ABSTRACT
TumorPrism3D software was developed to segment brain tumors with a straightforward and user-friendly graphical interface applied to two- and three-dimensional brain magnetic resonance (MR) images. The MR images of 185 patients (103 males, 82 females) with glioblastoma multiforme were downloaded from The Cancer Imaging Archive (TCIA) to test the tumor segmentation performance of this software. Regions of interest (ROIs) corresponding to contrast-enhancing lesions, necrotic portions, and non-enhancing T2 high signal intensity components were segmented for each tumor. TumorPrism3D demonstrated high accuracy in segmenting all three tumor components in cases of glioblastoma multiforme. They achieved a better Dice similarity coefficient (DSC) ranging from 0.83 to 0.91 than 3DSlicer with a DSC ranging from 0.80 to 0.84 for the accuracy of segmented tumors. Comparative analysis with the widely used 3DSlicer software revealed TumorPrism3D to be approximately 37.4% faster in the segmentation process from initial contour drawing to final segmentation mask determination. The semi-automated nature of TumorPrism3D facilitates reproducible tumor segmentation at a rapid pace, offering the potential for quantitative analysis of tumor characteristics and artificial intelligence-assisted segmentation in brain MR imaging.
ABSTRACT
OBJECTIVE: Few studies have explored the clinical feasibility of using deep-learning reconstruction to reduce the radiation dose of CT. We aimed to compare the image quality and lung nodule detectability between chest CT using a quarter of the low dose (QLD) reconstructed with vendor-agnostic deep-learning image reconstruction (DLIR) and conventional low-dose (LD) CT reconstructed with iterative reconstruction (IR). MATERIALS AND METHODS: We retrospectively collected 100 patients (median age, 61 years [IQR, 53-70 years]) who received LDCT using a dual-source scanner, where total radiation was split into a 1:3 ratio. QLD CT was generated using a quarter dose and reconstructed with DLIR (QLD-DLIR), while LDCT images were generated using a full dose and reconstructed with IR (LD-IR). Three thoracic radiologists reviewed subjective noise, spatial resolution, and overall image quality, and image noise was measured in five areas. The radiologists were also asked to detect all Lung-RADS category 3 or 4 nodules, and their performance was evaluated using area under the jackknife free-response receiver operating characteristic curve (AUFROC). RESULTS: The median effective dose was 0.16 (IQR, 0.14-0.18) mSv for QLD CT and 0.65 (IQR, 0.57-0.71) mSv for LDCT. The radiologists' evaluations showed no significant differences in subjective noise (QLD-DLIR vs. LD-IR, lung-window setting; 3.23 ± 0.19 vs. 3.27 ± 0.22; P = .11), spatial resolution (3.14 ± 0.28 vs. 3.16 ± 0.27; P = .12), and overall image quality (3.14 ± 0.21 vs. 3.17 ± 0.17; P = .15). QLD-DLIR demonstrated lower measured noise than LD-IR in most areas (P < .001 for all). No significant difference was found between QLD-DLIR and LD-IR for the sensitivity (76.4% vs. 72.2%; P = .35) or the AUFROCs (0.77 vs. 0.78; P = .68) in detecting Lung-RADS category 3 or 4 nodules. Under a noninferiority limit of -0.1, QLD-DLIR showed noninferior detection performance (95% CI for AUFROC difference, -0.04 to 0.06). CONCLUSION: QLD-DLIR images showed comparable image quality and noninferior nodule detectability relative to LD-IR images.
Subject(s)
Deep Learning , Lung Neoplasms , Humans , Middle Aged , Drug Tapering , Lung Neoplasms/diagnostic imaging , Retrospective Studies , Tomography, X-Ray ComputedABSTRACT
OBJECTIVE: To investigate performance of 1-mm, sharp kernel, low-dose chest computed tomography (LDCT) for coronary artery calcium scoring (CACS) using deep learning (DL)-based denoising technique. METHODS: This retrospective, intra-individual comparative study consisted of four image datasets of 131 participants who underwent LDCT and calcium CT on the same day between January and February 2020; 1-mm LDCT with DL, 1-mm LDCT with iterative reconstruction (IR), 3-mm LDCT, and calcium CT. CACS from calcium CT were considered as reference and CACS were categorized as 0, 1-10, 11-100, 101-400, and > 400. We compared CACS from LDCTs with that from calcium CT. RESULTS: Mean CACS was 104.8 ± 249.1 and proportion of positive CACS was 45% (59/131). CACS from LDCT images tended to be underestimated than those from calcium CT: 1-mm LDCT with DL (93.5 ± 249.6, p = 0.002), 1-mm LDCT with IR (94.7 ± 249.9, p < 0.001), and 3-mm LDCT (90.3 ± 245.3, p = 0.004). All LDCT datasets showed excellent agreement with calcium CT: intraclass correlation coefficient (ICC) = 0.961 (95% confidence interval (CI), 0.945-0.972) for DL, 0.969 (95% CI, 0.956-0.978) for IR, and 0.952 (95% CI, 0.932-0.966) for 3-mm LDCT; weighted kappa for CACS classification, 0.930 (95% CI, 0.893-0.966) for 1-mm LDCT with DL, 0.908 (95% CI, 0.866-0.950) for 1-mm LDCT with IR, and 0.846 (95% CI, 0.780-0.912) for 3-mm LDCT. The accuracy of CACS classification of 1-mm LDCT with DL (90%) tended to be better than 1-mm LDCT with IR (87%) and 3-mm LDCT (84.7%) (p = 0.10). CONCLUSION: DL-based noise reduction algorithm can offer reliable calcium scores in 1-mm LDCT reconstructed with sharp kernel. KEY POINTS: ⢠Deep learning (DL)-based noise reduction enables calcium scoring at 1-mm, sharp kernel reconstructed low-dose chest CT (LDCT). ⢠Both iterative reconstruction and DL-based noise reduction underestimated calcium score, but agreement were excellent with those from calcium CT. ⢠Accuracy of categorical classification of calcium scoring tended to be highest in 1-mm LDCT with DL compared to 1-mm LDCT with IR and 3-mm LDCT (90%, 87%, and 84.7%, p = 0.10).
Subject(s)
Coronary Artery Disease , Deep Learning , Humans , Coronary Artery Disease/diagnostic imaging , Calcium , Retrospective Studies , Tomography, X-Ray Computed/methods , Algorithms , Radiation Dosage , Radiographic Image Interpretation, Computer-Assisted/methodsABSTRACT
Gaining the ability to audit the behavior of deep learning (DL) denoising models is of crucial importance to prevent potential hallucinations and adversarial clinical consequences. We present a preliminary version of AntiHalluciNet, which is designed to predict spurious structural components embedded in the residual noise from DL denoising models in low-dose CT and assess its feasibility for auditing the behavior of DL denoising models. We created a paired set of structure-embedded and pure noise images and trained AntiHalluciNet to predict spurious structures in the structure-embedded noise images. The performance of AntiHalluciNet was evaluated by using a newly devised residual structure index (RSI), which represents the prediction confidence based on the presence of structural components in the residual noise image. We also evaluated whether AntiHalluciNet could assess the image fidelity of a denoised image by using only a noise component instead of measuring the SSIM, which requires both reference and test images. Then, we explored the potential of AntiHalluciNet for auditing the behavior of DL denoising models. AntiHalluciNet was applied to three DL denoising models (two pre-trained models, RED-CNN and CTformer, and a commercial software, ClariCT.AI [version 1.2.3]), and whether AntiHalluciNet could discriminate between the noise purity performances of DL denoising models was assessed. AntiHalluciNet demonstrated an excellent performance in predicting the presence of structural components. The RSI values for the structural-embedded and pure noise images measured using the 50% low-dose dataset were 0.57 ± 31 and 0.02 ± 0.02, respectively, showing a substantial difference with a p-value < 0.0001. The AntiHalluciNet-derived RSI could differentiate between the quality of the degraded denoised images, with measurement values of 0.27, 0.41, 0.48, and 0.52 for the 25%, 50%, 75%, and 100% mixing rates of the degradation component, which showed a higher differentiation potential compared with the SSIM values of 0.9603, 0.9579, 0.9490, and 0.9333. The RSI measurements from the residual images of the three DL denoising models showed a distinct distribution, being 0.28 ± 0.06, 0.21 ± 0.06, and 0.15 ± 0.03 for RED-CNN, CTformer, and ClariCT.AI, respectively. AntiHalluciNet has the potential to predict the structural components embedded in the residual noise from DL denoising models in low-dose CT. With AntiHalluciNet, it is feasible to audit the performance and behavior of DL denoising models in clinical environments where only residual noise images are available.
ABSTRACT
We propose a deep learning-assisted overscan decision algorithm in chest low-dose computed tomography (LDCT) applicable to the lung cancer screening. The algorithm reflects the radiologists' subjective evaluation criteria according to the Korea institute for accreditation of medical imaging (KIAMI) guidelines, where it judges whether a scan range is beyond landmarks' criterion. The algorithm consists of three stages: deep learning-based landmark segmentation, rule-based logical operations, and overscan determination. A total of 210 cases from a single institution (internal data) and 50 cases from 47 institutions (external data) were utilized for performance evaluation. Area under the receiver operating characteristic (AUROC), accuracy, sensitivity, specificity, and Cohen's kappa were used as evaluation metrics. Fisher's exact test was performed to present statistical significance for the overscan detectability, and univariate logistic regression analyses were performed for validation. Furthermore, an excessive effective dose was estimated by employing the amount of overscan and the absorbed dose to effective dose conversion factor. The algorithm presented AUROC values of 0.976 (95% confidence interval [CI]: 0.925-0.987) and 0.997 (95% CI: 0.800-0.999) for internal and external dataset, respectively. All metrics showed average performance scores greater than 90% in each evaluation dataset. The AI-assisted overscan decision and the radiologist's manual evaluation showed a statistically significance showing a p-value less than 0.001 in Fisher's exact test. In the logistic regression analysis, demographics (age and sex), data source, CT vendor, and slice thickness showed no statistical significance on the algorithm (each p-value > 0.05). Furthermore, the estimated excessive effective doses were 0.02 ± 0.01 mSv and 0.03 ± 0.05 mSv for each dataset, not a concern within slight deviations from an acceptable scan range. We hope that our proposed overscan decision algorithm enables the retrospective scan range monitoring in LDCT for lung cancer screening program, and follows an as low as reasonably achievable (ALARA) principle.
Subject(s)
Deep Learning , Lung Neoplasms , Accreditation , Algorithms , Early Detection of Cancer , Humans , Lung Neoplasms/diagnostic imaging , Retrospective Studies , Tomography, X-Ray ComputedABSTRACT
While the recent advancements of computed tomography (CT) technology have contributed in reducing radiation dose and image noise, an objective evaluation of image quality in patient scans has not yet been established. In this study, we present a patient-specific CT image quality evaluation method that includes fully automated measurements of noise level, structure sharpness, and alteration of structure. This study used the CT images of 120 patients from four different CT scanners reconstructed with three types of algorithm: filtered back projection (FBP), vendor-specific iterative reconstruction (IR), and a vendor-agnostic deep learning model (DLM, ClariCT.AI, ClariPi Inc.). The structure coherence feature (SCF) was used to divide an image into the homogeneous (RH) and structure edge (RS) regions, which in turn were used to localize the regions of interests (ROIs) for subsequent analysis of image quality indices. The noise level was calculated by averaging the standard deviations from five randomly selected ROIs on RH, and the mean SCFs on RS was used to estimate the structure sharpness. The structure alteration was defined by the standard deviation ratio between RS and RH on the subtraction image between FBP and IR or DLM, in which lower structure alterations indicate successful noise reduction without degradation of structure details. The estimated structure sharpness showed a high correlation of 0.793 with manually measured edge slopes. Compared to FBP, IR and DLM showed 34.38% and 51.30% noise reduction, 2.87% and 0.59% lower structure sharpness, and 2.20% and -12.03% structure alteration, respectively, on an average. DLM showed statistically superior performance to IR in all three image quality metrics. This study is expected to contribute to enhance the CT protocol optimization process by allowing a high throughput and quantitative image quality evaluation during the introduction or adjustment of lower-dose CT protocol into routine practice.
Subject(s)
Radiographic Image Interpretation, Computer-Assisted , Tomography, X-Ray Computed , Algorithms , Humans , Radiation Dosage , Radiographic Image Interpretation, Computer-Assisted/methods , Tomography, X-Ray Computed/methodsABSTRACT
OBJECTIVE: To investigate clinical applicability of deep learning(DL)-based reconstruction of virtual monoenergetic images(VMIs) of arterial phase liver CT obtained by rapid kVp-switching dual-energy CT for evaluation of hypervascular liver lesions. MATERIALS AND METHODS: We retrospectively included 109 patients who had available late arterial phase liver CT images of the liver obtained with a rapid switching kVp DECT scanner for suspicious intra-abdominal malignancies. Two VMIs of 70 keV and 40 keV were reconstructed using adaptive statistical iterative reconstruction (ASiR-V) for arterial phase scans. VMIs at 40 keV were additionally reconstructed with a vendor-agnostic DL-based reconstruction technique (ClariCT.AI, ClariPi, DL 40 keV). Qualitative, quantitative image quality and subjective diagnostic acceptability were compared according to reconstruction techniques. RESULTS: In qualitative analysis, DL 40 keV images showed less image noise (4.55 vs 3.11 vs 3.95, p < 0.001), better image sharpness (4.75 vs 4.16 vs 4.3, p < 0.001), better image contrast (4.98 vs 4.72 vs 4.19, p < 0.017), better lesion conspicuity (4.61 vs 4.23 vs 3.4, p < 0.001) and diagnostic acceptability (4.59 vs 3.88 vs 4.09, p < 0.001) compared with ASiR-V 40 keV or 70 keV image sets. In quantitative analysis, DL 40 keV significantly reduced image noise relative to ASiR-V 40 keV images (49.9%, p < 0.001) and ASiR-V 70 keV images (85.2%, p = 0.012). DL 40 keV images showed significantly higher CNRlesion to the liver and SNRliver than ASiR-V 40 keV image and 70 keV images (p < 0.001). CONCLUSION: DL-based reconstruction of 40 keV images using vendor-agnostic software showed greater noise reduction, better lesion conspicuity, image contrast, image sharpness, and higher overall image diagnostic acceptability than ASiR for 40 keV or 70 keV images in patients with hypervascular liver lesions.
Subject(s)
Deep Learning , Liver Neoplasms , Humans , Liver Neoplasms/blood supply , Liver Neoplasms/diagnostic imaging , Radiographic Image Interpretation, Computer-Assisted/methods , Retrospective Studies , Signal-To-Noise Ratio , Tomography, X-Ray Computed/methodsABSTRACT
OBJECTIVES: To evaluate the diagnostic value of deep learning model (DLM) reconstructed dual-energy CT (DECT) low-keV virtual monoenergetic imaging (VMI) for assessing hypoenhancing hepatic metastases. METHODS: This retrospective study included 131 patients who underwent contrast-enhanced DECT (80-kVp and 150-kVp with a tin filter) in the portal venous phase for hepatic metastasis surveillance. Linearly blended images simulating 100-kVp images (100-kVp), standard 40-keV VMI images (40-keV VMI), and post-processed 40-keV VMI using a vendor-agnostic DLM (i.e., DLM 40-keV VMI) were reconstructed. Lesion conspicuity and diagnostic acceptability were assessed by three independent reviewers and compared using the Wilcoxon signed-rank test. The contrast-to-noise ratios (CNRs) were also measured placing ROIs in metastatic lesions and liver parenchyma. The detection performance of hepatic metastases was assessed by using a jackknife alternative free-response ROC method. The consensus by two independent radiologists was used as the reference standard. RESULTS: DLM 40-keV VMI, compared to 40-keV VMI and 100-kVp, showed a higher lesion-to-liver CNR (8.25 ± 3.23 vs. 6.05 ± 2.38 vs. 5.99 ± 2.00), better lesion conspicuity (4.3 (4.0-4.7) vs. 3.7 (3.7-4.0) vs. 3.7 (3.3-4.0)), and better diagnostic acceptability (4.3 (4.0-4.3) vs. 3.0 (2.7-3.3) vs. 4.0 (4.0-4.3)) (p < 0.001 for all). For lesion detection (246 hepatic metastases in 68 patients), the figure of merit was significantly higher with DLM 40-keV VMI than with 40-keV VMI (0.852 vs. 0.822, p = 0.012), whereas no significant difference existed between DLM 40-keV VMI and 100-kVp (0.852 vs. 0.842, p = 0.31). CONCLUSIONS: DLM 40-keV VMI provided better image quality and comparable diagnostic performance for detecting hypoenhancing hepatic metastases compared to linearly blended images. KEY POINTS: ⢠DLM 40-keV VMI provides a superior image quality compared with 40-keV or 100-kVp for assessing hypoenhancing hepatic metastasis. ⢠DLM 40-keV VMI has the highest CNR and lesion conspicuity score for hypoenhancing hepatic metastasis due to noise reduction and structural preservation. ⢠DLM 40-keV VMI provides higher lesion detectability than standard 40-keV VMI (p = 0.012).
Subject(s)
Deep Learning , Liver Neoplasms , Radiography, Dual-Energy Scanned Projection , Humans , Image Processing, Computer-Assisted/methods , Liver Neoplasms/diagnostic imaging , Liver Neoplasms/secondary , Radiographic Image Interpretation, Computer-Assisted/methods , Radiography, Dual-Energy Scanned Projection/methods , Retrospective Studies , Signal-To-Noise Ratio , Tomography, X-Ray Computed/methodsABSTRACT
OBJECTIVES: To compare the dose reduction potential (DRP) of a vendor-agnostic deep learning model (DLM, ClariCT.AI) with that of a vendor-specific deep learning-based image reconstruction algorithm (DLR, TrueFidelity™). METHODS: Computed tomography (CT) images of a multi-sized image quality phantom (Mercury v4.0) were acquired under six radiation dose levels (0.48/0.97/1.93/3.87/7.74/15.47 mGy) and were reconstructed using filtered back projection (FBP) and three strength levels of the DLR (low/medium/high). The FBP images were denoised using the DLM. For all DLM and DLR images, the detectability index (d') (a task-based detection performance metric) was obtained, under various combinations of three target sizes (10/5/1 mm), five inlets (CT value difference with the background; -895/50/90/335/1000 HU), five phantom diameters (36/31/26/21/16 cm), and six radiation dose levels. Dose reduction potential (DRP) measures the dose reduction made by using DLM or DLR, while yielding d' equivalent to that of FBP at full dose. RESULTS: The DRPs of the DLM, DLR-low, DLR-medium, and DLR-high were 86% (81-88%), 60% (46-67%), 76% (60-81%), and 87% (78-92%), respectively. For 10-mm targets, the DRP of the DLM (87%) was higher than that of all DLR algorithms (58-86%). However, for smaller targets (5 mm/1 mm), the DRPs of the DLR-high (89/88%) were greater than those of the DLM (87/84%). CONCLUSION: The dose reduction potential of the vendor-agnostic DLM was shown to be comparable to that of the vendor-specific DLR at high strength and superior to those of the DLRs at medium and low strengths. KEY POINTS: ⢠DRP of the vendor-agnostic model was comparable to that of high-strength vendor-specific model and superior to those of medium- and low-strength models. ⢠Under various radiation dose levels, the deep learning model shows higher detectability indexes compared to FBP.
Subject(s)
Deep Learning , Algorithms , Drug Tapering , Humans , Image Processing, Computer-Assisted , Phantoms, Imaging , Radiation Dosage , Radiographic Image Interpretation, Computer-Assisted , Tomography, X-Ray ComputedABSTRACT
OBJECTIVES: To compare the overall image quality and detectability of significant (malignant and pre-malignant) liver lesions of low-dose liver CT (LDCT, 33.3% dose) using deep learning denoising (DLD) to standard-dose CT (SDCT, 100% dose) using model-based iterative reconstruction (MBIR). METHODS: In this retrospective study, CT images of 80 patients with hepatic focal lesions were included. For noninferiority analysis of overall image quality, a margin of - 0.5 points (scored in a 5-point scale) for the difference between scan protocols was pre-defined. Other quantitative or qualitative image quality assessments were performed. Additionally, detectability of significant liver lesions was compared, with 64 pairs of CT, using the jackknife alternative free-response ROC analysis, with noninferior margin defined by the lower limit of 95% confidence interval (CI) of the difference of figure-of-merit less than - 0.1. RESULTS: The mean overall image quality scores with LDCT and SDCT were 3.77 ± 0.38 and 3.94 ± 0.34, respectively, demonstrating a difference of - 0.17 (95% CI: - 0.21 to - 0.12), which did not cross the predefined noninferiority margin of - 0.5. Furthermore, LDCT showed significantly superior quantitative results of liver lesion contrast to noise ratio (p < 0.05). However, although LDCT scored higher than the average score in qualitative image quality assessments, they were significantly lower than those of SDCT (p < 0.05). Figure-of-merit for lesion detection was 0.859 for LDCT and 0.878 for SDCT, showing noninferiority (difference: - 0.019, 95% CI: - 0.058 to 0.021). CONCLUSION: LDCT using DLD with 67% radiation dose reduction showed non-inferior overall image quality and lesion detectability, compared to SDCT. KEY POINTS: ⢠Low-dose liver CT using deep learning denoising (DLD), at 67% dose reduction, provided non-inferior overall image quality compared to standard-dose CT using model-based iterative reconstruction (MBIR). ⢠Low-dose CT using DLD showed significantly less noise and higher CNR lesion to liver than standard-dose CT using MBIR and demonstrated at least average image quality score among all readers, albeit with lower scores than standard-dose CT using MBIR. ⢠Low-dose liver CT showed noninferior detectability for malignant and pre-malignant liver lesions, compared to standard-dose CT.
Subject(s)
Deep Learning , Liver Neoplasms , Algorithms , Humans , Liver Neoplasms/diagnostic imaging , Radiation Dosage , Radiographic Image Interpretation, Computer-Assisted/methods , Retrospective Studies , Tomography, X-Ray Computed/methodsABSTRACT
OBJECTIVE: The purpose of this study was to assess whether a deep learning (DL) algorithm could enable simultaneous noise reduction and edge sharpening in low-dose lumbar spine CT. MATERIALS AND METHODS: This retrospective study included 52 patients (26 male and 26 female; median age, 60.5 years) who had undergone CT-guided lumbar bone biopsy between October 2015 and April 2020. Initial 100-mAs survey images and 50-mAs intraprocedural images were reconstructed by filtered back projection. Denoising was performed using a vendor-agnostic DL model (ClariCT.AI™, ClariPI) for the 50-mAS images, and the 50-mAs, denoised 50-mAs, and 100-mAs CT images were compared. Noise, signal-to-noise ratio (SNR), and edge rise distance (ERD) for image sharpness were measured. The data were summarized as the mean ± standard deviation for these parameters. Two musculoskeletal radiologists assessed the visibility of the normal anatomical structures. RESULTS: Noise was lower in the denoised 50-mAs images (36.38 ± 7.03 Hounsfield unit [HU]) than the 50-mAs (93.33 ± 25.36 HU) and 100-mAs (63.33 ± 16.09 HU) images (p < 0.001). The SNRs for the images in descending order were as follows: denoised 50-mAs (1.46 ± 0.54), 100-mAs (0.99 ± 0.34), and 50-mAs (0.58 ± 0.18) images (p < 0.001). The denoised 50-mAs images had better edge sharpness than the 100-mAs images at the vertebral body (ERD; 0.94 ± 0.2 mm vs. 1.05 ± 0.24 mm, p = 0.036) and the psoas (ERD; 0.42 ± 0.09 mm vs. 0.50 ± 0.12 mm, p = 0.002). The denoised 50-mAs images significantly improved the visualization of the normal anatomical structures (p < 0.001). CONCLUSION: DL-based reconstruction may enable simultaneous noise reduction and improvement in image quality with the preservation of edge sharpness on low-dose lumbar spine CT. Investigations on further radiation dose reduction and the clinical applicability of this technique are warranted.
Subject(s)
Deep Learning , Algorithms , Female , Humans , Male , Middle Aged , Pilot Projects , Radiation Dosage , Radiographic Image Interpretation, Computer-Assisted , Retrospective Studies , Tomography, X-Ray ComputedABSTRACT
OBJECTIVE: To compare the image quality between the vendor-agnostic and vendor-specific algorithms on ultralow-dose chest CT. METHODS: Vendor-agnostic deep learning post-processing model (DLM), vendor-specific deep learning image reconstruction (DLIR, high level), and adaptive statistical iterative reconstruction (ASiR, 70%) algorithms were employed. One hundred consecutive ultralow-dose noncontrast CT scans (CTDIvol; mean, 0.33 ± 0.056 mGy) were reconstructed with five algorithms: DLM-stnd (standard kernel), DLM-shrp (sharp kernel), DLIR, ASiR-stnd, and ASiR-shrp. Three thoracic radiologists blinded to the reconstruction algorithms reviewed five sets of 100 images and assessed subjective noise, spatial resolution, distortion artifact, and overall image quality. They selected the most preferred algorithm among five image sets for each case. Image noise and signal-to-noise ratio were measured. Edge-rise-distance was measured at a pulmonary vessel, i.e., the distance between two points where attenuation was 10% and 90% of maximal intravascular intensity. The skewness of attenuation was calculated in homogeneous areas. RESULTS: DLM-stnd, followed by DLIR, showed the best subjective noise on both lung and mediastinal windows, while DLIR yielded the least measured noise (ps < .0001). Compared to DLM-stnd, DLIR showed inferior subjective spatial resolution on lung window and higher edge-rise-distance (ps < .0001). Additionally, DLIR showed the most frequent distortion artifacts and deviated skewness (ps < .0001). DLM-stnd scored the best overall image quality, followed by DLM-shrp and DLIR (mean score 3.89 ± 0.19, 3.68 ± 0.24, and 3.53 ± 0.33; ps < .001). Two among three readers preferred DLM-stnd on both windows. CONCLUSION: Although DLIR provided the best quantitative noise profile, DLM-stnd showed the best overall image quality with fewer artifacts and was preferred by two among three readers. KEY POINTS: ⢠A vendor-agnostic deep learning post-processing algorithm applied to ultralow-dose chest CT exhibited the best image quality compared to vendor-specific deep learning algorithm and ASiR techniques. ⢠Two out of three readers preferred a vendor-agnostic deep learning post-processing algorithm in comparison to vendor-specific deep learning algorithm and ASiR techniques. ⢠A vendor-specific deep learning reconstruction algorithm yielded the least image noise, but showed significantly more frequent specific distortion artifacts and increased skewness of attenuation compared to a vendor-agnostic algorithm.
Subject(s)
Deep Learning , Algorithms , Humans , Radiation Dosage , Radiographic Image Interpretation, Computer-Assisted , Thorax , Tomography, X-Ray ComputedABSTRACT
To evaluate the effect of radiation dose reduction on image quality and diagnostic confidence in contrast-enhanced whole-body computed tomography (WBCT) staging. We randomly selected March 2016 for retrospective inclusion of 18 consecutive patients (14 female, 60 ± 15 years) with clinically indicated WBCT staging on the same 3rd generation dual-source CT. Using low-dose simulations, we created data sets with 100, 80, 60, 40, and 20% of the original radiation dose. Each set was reconstructed using filtered back projection (FBP) and Advanced Modeled Iterative Reconstruction (ADMIRE®, Siemens Healthineers, Forchheim, Germany) strength 1-5, resulting in 540 datasets total. ADMIRE 2 was the reference standard for intraindividual comparison. The effective radiation dose was calculated using commercially available software. For comparison of objective image quality, noise assessments of subcutaneous adipose tissue regions were performed automatically using the software. Three radiologists blinded to the study evaluated image quality and diagnostic confidence independently on an equidistant 5-point Likert scale (1 = poor to 5 = excellent). At 100%, the effective radiation dose in our population was 13.3 ± 9.1 mSv. At 20% radiation dose, it was possible to obtain comparably low noise levels when using ADMIRE 5 (p = 1.000, r = 0.29). We identified ADMIRE 3 at 40% radiation dose (5.3 ± 3.6 mSv) as the lowest achievable radiation dose with image quality and diagnostic confidence equal to our reference standard (p = 1.000, r > 0.4). The inter-rater agreement for this result was almost perfect (ICC ≥ 0.958, 95% CI 0.909-0.983). On a 3rd generation scanner, it is feasible to maintain good subjective image quality, diagnostic confidence, and image noise in single-energy WBCT staging at dose levels as low as 40% of the original dose (5.3 ± 3.6 mSv), when using ADMIRE 3.
ABSTRACT
OBJECTIVES: To evaluate the image quality of low iodine concentration, dual-energy CT (DECT) combined with a deep learning-based noise reduction technique for pediatric abdominal CT, compared with standard iodine concentration single-energy polychromatic CT (SECT). METHODS: From December 2016 to May 2017, DECT with 300 mgâ¢I/mL contrast medium was performed in 29 pediatric patients (17 boys, 12 girls; age, 2-19 years). The DECT images were reconstructed using a noise-optimized virtual monoenergetic reconstruction image (VMI) with and without a deep learning method. SECT images with 350 mgâ¢I/mL contrast medium, performed within the last 3 months before the DECT, served as reference images. The quantitative and qualitative parameters were compared using paired t tests and Wilcoxon signed-rank tests, and the differences in radiation dose and total iodine administration were assessed. RESULTS: The linearly blended DECT showed lower attenuation and higher noise than SECT. The 60-keV VMI showed an increase in attenuation and higher noise than SECT. The combined 60-keV VMI plus deep learning images showed low noise, no difference in contrast-to-noise ratios, and overall image quality or diagnostic image quality, but showed a higher signal-to-noise ratio in the liver and lower enhancement of lesions than SECT. The overall image and diagnostic quality of lesions were maintained on the combined noise reduction approach. The CT dose index volume and total iodine administration in DECT were respectively 19.6% and 14.3% lower than those in SECT. CONCLUSION: Low iodine concentration DECT, combined with deep learning in pediatric abdominal CT, can maintain image quality while reducing the radiation dose and iodine load, compared with standard SECT. KEY POINTS: ⢠An image noise reduction approach combining deep learning and noise-optimized virtual monoenergetic image reconstruction can maintain image quality while reducing radiation dose and iodine load. ⢠The 60-keV virtual monoenergetic image reconstruction plus deep learning images showed low noise, no difference in contrast-to-noise ratio, and overall image quality, but showed a higher signal-to-noise ratio in the liver and a lower enhancement of lesion than single-energy polychromatic CT. ⢠This combination could offer a 19.6% reduction in radiation dose and a 14.3% reduction in iodine load, in comparison with a control group that underwent single-energy polychromatic CT with the standard protocol.
Subject(s)
Deep Learning , Radiography, Dual-Energy Scanned Projection , Adolescent , Adult , Child , Child, Preschool , Contrast Media , Female , Humans , Male , Retrospective Studies , Signal-To-Noise Ratio , Tomography, X-Ray Computed , Young AdultABSTRACT
OBJECTIVE: This study determined the effect of dose reduction and kernel selection on quantifying emphysema using low-dose computed tomography (LDCT) and evaluated the efficiency of a deep learning-based kernel conversion technique in normalizing kernels for emphysema quantification. METHODS: A sample of 131 participants underwent LDCT and standard-dose computed tomography (SDCT) at 1- to 2-year intervals. LDCT images were reconstructed with B31f and B50f kernels, and SDCT images were reconstructed with B30f kernels. A deep learning model was used to convert the LDCT image from a B50f kernel to a B31f kernel. Emphysema indices (EIs), lung attenuation at 15th percentile (perc15), and mean lung density (MLD) were calculated. Comparisons among the different kernel types for both LDCT and SDCT were performed using Friedman's test and Bland-Altman plots. RESULTS: All values of LDCT B50f were significantly different compared with the values of LDCT B31f and SDCT B30f (p < 0.05). Although there was a statistical difference, the variation of the values of LDCT B50f significantly decreased after kernel normalization. The 95% limits of agreement between the SDCT and LDCT kernels (B31f and converted B50f) ranged from - 2.9 to 4.3% and from - 3.2 to 4.4%, respectively. However, there were no significant differences in EIs and perc15 between SDCT and LDCT converted B50f in the non-chronic obstructive pulmonary disease (COPD) participants (p > 0.05). CONCLUSION: The deep learning-based CT kernel conversion of sharp kernel in LDCT significantly reduced variation in emphysema quantification, and could be used for emphysema quantification. KEY POINTS: ⢠Low-dose computed tomography with smooth kernel showed adequate performance in quantifying emphysema compared with standard-dose CT. ⢠Emphysema quantification is affected by kernel selection and the application of a sharp kernel resulted in a significant overestimation of emphysema. ⢠Deep learning-based kernel normalization of sharp kernel significantly reduced variation in emphysema quantification.
Subject(s)
Deep Learning , Diagnosis, Computer-Assisted/methods , Pulmonary Emphysema/diagnostic imaging , Tomography, X-Ray Computed , Aged , Biometry , Cross-Sectional Studies , Female , Humans , Lung/diagnostic imaging , Lung Diseases/diagnostic imaging , Male , Middle Aged , Radiation Dosage , Retrospective Studies , Treatment OutcomeABSTRACT
OBJECTIVE: To assess the feasibility of applying a deep learning-based denoising technique to coronary CT angiography (CCTA) along with iterative reconstruction for additional noise reduction. MATERIALS AND METHODS: We retrospectively enrolled 82 consecutive patients (male:female = 60:22; mean age, 67.0 ± 10.8 years) who had undergone both CCTA and invasive coronary artery angiography from March 2017 to June 2018. All included patients underwent CCTA with iterative reconstruction (ADMIRE level 3, Siemens Healthineers). We developed a deep learning based denoising technique (ClariCT.AI, ClariPI), which was based on a modified U-net type convolutional neural net model designed to predict the possible occurrence of low-dose noise in the originals. Denoised images were obtained by subtracting the predicted noise from the originals. Image noise, CT attenuation, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were objectively calculated. The edge rise distance (ERD) was measured as an indicator of image sharpness. Two blinded readers subjectively graded the image quality using a 5-point scale. Diagnostic performance of the CCTA was evaluated based on the presence or absence of significant stenosis (≥ 50% lumen reduction). RESULTS: Objective image qualities (original vs. denoised: image noise, 67.22 ± 25.74 vs. 52.64 ± 27.40; SNR [left main], 21.91 ± 6.38 vs. 30.35 ± 10.46; CNR [left main], 23.24 ± 6.52 vs. 31.93 ± 10.72; all p < 0.001) and subjective image quality (2.45 ± 0.62 vs. 3.65 ± 0.60, p < 0.001) improved significantly in the denoised images. The average ERDs of the denoised images were significantly smaller than those of originals (0.98 ± 0.08 vs. 0.09 ± 0.08, p < 0.001). With regard to diagnostic accuracy, no significant differences were observed among paired comparisons. CONCLUSION: Application of the deep learning technique along with iterative reconstruction can enhance the noise reduction performance with a significant improvement in objective and subjective image qualities of CCTA images.
Subject(s)
Computed Tomography Angiography/methods , Deep Learning , Adult , Aged , Aged, 80 and over , Coronary Artery Disease/diagnostic imaging , Female , Humans , Male , Middle Aged , Radiographic Image Interpretation, Computer-Assisted/methods , Retrospective Studies , Signal-To-Noise RatioABSTRACT
PURPOSE: To examine the potential effect of CT dose variation on radiomic features in vivo using simulated contrast-enhanced CT dose reduction in patients with non-small lung cell cancer (NSCLC). METHODS: In this retrospective study, we included 69 patients (25 females, 44 males, median age 66 years) with histologically proven NSCLC who underwent a whole contrast-enhanced body FDG-PET/CT for primary staging. To simulate different CT dose levels, we used an algorithm to simulate low-dose CT images based on a noise model derived from phantom experiments. The tumor lesions and reference regions in the paraspinal muscle were manually segmented to obtain three-dimensional regions of interest. Radiomic feature extraction was performed using the PyRadiomics toolbox. The median relative feature value deviation was assessed for each feature and each dose level. RESULTS: The mean segmented tumor volume was 340 ml. T-stages of the primary tumors were primarily T3/4. For NSCLCs, the median relative feature value deviation in the lowest dose images varied for the calculated features from 52.2% to -49.5%. In general, dose-dependent deviations of feature values showed a monotonous increase or decrease with decreasing dose levels. Statistical analyses revealed significant differences between the dose levels in 91% of features. CONCLUSIONS: We examined the effects of simulated CT dose reduction on the values of radiomic features in primary NSCLC and showed significant deviations of varying degrees in numerous feature classes. This is a theoretical indicator of potential influence under real conditions, which should be taken into account in clinical use.
Subject(s)
Carcinoma, Non-Small-Cell Lung/diagnostic imaging , Contrast Media , Lung Neoplasms/diagnostic imaging , Positron Emission Tomography Computed Tomography/methods , Radiation Dosage , Radiographic Image Enhancement/methods , Aged , Aged, 80 and over , Carcinoma, Non-Small-Cell Lung/pathology , Computer Simulation , Female , Humans , Lung Neoplasms/pathology , Male , Middle Aged , Neoplasm Staging , Retrospective Studies , Whole Body Imaging/methodsABSTRACT
OBJECTIVE: To compare image qualities between vendor-neutral and vendor-specific hybrid iterative reconstruction (IR) techniques for abdominopelvic computed tomography (CT) in young patients. MATERIALS AND METHODS: In phantom study, we used an anthropomorphic pediatric phantom, age-equivalent to 5-year-old, and reconstructed CT data using traditional filtered back projection (FBP), vendor-specific and vendor-neutral IR techniques (ClariCT; ClariPI) in various radiation doses. Noise, low-contrast detectability and subjective spatial resolution were compared between FBP, vendor-specific (i.e., iDose1 to 5; Philips Healthcare), and vendor-neutral (i.e., ClariCT1 to 5) IR techniques in phantom. In 43 patients (median, 14 years; age range 1-19 years), noise, contrast-to-noise ratio (CNR), and qualitative image quality scores of abdominopelvic CT were compared between FBP, iDose level 4 (iDose4), and ClariCT level 2 (ClariCT2), which showed most similar image quality to clinically used vendor-specific IR images (i.e., iDose4) in phantom study. Noise, CNR, and qualitative imaging scores were compared using one-way repeated measure analysis of variance. RESULTS: In phantom study, ClariCT2 showed noise level similar to iDose4 (14.68-7.66 Hounsfield unit [HU] vs. 14.78-6.99 HU at CT dose index volume range of 0.8-3.8 mGy). Subjective low-contrast detectability and spatial resolution were similar between ClariCT2 and iDose4. In clinical study, ClariCT2 was equivalent to iDose4 for noise (14.26-17.33 vs. 16.01-18.90) and CNR (3.55-5.24 vs. 3.20-4.60) (p > 0.05). For qualitative imaging scores, the overall image quality ([reader 1, reader 2]; 2.74 vs. 2.07, 3.02 vs. 2.28) and noise (2.88 vs. 2.23, 2.93 vs. 2.33) of ClariCT2 were superior to those of FBP (p < 0.05), and not different from those of iDose4 (2.74 vs. 2.72, 3.02 vs. 2.98; 2.88 vs. 2.77, 2.93 vs. 2.86) (p > 0.05). CONCLUSION: Vendor-neutral IR technique shows image quality similar to that of clinically used vendor-specific hybrid IR technique for abdominopelvic CT in young patients.
Subject(s)
Abdomen/diagnostic imaging , Tomography, X-Ray Computed , Adolescent , Child , Child, Preschool , Female , Humans , Infant , Male , Phantoms, Imaging , Radiation Dosage , Radiographic Image Interpretation, Computer-Assisted , Signal-To-Noise Ratio , Young AdultABSTRACT
Lung densitometry is being frequently adopted in CT-based emphysema quantification, yet known to be affected by the choice of reconstruction kernel. This study presents a two-step deep learning architecture that enables accurate normalization of reconstruction kernel effects on emphysema quantification in low-dose CT. Deep learning is used to convert a CT image of a sharp kernel to that of a standard kernel with restoration of truncation artifacts and smoothing-free pixel size normalization. We selected 353 scans reconstructed by both standard and sharp kernels from four different CT scanners from the United States National Lung Screening Trial program database. A truncation artifact correction model was constructed with a combination of histogram extrapolation and a deep learning model trained with truncated and non-truncated image sets. Then, we performed frequency domain zero-padding to normalize reconstruction field of view effects while preventing image smoothing effects. The kernel normalization model has a U-Net based architecture trained for each CT scanner dataset. Three lung density measurements including relative lung area under 950 HU (RA950), lower 15th percentile threshold (perc15), and mean lung density were obtained in the datasets from standard, sharp, and normalized kernels. The effect of kernel normalization was evaluated with pair-wise differences in lung density metrics. The mean of pair-wise differences in RA950 between standard and sharp kernel reconstructions was reduced from 10.75% to -0.07% using kernel normalization. The difference for perc15 decreased from -31.03 HU to -0.30 HU after kernel normalization. Our study demonstrated the feasibility of applying deep learning techniques for normalizing CT kernel effects, thereby reducing the kernel-induced variability in lung density measurements. The deep learning model could increase the accuracy of emphysema quantification, thereby allowing reliable surveillance of emphysema in lung cancer screening even when follow-up CT scans are acquired with different reconstruction kernels.