noninferior to those of standard-dose CT (SDCT, 100% dose) using model-based iterative reconstruction(MBIR) in patients with chronic liver disease focusing on arterial phase.

METHODS

Sixty-seven patients underwent hepatic multiphase CT using a dual-source scanner to obtain two different radiation dose CT scans (100%, SDCT and 33.3%, ULDCT). ULDCT using DLM and SDCT using MBIR were compared. A margin of -0.5 for the difference between the two protocols was pre-defined as noninferiority of the overall image quality of the arterial phase image. Quantitative image analysis (signal to noise ratio[SNR] and contrast to noise ratio[CNR]) was also conducted. The detectability of hepatic arterial focal lesions was compared using the Jackknife free-response receiver operating characteristic analysis. Non-inferiority was satisfied if the margin of the lower limit of 95%CI of the difference in figure-of-merit was less than -0.1.

RESULTS

Mean overall arterial phase image quality scores with ULDCT using DLM and SDCT using MBIR were 4.35 ± 0.57 and 4.08 ± 0.58, showing noninferiority (difference: -0.269; 95 %CI, -0.374 to -0.164). ULDCT using DLM showed a significantly superior contrast-to-noise ratio of arterial enhancing lesion (p < 0.05). Figure-of-merit for detectability of arterial hepatic focal lesion was 0.986 for ULDCT using DLM and 0.963 for SDCT using MBIR, showing noninferiority (difference: -0.023, 95 %CI: -0.016 to 0.063).

CONCLUSION

ULDCT using DLM with 66.7% dose reduction showed non-inferior overall image quality and detectability of arterial focal hepatic lesion compared to SDCT using MBIR.

Diagnostic Performance in Low- and High-Contrast Tasks of an Image-Based Denoising Algorithm Applied to Radiation Dose-Reduced Multiphase Abdominal CT Examinations

BACKGROUND. Anatomic redundancy between phases can be used to achieve denoising of multiphase CT examinations. A limitation of iterative reconstruction (IR) techniques is that they generally require use of CT projection data. A frequency-split multi-band-filtration algorithm applies denoising to the multiphase CT images themselves. This method does not require knowledge of the acquisition process or integration into the reconstruction system of the scanner, and it can be implemented as a supplement to commercially available IR algorithms. OBJECTIVE. The purpose of the present study is to compare radiologists' performance for low-contrast and high-contrast diagnostic tasks (i.e., tasks for which differences in CT attenuation between the imaging target and its anatomic background are subtle or large, respectively) evaluated on multiphase abdominal CT between routine-dose images and radiation dose-reduced images processed by a frequency-split multiband-filtration denoising algorithm. METHODS. This retrospective single-center study included 47 patients who underwent multiphase contrast-enhanced CT for known or suspected liver metastases (a low-contrast task) and 45 patients who underwent multiphase contrast-enhanced CT for pancreatic cancer staging (a high-contrast task). Radiation dose-reduced images corresponding to dose reduction of 50% or more were created using a validated noise insertion technique and then underwent denoising using the frequency-split multi-band-filtration algorithm. Images were independently evaluated in multiple sessions by different groups of abdominal radiologists for each task (three readers in the low-contrast arm and four readers in the high-contrast arm). The noninferiority of denoised radiation dose-reduced images to routine-dose images was assessed using the jackknife alternative free-response ROC (JAFROC) figure-of-merit (FOM; limit of noninferiority, -0.10) for liver metastases detection and using the Cohen kappa statistic and reader confidence scores (100-point scale) for pancreatic cancer vascular invasion. RESULTS. For liver metastases detection, the JAFROC FOM for denoised radiation dose-reduced images was 0.644 (95% CI, 0.510-0.778), and that for routine-dose images was 0.668 (95% CI, 0.543-0.792; estimated difference, -0.024 [95% CI, -0.084 to 0.037]). Intraobserver agreement for pancreatic cancer vascular invasion was substantial to near perfect when the two image sets were compared (κ = 0.53-1.00); the 95% CIs of all differences in confidence scores between image sets contained zero. CONCLUSION. Multiphase contrast-enhanced abdominal CT images with a radiation dose reduction of 50% or greater that undergo denoising by a frequency-split multiband-filtration algorithm yield performance similar to that of routine-dose images for detection of liver metastases and vascular staging of pancreatic cancer. CLINICAL IMPACT. The image-based denoising algorithm facilitates radiation dose reduction of multiphase examinations for both low- and high-contrast diagnostic tasks without requiring manufacturer-specific hardware or software.

Validation of Virtual Monochromatic Images and Effect of Body Size Obtained Using a Rapid kVp-switching Dual-energy Computed Tomography System: A Phantom Study

The objective of this paper is to validate virtual monochromatic computed tomography (CT) numbers and the effect of the body size of insert materials in phantoms on the findings of a dual-energy CT scanner. The material inserted in the phantom simulates human organs. This study investigated the effect of different body sizes on CT numbers to understand the accuracy of dual-energy CT. The effect of body size on virtual monochromatic CT numbers was investigated using a QRM phantom. The true monochromatic CT numbers of insert materials were calculated from coefficients obtained using NIST XCOM. The true Z eff values were supplied by phantom manufacturers or computed using Mayneord's equation. The virtual monochromatic CT numbers of insert materials in both the phantoms varied with energy. The CT numbers of materials with a Z eff of >7.42 (water Z eff ) and <7.42 decreased and increased with energy, respectively. The CT numbers were affected by phantom size as a function of energy. For water, tissues, and air, the CT numbers in the XL phantom were considerably larger than those in other phantom sizes at 40 keV. Body size affected the CT numbers, particularly for the XL size and at low energies. For all materials, the magnitude of difference between the measured and true CT numbers was related to the Z eff of the materials, potentially because the photoelectric effect is more prominent at low energies for materials with a higher Z eff . The difference in CT numbers appeared to be dependent on position. The true and measured Z eff agreed to within 6% for all the materials except the SR2 brain, for which the discrepancy was 25%.

A Review of Deep Learning CT Reconstruction: Concepts, Limitations, and Promise in Clinical Practice

Purpose of Review

Deep Learning reconstruction (DLR) is the current state-of-the-art method for CT image formation. Comparisons to existing filter back-projection, iterative, and model-based reconstructions are now available in the literature. This review summarizes the prior reconstruction methods, introduces DLR, and then reviews recent findings from DLR from a physics and clinical perspective.

Recent Findings

DLR has been shown to allow for noise magnitude reductions relative to filtered back-projection without suffering from “plastic” or “blotchy” noise texture that was found objectionable with most iterative and model-based solutions. Clinically, early reader studies have reported increases in subjective quality scores and studies have successfully implemented DLR-enabled dose reductions.

Summary

The future of CT image reconstruction is bright; deep learning methods have only started to tackle problems in this space via addressing noise reduction. Artifact mitigation and spectral applications likely be future candidates for DLR applications.

Comparative Assessment of Noise Properties for Two Deep Learning CT Image Reconstruction Techniques and Filtered Back Projection

BACKGROUND

Two deep-learning image reconstruction (DLIR) techniques from two different CT vendors have recently been introduced into clinical practice.

PURPOSE

To characterize the noise properties of two DLIR techniques with different training methods, using a phantom containing a simple uniform and a complex non-uniform region.

METHODS

A water-bath phantom with a diameter of 300 mm was used as a base phantom. A textured phantom with a diameter of 128 mm, which was made of two materials, one equivalent to water and the other being 12 mg/mL diluted iodine, irregularly mixed to create a complex texture (non-uniform region), was placed in the base phantom. Thirty repeated phantom scans were performed using two CT scanners (GE, Revolution CT with Apex Edition; Canon, Aquilion One PRISM Edition) at two dose levels (CTDI: 5 and 15 mGy). Images were reconstructed with each CT system's filtered back projection (FBP) and DLIR [GE, TrueFidelity (TF); Canon, Advanced intelligent Clear-IQ Engine Body Sharp (AC)] for three process strengths. For basic characteristics of noise, the standard deviation (SD) and noise power spectrum (NPS) were measured for the uniform (water) region. A noise magnitude map was generated by calculating the inter-image SD at each pixel position across the 30 images. Then, a noise reduction map (NRM), which visualizes the relative differences in noise magnitude between FBP and DLIR, was calculated. The NRM values ranged from 0.0 to 1.0. A low NRM value represents a less aggressive noise reduction. The histograms of the NRM value were analyzed for the uniform and non-uniform regions.

RESULTS

The reduction in noise magnitude compared with FBP tended to be greater with AC (45%-85%) than with TF (32%-65%). The average NPS frequencies of TF and AC were almost comparable to those of FBP, except for the low-dose condition and the high noise reduction strength for AC. The NRM values of TF and AC were higher in the uniform region than in the non-uniform region. In the non-uniform region, TF's average NRM values (0.21-0.48) tended to be lower than AC's (0.39-0.78). The histograms for TF showed a small overlap between the uniform and the non-uniform regions; in contrast, those for AC showed a greater overlap. This difference seems to indicate that TF processes the uniform and non-uniform regions more differently than AC does.

CONCLUSION

This study has revealed a distinct difference in characteristics between the two DLIR techniques: TF tends to offer less aggressive noise reduction in non-uniform regions and preserve the original signals, whereas AC tends to prioritize noise filtering over edge-preservation, especially at the low-dose condition and with the high noise reduction strength. This article is protected by copyright. All rights reserved.

Fully automated image quality evaluation on patient CT: Multi-vendor and multi-reconstruction study

While the recent advancements of computed tomography (CT) technology have contributed in reducing radiation dose and image noise, an objective evaluation of image quality in patient scans has not yet been established. In this study, we present a patient-specific CT image quality evaluation method that includes fully automated measurements of noise level, structure sharpness, and alteration of structure. This study used the CT images of 120 patients from four different CT scanners reconstructed with three types of algorithm: filtered back projection (FBP), vendor-specific iterative reconstruction (IR), and a vendor-agnostic deep learning model (DLM, ClariCT.AI, ClariPi Inc.). The structure coherence feature (SCF) was used to divide an image into the homogeneous (RH) and structure edge (RS) regions, which in turn were used to localize the regions of interests (ROIs) for subsequent analysis of image quality indices. The noise level was calculated by averaging the standard deviations from five randomly selected ROIs on RH, and the mean SCFs on RS was used to estimate the structure sharpness. The structure alteration was defined by the standard deviation ratio between RS and RH on the subtraction image between FBP and IR or DLM, in which lower structure alterations indicate successful noise reduction without degradation of structure details. The estimated structure sharpness showed a high correlation of 0.793 with manually measured edge slopes. Compared to FBP, IR and DLM showed 34.38% and 51.30% noise reduction, 2.87% and 0.59% lower structure sharpness, and 2.20% and -12.03% structure alteration, respectively, on an average. DLM showed statistically superior performance to IR in all three image quality metrics. This study is expected to contribute to enhance the CT protocol optimization process by allowing a high throughput and quantitative image quality evaluation during the introduction or adjustment of lower-dose CT protocol into routine practice.

Comparison of a Deep Learning-Based Reconstruction Algorithm with Filtered Back Projection and Iterative Reconstruction Algorithms for Pediatric Abdominopelvic CT

Objective

To compare a deep learning-based reconstruction (DLR) algorithm for pediatric abdominopelvic computed tomography (CT) with filtered back projection (FBP) and iterative reconstruction (IR) algorithms.

Materials and Methods

Post-contrast abdominopelvic CT scans obtained from 120 pediatric patients (mean age ± standard deviation, 8.7 ± 5.2 years; 60 males) between May 2020 and October 2020 were evaluated in this retrospective study. Images were reconstructed using FBP, a hybrid IR algorithm (ASiR-V) with blending factors of 50% and 100% (AV50 and AV100, respectively), and a DLR algorithm (TrueFidelity) with three strength levels (low, medium, and high). Noise power spectrum (NPS) and edge rise distance (ERD) were used to evaluate noise characteristics and spatial resolution, respectively. Image noise, edge definition, overall image quality, lesion detectability and conspicuity, and artifacts were qualitatively scored by two pediatric radiologists, and the scores of the two reviewers were averaged. A repeated-measures analysis of variance followed by the Bonferroni post-hoc test was used to compare NPS and ERD among the six reconstruction methods. The Friedman rank sum test followed by the Nemenyi-Wilcoxon-Wilcox all-pairs test was used to compare the results of the qualitative visual analysis among the six reconstruction methods.

Results

The NPS noise magnitude of AV100 was significantly lower than that of the DLR, whereas the NPS peak of AV100 was significantly higher than that of the high- and medium-strength DLR (p < 0.001). The NPS average spatial frequencies were higher for DLR than for ASiR-V (p < 0.001). ERD was shorter with DLR than with ASiR-V and FBP (p < 0.001). Qualitative visual analysis revealed better overall image quality with high-strength DLR than with ASiR-V (p < 0.001).

Conclusion

For pediatric abdominopelvic CT, the DLR algorithm may provide improved noise characteristics and better spatial resolution than the hybrid IR algorithm.

Evaluation of Abdominal CT Obtained Using a Deep Learning-Based Image Reconstruction Engine Compared with CT Using Adaptive Statistical Iterative Reconstruction

Purpose:

Materials and Methods:

Results:

Conclusions:

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

BACKGROUND

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

METHODS

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

RESULTS

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

CONCLUSIONS

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

Task-specific spatial resolution properties of iterative and deep learning-based reconstructions in computed tomography: Comparison using tasks assuming small and large enhanced vessels

PURPOSE

The present study aims to evaluate TTFs of deep-learning-based image reconstruction (DLIR) and iterative reconstruction (IR) in computed tomography (CT) using a conventional task with a rod object with a diameter of 30 mm and a newly-proposed task with a wire of 1 mm in diameter, simulating large and small enhanced vessels, respectively.

METHODS

The rod or wire phantom made of a material equivalent to diluted iodine that exhibits about 270 Hounsfield unit (HU) was placed inside a 30-cm water phantom. In-plane and z-directional TTFs were measured for the rod using the circular edge (CE) and plane edge (PE) methods, respectively. By using the wire (iodine wire: IW), in-plane and z-directional TTFs were measured using Fourier transform (IW method). TTFs of filtered back projection (FBP), IR, and DLIR of a 256-row CT system and FBP and IR of a 64-row CT system were evaluated with CT dose indices of 10 and 5 mGy.

RESULTS

For DLIR and IR, TTFs measured using the IW method were notably lower than those using the CE (or PE) method; moreover, they were also lower than those of corresponding FBP, indicating that the small enhanced vessels with a diameter of about 1 mm would be blurred with both DLIR and IR.

CONCLUSIONS

The proposed IW method has turned out to be effective to evaluate TTFs for small enhanced vessels, which have not been properly evaluated by the CE or PE method conventionally recommended.

Deep-learning reconstruction for ultra-low-dose lung CT: Volumetric measurement accuracy and reproducibility of artificial ground-glass nodules in a phantom study

OBJECTIVES

The lung nodule volume determined by CT is used for nodule diagnoses and monitoring tumor responses to therapy. Increased image noise on low-dose CT degrades the measurement accuracy of the lung nodule volume. We compared the volumetric accuracy among deep-learning reconstruction (DLR), model-based iterative reconstruction (MBIR), and hybrid iterative reconstruction (HIR) at an ultra-low-dose setting.

METHODS

Artificial ground-glass nodules (6 mm and 10 mm diameters, -660 HU) placed at the lung-apex and the middle-lung field in chest phantom were scanned by 320-row CT with the ultra-low-dose setting of 6.3 mAs. Each scan data set was reconstructed by DLR, MBIR, and HIR. The volumes of nodules were measured semi-automatically, and the absolute percent volumetric error (APEvol) was calculated. The APEvol provided by each reconstruction were compared by the Tukey-Kramer method. Inter- and intraobserver variabilities were evaluated by a Bland-Altman analysis with limits of agreements.

RESULTS

DLR provided a lower APEvol compared to MBIR and HIR. The APEvol of DLR (1.36%) was significantly lower than those of the HIR (8.01%, p = 0.0022) and MBIR (7.30%, p = 0.0053) on a 10-mm-diameter middle-lung nodule. DLR showed narrower limits of agreement compared to MBIR and HIR in the inter- and intraobserver agreement of the volumetric measurement.

CONCLUSIONS

DLR showed higher accuracy compared to MBIR and HIR for the volumetric measurement of artificial ground-glass nodules by ultra-low-dose CT.

ADVANCES IN KNOWLEDGE

DLR with ultra-low-dose setting allows a reduction of dose exposure, maintaining accuracy for the volumetry of lung nodule, especially in patients which deserve a long-term follow-up.

Evaluation of Apparent Noise on CT Images Using Moving Average Filters

This study aims to devise a simple method for evaluating the magnitude of texture noise (apparent noise) observed on computed tomography (CT) images scanned at a low radiation dose and reconstructed using iterative reconstruction (IR) and deep learning reconstruction (DLR) algorithms, and to evaluate the apparent noise in CT images reconstructed using the filtered back projection (FBP), IR, and two types of DLR (AiCE Body and AiCE Body Sharp) algorithms. We set a square region of interest (ROI) on CT images of standard- and obese-sized low-contrast phantoms, slid different-sized moving average filters in the ROI vertically and horizontally in steps of 1 pixel, and calculated the standard deviation (SD) of the mean CT values for each filter size. The SD of the mean CT values was fitted with a curve inversely proportional to the filter size, and an apparent noise index was determined from the curve-fitting formula. The apparent noise index of AiCE Body Sharp images for a given mAs value was approximately 58, 23, and 18% lower than that of the FBP, AIDR 3D, and AiCE Body images, respectively. The apparent noise index was considered to reflect noise power spectrum values at lower spatial frequency. Moreover, the apparent noise index was inversely proportional to the square roots of the mAs values. Thus, the apparent noise index could be a useful indicator to quantify and compare texture noise on CT images obtained with different scan parameters and reconstruction algorithms.

Denoising of pediatric low dose abdominal CT using deep learning based algorithm

OBJECTIVES

To evaluate standard dose-like computed tomography (CT) images generated by a deep learning method, trained using unpaired low-dose CT (LDCT) and standard-dose CT (SDCT) images.

MATERIALS AND METHODS

LDCT (80 kVp, 100 mAs, n = 83) and SDCT (120 kVp, 200 mAs, n = 42) images were divided into training (42 LDCT and 42 SDCT) and validation (41 LDCT) sets. A generative adversarial network framework was used to train unpaired datasets. The trained deep learning method generated virtual SDCT images (VIs) from the original LDCT images (OIs). To test the proposed method, LDCT images (80 kVp, 262 mAs, n = 33) were collected from another CT scanner using iterative reconstruction (IR). Image analyses were performed to evaluate the qualities of VIs in the validation set and to compare the performance of deep learning and IR in the test set.

RESULTS

The noise of the VIs was the lowest in both validation and test sets (all p<0.001). The mean CT number of the VIs for the portal vein and liver was lower than that of OIs in both validation and test sets (all p<0.001) and was similar to those of SDCT. The contrast-to-noise ratio of portal vein and the signal-to-noise ratio (SNR) of portal vein and liver of VIs were higher than those of SDCT (all p<0.05). The SNR of VIs in test sets was the highest among three images.

CONCLUSION

The deep learning method trained by unpaired datasets could reduce noise of LDCT images and showed similar performance to SAFIRE. It can be applied to LDCT images of older CT scanners without IR.

Quantitative measurements of emphysema in ultra-high resolution computed tomography using model-based iterative reconstruction in comparison to that using hybrid iterative reconstruction

The percentage of low attenuation volume ratio (LAVR), which is measured using computed tomography (CT), is an index of the severity of emphysema. For LAVR evaluation, ultra-high-resolution (U-HR) CT images are useful. To improve the image quality of U-HRCT, iterative reconstruction is used. There are two types of iterative reconstruction: hybrid iterative reconstruction (HIR) and model-based iterative reconstruction (MBIR). In this study, we physically and clinically evaluated U-HR images reconstructed with HIR and MBIR, and demonstrated the usefulness of U-HR images with MBIR for quantitative measurements of emphysema. Both images were reconstructed with a slice thickness of 0.25 mm and an image matrix size of 1024 × 1024 pixels. For physical evaluation, the modulation transfer function (MTF) and noise power spectrum (NPS) of HIR and MBIR were compared. For clinical evaluation, LAVR calculated from HIR and MBIR were compared using the Wilcoxon matched-pairs signed-rank test. In addition, the correlation between LAVR and forced expiratory volume in one second (FEV₁%) was evaluated using the Spearman rank correlation test. The MTFs of HIR and MBIR were comparable. The NPS of MBIR was lower than that of HIR. The mean LAVR values calculated from HIR and MBIR were 19.5 ± 12.6% and 20.4 ± 11.7%, respectively (p = 0.84). The correlation coefficients between LAVR and FEV₁% that were taken from HIR and MBIR were 0.64 and 0.74, respectively (p < 0.01). MBIR is more useful than HIR for the quantitative measurements of emphysema with U-HR images.

Deep-learning image reconstruction for image quality evaluation and accurate bone mineral density measurement on quantitative CT: A phantom-patient study

BACKGROUND AND PURPOSE

To investigate the image quality and accurate bone mineral density (BMD) on quantitative CT (QCT) for osteoporosis screening by deep-learning image reconstruction (DLIR) based on a multi-phantom and patient study.

MATERIALS AND METHODS

High-contrast spatial resolution, low-contrast detectability, modulation function test (MTF), noise power spectrum (NPS), and image noise were evaluated for physical image quality on Caphan 500 phantom. Three calcium hydroxyapatite (HA) inserts were used for accurate BMD measurement on European Spine Phantom (ESP). CT images were reconstructed with filtered back projection (FBP), adaptive statistical iterative reconstruction-veo 50% (ASiR-V50%), and three levels of DLIR(L/M/H). Subjective evaluation of the image high-contrast spatial resolution and low-contrast detectability were compared visually by qualified radiologists, whilst the statistical difference in the objective evaluation of the image high-contrast spatial resolution and low-contrast detectability, image noise, and relative measurement error were compared using one-way analysis of variance (ANOVA). Cohen's kappa coefficient (k) was performed to determine the interobserver agreement in qualitative evaluation between two radiologists.

RESULTS

Overall, for three levels of DLIR, 50% MTF was about 4.50 (lp/cm), better than FBP (4.12 lp/cm) and ASiR-V50% (4.00 lp/cm); the 2 mm low-contrast object was clearly resolved at a 0.5% contrast level, while 3mm at FBP and ASiR-V50%. As the strength level decreased and radiation dose increased, DLIR at three levels showed a higher NPS peak frequency and lower noise level, leading to leftward and rightward shifts, respectively. Measured L1, L2, and L3 were slightly lower than that of nominal HA inserts (44.8, 95.9, 194.9 versus 50.2, 100.6, 199.2mg/cm³) with a relative measurement error of 9.84%, 4.08%, and 2.60%. Coefficients of variance for the L1, L2, and L3 HA inserts were 1.51%, 1.41%, and 1.18%. DLIR-M and DLIR-H scored significantly better than ASiR-V50% in image noise (4.83 ± 0.34, 4.50 ± 0.50 versus 4.17 ± 0.37), image contrast (4.67 ± 0.73, 4.50 ± 0.70 versus 3.80 ± 0.99), small structure visibility (4.83 ± 0.70, 4.17 ± 0.73 versus 3.83 ± 1.05), image sharpness (3.83 ± 1.12, 3.53 ± 0.90 versus 3.27 ± 1.16), and artifacts (3.83 ± 0.90, 3.42 ± 0.37 versus 3.10 ± 0.83). The CT value, image noise, contrast noise ratio, and image artifacts in DLIR-M and DLIR-H outperformed ASiR-V50% and FBP (P<0.001), whilst it showed no statistically significant between DLIR-L and ASiR-V50% (P>0.05). The prevalence of osteoporosis was 74 (24.67%) in women and 49 (11.79%) in men, whilst the osteoporotic vertebral fracture rate was 26 (8.67%) in women and (5.29%) in men.

CONCLUSION

Image quality with DLIR was high-qualified without affecting the accuracy of BMD measurement. It has a potential clinical utility in osteoporosis screening.

A phantom study comparing low-dose CT physical image quality from five different CT scanners

BACKGROUND

To systematically evaluate the physical image quality of low-dose computed tomography (LDCT) on CT scanners from 5 different manufacturers using a phantom model.

METHODS

CT images derived from a Catphan 500 phantom were acquired using manufacturer-specific iterative reconstruction (IR) algorithms and deep learning image reconstruction (DLIR) on CT scanners from 5 different manufacturers and compared using filtered back projection with 2 radiation doses of 0.25 and 0.75 mGy. Image high-contrast spatial resolution and image noise were objectively characterized by modulation transfer function (MTF) and noise power spectrum (NPS). Image high-contrast spatial resolution and image low-contrast detectability were compared directly by visual evaluation. CT number linearity and image uniformity were compared with intergroup differences using one-way analysis of variance (ANOVA).

RESULTS

The CT number linearity of 4 insert materials were as follows: acrylic (95% CI: 120.35 to 121.27; P=0.134), low-density polyethylene (95% CI: -98.43 to -97.43; P=0.070), air (95% CI: -996.16 to -994.51; P=0.018), and Teflon (95% CI: 984.40 to 986.87; P=0.883). The image uniformity values of GE Healthcare (95% CI: 3.24 to 3.83; P=0.138), Philips (95% CI: 2.62 to 3.70; P=0.299), Siemens (95% CI: 2.10 to 3.59; P=0.054), Minfound (95% CI: 2.35 to 3.65; P=0.589), and Neusoft (95% CI: 2.63 to 3.37; P=0.900) were evaluated and found to be within ±4 Hounsfield units (HU), with a range of 0.99-2.76 HU for standard deviations. There was no statistically significant difference in CT number linearity and image uniformity across the 5 CT scanners under different radiation doses with IR and DLIR algorithms (P>0.05). The resolution level at 10% MTF was 6.98 line-pairs-per-centimeter (lp/cm) on average, which was similar to the subjective evaluation results (mostly up to 7 lp/cm). DLIR at all 3 levels had the highest 50% MTF values among all reconstruction algorithms. For image low-contrast detectability, the minimum diameter of distinguishable contrast holes reached 4 mm at a 0.5% resolution. Increasing the radiation dose and IR strength reduced the image noise and NPS curve peak frequency while improving image low-contrast detectability.

CONCLUSIONS

This study demonstrated that the image quality of CT scanners from 5 different manufacturers in LDCT is comparable and that the CT number linearity is unbiased and can contribute to accurate bone mineral density quantification.

Noise reduction profile: A new method for evaluation of noise reduction techniques in CT

Purpose

Noise power spectrum (NPS) is a commonly used performance metric to evaluate noise‐reduction techniques (NRT) in imaging systems. The images reconstructed with and without an NRT can be compared via their NPS to better understand the NRT's effects on image noise. However, when comparing NPSs, simple visual assessments or a comparison of NPS peaks or medians are often used. These assessments make it difficult to objectively evaluate the effect of noise reduction across all spatial frequencies. In this work, we propose a new noise reduction profile (NRP) to facilitate a more complete and objective evaluation of NPSs for a range of NRTs used specifically in computed tomography (CT).

Methods and materials

The homogeneous section of the ACR or Catphan phantoms was scanned on different CT scanners equipped with the following NRTs: AIDR3D, AiCE, ASiR, ASiR‐V, TrueFidelity, iDose, SAFIRE, and ADMIRE. The images were then reconstructed with all strengths of each NRT in reference to the baseline filtered back projection (FBP) images. One set of the baseline FBP images was also processed with PixelShine, an NRT based on artificial intelligence. The NPSs of the images before and after noise reduction were calculated in both the xy‐plane and along the z‐direction. The difference in the logarithmic scale between each NPS (baseline FBP and NRT) was then calculated and deemed the NRP. Furthermore, the relationship between the NRP and NPS peak positions was mathematically analyzed.

Results

Each NRT has its own unique NRP. By comparing the NPS and NRP for each NRT, it was found that NRP is related to the peak shift of NPS. Additionally, under the assumption that the NPS has one peak and is differentiable, a relationship was mathematically derived between the slope of the NRP at the peak position of the NPS before noise reduction and the shift of the NPS peak position after noise reduction.

Conclusions

A new metric, NRP, was proposed based on NPS to objectively evaluate and compare methods for noise reduction in CT. The NRP can be used to compare the effects of various NRTs on image noise in both the xy‐plane and z‐direction. It also enables unbiased assessment of the detailed noise reduction properties of each NRT over all relevant spatial frequencies.

Effects of Adaptive Statistical Iterative Reconstruction-V Technology on the Image Quality and Radiation Dose of Unenhanced and Enhanced CT Scans of the Piglet Abdomen

To investigate the optimal pre- and post-adaptive statistical iterative reconstruction-V (ASiR-V) levels in pediatric abdominal computed tomography (CT) to minimize radiation exposure and maintain image quality using an animal model. A total of 10 standard piglets were selected and scanned to obtain unenhanced and enhanced images under different pre-ASiR-V conditions. The corresponding images were obtained using ASiR-V algorithm at different post-ASiR-V levels. CT value, signal-to-noise ratio (SNR), contrast noise ratio (CNR) of abdominal tissues, subjective image score, and radiation dose of unenhanced and enhanced scans were analyzed. With the increase of pre-ASiR-V level, the radiation dose in piglets gradually decreased (P < 0.05). Within the same group of pre-ASiR-V, the image noise was decreased (P < 0.05) by increasing post-ASiR-V level. There was no statistical difference between SNR and CNR values. In unenhanced CT, the subjective score of the images with the combination of 40% pre- and 60% post-ASiR-V levels had no statistical difference compared to the combination of 0% pre- and 60% post-ASiR-V levels, while the radiation dose decreased by 31.6%. In the enhanced CT, the subjective image score with the 60% pre- and 60% post-ASiR-V combination had no statistical difference compared to the 0% pre- and 60% post-ASiR-V combination, while the radiation dose was reduced by 48.9%. The combined use of pre- and post-ASiR-V maintains image quality at the reduced radiation dose. The optimal level for unenhanced CT is 40% pre-combined with 60% post-ASiR-V, while that for enhanced CT is 60% pre- combined with 60% post-ASiR-V in pediatric abdominal CT.

Deep Learning-based Reconstruction for Lower-Dose Pediatric CT: Technical Principles, Image Characteristics, and Clinical Implementations

Optimizing the CT acquisition parameters to obtain diagnostic image quality at the lowest possible radiation dose is crucial in the radiosensitive pediatric population. The image quality of low-dose CT can be severely degraded by increased image noise with filtered back projection (FBP) reconstruction. Iterative reconstruction (IR) techniques partially resolve the trade-off relationship between noise and radiation dose but still suffer from degraded noise texture and low-contrast detectability at considerably low-dose settings. Furthermore, sophisticated model-based IR usually requires a long reconstruction time, which restricts its clinical usability. With recent advances in artificial intelligence technology, deep learning-based reconstruction (DLR) has been introduced to overcome the limitations of the FBP and IR approaches and is currently available clinically. DLR incorporates convolutional neural networks-which comprise multiple layers of mathematical equations-into the image reconstruction process to reduce image noise, improve spatial resolution, and preserve preferable noise texture in the CT images. For DLR development, numerous network parameters are iteratively optimized through an extensive learning process to discriminate true attenuation from noise by using low-dose training and high-dose teaching image data. After rigorous validations of network generalizability, the DLR engine can be used to generate high-quality images from low-dose projection data in a short reconstruction time in a clinical environment. Application of the DLR technique allows substantial dose reduction in pediatric CT performed for various clinical indications while preserving the diagnostic image quality. The authors present an overview of the basic concept, technical principles, and image characteristics of DLR and its clinical feasibility for low-dose pediatric CT. ^©RSNA, 2021.

Advances in deep learning for computed tomography denoising

Computed tomography (CT) has seen a rapid increase in use in recent years. Radiation from CT accounts for a significant proportion of total medical radiation. However, given the known harmful impact of radiation exposure to the human body, the excessive use of CT in medical environments raises concerns. Concerns over increasing CT use and its associated radiation burden have prompted efforts to reduce radiation dose during the procedure. Therefore, low-dose CT has attracted major attention in the radiology, since CT-associated x-ray radiation carries health risks for patients. The reduction of the CT radiation dose, however, compromises the signal-to-noise ratio, which affects image quality and diagnostic performance. Therefore, several denoising methods have been developed and applied to image processing technologies with the goal of reducing image noise. Recently, deep learning applications that improve image quality by reducing the noise and artifacts have become commercially available for diagnostic imaging. Deep learning image reconstruction shows great potential as an advanced reconstruction method to improve the quality of clinical CT images. These improvements can provide significant benefit to patients regardless of their disease, and further advances are expected in the near future.

CT Noise-Reduction Methods for Lower-Dose Scanning: Strengths and Weaknesses of Iterative Reconstruction Algorithms and New Techniques

Iterative reconstruction (IR) algorithms are the most widely used CT noise-reduction method to improve image quality and have greatly facilitated radiation dose reduction within the radiology community. Various IR methods have different strengths and limitations. Because IR algorithms are typically nonlinear, they can modify spatial resolution and image noise texture in different regions of the CT image; hence traditional image-quality metrics are not appropriate to assess the ability of IR to preserve diagnostic accuracy, especially for low-contrast diagnostic tasks. In this review, the authors highlight emerging IR algorithms and CT noise-reduction techniques and summarize how these techniques can be evaluated to help determine the appropriate radiation dose levels for different diagnostic tasks in CT. In addition to advanced IR techniques, we describe novel CT noise-reduction methods based on convolutional neural networks (CNNs). CNN-based noise-reduction techniques may offer the ability to reduce image noise while maintaining high levels of image detail but may have unique drawbacks. Other novel CT noise-reduction methods are being developed to leverage spatial and/or spectral redundancy in multiphase or multienergy CT. Radiologists and medical physicists should be familiar with these different alternatives to adapt available CT technology for different diagnostic tasks. The scope of this article is (a) to review the clinical applications of IR algorithms as well as their strengths, weaknesses, and methods of assessment and (b) to explore new CT image reconstruction and noise-reduction techniques that promise to facilitate radiation dose reduction. ^©RSNA, 2021.

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.

Scanner dependence of adaptive statistical iterative reconstruction with 3D noise power spectrum central frequency and noise magnitude ratios

PURPOSE

In this study, the noise reduction properties of the adaptive statistical iterative reconstruction (IR) on two different CT scanners of 64 and 256-slice were compared and their differences were assessed.

METHODS AND MATERIALS

The homogeneous module of the ACR CT phantom was scanned on the 64 and 256 slices CT scanners from the same vendor in the range of 15-40 mA. On each scanner, the data were reconstructed using filtered back projection (FBP) and at all strengths of IR with the STANDARD kernel. For each reconstruction, a 3D noise power spectrum (NPS) was calculated and the central frequency ratio in the xy plane (CFR_xy ), CFR in the z-direction (CFR_z ), and noise magnitude ratio (NMR) were derived. CFR is the central frequency ratio of NPS between the denoised image and the FBP image, and NMR is the ratio of the areas under the NPS curves. Ideally, both CFR_xy and CFR_z should be near 1, indicating minimal texture changes in both xy and z directions, while NMR should be as close to 0 as possible, indicating more noise reduction.

RESULTS

When comparing strengths with equivalent impact on noise texture, IR on the 64-slice reduced the noise magnitude in the xy plane more than that on the 256-slice. In the z-direction, the IR on the 256-slice produced a central frequency shift on the 256-slice but not on the 64-slice. In addition, the noise reduction effects of the IR on the 256-slice were affected when radiation exposure was below 2.0 mGy, but there was no observable dose-dependence on the 64-slice.

CONCLUSIONS

Our noise property analysis revealed that iterative reconstructions on different scanner platforms from the same vendor can be distinct, with unique effects on the noise texture and magnitude in CT images. The IR on a 64-slice scanner provides slightly enhanced noise reduction and maintains a noise reduction rate independent of dose, unlike the one on a 256-slice scanner. Notably, the IR on the 64-slice scanner was a 2D noise reduction technique (NRT), while the one on the 256-slice was a 3D NRT. These observations showcase the impact of different NRTs on clinical CT images, even when comparing the same NRT on different scanners.

Usefulness of Model-Based Iterative Reconstruction in Brain CT as Compared With Hybrid Iterative Reconstruction

OBJECTIVE

The aim of this study was to compare the contrast of gray to white matter between forward-projected model-based iterative reconstruction solution (FIRST) and hybrid iterative reconstruction (IR) by measuring computed tomography value of brain parenchyma.

METHODS

Computed tomography values of the gray and white matter in 15 areas of 21 patients (7 males, 14 females; average age, 49.5 ± 10.7 years) were measured and compared between FIRST and hybrid IR with filtered back projection (FBP) using 2 different reconstruction kernels FC21 and FC26.

RESULTS

The ratio of gray to white matter obtained using FIRST (1.25 ± 0.08) was significantly higher than that obtained using FBP with both kernel FC21 (1.13 ± 0.03) and kernel FC26 (1.22 ± 0.06).

CONCLUSIONS

FIRST increases the contrast between the gray and white matter, and decreases noise in brain computed tomography compared with FBP with hybrid IR.

Reduction of streak artifacts caused by low photon counts utilizing an image-based forward projection in computed tomography

BACKGROUND

The streak artifacts in computed tomography (CT) images caused by low photon counts are known to be effectively suppressed by raw-data-based techniques. This study aims to propose a technique to reduce the streak artifact without accessing the raw data.

METHODS

The proposed streak artifact reduction (SAR) technique consists of three steps: numerical forward projection to a CT image, adaptive filtering of the generated sinogram, and image reconstruction from the processed sinogram. The authors have expanded the two-dimensional method (2D-SAR) to three dimensions (3D-SAR) by using consecutive CT images. The modulation transfer function (MTF), the image noise (standard deviation), and the visibility of comb-shaped objects were evaluated at a low dose of 5 mGy. Using anthropomorphic abdominal and chest phantoms, CT images and the artifact index (AI) were compared between 3D-SAR and two types of iterative reconstruction (IR).

RESULTS

Sufficient artifact reductions associated with 54% and 61% reduction of noise for 2D- and 3D-SAR, respectively, were obtained in the phantom images, although the 50%MTF decreased by 28%. The visibility of the combs was improved with both the 2D- and 3D-SAR methods. The AI results of 3D-SAR were better than one type of IR and almost equal to the other type of IR, which was consistent with observed artifacts.

CONCLUSION

Both 2D-SAR and 3D-SAR have turned out to be effective in reducing streak artifacts. The proposed technique will be an effective tool since it needs no raw data, and thus can be applied to any CT images produced by a wide variety of CT systems.

Protocol Optimization Considerations for Implementing Deep Learning CT Reconstruction

OBJECTIVE. Previous advances over filtered back projection (FBP) have incorporated model-based iterative reconstruction. The purpose of this study was to characterize the latest advance in image reconstruction, that is, deep learning. The focus was on applying characterization results of a deep learning approach to decisions about clinical CT protocols. MATERIALS AND METHODS. A proprietary deep learning image reconstruction (DLIR) method was characterized against an existing advanced adaptive statistical iterative reconstruction method (ASIR-V) and FBP from the same vendor. The metrics used were contrast-to-noise ratio, spatial resolution as a function of contrast level, noise texture (i.e., noise power spectra [NPS]), noise scaling as a function of slice thickness, and CT number consistency. The American College of Radiology accreditation phantom and a uniform water phantom were used at a range of doses and slice thicknesses for both axial and helical acquisition modes. RESULTS. ASIR-V and DLIR were associated with improved contrast-to-noise ratio over FBP for all doses and slice thicknesses. No dose or contrast dependencies of spatial resolution were observed for ASIR-V or DLIR. NPS results showed DLIR maintained an FBP-like noise texture whereas ASIR-V shifted the NPS to lower frequencies. Noise changed with dose and slice thickness in the same manner for ASIR-V and FBP. DLIR slice thickness noise scaling differed from FBP, exhibiting less noise penalty with decreasing slice thickness. No clinically significant changes were observed in CT numbers for any measurement condition. CONCLUSION. In a phantom model, DLIR does not suffer from the concerns over reduction in spatial resolution and introduction of poor noise texture associated with previous methods.

Improved image quality in abdominal computed tomography reconstructed with a novel Deep Learning Image Reconstruction technique - initial clinical experience

BACKGROUND

A novel Deep Learning Image Reconstruction (DLIR) technique for computed tomography has recently received clinical approval.

PURPOSE

To assess image quality in abdominal computed tomography reconstructed with DLIR, and compare with standardly applied iterative reconstruction.

MATERIAL AND METHODS

Ten abdominal computed tomography scans were reconstructed with iterative reconstruction and DLIR of medium and high strength, with 0.625 mm and 2.5 mm slice thickness. Image quality was assessed using eight visual grading criteria in a side-by-side comparative setting. All series were presented twice to evaluate intraobserver agreement. Reader scores were compared using univariate logistic regression. Image noise and contrast-to-noise ratio were calculated for quantitative analyses.

RESULTS

For 2.5 mm slice thickness, DLIR images were more frequently perceived as equal or better than iterative reconstruction across all visual grading criteria (for both DLIR of medium and high strength, p < 0.001). Correspondingly, DLIR images were more frequently perceived as better (as opposed to equal or in favor of iterative reconstruction) for visual reproduction of liver parenchyma, intrahepatic vascular structures as well as overall impression of image noise and texture (p < 0.001). This improved image quality was also observed for 0.625 mm slice images reconstructed with DLIR of high strength when directly comparing to traditional iterative reconstruction in 2.5 mm slices. Image noise was significantly lower and contrast-to-noise ratio measurements significantly higher for images reconstructed with DLIR compared to iterative reconstruction (p < 0.01).

CONCLUSIONS

Abdominal computed tomography images reconstructed using a DLIR technique shows improved image quality when compared to standardly applied iterative reconstruction across a variety of clinical image quality criteria.

Deep learning image reconstruction for improvement of image quality of abdominal computed tomography: comparison with hybrid iterative reconstruction

PURPOSE

To evaluate the usefulness of the deep learning image reconstruction (DLIR) to enhance the image quality of abdominal CT, compared to iterative reconstruction technique.

METHOD

Pre and post-contrast abdominal CT images in 50 patients were reconstructed with 2 different algorithms: hybrid iterative reconstruction (hybrid IR: ASiR-V 50%) and DLIR (TrueFidelity). Standard deviation of attenuation in normal liver parenchyma was measured as the image noise on pre and post-contrast CT. The contrast-to-noise ratio (CNR) for the aorta, and the signal-to-noise ratio (SNR) of the liver were calculated on post-contrast CT. The overall image quality was graded on a 5-point scale ranging from 1 (poor) to 5 (excellent).

RESULTS

The image noise was significantly decreased by DLIR compared to hybrid-IR [hybrid IR, median 8.3 Hounsfield unit (HU) (interquartile range (IQR) 7.6-9.2 HU); DLIR, median 5.2 HU (IQR 4.6-5.8), P < 0.0001 for post-contrast CT]. The CNR and SNR were significantly improved by DLIR [CNR, median 4.5 (IQR 3.8-5.6) vs 7.3 (IQR 6.2-8.8), P < 0.0001; SNR, median 9.4 (IQR 8.3-10.1) vs 15.0 (IQR 13.2-16.4), P < 0.0001]. The overall image quality score was also higher for DLIR compared to hybrid-IR (hybrid IR 3.1 ± 0.6 vs DLIR 4.6 ± 0.5, P < 0.0001 for post-contrast CT).

CONCLUSIONS

Image noise, overall image quality, CNR and SNR for abdominal CT images are improved with DLIR compared to hybrid IR.

Performance of clinically available deep learning image reconstruction in computed tomography: a phantom study

Purpose: To assess the physical performance of deep learning image reconstruction (DLIR) compared with those of filtered back projection (FBP) and iterative reconstruction (IR) and to estimate the dose reduction potential of the technique. Approach: A cylindrical water bath phantom with a diameter of 300 mm including two rods composed of acrylic and soft tissue-equivalent material was scanned using a clinical computed tomography (CT) scanner at four dose levels (CT dose index of 20, 15, 10, and 5 mGy). Phantom images were reconstructed using FBP, DLIR, and IR. The in-plane and z axis task transfer functions (TTFs) and in-plane noise power spectrum (NPS) were measured. The dose reduction potential was estimated by evaluating the system performance function calculated from TTF and NPS. The visibilities of a bar pattern phantom placed in the same water bath phantom were compared. Results: The use of DLIR resulted in a notable decrease in noise magnitude. The shift in peak NPS frequency was reduced compared with IR. Preservation of in-plane TTF was superior using DLIR than using IR. The estimated dose reduction potentials of DLIR and IR were 39% to 54% and 19% to 29%, respectively. However, the z axis resolution was decreased with DLIR by 6% to 21% compared with FBP. The bar pattern visibilities were approximately consistent with the TTF results in both planes. Conclusions: The in-plane edge-preserving noise reduction performance of DLIR is superior to that of IR. Moreover, DLIR enables approximately half-dose acquisitions with no deterioration in noise texture in cases that permit some z axis resolution reduction.

Low-Dose Abdominal CT Using a Deep Learning-Based Denoising Algorithm: A Comparison with CT Reconstructed with Filtered Back Projection or Iterative Reconstruction Algorithm

Objective

To compare the image quality of low-dose (LD) computed tomography (CT) obtained using a deep learning-based denoising algorithm (DLA) with LD CT images reconstructed with a filtered back projection (FBP) and advanced modeled iterative reconstruction (ADMIRE).

Materials and Methods

One hundred routine-dose (RD) abdominal CT studies reconstructed using FBP were used to train the DLA. Simulated CT images were made at dose levels of 13%, 25%, and 50% of the RD (DLA-1, -2, and -3) and reconstructed using FBP. We trained DLAs using the simulated CT images as input data and the RD CT images as ground truth. To test the DLA, the American College of Radiology CT phantom was used together with 18 patients who underwent abdominal LD CT. LD CT images of the phantom and patients were processed using FBP, ADMIRE, and DLAs (LD-FBP, LD-ADMIRE, and LD-DLA images, respectively). To compare the image quality, we measured the noise power spectrum and modulation transfer function (MTF) of phantom images. For patient data, we measured the mean image noise and performed qualitative image analysis. We evaluated the presence of additional artifacts in the LD-DLA images.

Results

LD-DLAs achieved lower noise levels than LD-FBP and LD-ADMIRE for both phantom and patient data (all p < 0.001). LD-DLAs trained with a lower radiation dose showed less image noise. However, the MTFs of the LD-DLAs were lower than those of LD-ADMIRE and LD-FBP (all p < 0.001) and decreased with decreasing training image dose. In the qualitative image analysis, the overall image quality of LD-DLAs was best for DLA-3 (50% simulated radiation dose) and not significantly different from LD-ADMIRE. There were no additional artifacts in LD-DLA images.

Conclusion

DLAs achieved less noise than FBP and ADMIRE in LD CT images, but did not maintain spatial resolution. The DLA trained with 50% simulated radiation dose showed the best overall image quality.

Noise reduction in CT image using prior knowledge aware iterative denoising

The clinical demand for low image noise often limits the slice thickness used in many CT applications. However, a thick-slice image is more susceptible to longitudinal partial volume effects, which can blur key anatomic structures and pathologies of interest. In this work, we develop a prior-knowledge-aware iterative denoising (PKAID) framework that utilizes spatial data redundancy in the slice increment direction to generate low-noise, thin-slice images, and demonstrate its application in non-contrast head CT exams. The proposed technique takes advantage of the low-noise of thicker images and exploits the structural similarity between the thick- and thin-slice images to reduce noise in the thin-slice image. Phantom data and patient cases (n=3) of head CT were used to assess performance of this method. Images were reconstructed at clinically-utilized slice thickness (5 mm) and thinner slice thickness (2 mm). PKAID was used to reduce image noise in 2 mm images using the 5 mm images as low-noise prior. Noise amplitude, noise power spectra (NPS), modulation transfer function (MTF), and slice sensitivity profiles (SSP) of images before/after denoising were analyzed. The NPS and MTF analysis showed that PKAID preserved noise texture and resolution of the original thin-slice image, while reducing noise to the level of thick-slice image. The SSP analysis showed that the slice thickness of the original thin-slice image was retained. Patient examples demonstrated that PKAID-processed, thin-slice images better delineated brain structures and key pathologies such as subdural hematoma compared to the clinical 5 mm images, while additionally reducing image noise. To test an alternative PKAID utilization for dose reduction, a head exam with 40% dose reduction was simulated using projection-domain noise insertion. The image of 5 mm slice thickness was then denoised using PKAID. The results showed that the PKAID-processed reduced-dose images maintained similar noise and image quality compared to the full-dose images.

Evaluation of Image Quality for 7 Iterative Reconstruction Algorithms in Chest Computed Tomography Imaging: A Phantom Study

OBJECTIVES

This study aimed to evaluate the image quality of 7 iterative reconstruction (IR) algorithms in comparison to filtered back-projection (FBP) algorithm.

METHODS

An anthropomorphic chest phantom was scanned on 4 computed tomography scanners and reconstructed with FBP and IR algorithms. Image quality of anatomical details-large/medium-sized pulmonary vessels, small pulmonary vessels, thoracic wall, and small and large lesions-was scored. Furthermore, general impression of noise, image contrast, and artifacts were evaluated. Visual grading regression was used to analyze the data. Standard deviations were measured, and the noise power spectrum was calculated.

RESULTS

Iterative reconstruction algorithms showed significantly better results when compared with FBP for these criteria (regression coefficients/P values in parentheses): vessels (FIRST: -1.8/0.05, AIDR Enhanced: <-2.3/0.01, Veo: <-0.1/0.03, ADMIRE: <-2.1/0.04), lesions (FIRST: <-2.6/0.01, AIDR Enhanced: <-1.9/0.03, IMR1: <-2.7/0.01, Veo: <-2.4/0.02, ADMIRE: -2.3/0.02), image noise (FIRST: <-3.2/0.004, AIDR Enhanced: <-3.5/0.002, IMR1: <-6.1/0.001, iDose: <-2.3/0.02, Veo: <-3.4/0.002, ADMIRE: <-3.5/0.02), image contrast (FIRST: -2.3/0.01, AIDR Enhanced: -2.5/0.01, IMR1: -3.7/0.001, iDose: -2.1/0.02), and artifacts (FIRST: <-3.8/0.004, AIDR Enhanced: <-2.7/0.02, IMR1: <-2.6/0.02, iDose: -2.1/0.04, Veo: -2.6/0.02). The iDose algorithm was the only IR algorithm that maintained the noise frequencies.

CONCLUSIONS

Iterative reconstruction algorithms performed differently on all evaluated criteria, showing the importance of careful implementation of algorithms for diagnostic purposes.

Size and volume of kidney stones in computed tomography: Influence of acquisition techniques and image reconstruction parameters

PURPOSE

Computed tomography (CT) is routinely used to assess suspected urolithiasis. Information obtained from CT include presence, location and size of stones, with the latter frequently determining treatment strategy. While there is consensus regarding measurements procedures of kidney stones, influence of radiation dose and reconstruction techniques on stone measurements are unknown. The purpose of this study was to systematically evaluate the influence of these technical determinants on kidney stone size measurements.

METHOD

47 kidney stones of different composition were scanned using a 64-row-multi-detector CT in a 3D-printed, semi-anthropomorphic phantom. Reference stone sizes were measured manually with a digital caliper (Man-M). Stones were imaged with 2 and 10 mGy CTDI. Images were reconstructed using filtered-back-projection, hybrid-iterative and model-based-iterative reconstruction algorithms (FBP, HIR, MBIR) in combination with different kernels and denoising levels. All stones underwent semi-automatic, threshold-based segmentation for computation of maximum diameter and volume. Statistics were conducted using ANOVA ± correction for multiple comparisons.

RESULTS

Overall stone size as compared to manual measurements was overestimated in CT (10.0 ± 3.1 vs. 8.8 ± 2.9 mm, p < 0.05) yet showing a good correlation (R² = 0.66). Radiation dose and denoising levels did not significantly influence measurements (p > 0.05). MBIR and sharp kernels showed closest agreement with Man-M (9.3 ± 3.1 vs. 8.8 ± 2.9 mm, p < 0.05). Differences within single stones were as high as 40 % (e.g. Man-M: 5.9 mm, CT: 7.3-12.0 mm).

CONCLUSIONS

CT-based measurements of kidney stone size appear unaffected by radiation dose and denoising technique, whereas reconstruction algorithms and kernels demonstrate a relevant impact on size measurements. Smallest differences were found using MBIR with a sharp kernel.

Protocol Optimization for Renal Mass Detection and Characterization

Renal masses increasingly are found incidentally, largely due to the frequent use of medical imaging. Computed tomography (CT) and MR imaging are mainstays for renal mass characterization, presurgical planning of renal tumors, and surveillance after surgery or systemic therapy for advanced renal cell carcinomas. CT protocols should be tailored to different clinical indications, balancing diagnostic accuracy and radiation exposure. MR imaging protocols should take advantage of the improved soft tissue contrast for renal tumor diagnosis and staging. Optimized imaging protocols enable analysis of imaging features that help narrow the differential diagnoses and guide management in patients with renal masses.

Evaluation of three-dimensional iterative image reconstruction in virtual monochromatic imaging at 40 kilo-electron volts: phantom and clinical studies to assess the image noise and image quality in comparison with other reconstruction techniques

OBJECTIVE

The purpose of this study was to evaluate the image quality in virtual monochromatic imaging (VMI) at 40 kilo-electron volts (keV) with three-dimensional iterative image reconstruction (3D-IIR).

METHODS

A phantom study and clinical study (31 patients) were performed with dual-energy CT (DECT). VMI at 40 keV was obtained and the images were reconstructed using filtered back projection (FBP), 50% adaptive statistical iterative reconstruction (ASiR), and 3D-IIR. We conducted subjective and objective evaluations of the image quality with each reconstruction technique.

RESULTS

The image contrast-to-noise ratio and image noise in both the clinical and phantom studies were significantly better with 3D-IIR than with 50% ASiR, and with 50% ASiR than with FBP (all, p < 0.05). The standard deviation and noise power spectra of the reconstructed images decreased in the order of 3D-IIR to 50% ASiR to FBP, while the modulation transfer function was maintained across the three reconstruction techniques. In most subjective evaluations in the clinical study, the image quality was significantly better with 3D-IIR than with 50% ASiR, and with 50% ASiR than with FBP (all, p < 0.001). Regarding the diagnostic acceptability, all images using 3D-IIR were evaluated as being fully or probably acceptable.

CONCLUSIONS

The quality of VMI at 40 keV is improved by 3D-IIR, which allows the image noise to be reduced and structural details to be maintained.

ADVANCES IN KNOWLEDGE

The improvement of the image quality of VMI at 40 keV by 3D-IIR may increase the subjective acceptance in the clinical setting.

Spatial resolution compensation by adjusting the reconstruction kernels for iterative reconstruction images of computed tomography

PURPOSE

Hybrid iterative reconstruction (IR) is useful to reduce noise in computed tomography (CT) images. However, it often decreases the spatial resolution. The ability of high spatial resolution kernels (harder kernels) to compensate for the decrease in the spatial resolution of hybrid IRs was investigated.

METHODS

An elliptic cylindrical phantom simulating an adult abdomen was used. Two types of rod-shaped objects with ~330 and ~130 HU were inserted to simulate contrasts of arteries in CT angiography. Two multi-slice CT systems were used to scan the phantoms with 120 kVp and scan doses of 20 and 10 mGy. The task transfer functions (TTFs) were measured from the circular edges of the rod images. The noise power spectrum (NPS) was measured from the images of the water-only section. The CT images were reconstructed using a filtered back projection (FBP) with baseline kernels and two levels of hybrid IRs with harder kernels. The profiles of the clinical images across the aortic dissection flaps were measured to evaluate actual spatial resolutions.

RESULTS

The TTF degradation of each hybrid IR was recovered by the harder kernels, whereas the noise reduction effect was retained, for both the 20 and 10 mGy. The profiles of the dissection flaps for the FBP were maintained by using the harder kernels. Even with the best combination of hybrid IR and harder kernel, the noise level at 10 mGy was not reduced to the level of FBP at 20 mGy, suggesting no capability of a 50% dose reduction while maintaining noise.

Image Quality Assessment of Abdominal CT by Use of New Deep Learning Image Reconstruction: Initial Experience

OBJECTIVE. The purpose of this study was to perform quantitative and qualitative evaluation of a deep learning image reconstruction (DLIR) algorithm in contrast-enhanced oncologic CT of the abdomen. MATERIALS AND METHODS. Retrospective review (April-May 2019) of the cases of adults undergoing oncologic staging with portal venous phase abdominal CT was conducted for evaluation of standard 30% adaptive statistical iterative reconstruction V (30% ASIR-V) reconstruction compared with DLIR at low, medium, and high strengths. Attenuation and noise measurements were performed. Two radiologists, blinded to examination details, scored six categories while comparing reconstructions for overall image quality, lesion diagnostic confidence, artifacts, image noise and texture, lesion conspicuity, and resolution. RESULTS. DLIR had a better contrast-to-noise ratio than 30% ASIR-V did; high-strength DLIR performed the best. High-strength DLIR was associated with 47% reduction in noise, resulting in a 92-94% increase in contrast-to-noise ratio compared with that of 30% ASIR-V. For overall image quality and image noise and texture, DLIR scored significantly higher than 30% ASIR-V with significantly higher scores as DLIR strength increased. A total of 193 lesions were identified. The lesion diagnostic confidence, conspicuity, and artifact scores were significantly higher for all DLIR levels than for 30% ASIR-V. There was no significant difference in perceived resolution between the reconstruction methods. CONCLUSION. Compared with 30% ASIR-V, DLIR improved CT evaluation of the abdomen in the portal venous phase. DLIR strength should be chosen to balance the degree of desired denoising for a clinical task relative to mild blurring, which increases with progressively higher DLIR strengths.