Task-based characterization of a deep learning image reconstruction and comparison with filtered back-projection and a partial model-based iterative reconstruction in abdominal CT: A phantom study

Super-resolution deep learning image reconstruction: image quality and myocardial homogeneity in coronary computed tomography angiography

BACKGROUND

The recently introduced super-resolution (SR) deep learning image reconstruction (DLR) is potentially effective in reducing noise level and enhancing the spatial resolution. We aimed to investigate whether SR-DLR has advantages in the overall image quality and intensity homogeneity on coronary computed tomography (CT) angiography with four different approaches: filtered-back projection (FBP), hybrid iterative reconstruction (IR), DLR, and SR-DLR.

METHODS

Sixty-three patients (mean age, 61 ± 11 years; range, 18-81 years; 40 men) who had undergone coronary CT angiography between June and October 2022 were retrospectively included. Image noise, signal to noise ratio, and contrast to noise ratio were quantified in both proximal and distal segments of the major coronary arteries. The left ventricle myocardium contrast homogeneity was analyzed. Two independent reviewers scored overall image quality, image noise, image sharpness, and myocardial homogeneity.

RESULTS

Image noise in Hounsfield units (HU) was significantly lower (P < 0.001) for the SR-DLR (11.2 ± 2.0 HU) compared to those associated with other image reconstruction methods including FBP (30.5 ± 10.5 HU), hybrid IR (20.0 ± 5.4 HU), and DLR (14.2 ± 2.5 HU) in both proximal and distal segments. SR-DLR significantly improved signal to noise ratio and contrast to noise ratio in both the proximal and distal segments of the major coronary arteries. No significant difference was observed in the myocardial CT attenuation with SR-DLR among different segments of the left ventricle myocardium (P = 0.345). Conversely, FBP and hybrid IR resulted in inhomogeneous myocardial CT attenuation (P < 0.001). Two reviewers graded subjective image quality with SR-DLR higher than other image reconstruction techniques (P < 0.001).

CONCLUSIONS

SR-DLR improved image quality, demonstrated clearer delineation of distal segments of coronary arteries, and was seemingly accurate for quantifying CT attenuation in the myocardium.

Optimizing computed tomography image reconstruction for focal hepatic lesions: Deep learning image reconstruction vs iterative reconstruction

Background

Deep learning image reconstruction (DLIR) is a novel computed tomography (CT) reconstruction technique that minimizes image noise, enhances image quality, and enables radiation dose reduction. This study aims to compare the diagnostic performance of DLIR and iterative reconstruction (IR) in the evaluation of focal hepatic lesions.

Methods

We conducted a retrospective study of 216 focal hepatic lesions in 109 adult participants who underwent abdominal CT scanning at our institution. We used DLIR (low, medium, and high strength) and IR (0 %, 10 %, 20 %, and 30 %) techniques for image reconstruction. Four experienced abdominal radiologists independently evaluated focal hepatic lesions based on five qualitative aspects (lesion detectability, lesion border, diagnostic confidence level, image artifact, and overall image quality). Quantitatively, we measured and compared the level of image noise for each technique at the liver and aorta.

Results

There were significant differences (p < 0.001) among the seven reconstruction techniques in terms of lesion borders, image artifacts, and overall image quality. Low-strength DLIR (DLIR-L) exhibited the best overall image quality. Although high-strength DLIR (DLIR-H) had the least image noise and fewest artifacts, it also had the lowest scores for lesion borders and overall image quality. Image noise showed a weak to moderate positive correlation with participants' body mass index and waist circumference.

Conclusions

The optimal-strength DLIR significantly improved overall image quality for evaluating focal hepatic lesions compared to the IR technique. DLIR-L achieved the best overall image quality while maintaining acceptable levels of image noise and quality of lesion borders.

Low-contrast lesion detection in neck CT: a multireader study comparing deep learning, iterative, and filtered back projection reconstructions using realistic phantoms

BACKGROUND

Computed tomography (CT) reconstruction algorithms can improve image quality, especially deep learning reconstruction (DLR). We compared DLR, iterative reconstruction (IR), and filtered back projection (FBP) for lesion detection in neck CT.

METHODS

Nine patient-mimicking neck phantoms were examined with a 320-slice scanner at six doses: 0.5, 1, 1.6, 2.1, 3.1, and 5.2 mGy. Each of eight phantoms contained one circular lesion (diameter 1 cm; contrast -30 HU to the background) in the parapharyngeal space; one phantom had no lesions. Reconstruction was made using FBP, IR, and DLR. Thirteen readers were tasked with identifying and localizing lesions in 32 images with a lesion and 20 without lesions for each dose and reconstruction algorithm. Receiver operating characteristic (ROC) and localization ROC (LROC) analysis were performed.

RESULTS

DLR improved lesion detection with ROC area under the curve (AUC) 0.724 ± 0.023 (mean ± standard error of the mean) using DLR versus 0.696 ± 0.021 using IR (p = 0.037) and 0.671 ± 0.023 using FBP (p < 0.001). Likewise, DLR improved lesion localization, with LROC AUC 0.407 ± 0.039 versus 0.338 ± 0.041 using IR (p = 0.002) and 0.313 ± 0.044 using FBP (p < 0.001). Dose reduction to 0.5 mGy compromised lesion detection in FBP-reconstructed images compared to doses ≥ 2.1 mGy (p ≤ 0.024), while no effect was observed with DLR or IR (p ≥ 0.058).

CONCLUSION

DLR improved the detectability of lesions in neck CT imaging. Dose reduction to 0.5 mGy maintained lesion detectability when denoising reconstruction was used.

RELEVANCE STATEMENT

Deep learning enhances lesion detection in neck CT imaging compared to iterative reconstruction and filtered back projection, offering improved diagnostic performance and potential for x-ray dose reduction.

KEY POINTS

Low-contrast lesion detectability was assessed in anatomically realistic neck CT phantoms. Deep learning reconstruction (DLR) outperformed filtered back projection and iterative reconstruction. Dose has little impact on lesion detectability against anatomical background structures.

[Validation of Optimal Imaging Conditions for Coronary Computed Tomography Angiography Using High-definition Mode and Deep Learning Image Reconstruction Algorithm]

PURPOSE

To verify the optimal imaging conditions for coronary computed tomography angiography (CCTA) examinations when using high-definition (HD) mode and deep learning image reconstruction (DLIR) in combination.

METHOD

A chest phantom and an in-house phantom using 3D printer were scanned with a 256-row detector CT scanner. The scan parameters were as follows - acquisition mode: ON (HD mode) and OFF (normal resolution [NR] mode), rotation time: 0.28 s/rotation, beam coverage width: 160 mm, and the radiation dose was adjusted based on CT-AEC. Image reconstruction was performed using ASiR-V (Hybrid-IR), TrueFidelity Image (DLIR), and HD-Standard (HD mode) and Standard (NR mode) reconstruction kernels. The task-based transfer function (TTF) and noise power spectrum (NPS) were measured for image evaluation, and the detectability index (d') was calculated. Visual evaluation was also performed on an in-house coronary phantom.

RESULT

The in-plane TTF was better for the HD mode than for the NR mode, while the z-axis TTF was lower for DLIR than for Hybrid-IR. The NPS values in the high-frequency region were higher for the HD mode compared to those for the NR mode, and the NPS was lower for DLIR than for Hybrid-IR. The combination of HD mode and DLIR showed the best value for in-plane d', whereas the combination of NR mode and DLIR showed the best value for z-axis d'. In the visual evaluation, the combination of NR mode and DLIR showed the best values from a noise index of 45 HU.

CONCLUSION

The optimal combination of HD mode and DLIR depends on the image noise level, and the combination of NR mode and DLIR was the best imaging condition under noisy conditions.

Optimization of CT pulmonary angiography for pulmonary embolism using task-based image quality assessment and diagnostic reference levels: A multicentric study

PURPOSE

To establish size-specific diagnostic reference levels (DRLs) for pulmonary embolism (PE) based on patient CT examinations performed on 74 CT devices. To assess task-based image quality (IQ) for each device and to investigate the variability of dose and IQ across different CTs. To propose a dose/IQ optimization.

METHODS

1051 CT pulmonary angiography dose data were collected. DRLs were calculated as the 75th percentile of CT dose index (CTDI) for two patient categories based on the thoracic perimeters. IQ was assessed with two thoracic phantom sizes using local acquisition parameters and three other dose levels. The area under the ROC curve (AUC) of a 2 mm low perfused vessel was assessed with a non-prewhitening with eye-filter model observer. The optimal IQ-dose point was mathematically assessed from the relationship between IQ and dose.

RESULTS

The DRLs of CTDIvol were 6.4 mGy and 10 mGy for the two patient categories. 75th percentiles of phantom CTDIvol were 6.3 mGy and 10 mGy for the two phantom sizes with inter-quartile AUC values of 0.047 and 0.066, respectively. After the optimization, 75th percentiles of phantom CTDIvol decreased to 5.9 mGy and 7.55 mGy and the interquartile AUC values were reduced to 0.025 and 0.057 for the two phantom sizes.

CONCLUSION

DRLs for PE were proposed as a function of patient thoracic perimeters. This study highlights the variability in terms of dose and IQ. An optimization process can be started individually and lead to a harmonization of practice throughout multiple CT sites.

Gated cardiac CT in infants: What can we expect from deep learning image reconstruction algorithm?

BACKGROUND

ECG-gated cardiac CT is now widely used in infants with congenital heart disease (CHD). Deep Learning Image Reconstruction (DLIR) could improve image quality while minimizing the radiation dose.

OBJECTIVES

To define the potential dose reduction using DLIR with an anthropomorphic phantom.

METHOD

An anthropomorphic pediatric phantom was scanned with an ECG-gated cardiac CT at four dose levels. Images were reconstructed with an iterative and a deep-learning reconstruction algorithm (ASIR-V and DLIR). Detectability of high-contrast vessels were computed using a mathematical observer. Discrimination between two vessels was assessed by measuring the CT spatial resolution. The potential dose reduction while keeping a similar level of image quality was assessed.

RESULTS

DLIR-H enhances detectability by 2.4% and discrimination performances by 20.9% in comparison with ASIR-V 50. To maintain a similar level of detection, the dose could be reduced by 64% using high-strength DLIR in comparison with ASIR-V50.

CONCLUSION

DLIR offers the potential for a substantial dose reduction while preserving image quality compared to ASIR-V.

A CT deep learning reconstruction algorithm: Image quality evaluation for brain protocol at decreasing dose indexes in comparison with FBP and statistical iterative reconstruction algorithms

PURPOSE

To characterise the impact of Precise Image (PI) deep learning reconstruction algorithm on image quality, compared to filtered back-projection (FBP) and iDose4 iterative reconstruction for brain computed tomography (CT) phantom images.

METHODS

Catphan-600 phantom was acquired with an Incisive CT scanner using a dedicated brain protocol, at six different dose levels (volume computed tomography dose index (CTDIvol): 7/14/29/49/56/67 mGy). Images were reconstructed using FBP, levels 2/5 of iDose4, and PI algorithm (Sharper/Sharp/Standard/Smooth/Smoother). Image quality was assessed by evaluating CT numbers, image histograms, noise, image non-uniformity (NU), noise power spectrum, target transfer function, and detectability index.

RESULTS

The five PI levels did not significantly affect the mean CT number. For a given CTDIvol using Sharper-to-Smoother levels, the spatial resolution for all the investigated materials and the detectability index increased while the noise magnitude decreased, slightly affecting noise texture. For a fixed PI level increasing the CTDIvol the detectability index increased, the noise magnitude decreased. From 29 mGy, NU values converged within 1 Hounsfield Unit from each other without a substantial improvement at higher CTDIvol values.

CONCLUSIONS

The improved performances of intermediate PI levels in brain protocols compared to conventional algorithms seem to suggest a potential reduction of CTDIvol.

Performance evaluation of deep learning image reconstruction algorithm for dual-energy spectral CT imaging: A phantom study

OBJECTIVES

To evaluate the performance of deep learning image reconstruction (DLIR) algorithm in dual-energy spectral CT (DEsCT) as a function of radiation dose and image energy level, in comparison with filtered-back-projection (FBP) and adaptive statistical iterative reconstruction-V (ASIR-V) algorithms.

METHODS

An ACR464 phantom was scanned with DEsCT at four dose levels (3.5 mGy, 5 mGy, 7.5 mGy, and 10 mGy). Virtual monochromatic images were reconstructed at five energy levels (40 keV, 50 keV, 68 keV, 74 keV, and 140 keV) using FBP, 50% and 100% ASIR-V, DLIR at low (DLIR-L), medium (DLIR-M), and high (DLIR-H) settings. The noise power spectrum (NPS), task-based transfer function (TTF) and detectability index (d') were computed and compared among reconstructions.

RESULTS

NPS area and noise increased as keV decreased, with DLIR having slower increase than FBP and ASIR-V, and DLIR-H having the lowest values. DLIR had the best 40 keV/140 keV noise ratio at various energy levels, DLIR showed higher TTF (50%) than ASIR-V for all materials, especially for the soft tissue-like polystyrene insert, and DLIR-M and DLIR-H provided higher d' than DLIR-L, ASIR-V and FBP in all dose and energy levels. As keV increases, d' increased for acrylic insert, and d' of the 50 keV DLIR-M and DLIR-H images at 3.5 mGy (7.39 and 8.79, respectively) were higher than that (7.20) of the 50 keV ASIR-V50% images at 10 mGy.

CONCLUSIONS

DLIR provides better noise containment for low keV images in DEsCT and higher TTF(50%) for the polystyrene insert over ASIR-V. DLIR-H has the lowest image noise and highest detectability in all dose and energy levels. DEsCT 50 keV images with DLIR-M and DLIR-H show potential for 65% dose reduction over ASIR-V50% withhigher d'.

Synthetization of high-dose images using low-dose CT scans

BACKGROUND

Radiation dose reduction has been the focus of many research activities in x-ray CT. Various approaches were taken to minimize the dose to patients, ranging from the optimization of clinical protocols, refinement of the scanner hardware design, and development of advanced reconstruction algorithms. Although significant progress has been made, more advancements in this area are needed to minimize the radiation risks to patients.

PURPOSE

Reconstruction algorithm-based dose reduction approaches focus mainly on the suppression of noise in the reconstructed images while preserving detailed anatomical structures. Such an approach effectively produces synthesized high-dose images (SHD) from the data acquired with low-dose scans. A representative example is the model-based iterative reconstruction (MBIR). Despite its widespread deployment, its full adoption in a clinical environment is often limited by an undesirable image texture. Recent studies have shown that deep learning image reconstruction (DLIR) can overcome this shortcoming. However, the limited availability of high-quality clinical images for training and validation is often the bottleneck for its development. In this paper, we propose a novel approach to generate SHD with existing low-dose clinical datasets that overcomes both the noise texture issue and the data availability issue.

METHODS

Our approach is based on the observation that noise in the image can be effectively reduced by performing image processing orthogonal to the imaging plane. This process essentially creates an equivalent thick-slice image (TSI), and the characteristics of TSI depend on the nature of the image processing. An advantage of this approach is its potential to reduce impact on the noise texture. The resulting image, however, is likely corrupted by the anatomical structural degradation due to partial volume effects. Careful examination has shown that the differential signal between the original and the processed image contains sufficient information to identify regions where anatomical structures are modified. The differential signal, unfortunately, contains significant noise and has to be removed. The noise removal can be accomplished by performing iterative noise reduction to preserve structural information. The processed differential signal is subsequently subtracted from TSI to arrive at SHD.

RESULTS

The algorithm was evaluated extensively with phantom and clinical datasets. For better visual inspection, difference images between the original and SHD were generated and carefully examined. Negligible residual structure could be observed. In addition to the qualitative inspection, quantitative analyses were performed on clinical images in terms of the CT number consistency and the noise reduction characteristics. Results indicate that no CT number bias is introduced by the proposed algorithm. In addition, noise reduction capability is consistent across different patient anatomical regions. Further, simulated water phantom scans were utilized in the generation of the noise power spectrum (NPS) to demonstrate the preservation of the noise-texture.

CONCLUSIONS

We present a method to generate SHD datasets from regularly acquired low-dose CT scans. Images produced with the proposed approach exhibit excellent noise-reduction with the desired noise-texture. Extensive clinical and phantom studies have demonstrated the efficacy and robustness of our approach. Potential limitations of the current implementation are discussed and further research topics are outlined.

Physical characteristics of deep learning-based image processing software in computed tomography: a phantom study

PURPOSE

This study aimed to assess the image characteristics of deep-learning-based image processing software (DLIP; FCT PixelShine, FUJIFILM, Tokyo, Japan) and compare it with filtered back projection (FBP), model-based iterative reconstruction (MBIR), and deep-learning-based reconstruction (DLR).

METHODS

This phantom study assessed the object-specific spatial resolution (task-based transfer function [TTF]), noise characteristics (noise power spectrum [NPS]), and low-contrast detectability (low-contrast object-specific contrast-to-noise ratio [CNRLO]) at three different output doses (standard: 10 mGy; low: 3.9 mGy; ultralow: 2.0 mGy). The processing strength of DLIPFBP with A1, A4, and A9 was compared with those of FBP, MBIR, and DLR.

RESULT

The standard dose with high-contrast TTFs of DLIPFBP exceeded that of FBP. Low-contrast TTFs were comparable to or lower than that of FBP. The NPS peak frequency (fP) of DLIPFBP shifts to low spatial frequencies of up to 8.6% at ultralow doses compared to the standard FBP dose. MBIR shifted the most fP compared to FBP-a marked shift of up to 49%. DLIPFBP showed a CNRLO equal to or greater than that of DLR in standard or low doses. In contrast, the CNRLO of the DLIPFBP was equal to or lower than that of the DLR in ultralow doses.

CONCLUSION

DLIPFBP reduced image noise while maintaining a resolution similar to commercially available MBIR and DLR. The slight spatial frequency shift of fP in DLIPFBP contributed to the noise texture degradation suppression. The NPS suppression in the low spatial frequency range effectively improved the low-contrast detectability.

Improving image quality with super-resolution deep-learning-based reconstruction in coronary CT angiography

OBJECTIVES

To evaluate the effect of super-resolution deep-learning-based reconstruction (SR-DLR) on the image quality of coronary CT angiography (CCTA).

METHODS

Forty-one patients who underwent CCTA using a 320-row scanner were retrospectively included. Images were reconstructed with hybrid (HIR), model-based iterative reconstruction (MBIR), normal-resolution deep-learning-based reconstruction (NR-DLR), and SR-DLR algorithms. For each image series, image noise, and contrast-to-noise ratio (CNR) at the left main trunk, right coronary artery, left anterior descending artery, and left circumflex artery were quantified. Blooming artifacts from calcified plaques were measured. Image sharpness, noise magnitude, noise texture, edge smoothness, overall quality, and delineation of the coronary wall, calcified and noncalcified plaques, cardiac muscle, and valves were subjectively ranked on a 4-point scale (1, worst; 4, best). The quantitative parameters and subjective scores were compared among the four reconstructions. Task-based image quality was assessed with a physical evaluation phantom. The detectability index for the objects simulating the coronary lumen, calcified plaques, and noncalcified plaques was calculated from the noise power spectrum (NPS) and task-based transfer function (TTF).

RESULTS

SR-DLR yielded significantly lower image noise and blooming artifacts with higher CNR than HIR, MBIR, and NR-DLR (all p < 0.001). The best subjective scores for all the evaluation criteria were attained with SR-DLR, with significant differences from all other reconstructions (p < 0.001). In the phantom study, SR-DLR provided the highest NPS average frequency, TTF50%, and detectability for all task objects.

CONCLUSION

SR-DLR considerably improved the subjective and objective image qualities and object detectability of CCTA relative to HIR, MBIR, and NR-DLR algorithms.

CLINICAL RELEVANCE STATEMENT

The novel SR-DLR algorithm has the potential to facilitate accurate assessment of coronary artery disease on CCTA by providing excellent image quality in terms of spatial resolution, noise characteristics, and object detectability.

KEY POINTS

• SR-DLR designed for CCTA improved image sharpness, noise property, and delineation of cardiac structures with reduced blooming artifacts from calcified plaques relative to HIR, MBIR, and NR-DLR. • In the task-based image-quality assessments, SR-DLR yielded better spatial resolution, noise property, and detectability for objects simulating the coronary lumen, coronary calcifications, and noncalcified plaques than other reconstruction techniques. • The image reconstruction times of SR-DLR were shorter than those of MBIR, potentially serving as a novel standard-of-care reconstruction technique for CCTA performed on a 320-row CT scanner.

Accuracy of liver metastasis detection and characterization: Dual-energy CT versus single-energy CT with deep learning reconstruction

PURPOSE

To assess whether image quality differences between SECT (single-energy CT) and DECT (dual-energy CT 70 keV) with equivalent radiation doses result in altered detection and characterization accuracy of liver metastases when using deep learning image reconstruction (DLIR), and whether DECT spectral curve usage improves accuracy of indeterminate lesion characterization.

METHODS

In this prospective Health Insurance Portability and Accountability Act-compliant study (March through August 2022), adult men and non-pregnant adult women with biopsy-proven colorectal cancer and liver metastases underwent SECT (120 kVp) and a DECT (70 keV) portovenous abdominal CT scan using DLIR in the same breath-hold (Revolution CT ES; GE Healthcare). Participants were excluded if consent could not be obtained, if there were nonequivalent radiation doses between the two scans, or if the examination was cancelled/rescheduled. Three radiologists independently performed lesion detection and characterization during two separate sessions (SECT DLIRmedium and DECT DLIRhigh) as well as reported lesion confidence and overall image quality. Hounsfield units were measured. Spectral HU curves were provided for any lesions rated as indeterminate. McNemar's test was used to test the marginal homogeneity in terms of diagnostic sensitivity, accuracy and lesion detection. A generalized estimating equation method was used for categorical outcomes.

RESULTS

30 participants (mean age, 58 years ± 11, 21 men) were evaluated. Mean CTDIvol was 34 mGy for both scans. 141 lesions (124 metastases, 17 benign) with a mean size of 0.8 cm ± 0.3 cm were identified. High scores for image quality (scores of 4 or 5) were not significantly different between DECT (N = 71 out of 90 total scores from the three readers) and SECT (N = 62) (OR, 2.01; 95% CI:0.89, 4.57; P = 0.093). Equivalent image noise to SECT DLIRmed (HU SD 10 ± 2) was obtained with DECT DLIRhigh (HU SD 10 ± 3) (P = 1). There was no significant difference in lesion detection between DECT and SECT (140/141 lesions) (99.3%; 95% CI:96.1%, 100%).The mean lesion confidence scores by each reader were 4.2 ± 1.3, 3.9 ± 1.0, and 4.8 ± 0.8 for SECT and 4.1 ± 1.4, 4.0 ± 1.0, and 4.7 ± 0.8 for DECT (odds ratio [OR], 0.83; 95% CI: 0.62, 1.11; P = 0.21). Small lesion (≤5mm) characterization accuracy on SECT and DECT was 89.1% (95% CI:76.4%, 96.4%; 41/46) and 84.8% (71.1%, 93.7%; 39/46), respectively (P = 0.41). Use of spectral HU lesion curves resulted in 34 correct changes in characterizations and no mischaracterizations.

CONCLUSION

DECT required a higher strength of DLIR to obtain equivalent noise compared to SECT DLIR. At equivalent radiation doses and image noise, there was no significant difference in subjective image quality or observer lesion performance between DECT (70 keV) and SECT. However, DECT spectral HU curves of indeterminate lesions improved characterization.

Implementation of AI image reconstruction in CT-how is it validated and what dose reductions can be achieved

CT reconstruction has undergone a substantial change over the last decade with the introduction of iterative reconstruction (IR) and now with deep learning reconstruction (DLR). In this review, DLR will be compared to IR and filtered back-projection (FBP) reconstructions. Comparisons will be made using image quality metrics such as noise power spectrum, contrast-dependent task-based transfer function, and non-prewhitening filter detectability index (dNPW'). Discussion on how DLR has impacted CT image quality, low-contrast detectability, and diagnostic confidence will be provided. DLR has shown the ability to improve in areas that IR is lacking, namely: noise magnitude reduction does not alter noise texture to the degree that IR did, and the noise texture found in DLR is more aligned with noise texture of an FBP reconstruction. Additionally, the dose reduction potential for DLR is shown to be greater than IR. For IR, the consensus was dose reduction should be limited to no more than 15-30% to preserve low-contrast detectability. For DLR, initial phantom and patient observer studies have shown acceptable dose reduction between 44 and 83% for both low- and high-contrast object detectability tasks. Ultimately, DLR is able to be used for CT reconstruction in place of IR, making it an easy "turnkey" upgrade for CT reconstruction. DLR for CT is actively being improved as more vendor options are being developed and current DLR options are being enhanced with second generation algorithms being released. DLR is still in its developmental early stages, but is shown to be a promising future for CT reconstruction.

Impact of Phantom Size on Low-Energy Virtual Monoenergetic Images of Three Dual-Energy CT Platforms

The purpose of this study was to compare the quality of low-energy virtual monoenergetic images (VMIs) obtained with three Dual-Energy CT (DECT) platforms according to the phantom diameter. Three sections of the Mercury Phantom 4.0 were scanned on two generations of split-filter CTs (SFCT-1st and SFCT-2nd) and on one Dual-source CT (DSCT). The noise power spectrum (NPS), task-based transfer function (TTF), and detectability index (d') were assessed on VMIs from 40 to 70 keV. The highest noise magnitude values were found with SFCT-1st and noise magnitude was higher with DSCT than with SFCT-2nd for 26 cm (10.2% ± 1.3%) and 31 cm (7.0% ± 2.5%), and the opposite for 36 cm (-4.2% ± 2.5%). The highest average NPS spatial frequencies and TTF values at 50% (f50) values were found with DSCT. For all energy levels, the f50 values were higher with SFCT-2nd than SFCT-1st for 26 cm (3.2% ± 0.4%) and the opposite for 31 cm (-6.9% ± 0.5%) and 36 cm (-5.6% ± 0.7%). The lowest d' values were found with SFCT-1st. For all energy levels, the d' values were lower with DSCT than with SFCT-2nd for 26 cm (-6.2% ± 0.7%), similar for 31 cm (-0.3% ± 1.9%) and higher for 36 cm (5.4% ± 2.7%). In conclusion, compared to SFCT-1st, SFCT-2nd exhibited a lower noise magnitude and higher detectability values. Compared with DSCT, SFCT-2nd had a lower noise magnitude and higher detectability for the 26 cm, but the opposite was true for the 36 cm.

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

PURPOSE

To evaluate the efficacy of low-dose CT (LDCT) with deep learning image reconstruction (DLIR) for the surveillance of pancreatic cystic lesions (PCLs) compared with standard-dose CT (SDCT) with adaptive statistical iterative reconstruction (ASIR-V).

METHODS

The study enrolled 103 patients who underwent pancreatic CT for follow-up of incidentally detected PCLs. The CT protocol included LDCT in the pancreatic phase with 40% ASIR-V, DLIR at medium (DLIR-M) and high levels (DLIR-H), and SDCT in the portal-venous phase with 40% ASIR-V. The overall image quality and conspicuity of PCLs were qualitatively assessed using five-point scales by two radiologists. The size of PCLs, presence of thickened/enhancing walls, enhancing mural nodules, and main pancreatic duct dilatation were reviewed. CT noise and cyst-to-pancreas contrast-to-noise ratio (CNR) were measured. Qualitative and quantitative parameters were analyzed using the chi-squared test, one-way ANOVA, and t-test. Additionally, interobserver agreement was analyzed using the kappa and weighted-kappa statistics.

RESULTS

The volume CT dose-indexes in LDCT and SDCT were 3.0 ± 0.6 mGy and 8.4 ± 2.9 mGy, respectively. LDCT with DLIR-H showed the highest overall image quality, the lowest noise, and the highest CNR. The PCL conspicuity in LDCT with either DLIR-M or DLIR-H was not significantly different from that in SDCT with ASIR-V. Other findings depicting PCLs also revealed no significant differences between LDCT with DLIR and SDCT with ASIR-V. Moreover, the results revealed good or excellent interobserver agreement.

CONCLUSION

LDCT with DLIR has a comparable performance with SDCT for the follow-up of incidentally detected PCLs.

Deep-learning CT reconstruction in clinical scans of the abdomen: a systematic review and meta-analysis

OBJECTIVE

To perform a systematic literature review and meta-analysis of the two most common commercially available deep-learning algorithms for CT.

METHODS

We used PubMed, Scopus, Embase, and Web of Science to conduct systematic searches for studies assessing the most common commercially available deep-learning CT reconstruction algorithms: True Fidelity (TF) and Advanced intelligent Clear-IQ Engine (AiCE) in the abdomen of human participants since only these two algorithms currently have adequate published data for robust systematic analysis.

RESULTS

Forty-four articles fulfilled inclusion criteria. 32 studies evaluated TF and 12 studies assessed AiCE. DLR algorithms produced images with significantly less noise (22-57.3% less than IR) but preserved a desirable noise texture with increased contrast-to-noise ratios and improved lesion detectability on conventional CT. These improvements with DLR were similarly noted in dual-energy CT which was only assessed for a single vendor. Reported radiation reduction potential was 35.1-78.5%. Nine studies assessed observer performance with the two dedicated liver lesion studies being performed on the same vendor reconstruction (TF). These two studies indicate preserved low contrast liver lesion detection (> 5 mm) at CTDIvol 6.8 mGy (BMI 23.5 kg/m2) to 12.2 mGy (BMI 29 kg/m2). If smaller lesion detection and improved lesion characterization is needed, a CTDIvol of 13.6-34.9 mGy is needed in a normal weight to obese population. Mild signal loss and blurring have been reported at high DLR reconstruction strengths.

CONCLUSION

Deep learning reconstructions significantly improve image quality in CT of the abdomen. Assessment of other dose levels and clinical indications is needed. Careful choice of radiation dose levels is necessary, particularly for small liver lesion assessment.

Clinical Feasibility of Deep Learning-Based Image Reconstruction on Coronary Computed Tomography Angiography

This study evaluated the feasibility of deep-learning-based image reconstruction (DLIR) on coronary computed tomography angiography (CCTA). By using a 20 cm water phantom, the noise reduction ratio and noise power spectrum were evaluated according to the different reconstruction methods. Then 46 patients who underwent CCTA were retrospectively enrolled. CCTA was performed using the 16 cm coverage axial volume scan technique. All CT images were reconstructed using filtered back projection (FBP); three model-based iterative reconstructions (MBIR) of 40%, 60%, and 80%; and three DLIR algorithms: low (L), medium (M), and high (H). Quantitative and qualitative image qualities of CCTA were compared according to the reconstruction methods. In the phantom study, the noise reduction ratios of MBIR-40%, MBIR-60%, MBIR-80%, DLIR-L, DLIR-M, and DLIR-H were 26.7 ± 0.2%, 39.5 ± 0.5%, 51.7 ± 0.4%, 33.1 ± 0.8%, 43.2 ± 0.8%, and 53.5 ± 0.1%, respectively. The pattern of the noise power spectrum of the DLIR images was more similar to FBP images than MBIR images. In a CCTA study, CCTA yielded a significantly lower noise index with DLIR-H reconstruction than with the other reconstruction methods. DLIR-H showed a higher SNR and CNR than MBIR (p < 0.05). The qualitative image quality of CCTA with DLIR-H was significantly higher than that of MBIR-80% or FBP. The DLIR algorithm was feasible and yielded a better image quality than the FBP or MBIR algorithms on CCTA.

Impact of deep learning image reconstruction algorithms on CT radiomic features in patients with liver tumors

Objective

To evaluate the impact of deep learning image reconstruction (DLIR) and adaptive statistical iterative reconstruction-Veo (ASIR-V) on abdominal CT radiomic features acquired in portal venous phase in liver tumor patients.

Methods

Sixty patients with liver tumors who underwent contrast-enhanced abdominal CT were retrospectively enrolled. Six groups including filtered back projection (FBP), ASIR-V (30%, 70%) and DLIR at low (DLIR-L), medium (DLIR-M and high (DLIR-H), were reconstructed using portal venous phase data. CT-based radiomic features (first-order, texture and wavelet features) were extracted from 2D and 3D liver tumors, peritumor and liver parenchyma. All features were analyzed for comparison. P &lt; 0.05 indicated statistically different. The consistency of 3D lesion feature extraction was assessed by calculating intraclass correlation coefficient (ICC).

Results

Different reconstruction algorithms influenced most radiomic features. The percentages of first-order, texture and wavelet features without statistical difference among 2D and 3D lesions, peritumor and liver parenchyma for all six groups were 27.78% (5/18), 5.33% (4/75) and 5.56% (1/18), respectively (all p &gt; 0.05), and they decreased while the level of reconstruction strengthened for both ASIR-V and DLIR. Compared with FBP, the features of ASIR-V30% and 70% without statistical difference decreased from 71.31% to 23.95%, and DLIR-L, DLIR-M, and DLIR-H decreased from 31.65% to 27.11% and 23.73%. Among texture features, unaffected features of peritumor were larger than those of lesions and liver parenchyma, and unaffected 3D lesions features were larger than those of 2D lesions. The consistency of 3D lesion first-order features was excellent, with intra- and inter-observer ICCs ranging from 0.891 to 0.999 and 0.880 to 0.998.

Conclusions

Both ASIR-V and DLIR algorithms with different strengths influenced the radiomic features of abdominal CT images in portal venous phase, and the influences aggravated as reconstruction strength increased.

Applications of Artificial Intelligence in the Radiology Roundtrip: Process Streamlining, Workflow Optimization, and Beyond

There are many impactful applications of artificial intelligence (AI) in the electronic radiology roundtrip and the patient's journey through the healthcare system that go beyond diagnostic applications. These tools have the potential to improve quality and safety, optimize workflow, increase efficiency, and increase patient satisfaction. In this article, we review the role of AI for process improvement and workflow enhancement which includes applications beginning from the time of order entry, scan acquisition, applications supporting the image interpretation task, and applications supporting tasks after image interpretation such as result communication. These non-diagnostic workflow and process optimization tasks are an important part of the arsenal of potential AI tools that can streamline day to day clinical practice and patient care.

Image-quality characteristics in the longitudinal direction from different image-reconstruction algorithms during single-rotation volume acquisition on head computed tomography: A phantom study

Background

A multi detector computed tomography (CT) scanner with wide-area coverage enables whole-brain volumetric scanning in a single rotation.

Purpose

To investigate variations in image-quality characteristics in the longitudinal direction for different image-reconstruction algorithms and strengths with phantoms.

Material and methods

Single-rotation volume scans were performed on a 320-row multidetector CT volume scanner using three types of phantoms. Tube current was set to 200 mA (standard dose) and 50 mA (low dose). All images were reconstructed with filtered back projection (FBP), mild and strong levels with hybrid iterative reconstruction (HIR), and model-based IR (MBIR). Computed tomography numbers, image noise, noise power spectrum (NPS), task-based transfer function (TTF), and visual spatial resolution were used to evaluate uniformity of image quality in the longitudinal direction ( Z-axis).

Results

The MBIR images showed smaller variation in CT numbers in the Z-axis. The difference in the highest and lowest CT numbers was smaller (5.0 Hounsfield units [HU]) for MBIR than for FBP (6.6 HU) and HIR (6.8 HU). The variations in image noise were the smallest for strong MBIR and the largest for FBP. The low-frequency component at NPS0.2 was lower for strong MBIR than for other algorithms. The high-frequency component at NPS0.8 was low in all reconstructions. For MBIR, the image resolution and TTFs were higher in the outer portion than in the center.

Conclusion

Model-based IR is the optimal image-reconstruction algorithm for single-volume scan of spherical subjects owing to its high in-plane resolution and uniformity of CT numbers, image noise, and NPS in the Z-axis.

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.

Brain image quality according to beam collimation width and image reconstruction algorithm: A phantom study

PURPOSE

To compare quantitatively and qualitatively brain image quality acquired in helical and axial modes on two wide collimation CT systems according to the dose level and algorithm used.

METHODS

Acquisitions were performed on an image quality and an anthropomorphic phantoms at three dose levels (CTDI_vol: 45/35/25 mGy) on two wide collimation CT systems (GE Healthcare and Canon Medical Systems) in axial and helical modes. Raw data were reconstructed using iterative reconstruction (IR) and deep-learning image reconstruction (DLR) algorithms. The noise power spectrum (NPS) was computed on both phantoms and the task-based transfer function (TTF) on the image quality phantom. The subjective quality of images from an anthropomorphic brain phantom was evaluated by two radiologists including overall image quality.

RESULTS

For the GE system, noise magnitude and noise texture (average NPS spatial frequency) were lower with DLR than with IR. For the Canon system, noise magnitude values were lower with DLR than with IR for similar noise texture but the opposite was true for spatial resolution. For both CT systems, noise magnitude was lower with the axial mode than with the helical mode for similar noise texture and spatial resolution. Radiologists rated the overall quality of all brain images as "satisfactory for clinical use", whatever the dose level, algorithm or acquisition mode.

CONCLUSIONS

Using 16-cm axial acquisition reduces image noise without changing the spatial resolution and image texture compared to helical acquisitions. Axial acquisition can be used in clinical routine for brain CT examinations with an explored length of less than 16 cm.

Evaluation of deep-learning image reconstruction for chest CT examinations at two different dose levels

AIMS

The aims of the present study were to, for both a full-dose protocol and an ultra-low dose (ULD) protocol, compare the image quality of chest CT examinations reconstructed using TrueFidelity (Standard kernel) with corresponding examinations reconstructed using ASIR-V (Lung kernel) and to evaluate if post-processing using an edge-enhancement filter affects the noise level, spatial resolution and subjective image quality of clinical images reconstructed using TrueFidelity.

METHODS

A total of 25 patients were examined with both a full-dose protocol and an ULD protocol using a GE Revolution APEX CT system (GE Healthcare, Milwaukee, USA). Three different reconstructions were included in the study: ASIR-V 40%, DLIR-H, and DLIR-H with additional post-processing using an edge-enhancement filter (DLIR-H + E2). Five observers assessed image quality in two separate visual grading characteristics (VGC) studies. The results from the studies were statistically analyzed using VGC Analyzer. Quantitative evaluations were based on determination of two-dimensional power spectrum (PS), contrast-to-noise ratio (CNR), and spatial resolution in the reconstructed patient images.

RESULTS

For both protocols, examinations reconstructed using TrueFidelity were statistically rated equal to or significantly higher than examinations reconstructed using ASIR-V 40%, but the ULD protocol benefitted more from TrueFidelity. In general, no differences in observer ratings were found between DLIR-H and DLIR-H + E2. For the three investigated image reconstruction methods, ASIR-V 40% showed highest noise and spatial resolution and DLIR-H the lowest, while the CNR was highest in DLIR-H and lowest in ASIR-V 40%.

CONCLUSION

The use of TrueFidelity for image reconstruction resulted in higher ratings on subjective image quality than ASIR-V 40%. The benefit of using TrueFidelity was larger for the ULD protocol than for the full-dose protocol. Post-processing of the TrueFidelity images using an edge-enhancement filter resulted in higher image noise and spatial resolution but did not affect the subjective image quality.

Deep Learning Image Reconstruction for CT: Technical Principles and Clinical Prospects

Filtered back projection (FBP) has been the standard CT image reconstruction method for 4 decades. A simple, fast, and reliable technique, FBP has delivered high-quality images in several clinical applications. However, with faster and more advanced CT scanners, FBP has become increasingly obsolete. Higher image noise and more artifacts are especially noticeable in lower-dose CT imaging using FBP. This performance gap was partly addressed by model-based iterative reconstruction (MBIR). Yet, its "plastic" image appearance and long reconstruction times have limited widespread application. Hybrid iterative reconstruction partially addressed these limitations by blending FBP with MBIR and is currently the state-of-the-art reconstruction technique. In the past 5 years, deep learning reconstruction (DLR) techniques have become increasingly popular. DLR uses artificial intelligence to reconstruct high-quality images from lower-dose CT faster than MBIR. However, the performance of DLR algorithms relies on the quality of data used for model training. Higher-quality training data will become available with photon-counting CT scanners. At the same time, spectral data would greatly benefit from the computational abilities of DLR. This review presents an overview of the principles, technical approaches, and clinical applications of DLR, including metal artifact reduction algorithms. In addition, emerging applications and prospects are discussed.

Photon-counting detector coronary CT angiography: impact of virtual monoenergetic imaging and iterative reconstruction on image quality

OBJECTIVES

To assess the impact of low kilo-electronvolt (keV) virtual monoenergetic image (VMI) energies and iterative reconstruction on image quality of clinical photon-counting detector coronary CT angiography (CCTA).

METHODS

CCTA with PCD-CT (prospective ECG-triggering, 120 kVp, automatic tube current modulation) was performed in a high-end cardiovascular phantom with dynamic flow, pulsatile heart motion, and including different calcified plaques with various stenosis grades and in 10 consecutive patients. VMI at 40,50,60 and 70 keV were reconstructed without (QIR-off) and with all quantum iterative reconstruction (QIR) levels (QIR-1 to 4). In the phantom, noise power spectrum, vessel attenuation, contrast-to-noise-ratio (CNR), and vessel sharpness were measured. Two readers graded stenoses in the phantom and graded overall image quality, subjective noise, vessel sharpness, vascular contrast, and coronary artery plaque delineation on 5-point Likert scales in patients.

RESULTS

In the phantom, noise texture was only slightly affected by keV and QIR while noise increased by 69% from 70 keV QIR-4 to 40 keV QIR-off. Reconstructions at 40 keV QIR-4 exhibited the highest CNR (46.1 ± 1.8), vessel sharpness (425 ± 42 ∆HU/mm), and vessel attenuation (1098 ± 14 HU). Stenosis measurements were not affected by keV or QIR level (p > 0.12) with an average error of 3%/6% for reader 1/reader 2, respectively. In patients, across all subjective categories and both readers, 40 keV QIR-3 and QIR-4 images received the best scores (p < 0.001).

CONCLUSION

Forty keV VMI with QIR-4 significantly improved image quality of CCTA with PCD-CT.

ADVANCES IN KNOWLEDGE

PCD-CT at 40 keV and QIR-4 improves image quality of CCTA.

Abdominopelvic CT Image Quality: Evaluation of Thin (0.5-mm) Slices Using Deep Learning Reconstruction

BACKGROUND. Because thick-section images (typically 3-5 mm) have low image noise, radiologists typically use them to perform clinical interpretation, although they may additionally refer to thin-section images (typically 0.5-0.625 mm) for problem solving. Deep learning reconstruction (DLR) can yield thin-section images with low noise. OBJECTIVE. The purpose of this study is to compare abdominopelvic CT image quality between thin-section DLR images and thin- and thick-section hybrid iterative reconstruction (HIR) images. METHODS. This retrospective study included 50 patients (31 men and 19 women; median age, 64 years) who underwent abdominopelvic CT between June 15, 2020, and July 29, 2020. Images were reconstructed at 0.5-mm section using DLR and at 0.5-mm and 3.0-mm sections using HIR. Five radiologists independently performed pairwise comparisons (0.5-mm DLR and either 0.5-mm or 3.0-mm HIR) and recorded the preferred image for subjective image quality measures (scale, -2 to 2). The pooled scores of readers were compared with a score of 0 (denoting no preference). Image noise was quantified using the SD of ROIs on regions of homogeneous liver. RESULTS. For comparison of 0.5-mm DLR images and 0.5-mm HIR images, the median pooled score was 2 (indicating a definite preference for DLR) for noise and overall image quality and 1 (denoting a slight preference for DLR) for sharpness and natural appearance. For comparison of 0.5-mm DLR and 3.0-mm HIR, the median pooled score was 1 for the four previously mentioned measures. These assessments were all significantly different (p < .001) from 0. For artifacts, the median pooled score for both comparisons was 0, which was not significant for comparison with 3.0-mm HIR (p = .03) but was significant for comparison with 0.5-mm HIR (p < .001) due to imbalance in scores of 1 (n = 28) and -1 (slight preference for HIR, n = 1). Noise for 0.5-mm DLR was lower by mean differences of 12.8 HU compared with 0.5-mm HIR and 4.4 HU compared with 3.0-mm HIR (both p < .001). CONCLUSION. Thin-section DLR improves subjective image quality and reduces image noise compared with currently used thin- and thick-section HIR, without causing additional artifacts. CLINICAL IMPACT. Although further diagnostic performance studies are warranted, the findings suggest the possibility of replacing current use of both thin- and thick-section HIR with the use of thin-section DLR only during clinical interpretations.

Improved image quality with deep learning reconstruction - a study on a semi-anthropomorphic upper-abdomen phantom

Purpose

To assess image quality of a deep learning reconstruction (DLR) algorithm across dose levels using a semi-anthropomorphic upper-abdominal phantom, and compare with filtered back projection (FBP) and hybrid iterative reconstruction (IR).

Material and methods

CT scans obtained at five dose levels (CTDIvol 5, 10, 15, 20 and 25 mGy) were reconstructed with FBP, hybrid IR (IR50, IR70 and IR90) and DLR of low (DLL), medium (DLM) and high strength (DLH) in 0.625 mm and 2.5 mm slices. CT number, homogeneity, noise, contrast, contrast-to-noise ratio (CNR), noise texture deviation (NTD; a measure of IR-specific artifacts), noise power spectrum (NPS) and task-based transfer function (TTF) were compared between reconstruction algorithms.

Results

CT numbers were highly consistent across reconstruction algorithms. Image noise was significantly reduced with higher levels of DLR. Noise texture (NPS and NTD) was with DLR maintained at comparable levels to FBP, contrary to increasing levels of hybrid IR. Images reconstructed with DLR of low and high strength in 0.625 mm slices showed similar noise characteristics to 2.5 mm slice FBP and IR50, respectively. Dose-reduction potential based on image noise with IR50 as reference was estimated to 35% for DLM and 74% for DLH.

Conclusions

The novel DLR algorithm demonstrates robust noise reduction with maintained noise texture characteristics despite higher algorithm strength, and may have overcome important limitations of IR. There may be potential for dose reduction and additional benefit from thin-slice reconstruction.

Is it possible to use low-dose deep learning reconstruction for the detection of liver metastases on CT routinely?

OBJECTIVES

To compare the image quality and hepatic metastasis detection of low-dose deep learning image reconstruction (DLIR) with full-dose filtered back projection (FBP)/iterative reconstruction (IR).

METHODS

A contrast-detail phantom consisting of low-contrast objects was scanned at five CT dose index levels (10, 6, 3, 2, and 1 mGy). A total of 154 participants with 305 hepatic lesions who underwent abdominal CT were enrolled in a prospective non-inferiority trial with a three-arm design based on phantom results. Data sets with full dosage (13.6 mGy) and low dosages (9.5, 6.8, or 4.1 mGy) were acquired from two consecutive portal venous acquisitions, respectively. All images were reconstructed with FBP (reference), IR (control), and DLIR (test). Eleven readers evaluated phantom data sets for object detectability using a two-alternative forced-choice approach. Non-inferiority analyses were performed to interpret the differences in image quality and metastasis detection of low-dose DLIR relative to full-dose FBP/IR.

RESULTS

The phantom experiment showed the dose reduction potential from DLIR was up to 57% based on the reference FBP dose index. Radiation decreases of 30% and 50% resulted in non-inferior image quality and hepatic metastasis detection with DLIR compared to full-dose FBP/IR. Radiation reduction of 70% by DLIR performed inferiorly in detecting small metastases (< 1 cm) compared to full-dose FBP (difference: -0.112; 95% confidence interval [CI]: -0.178 to 0.047) and full-dose IR (difference: -0.123; 95% CI: -0.182 to 0.053) (p < 0.001).

CONCLUSION

DLIR enables a 50% dose reduction for detecting low-contrast hepatic metastases while maintaining comparable image quality to full-dose FBP and IR.

KEY POINTS

• Non-inferiority study showed that deep learning image reconstruction (DLIR) can reduce the dose to oncological patients with low-contrast lesions without compromising the diagnostic information. • Radiation dose levels for DLIR can be reduced to 50% of full-dose FBP and IR for detecting low-contrast hepatic metastases, while maintaining comparable image quality. • The reduction of radiation by 70% by DLIR is clinically acceptable but insufficient for detecting small low-contrast hepatic metastases (< 1 cm).

Deep image and feature prior algorithm based on U-ConformerNet structure

PURPOSE

The reconstruction performance of the deep image prior (DIP) approach is limited by the conventional convolutional layer structure and it is difficult to enhance its potential. In order to improve the quality of image reconstruction and suppress artifacts, we propose a DIP algorithm with better performance, and verify its superiority in the latest case.

METHODS

We construct a new U-ConformerNet structure as the DIP algorithm's network, replacing the traditional convolutional layer-based U-net structure, and introduce the 'lpips' deep network based feature distance regularization method. Our algorithm can switch between supervised and unsupervised modes at will to meet different needs.

RESULTS

The reconstruction was performed on the low dose CT dataset (LoDoPaB). Our algorithm attained a PSNR of more than 35 dB under unsupervised conditions, and the PSNR under the supervised condition is greater than 36 dB. Both of which are better than the performance of the DIP-TV. Furthermore, the accuracy of this method is positively connected with the quality of the a priori image with the help of deep networks. In terms of noise eradication and artifact suppression, the DIP algorithm with U-ConformerNet structure outperforms the standard DIP method based on convolutional structure.

CONCLUSIONS

It is known by experimental verification that, in unsupervised mode, the algorithm improves the output PSNR by at least 2-3 dB when compared to the DIP-TV algorithm (proposed in 2020). In supervised mode, our algorithm approaches that of the state-of-the-art end-to-end deep learning algorithms.

Lung-Optimized Deep-Learning-Based Reconstruction for Ultralow-Dose CT

RATIONALE AND OBJECTIVES

To evaluate the image properties of lung-specialized deep-learning-based reconstruction (DLR) and its applicability in ultralow-dose CT (ULDCT) relative to hybrid- (HIR) and model-based iterative-reconstructions (MBIR).

MATERIALS AND METHODS

An anthropomorphic chest phantom was scanned on a 320-row scanner at 50-mA (low-dose-CT 1 [LDCT-1]), 25-mA (LDCT-2), and 10-mA (ULDCT). LDCT were reconstructed with HIR; ULDCT images were reconstructed with HIR (ULDCT-HIR), MBIR (ULDCT-MBIR), and DLR (ULDCT-DLR). Image noise and contrast-to-noise ratio (CNR) were quantified. With the LDCT images as reference standards, ULDCT image qualities were subjectively scored on a 5-point scale (1 = substantially inferior to LDCT-2, 3 = comparable to LDCT-2, 5 = comparable to LDCT-1). For task-based image quality analyses, a physical evaluation phantom was scanned at seven doses to achieve the noise levels equivalent to chest phantom; noise power spectrum (NPS) and task-based transfer function (TTF) were evaluated. Clinical ULDCT (10-mA) images obtained in 14 nonobese patients were reconstructed with HIR, MBIR, and DLR; the subjective acceptability was ranked.

RESULTS

Image noise was lower and CNR was higher in ULDCT-DLR and ULDCT-MBIR than in LDCT-1, LDCT-2, and ULDCT-HIR (p < 0.01). The overall quality of ULDCT-DLR was higher than of ULDCT-HIR and ULDCT-MBIR (p < 0.01), and almost comparable with that of LDCT-2 (mean score: 3.4 ± 0.5). DLR yielded the highest NPS peak frequency and TTF_50% for high-contrast object. In clinical ULDCT images, the subjective acceptability of DLR was higher than of HIR and MBIR (p < 0.01).

CONCLUSION

DLR optimized for lung CT improves image quality and provides possible greater dose optimization opportunity than HIR and MBIR.

Deep learning image reconstruction algorithm reduces image noise while alters radiomics features in dual-energy CT in comparison with conventional iterative reconstruction algorithms: a phantom study

OBJECTIVES

To compare image quality between a deep learning image reconstruction (DLIR) algorithm and conventional iterative reconstruction (IR) algorithms in dual-energy CT (DECT) and to assess the impact of these algorithms on radiomics robustness.

METHODS

A phantom with clinical-relevant densities was imaged on seven DECT scanners with the same voxel size using typical abdominal-pelvis examination protocols. On one DECT scanner, raw data were reconstructed using both conventional IR (adaptive statistical iterative reconstruction-V, ASIR-V) and DLIR. Nine sets of corresponding images were generated on other six DECT scanners using scanner-equipped conventional IR. Regions of interest were delineated through rigid registrations. Image quality was compared. Pyradiomics platform was used for radiomics feature extraction. Test-retest repeatability was assessed by Bland-Altman analysis for repeated scans. Inter-reconstruction algorithm reproducibility between conventional IR and DLIR was tested by intraclass correlation coefficient (ICC) and concordance correlation coefficient (CCC). Inter-scanner reproducibility was evaluated by coefficient of variation (CV) and quartile coefficient of dispersion (QCD). Robust features were identified.

RESULTS

DLIR significantly improved image quality. Ninety-four radiomics features were extracted and nine features were considered as robust. 93.87% features were repeatable between repeated scans. ASIR-V images showed higher reproducibility to other conventional IR than DLIR (ICC mean, 0.603 vs 0.558, p = 0.001; CCC mean, 0.554 vs 0.510, p = 0.004). 7.45% and 26.83% features were reproducible among scanners evaluated by CV and QCD, respectively.

CONCLUSIONS

DLIR improves quality of DECT images but may alter radiomics features compared to conventional IR. Nine robust DECT radiomics features were identified.

KEY POINTS

• DLIR improves DECT image quality in terms of signal-to-noise ratio and contrast-to-noise ratio compared with ASIR-V and showed the highest noise reduction rate and lowest peak frequency shift. • Most of radiomics features are repeatable between repeated DECT scans, while inter-reconstruction algorithm reproducibility between conventional IR and DLIR, and inter-scanner reproducibility, are low. • Although DLIR may alter radiomics features compared to IR algorithms, nine radiomics features survived repeatability and reproducibility analysis among DECT scanners and reconstruction algorithms, which allows further validation and clinical-relevant analysis.

Deep learning image reconstruction to improve accuracy of iodine quantification and image quality in dual-energy CT of the abdomen: a phantom and clinical study

OBJECTIVES

To investigate the effect of deep learning image reconstruction (DLIR) on the accuracy of iodine quantification and image quality of dual-energy CT (DECT) compared to that of other reconstruction algorithms in a phantom experiment and an abdominal clinical study.

METHODS

An elliptical phantom with five different iodine concentrations (1-12 mgI/mL) was imaged five times with fast-kilovoltage-switching DECT for three target volume CT dose indexes. All images were reconstructed using filtered back-projection, iterative reconstruction (two levels), and DLIR algorithms. Measured and nominal iodine concentrations were compared among the algorithms. Contrast-enhanced CT of the abdomen with the same scanner was acquired in clinical patients. In arterial and portal venous phase images, iodine concentration, image noise, and coefficients of variation for four locations were retrospectively compared among the algorithms. One-way repeated-measures analyses of variance were used to evaluate differences in the iodine concentrations, standard deviations, coefficients of variation, and percentages of error among the algorithms.

RESULTS

In the phantom study, the measured iodine concentrations were equivalent among the algorithms: within ± 8% of the nominal values, with root-mean-square deviations of 0.08-0.36 mgI/mL, regardless of radiation dose. In the clinical study (50 patients; 35 men; mean age, 68 ± 11 years), iodine concentrations were equivalent among the algorithms for each location (all p > .99). Image noise and coefficients of variation were lower with DLIR than with the other algorithms (all p < .01).

CONCLUSIONS

The DLIR algorithm reduced image noise and variability of iodine concentration values compared with other reconstruction algorithms in the fast-kilovoltage-switching dual-energy CT.

KEY POINTS

• In the phantom study, standard deviations and coefficients of variation in iodine quantification were lower on images with the deep learning image reconstruction algorithm than on those with other algorithms. • In the clinical study, iodine concentrations of measurement location in the upper abdomen were consistent across four reconstruction algorithms, while image noise and variability of iodine concentrations were lower on images with the deep learning image reconstruction algorithm.

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.

"Image quality evaluation of the Precise image CT deep learning reconstruction algorithm compared to Filtered Back-projection and iDose4: a phantom study at different dose levels"

PURPOSE

To characterize the performance of the Precise Image (PI) deep learning reconstruction (DLR) algorithm for abdominal Computed Tomography (CT) imaging.

METHODS

CT images of the Catphan-600 phantom (equipped with an external annulus) were acquired using an abdominal protocol at four dose levels and reconstructed using FBP, iDose⁴ (levels 2,5) and PI ('Soft Tissue' definition, levels 'Sharper','Sharp','Standard','Smooth','Smoother'). Image noise, image non-uniformity, noise power spectrum (NPS), target transfer function (TTF), detectability index (d'), CT numbers accuracy and image histograms were analyzed.

RESULTS

The behavior of the PI algorithm depended strongly on the selected level of reconstruction. The phantom analysis suggested that the PI image noise decreased linearly by varying the level of reconstruction from Sharper to Smoother, expressing a noise reduction up to 80% with respect to FBP. Additionally, the non-uniformity decreased, the histograms became narrower, and d' values increased as PI reconstruction levels changed from Sharper to Smoother. PI had no significant impact on the average CT number of different contrast objects. The conventional FBP NPS was deeply altered only by Smooth and Smoother levels of reconstruction. Furthermore, spatial resolution was found to be dose- and contrast-dependent, but in each analyzed condition it was greater than or comparable to FBP and iDose⁴ TTFs.

CONCLUSIONS

The PI algorithm can reduce image noise with respect to FBP and iDose⁴; spatial resolution, CT numbers and image uniformity are generally preserved by the algorithm but changes in NPS for the Smooth and Smoother levels need to be considered in protocols implementation.

Task-Based Image Quality Assessment Comparing Classical and Iterative Cone Beam CT Images on Halcyon®

BACKGROUND

Despite the development of iterative reconstruction (IR) in diagnostic imaging, CBCT are generally reconstructed with filtered back projection (FBP) in radiotherapy. Varian medical systems, recently released with their latest Halcyon^® V2.0 accelerator, a new IR algorithm for CBCT reconstruction.

PURPOSE

To assess the image quality of radiotherapy CBCT images reconstructed with FBP and an IR algorithm.

METHODS

Three CBCT acquisition modes (head, thorax and pelvis large) available on a Halcyon^® were assessed. Five acquisitions were performed for all modes on an image quality phantom and reconstructed with FBP and IR. Task-based image quality assessment was performed with noise power spectrum (NPS), task-based transfer function (TTF) and detectability index (d'). To illustrate the image quality obtained with both reconstruction types, CBCT acquisitions were made on 6 patients.

RESULTS

The noise magnitude and the spatial frequency of the NPS peak was lower with IR than with FBP for all modes. For all low and high-contrast inserts, the values for TTF at 50% were higher with IR than with FBP. For all inserts and all modes, the contrast values were similar with FBP and IR. For all low and high-contrast simulated lesions, d' values were higher with IR than with FBP for all modes. These results were also found on the 6 patients where the images were less noisy but smoother with IR-CBCT.

CONCLUSIONS

Using the IR algorithm for CBCT images in radiotherapy improve image quality and thus could increase the accuracy of online registration and limit positioning errors during processing.

Comparison of image quality of two versions of deep-learning image reconstruction algorithm on a rapid kV-switching CT: a phantom study

BACKGROUND

To assess the impact of the new version of a deep learning (DL) spectral reconstruction on image quality of virtual monoenergetic images (VMIs) for contrast-enhanced abdominal computed tomography in the rapid kV-switching platform.

METHODS

Two phantoms were scanned with a rapid kV-switching CT using abdomen-pelvic CT examination parameters at dose of 12.6 mGy. Images were reconstructed using two versions of DL spectral reconstruction algorithms (DLSR V1 and V2) for three reconstruction levels. The noise power spectrum (NSP) and task-based transfer function at 50% (TTF₅₀) were computed at 40/50/60/70 keV. A detectability index (d') was calculated for enhanced lesions at low iodine concentrations: 2, 1, and 0.5 mg/mL.

RESULTS

The noise magnitude was significantly lower with DLSR V2 compared to DLSR V1 for energy levels between 40 and 60 keV by -36.5% ± 1.4% (mean ± standard deviation) for the standard level. The average NPS frequencies increased significantly with DLSR V2 by 23.7% ± 4.2% for the standard level. The highest difference in TTF₅₀ was observed at the mild level with a significant increase of 61.7% ± 11.8% over 40-60 keV energy with DLSR V2. The d' values were significantly higher for DLSR V2 versus DLSR V1.

CONCLUSIONS

The DLSR V2 improves image quality and detectability of low iodine concentrations in VMIs compared to DLSR V1. This suggests a great potential of DLSR V2 to reduce iodined contrast doses.

Deep learning image reconstruction algorithm for carotid dual-energy computed tomography angiography: evaluation of image quality and diagnostic performance

Objectives

To evaluate image quality and diagnostic performance of carotid dual-energy computed tomography angiography (DECTA) using deep learning image reconstruction (DLIR) compared with images using adaptive statistical iterative reconstruction-Veo (ASIR-V).

Methods

Carotid DECTA datasets of 28 consecutive patients were reconstructed at 50 keV using DLIR at low, medium, and high levels (DLIR-L, DLIR-M, and DLIR-H) and 80% ASIR-V algorithms. Mean attenuation, image noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) at different levels of arteries were measured and calculated. Image quality for noise and texture, depiction of arteries, and diagnostic performance toward carotid plaques were assessed subjectively by two radiologists. Quantitative and qualitative parameters were compared between the ASIR-V, DLIR-L, DLIR-M, and DLIR-H groups.

Results

The image noise at aorta and common carotid artery, SNR, and CNR at all level arteries of DLIR-H images were significantly higher than those of ASIR-V images (p = 0.000–0.040). The quantitative analysis of DLIR-L and DLIR-M showed comparable denoise capability with ASIR-V. The overall image quality (p = 0.000) and image noise (p = 0.000–0.014) were significantly better in the DLIR-M and DLIR-H images. The image texture was improved by DLR at all level compared to ASIR-V images (p = 0.000–0.008). Depictions of head and neck arteries and diagnostic performance were comparable between four groups (p > 0.05).

Conclusions

Compared with 80% ASIR-V, we recommend DLIR-H for clinical carotid DECTA reconstruction, which can significantly improve the image quality of carotid DECTA at 50 keV but maintain a desirable diagnostic performance and arterial depiction.

Using Artificial Intelligence for Optimization of the Processes and Resource Utilization in Radiotherapy

The radiotherapy (RT) process from planning to treatment delivery is a multistep, complex operation involving numerous levels of human-machine interaction and requiring high precision. These steps are labor-intensive and time-consuming and require meticulous coordination between professionals with diverse expertise. We reviewed and summarized the current status and prospects of artificial intelligence and machine learning relevant to the various steps in RT treatment planning and delivery workflow specifically in low- and middle-income countries (LMICs). We also searched the PubMed database using the search terms (Artificial Intelligence OR Machine Learning OR Deep Learning OR Automation OR knowledge-based planning AND Radiotherapy) AND (list of Low- and Middle-Income Countries as defined by the World Bank at the time of writing this review). The search yielded a total of 90 results, of which results with first authors from the LMICs were chosen. The reference lists of retrieved articles were also reviewed to search for more studies. No language restrictions were imposed. A total of 20 research items with unique study objectives conducted with the aim of enhancing RT processes were examined in detail. Artificial intelligence and machine learning can improve the overall efficiency of RT processes by reducing human intervention, aiding decision making, and efficiently executing lengthy, repetitive tasks. This improvement could permit the radiation oncologist to redistribute resources and focus on responsibilities such as patient counseling, education, and research, especially in resource-constrained LMICs.

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.

Improved image quality and dose reduction in abdominal CT with deep-learning reconstruction algorithm: a phantom study

OBJECTIVES

To assess the impact of a new artificial intelligence deep-learning reconstruction (Precise Image; AI-DLR) algorithm on image quality against a hybrid iterative reconstruction (IR) algorithm in abdominal CT for different clinical indications.

METHODS

Acquisitions on phantoms were performed at 5 dose levels (CTDI_vol: 13/11/9/6/1.8 mGy). Raw data were reconstructed using level 4 of iDose⁴ (i4) and 3 levels of AI-DLR (Smoother/Smooth/Standard). Noise power spectrum (NPS), task-based transfer function (TTF) and detectability index (d') were computed: d' modelled detection of a liver metastasis (LM) and hepatocellular carcinoma at portal (HCCp) and arterial (HCCa) phases. Image quality was subjectively assessed on an anthropomorphic phantom by 2 radiologists.

RESULTS

From Standard to Smoother levels, noise magnitude and average NPS spatial frequency decreased and the detectability (d') of all simulated lesions increased. For both inserts, TTF values were similar for all three AI-DLR levels from 13 to 6 mGy but decreased from Standard to Smoother levels at 1.8 mGy. Compared to the i4 used in clinical practice, d' values were higher using the Smoother and Smooth levels and close for the Standard level. For all dose levels, except at 1.8 mGy, radiologists considered images satisfactory for clinical use for the 3 levels of AI-DLR, but rated images too smooth using the Smoother level.

CONCLUSION

Use of the Smooth and Smoother levels of AI-DLR reduces the image noise and improves the detectability of lesions and spatial resolution for standard and low-dose levels. Using the Smooth level is apparently the best compromise between the lowest dose level and adequate image quality.

KEY POINTS

• Evaluation of the impact of a new artificial intelligence deep-learning reconstruction (AI-DLR) on image quality and dose compared to a hybrid iterative reconstruction (IR) algorithm. • The Smooth and Smoother levels of AI-DLR reduced the image noise and improved the detectability of lesions and spatial resolution for standard and low-dose levels. • The Smooth level seems the best compromise between the lowest dose level and adequate image quality.

Impact of an artificial intelligence deep-learning reconstruction algorithm for CT on image quality and potential dose reduction: A phantom study

Background

Recently, computed tomography (CT) manufacturers have developed deep‐learning‐based reconstruction algorithms to compensate for the limitations of iterative reconstruction (IR) algorithms, such as image smoothing and the spatial resolution's dependence on contrast and dose levels.

Purpose

To assess the impact of an artificial intelligence deep‐learning reconstruction (AI‐DLR) algorithm on image quality and dose reduction compared with a hybrid IR algorithm in chest CT for different clinical indications.

Methods

Acquisitions on the CT American College of Radiology (ACR) 464 and CT Torso CTU‐41 phantoms were performed at five dose levels (CTDI_vol: 9.5/7.5/6/2.5/0.4 mGy) used for chest CT conditions. Raw data were reconstructed using filtered backprojection, two levels of IR (iDose⁴ levels 4 (i4) and 7 (i7)), and five levels of AI‐DLR (Precise Image; Smoother, Smooth, Standard, Sharp, Sharper). Noise power spectrum (NPS), task‐based transfer function, and detectability index (d′) were computed: d′‐modeled detection of a soft tissue mediastinal nodule (low‐contrast soft tissue chest nodule within the mediastinum [LCN]), ground‐glass opacity (GGO), or high‐contrast pulmonary (HCP) lesion. The subjective image quality of chest anthropomorphic phantom images was independently evaluated by two radiologists. They assessed image noise, image smoothing, contrast between vessels and fat in the mediastinum for mediastinal images, visual border detection between bronchus and lung parenchyma for parenchymal images, and overall image quality using a commonly used four‐ or five‐point scale.

Results

From Standard to Smoother levels, on average, the noise magnitude decreased (for all dose levels: −66.3% ± 0.5% for mediastinal images and −63.1% ± 0.1% for parenchymal images), the average NPS spatial frequency decreased (for all dose levels: −35.3% ± 2.2% for mediastinal images and −13.3% ± 2.2% for parenchymal images), and the detectability (d′) of the three lesions increased. The opposite pattern was found from Standard to Sharper levels. From Smoother to Sharper levels, the spatial resolution increased for the low‐contrast polyethylene insert and the opposite for the high‐contrast air insert. Compared to the i4 used in clinical practice, d′ values were higher using Smoother (mean for all dose levels: 338.7% ± 29.4%), Smooth (103.4% ± 11.2%), and Standard (34.1% ± 6.6%) levels for the LCN on mediastinal images and Smoother (169.5% ± 53.2% for GGO and 136.9% ± 1.6% for HCP) and Smooth (36.4% ± 22.1% and 24.1% ± 0.9%, respectively) levels for parenchymal images. Radiologists considered the images satisfactory for clinical use at these levels, but adaptation to the dose level of the protocol is required.

Conclusion

With AI‐DLR, the smoothest levels reduced the noise and improved the detectability of chest lesions but increased the image smoothing. The opposite was found with the sharpest levels. The choice of level depends on the dose level and type of image: mediastinal or parenchymal.

Virtual Noncontrast Abdominal Imaging with Photon-counting Detector CT

Background Accurate CT attenuation and diagnostic quality of virtual noncontrast (VNC) images acquired with photon-counting detector (PCD) CT are needed to replace true noncontrast (TNC) scans. Purpose To assess the attenuation errors and image quality of VNC images from abdominal PCD CT compared with TNC images. Materials and Methods In this retrospective study, consecutive adult patients who underwent a triphasic examination with PCD CT from July 2021 to October 2021 were included. VNC images were reconstructed from arterial and portal venous phase CT. The absolute attenuation error of VNC compared with TNC images was measured in multiple structures by two readers. Then, two readers blinded to image reconstruction assessed the overall image quality, image noise, noise texture, and delineation of small structures using five-point discrete visual scales (5 = excellent, 1 = nondiagnostic). Overall image quality greater than or equal to 3 was deemed diagnostic. In a phantom, noise texture, spatial resolution, and detectability index were assessed. A detectability index greater than or equal to 5 indicated high diagnostic accuracy. Interreader agreement was evaluated using the Krippendorff α coefficient. The paired t test and Friedman test were applied to compare objective and subjective results. Results Overall, 100 patients (mean age, 72 years ± 10 [SD]; 81 men) were included. In patients, VNC image attenuation values were consistent between readers (α = .60), with errors less than 5 HU in 76% and less than 10 HU in 95% of measurements. There was no evidence of a difference in error of VNC images from arterial or portal venous phase CT (3.3 HU vs 3.5 HU, P = .16). Subjective image quality was rated lower in VNC images for all categories (all, P < .001). Diagnostic quality of VNC images was reached in 99% and 100% of patients for readers 1 and 2, respectively. In the phantom, VNC images exhibited 33% higher noise, blotchier noise texture, similar spatial resolution, and inferior but overall good image quality (detectability index >20) compared with TNC images. Conclusion Abdominal virtual noncontrast images from the arterial and portal venous phase of photon-counting detector CT yielded accurate CT attenuation and good image quality compared with true noncontrast images. © RSNA, 2022 Online supplemental material is available for this article See also the editorial by Sosna in this issue.

Deep learning image reconstruction in pediatric abdominal and chest computed tomography: a comparison of image quality and radiation dose

BACKGROUND

Studies on the application of deep learning image reconstruction (DLIR) in pediatric computed tomography (CT) are limited and have so far been mostly based on phantom. The purpose of this study was to compare the image quality and radiation dose of DLIR with that of adaptive statistical iterative reconstruction-Veo (ASiR-V) during abdominal and chest CT for the pediatric population.

METHODS

A pediatric phantom was used for the pilot study, and 20 children were recruited for clinical verification. The preset scan parameter noise index (NI) was 5, 8, 11, 13, 15, and 18 for the phantom study, and 8 and 13 for the clinical pediatric study. We reconstructed CT images with ASiR-V 30%, ASiR-V 70%, DLIR-M (medium) and DLIR-H (high). The regions of interest (ROI) were marked on the organs of the abdomen (liver, kidney, and subcutaneous fat) and the chest (lung, mediastinum, and spine). The CT dose index volume (CTDIvol), CT value, image noise (N), signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were measured and calculated. The subjective image quality was assessed by 3 radiologists blindly using a 5-point scale. The dose reduction efficiency of DLIR was estimated.

RESULTS

In the phantom study, the interobserver assessment of the data measurement demonstrated good agreement [intraclass correlation coefficient (ICC) =0.814 for abdomen, 0.801 for chest]. Within the same dose level, the N, SNR, and CNR were statistically different among reconstructions, while the CT value remained the same. The N increased and SNR decreased as the radiation dose decreased. The DLIR-H performed better than ASiR-V when the radiation dose was reduced, without sacrificing image quality. In the patient study, the interobserver assessment of the data measurement demonstrated good agreement (ICC =0.774 for abdomen, 0.751 for chest). DLIR-H had the highest subjective and objective scores in the abdomen.

CONCLUSIONS

Application of DLIR could help to reduce radiation dose without sacrificing the image quality of pediatric CT scans. Further clinical validation is required.

Quantum Iterative Reconstruction for Low-Dose Ultra-High-Resolution Photon-Counting Detector CT of the Lung

The aim of this study was to characterize image quality and to determine the optimal strength levels of a novel iterative reconstruction algorithm (quantum iterative reconstruction, QIR) for low-dose, ultra-high-resolution (UHR) photon-counting detector CT (PCD-CT) of the lung. Images were acquired on a clinical dual-source PCD-CT in the UHR mode and reconstructed with a sharp lung reconstruction kernel at different strength levels of QIR (QIR-1 to QIR-4) and without QIR (QIR-off). Noise power spectrum (NPS) and target transfer function (TTF) were analyzed in a cylindrical phantom. 52 consecutive patients referred for low-dose UHR chest PCD-CT were included (CTDIvol: 1 ± 0.6 mGy). Quantitative image quality analysis was performed computationally which included the calculation of the global noise index (GNI) and the global signal-to-noise ratio index (GSNRI). The mean attenuation of the lung parenchyma was measured. Two readers graded images qualitatively in terms of overall image quality, image sharpness, and subjective image noise using 5-point Likert scales. In the phantom, an increase in the QIR level slightly decreased spatial resolution and considerably decreased noise amplitude without affecting the frequency content. In patients, GNI decreased from QIR-off (202 ± 34 HU) to QIR-4 (106 ± 18 HU) (p < 0.001) by 48%. GSNRI increased from QIR-off (4.4 ± 0.8) to QIR-4 (8.2 ± 1.6) (p < 0.001) by 87%. Attenuation of lung parenchyma was highly comparable among reconstructions (QIR-off: -849 ± 53 HU to QIR-4: -853 ± 52 HU, p < 0.001). Subjective noise was best in QIR-4 (p < 0.001), while QIR-3 was best for sharpness and overall image quality (p < 0.001). Thus, our phantom and patient study indicates that QIR-3 provides the optimal iterative reconstruction level for low-dose, UHR PCD-CT of the lungs.

Deep learning reconstruction improves radiomics feature stability and discriminative power in abdominal CT imaging: a phantom study

OBJECTIVES

To compare image quality of deep learning reconstruction (AiCE) for radiomics feature extraction with filtered back projection (FBP), hybrid iterative reconstruction (AIDR 3D), and model-based iterative reconstruction (FIRST).

METHODS

Effects of image reconstruction on radiomics features were investigated using a phantom that realistically mimicked a 65-year-old patient's abdomen with hepatic metastases. The phantom was scanned at 18 doses from 0.2 to 4 mGy, with 20 repeated scans per dose. Images were reconstructed with FBP, AIDR 3D, FIRST, and AiCE. Ninety-three radiomics features were extracted from 24 regions of interest, which were evenly distributed across three tissue classes: normal liver, metastatic core, and metastatic rim. Features were analyzed in terms of their consistent characterization of tissues within the same image (intraclass correlation coefficient ≥ 0.75), discriminative power (Kruskal-Wallis test p value < 0.05), and repeatability (overall concordance correlation coefficient ≥ 0.75).

RESULTS

The median fraction of consistent features across all doses was 6%, 8%, 6%, and 22% with FBP, AIDR 3D, FIRST, and AiCE, respectively. Adequate discriminative power was achieved by 48%, 82%, 84%, and 92% of features, and 52%, 20%, 17%, and 39% of features were repeatable, respectively. Only 5% of features combined consistency, discriminative power, and repeatability with FBP, AIDR 3D, and FIRST versus 13% with AiCE at doses above 1 mGy and 17% at doses ≥ 3 mGy. AiCE was the only reconstruction technique that enabled extraction of higher-order features.

CONCLUSIONS

AiCE more than doubled the yield of radiomics features at doses typically used clinically. Inconsistent tissue characterization within CT images contributes significantly to the poor stability of radiomics features.

KEY POINTS

• Image quality of CT images reconstructed with filtered back projection and iterative methods is inadequate for the majority of radiomics features due to inconsistent tissue characterization, low discriminative power, or low repeatability. • Deep learning reconstruction enhances image quality for radiomics and more than doubled the feature yield at doses that are typically used in clinical CT imaging. • Image reconstruction algorithms can optimize image quality for more reliable quantification of tissues in CT images.