Levels Of Iterative Reconstruction Research Articles

Reproducible spectral CT thermometry with liver-mimicking phantoms for image-guided thermal ablation

Objectives. Evaluate the reproducibility, temperature tolerance, and radiation dose requirements of spectral CT thermometry in tissue-mimicking phantoms to establish its utility for non-invasive temperature monitoring of thermal ablations. Methods. Three liver mimicking phantoms embedded with temperature sensors were individually scanned with a dual-layer spectral CT at different radiation dose levels during heating (35 °C–80 °C). Physical density maps were reconstructed from spectral results using varying reconstruction parameters. Thermal volumetric expansion was then measured at each temperature sensor every 5 °C in order to establish a correlation between physical density and temperature. Linear regressions were applied based on thermal volumetric expansion for each phantom, and coefficient of variation for fit parameters was calculated to characterize reproducibility of spectral CT thermometry. Additionally, temperature tolerance was determined to evaluate effects of acquisition and reconstruction parameters. The resulting minimum radiation dose to meet the clinical temperature accuracy requirement was determined for each slice thickness with and without additional denoising. Results. Thermal volumetric expansion was robustly replicated in all three phantoms, with a correlation coefficient variation of only 0.43%. Similarly, the coefficient of variation for the slope and intercept were 9.6% and 0.08%, respectively, indicating reproducibility of the spectral CT thermometry. Temperature tolerance ranged from 2 °C to 23 °C, decreasing with increased radiation dose, slice thickness, and iterative reconstruction level. To meet the clinical requirement for temperature tolerance, the minimum required radiation dose ranged from 20, 30, and 57 mGy for slice thickness of 2, 3, and 5 mm, respectively, but was reduced to 2 mGy with additional denoising. Conclusions. Spectral CT thermometry demonstrated reproducibility across three liver-mimicking phantoms and illustrated the clinical requirement for temperature tolerance can be met for different slice thicknesses. The reproducibility and temperature accuracy of spectral CT thermometry enable its clinical application for non-invasive temperature monitoring of thermal ablation.

Image quality assessment of coronary artery segments using ultra-high resolution dual source photon-counting detector computed tomography

PurposeThe study is intended to assess the image quality of ultra-high resolution (UHR) coronary computed tomography angiography (CCTA) performed on dual source photon-counting detector CT (PCD-CT). MethodConsecutive patients, who underwent clinically indicated CCTA on PCD-CT (UHR 120x 0.2 mm collimation), were included. CCTA images were reconstructed at 0.2 mm slice thickness with Bv40, Bv44, Bv48 and Bv56 kernels and quantum iterative reconstruction level 4. Contrast-to-noise (CNR) and signal-to-noise ratios (SNR) were quantified from contrast-enhanced blood and subcutaneous adipose tissue. All reconstructions were scored per coronary segment (18-segment model) for presence, image quality, motion artefacts, stack artefacts, plaque presence and composition, and stenosis degree. Image quality was scored by two independent observers. ResultsSixty patients were included (median age 62 [25th – 75th percentile: 53–67] years, 45% male, median calcium score 62 [0–217]). The mean heart rate during scanning was 71 ± 11 bpm. Median CTDIvol was 19 [16–22]mGy and median DLP 243 [198–327]mGy.cm. The SNR was 9.3 ± 2.3 and the CNR was 11.7 ± 2.6. Of the potential 1080 coronary segments (60 patients x 18 segments), 255/256 (reader1/reader2) segments could not be assessed for being absent or non-evaluable due to size. Both readers scored 85% of the segments as excellent or very good (Intraclass Correlation Coefficient: 0.88 (95% CI: 0.87–0.90). Motion artefacts were present in 45(5%) segments, stack artefacts in 60(7%) segments and metal artefacts in 9(1%) segments. ConclusionUHR dual-source PCD-CT CCTA provides excellent or very good image quality in 85% of coronary segments at relatively high heart rates at moderate radiation dose with only limited stack artefacts.

Intra-individual comparison of coronary artery stenosis measurements between energy-integrating detector CT and photon-counting detector CT

AbstractPurposeTo compare intra-individual percentage diameter stenosis (PDS) measurements of coronary artery stenoses between energy-integrating detector computed tomography (EID-CT) and a clinical photon-counting detector computed tomography (PCD-CT) systems using similar acquisition and reconstruction settings.MethodsPatients (n = 23, mean age of 65 ± 12.1 years, out of these 16 (69.6%) male) were imaged on a conventional EID- and a clinical PCD-CT system with a median of 5.5 (3.0–12.5) days apart. Sequential CCTA scans were acquired and reconstructed using similar settings, including a vascular Bv36 kernel, a tube voltage of 110 kVp for EID-CT vs 120 kVp for PCD-CT, a slice thickness of 0.5 for EID-CT vs 0.6 for PCD-CT, and an iterative reconstruction strength of 3 on EID-CT vs a virtual monoenergetic reconstruction at 55 keV and quantum iterative reconstruction level of 3 on PCD-CT. Radiation dose, contrast volume, and injection parameters were matched as similarly as possible between the systems. PDS measurements were performed according to the coronary artery disease reporting and data system (CAD-RADS) by two trained readers and compared between the different modalities using the Wilcoxon rank sum test, Spearman correlation, and Bland-Altman analysis.ResultsPCD-CT measured significantly lower PDS values than EID-CT [PDSEID-CT: 45.1% (35.1%–64.0%) vs. PDSPCD-CT 44.2% (32.4%–61.0%), P < 0.0001]. This difference led to a mean bias of 1.8 (LoA −3.0/6.5) with an excellent ICC (0.99) value among EID- and PCD-CT. The mean intra-individual deviation between the examinations was 1.8% between the scanners. This led to CAD-RADS re-classification in 3/23 cases (13.0%, new-lower class) for the first reader, and in 4/23 cases (13.0%, new-lower and 4.4%, new-higher class) for the second reader. Inter-reader agreement between the two readers for each stenosis was very strong (ICC = 0.98).ConclusionsCoronary artery stenosis measurements from PCD-CT correlate strongly to EID-CT-based measurements, despite the tendency of the measurement from PCD-CT being lower. This difference led to a change in CAD-RADS classification in 17.4% of patients. The effects on clinical decision-making, downstream testing, and prognosis have to be evaluated in future studies.

Intra-individual comparison of coronary CT angiography-based FFR between energy-integrating and photon-counting detector CT systems

BackgroundCoronary computed tomography angiography (CCTA)-based fractional flow reserve (CT-FFR) allows for noninvasive determination of the functional severity of anatomic lesions in patients with coronary artery disease. The aim of this study was to intra-individually compare CT-FFR between photon-counting detector (PCD) and conventional energy-integrating detector (EID) CT systems. MethodsIn this single-center prospective study, subjects who underwent clinically indicated CCTA on an EID-CT system were recruited for a research CCTA on PCD-CT within 30 days. Image reconstruction settings were matched as closely as possible between EID-CT (Bv36 kernel, iterative reconstruction strength level 3, slice thickness 0.5 mm) and PCD-CT (Bv36 kernel, quantum iterative reconstruction level 3, virtual monoenergetic level 55 keV, slice thickness 0.6 mm). CT-FFR was measured semi-automatically using a prototype on-site machine learning algorithm by two readers. CT-FFR analysis was performed per-patient and per-vessel, and a CT-FFR ≤ 0.75 was considered hemodynamically significant. ResultsA total of 22 patients (63.3 ± 9.2 years; 7 women) were included. Median time between EID-CT and PCD-CT was 5.5 days. Comparison of CT-FFR values showed no significant difference and strong agreement between EID-CT and PCD-CT in the per-vessel analysis (0.88 [0.74–0.94] vs. 0.87 [0.76–0.93], P = 0.096, mean bias 0.02, limits of agreement [LoA] −0.14/0.19, r = 0.83, ICC = 0.92), and in the per-patient analysis (0.81 [0.60–0.86] vs. 0.76 [0.64–0.86], P = 0.768, mean bias 0.02, LoA −0.15/0.19, r = 0.90, ICC = 0.93). All included patients were classified into the same category (CT-FFR > 0.75 vs ≤0.75) with both CT systems. ConclusionsCT-FFR evaluation is feasible with PCD-CT and it shows a strong agreement with EID-CT-based evaluation when images are similarly reconstructed.

Optimization of chest CT protocols based on pixel image matrix, kernels and iterative reconstruction levels – A phantom study

IntroductionThis study investigated the impact of high matrix image reconstruction in combination with different reconstruction kernels and levels of iterative reconstructions on image quality in chest CT. MethodsAn anthropomorphic chest phantom (Kyoto Kagaku Co., Ltd., Kyoto, Japan), and a Catphan® 600 (The Phantom Laboratory, Greenwich, NY, USA) phantom were scanned using a dual source scanner. Standard institutional protocol with 512 × 512 matrix was used as a reference. Reconstructions were performed for 768 × 768 and 1024 × 1024 matrices and all possible combinations of three different kernels and five levels of iterative reconstructions were included. Signal difference to noise ratio (SdNR) and line pairs per cm (lp/cm) were manually measured. A Linear regression model was applied for objective image analysis (SdNR) and inter–and intra-reader agreement was given as Cohen's kappa for the visual image assessment. ResultsMatrix size did not have a significant impact on SdNR (p = 0.595). Kernel (p = 0.014) and ADMIRE level (p = 0.001) had a statistically significant impact on SdNR. The spatial resolution ranged from 7 lp/cm to 9 lp/cm. The highest spatial resolution was achieved using kernel Br64 and ADMIRE 1, 2 and 3 in both 768- and 1024-matrices, and with Br59 with ADMIRE 2 and 4 and 768-matrix, all visualizing 9 lp/cm. Both readers scored kernel Br59 highest, and the scoring increased with increasing levels of Iterative Reconstruction. ConclusionMatrix size did not influence image quality, however, the choice of kernel and degree of IR had an impact on objective and visual image quality in 768 - and 1024-matrices, suggesting that increased degree of IR may improve diagnostic image quality in chest CT. Implications for practiceImage quality in CT of the lung may be improved by increasing the level of IR.

Image Quality Analysis of Photon-Counting CT Compared with Dual-Source CT: A Phantom Study for Chest CT Examinations.

A comparison was made between the image quality of a photon-counting CT (PCCT) and a dual-source CT (DSCT). The evaluation of image quality was performed using a Catphan CT phantom, and the physical metrics, such as the noise power spectrum and task transfer function, were measured for both PCCT and DSCT at three CT dose indices (1, 5 and 10 mGy). Polyenergetic and virtual monoenergetic reconstructions were used to evaluate the performance differences by simulating a Gaussian spot with a radius of 5 mm and calculating the detectability index. The highest iterative reconstruction level was able to decrease the noise by about 70% compared with the filtered back projection using a parenchyma reconstruction kernel. The PCCT task transfer functions remained constant, while those of the DSCT increased with the reconstruction strength level. At monoenergetic 70 keV, a 50% decrease in noise was observed for DSCT with image smoothing, while PCCT had the same 50% decrease in noise without any smoothing. The PCCT detectability index at a reconstruction strength level of two was equivalent to the highest level of ADMIRE 5 for DSCT. The PCCT showed its superiority over the DSCT, especially for lung nodule detection.

Observer Sensitivity for Detection of Pulmonary Nodules in Ultra-Low Dose Computed Tomography Protocols Using a Third-Generation Dual-Source CT with Ultra-High Pitch—A Phantom Study

This study evaluates ultra-low-dose computed tomography (ULDCT) protocols concerning the detectability of pulmonary nodules. The influence of tube current settings, kernels, strength levels of third-generation iterative reconstruction algorithms, and pitch was investigated. A chest phantom with artificial spherical nodules of different densities and diameters was examined with a third-generation dual-source CT. Scanning and post-processing protocols, tube current levels, and ultra-high and non-high pitch modes were applied. Images were reconstructed with filtered back-projection (FBP) or advanced model-based iterative reconstruction (ADMIRE) algorithms. Sharp (Bl57) or medium-soft (Br36) convolution kernels were applied. The reading was performed by an experienced and an inexperienced reader. The highest observer sensitivity was found using a non-high pitch protocol at tube currents of 120 mAs and 90 mAs with the sharp kernel and iterative reconstruction level of 5. Non-high pitch protocols showed better detectability of solid nodules. Combinations with the medium-soft kernel achieved slightly higher observer sensitivity than with the sharp kernel. False positives (FP) occurred more often for subsolid nodules, at a tube current level of 120 mAs, and with the sharp kernel. A tube current level of 90 mAs combined with the highest iterative reconstruction level achieved the highest accuracy in lung nodule detection regardless of size, density, and reader experience.

Potential of employing a quantum iterative reconstruction algorithm for ultra-high-resolution photon-counting detector CT of the hip

IntroductionThis study investigated the image quality of a new quantum iterative reconstruction algorithm (QIR) for high resolution photon-counting CT of the hip. MethodsUsing a first-generation photon-counting CT scanner, five cadaveric specimens were examined with ultra-high-resolution protocols matched for radiation dose. Images were post-processed with a sharp convolution kernel and five different strength levels of iterative reconstruction (QIR 0 – QIR 4). Subjective image quality was rated independently by three radiologists on a five-point scale. Intraclass correlation coefficients (ICC) were computed for assessing interrater agreement. Objective image quality was evaluated by means of contrast-to-noise-ratios (CNR) in bone and muscle tissue. ResultsFor osseous tissue, subjective image quality was rated best for QIR 2 reformatting (median 5 [interquartile range 5–5]). Contrarily, for soft tissue, QIR 4 received the highest ratings among compared strength levels (3 [3–4]). Both ICCbone (0.805; 95% confidence interval 0.711–0.877; p < 0.001) and ICCmuscle (0.885; 0.824–0.929; p < 0.001) suggested good interrater agreement. CNR in bone and muscle tissue increased with ascending strength levels of iterative reconstruction with the highest results recorded for QIR 4 (CNRbone 29.43 ± 2.61; CNRmuscle 8.09 ± 0.77) and lowest results without QIR (CNRbone 3.90 ± 0.29; CNRmuscle 1.07 ± 0.07) (all p < 0.001). ConclusionReconstructing photon-counting CT data with an intermediate QIR strength level appears optimal for assessment of osseous tissue, whereas soft tissue analysis benefitted from applying the highest strength level available. Implications for practiceQuantum iterative reconstruction technique can enhance image quality by significantly reducing noise and improving CNR in ultra-high resolution CT imaging of the hip.

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

BackgroundRecently, 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.PurposeTo 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.MethodsAcquisitions on the CT American College of Radiology (ACR) 464 and CT Torso CTU‐41 phantoms were performed at five dose levels (CTDIvol: 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.ResultsFrom 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.ConclusionWith 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.

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.

Classification of moving coronary calcified plaques based on motion artifacts using convolutional neural networks: a robotic simulating study on influential factors

BackgroundMotion artifacts affect the images of coronary calcified plaques. This study utilized convolutional neural networks (CNNs) to classify the motion-contaminated images of moving coronary calcified plaques and to determine the influential factors for the classification performance.MethodsTwo artificial coronary arteries containing four artificial plaques of different densities were placed on a robotic arm in an anthropomorphic thorax phantom. Each artery moved linearly at velocities ranging from 0 to 60 mm/s. CT examinations were performed with four state-of-the-art CT systems. All images were reconstructed with filtered back projection and at least three levels of iterative reconstruction. Each examination was performed at 100%, 80% and 40% radiation dose. Three deep CNN architectures were used for training the classification models. A five-fold cross-validation procedure was applied to validate the models.ResultsThe accuracy of the CNN classification was 90.2 ± 3.1%, 90.6 ± 3.5%, and 90.1 ± 3.2% for the artificial plaques using Inception v3, ResNet101 and DenseNet201 CNN architectures, respectively. In the multivariate analysis, higher density and increasing velocity were significantly associated with higher classification accuracy (all P < 0.001). The classification accuracy in all three CNN architectures was not affected by CT system, radiation dose or image reconstruction method (all P > 0.05).ConclusionsThe CNN achieved a high accuracy of 90% when classifying the motion-contaminated images into the actual category, regardless of different vendors, velocities, radiation doses, and reconstruction algorithms, which indicates the potential value of using a CNN to correct calcium scores.

Design of a 3D printed coronary artery model for CT optimization

IntroductionTo design a custom phantom of the coronary arteries to optimize CT coronary angiography (CCTA) protocols. MethodsCharacteristics of the left and right coronary arteries (mean Hounsfield Unit (HU) values and diameters) were collected from consecutive CCTA examinations (n = 43). Four different materials (two mixtures of glycerine, gelatine and water, pig hearts, Ecoflex™ silicone) were scanned inside a Lungman phantom using the CCTA protocol to find the closest model to in vivo data. A 3D printed model of the coronary artery tree was created using CCTA data by exporting a CT volume rendering into Autodesk Meshmixer™ software. The model was placed in an acid bath for 5 h, then covered in Ecoflex™, which was removed after drying. Both the Ecoflex™ and pig heart were later filled with a mixture of contrast (Visipaque 320 mg I/ml), NaCl and gelatin and scanned with different levels of tube current and iterative reconstruction (ASiR-V). Objective (HU, noise and size (vessel diameter) and subjective analysis were performed on all scans. ResultsThe gelatine mixtures had HU values of 130 and 129, Ecoflex™ 65 and the pig heart 56. At the different mA/ASiR-V levels the contrast filled Ecoflex™ had a mean HU 318 ± 4, noise 47±7HU and diameter of 4.4 mm. The pig heart had a mean HU of 209 ± 5, noise 38±4HU and a diameter of 4.4 mm. With increasing iterative reconstruction level the visualisation of the pig heart arteries decreased so no measurements could be performed. ConclusionThe use of a 3D printed model of the arteries and casting with the Ecoflex™ silicone is the most suitable solution for a custom-designed phantom. Implications for practiceCustom designed phantoms using 3D printing technology enable cost effective optimisation of CT protocols.

Coronary Artery Calcium Scoring: Toward a New Standard.

Although the Agatston score is a commonly used quantification method, rescan reproducibility is suboptimal, and different CT scanners result in different scores. In 2007, McCollough et al (Radiology 2007;243:527-538) proposed a standard for coronary artery calcium quantification. Advancements in CT technology over the last decade, however, allow for improved acquisition and reconstruction methods. This study aims to investigate the feasibility of a reproducible reduced dose alternative of the standardized approach for coronary artery calcium quantification on state-of-the-art CT systems from 4 major vendors. An anthropomorphic phantom containing 9 calcifications and 2 extension rings were used. Images were acquired with 4 state-of-the-art CT systems using routine protocols and a variety of tube voltages (80-120 kV), tube currents (100% to 25% dose levels), slice thicknesses (3/2.5 and 1/1.25 mm), and reconstruction techniques (filtered back projection and iterative reconstruction). Every protocol was scanned 5 times after repositioning the phantom to assess reproducibility. Calcifications were quantified as Agatston scores. Reducing tube voltage to 100 kV, dose to 75%, and slice thickness to 1 or 1.25 mm combined with higher iterative reconstruction levels resulted in an on average 36% lower intrascanner variability (interquartile range) compared with the standard 120 kV protocol. Interscanner variability per phantom size decreased by 34% on average. With the standard protocol, on average, 6.2 ± 0.4 calcifications were detected, whereas 7.0 ± 0.4 were detected with the proposed protocol. Pairwise comparisons of Agatston scores between scanners within the same phantom size demonstrated 3 significantly different comparisons at the standard protocol (P < 0.05), whereas no significantly different comparisons arose at the proposed protocol (P > 0.05). On state-of-the-art CT systems of 4 different vendors, a 25% reduced dose, thin-slice calcium scoring protocol led to improved intrascanner and interscanner reproducibility and increased detectability of small and low-density calcifications in this phantom. The protocol should be extensively validated before clinical use, but it could potentially improve clinical interscanner/interinstitutional reproducibility and enable more consistent risk assessment and treatment strategies.

Comparison of image quality between spectral photon-counting CT and dual-layer CT for the evaluation of lung nodules: a phantom study.

To evaluate the image quality (IQ) of a spectral photon-counting CT (SPCCT) using filtered back projection (FBP) and hybrid iterative reconstruction (IR) algorithms (iDose4), in comparison with a dual-layer CT (DLCT) system, and to choose the best image quality according to the IR level for SPCCT. Two phantoms were scanned using a standard lung protocol (120 kVp, 40 mAs) with SPCCT and DLCT systems. Raw data were reconstructed using FBP and 9 iDose4 levels (i1/i2/i3/i4/i5/i6/i7/i9/i11) for SPCCT and 7 for DLCT (i1/i2/i3/i4/i5/i6/i7). Noise power spectrum and task-based transfer function (TTF) were computed. Detectability index (d') was computed for detection of 4 mm ground-glass nodule (GGN) and solid nodule. Two chest radiologists performed an IQ evaluation (noise/nodule sharpness/nodule conspicuity/overall IQ) in consensus, and chose the best image for SPCCT. Noise magnitude was -47% ± 2% lower on average with SPCCT than with DLCT for iDose4 range from i1 to i6. Average NPS spatial frequencies increased for SPCCT in comparison with DLCT. TTF also increased, except for the air insert with FBP, and i1/i2/i3. Higher detectability was found for SPCCT for both GGN and solid nodules. IQ for both types of nodule was rated consistently higher with SPCCT than with DLCT for the same iDose4 level. For SPCCT and both nodules, the scores for noise and conspicuity improved with increasing iDose4 level. iDose4 level 6 provided the best subjective IQ for both types of nodule. Higher IQ for GGN and solid nodules was demonstrated with SPCCT compared with DLCT with better detectability using iDose4. Using spectral photon-counting CT compared with dual-layer CT, noise magnitude was reduced with improvements in spatial resolution and detectability of ground-glass nodules and solid lung nodules. As the iDose4 level increased, noise magnitude was reduced and detectability of ground-glass and solid lung nodules was better for both CT systems. For spectral photon-counting CT imaging, two chest radiologists determined iDose4 level 6 as the best image quality for detecting ground-glass nodules and solid lung nodules.

Intraindividual variation of dose parameters in oncologic CT imaging.

The objective of this study is to identify essential aspects influencing radiation dose in computed tomography [CT] of the chest, abdomen and pelvis by intraindividual comparison of imaging parameters and patient related factors. All patients receiving at least two consecutive CT examinations for tumor staging or follow-up within a period of 22 months were included in this retrospective study. Different CT dose estimates (computed tomography dose index [CTDIvol], dose length product [DLP], size-specific dose estimate [SSDE]) were correlated with patient's body mass index [BMI], scan length and technical parameters (tube current, tube voltage, pitch, noise level, level of iterative reconstruction). Repeated-measures-analysis was initiated with focus on response variables (CTDIvol, DLP, SSDE) and possible factors (age, BMI, noise, scan length, peak kilovoltage [kVp], tube current, pitch, adaptive statistical iterative reconstruction [ASIR]). A univariate-linear-mixed-model with repeated-measures-analysis followed by Bonferroni adjustments was used to find associations between CT imaging parameters, BMI and dose estimates followed by a subsequent multivariate-mixed-model with repeated-measures-analysis with Bonferroni adjustments for significant parameters. A p-value <0.05 was considered statistically significant. We found all dose estimates in all imaging regions were substantially affected by tube current. The iterative reconstruction significantly influenced all dose estimates in the thoracoabdominopelvic scans as well as DLP and SSDE in chest-CT. Pitch factor affected all dose parameters in the thoracoabdominopelvic CT group. These results provide further evidence that tube current has a pivotal role and potential in radiation dose management. The use of iterative reconstruction algorithms can substantially decrease radiation dose especially in thoracoabdominopelvic and chest-CT-scans. Pitch factor should be kept at a level of ≥1.0 in order to reduce radiation dose.

Impact of increasing levels of adaptive statistical iterative reconstruction on image quality in oil-based postmortem CT angiography in coronary arteries

IntroductionPostmortem multi-detector computed tomography (PMCT) has become an important part in forensic imaging. Modern reconstruction techniques such as iterative reconstruction (IR) are frequently used in postmortem CT angiography (PMCTA). The image quality of PMCTA depends on the strength of IR. For this purpose, we aimed to investigate the impact of different advanced IR levels on the objective and subjective PMCTA image quality.Material and methodsWe retrospectively analyzed the coronary arteries of 27 human cadavers undergoing whole-body postmortem CT angiography between July 2017 and March 2018 in a single center. Iterative reconstructions of the coronary arteries were processed in five different level settings (0%; 30%; 50%; 70%; 100%) by using an adaptive statistical IR method. We evaluated the objective (contrast-to-noise ratio (CNR)) and subjective image quality in several anatomical locations.ResultsOur results demonstrate that the increasing levels of an IR technique have relevant impact on the image quality in PMCTA scans in forensic postmortem examinations. Higher levels of IR have led to a significant reduction of image noise and therefore to a significant improvement of objective image quality (+ 70%). However, subjective image quality is inferior at higher levels of IR due to plasticized image appearance.ConclusionObjective image quality in PMCTA progressively improves with increasing level of IR with the best CNR at the highest IR level. However, subjective image quality is best at low to medium levels of IR. To obtain a “classic” image appearance with optimal image quality, PMCTAs should be reconstructed at medium levels of IR.

Feasibility of lung imaging with a large field-of-view spectral photon-counting CT system

PurposeThe purpose of this study was to characterize the technical capabilities and feasibility of a large field-of-view clinical spectral photon-counting computed tomography (SPCCT) prototype for high-resolution (HR) lung imaging. Materials and methodsMeasurement of modulation transfer function (MTF) and acquisition of a line pairs phantom were performed. An anthropomorphic lung nodule phantom was scanned with standard (120kVp, 62mAs), low (120kVp, 11mAs), and ultra-low (80kVp, 3mAs) radiation doses. A human volunteer underwent standard (120kVp, 63mAs) and low (120kVp, 11mAs) dose scans after approval by the ethics committee. HR images were reconstructed with 1024 matrix, 300mm field of view and 0.25mm slice thickness using a filtered-back projection (FBP) and two levels of iterative reconstruction (iDose 5 and 9). The conspicuity and sharpness of various lung structures (distal airways, vessels, fissures and proximal bronchial wall), image noise, and overall image quality were independently analyzed by three radiologists and compared to a previous HR lung CT examination of the same volunteer performed with a conventional CT equipped with energy integrating detectors (120kVp, 10mAs, FBP). ResultsTen percent MTF was measured at 22.3lp/cm with a cut-off at 31lp/cm. Up to 28lp/cm were depicted. While mixed and solid nodules were easily depicted on standard and low-dose phantom images, higher iDose levels and slice thicknesses (1mm) were needed to visualize ground-glass components on ultra-low-dose images. Standard dose SPCCT images of in vivo lung structures were of greater conspicuity and sharpness, with greater overall image quality, and similar image noise (despite a flux reduction of 23%) to conventional CT images. Low-dose SPCCT images were of greater or similar conspicuity and sharpness, similar overall image quality, and lower but acceptable image noise (despite a flux reduction of 89%). ConclusionsA large field-of-view SPCCT prototype demonstrates HR technical capabilities and high image quality for high resolution lung CT in human.

Effect of CT reconstruction settings on the performance of a deep learning based lung nodule CAD system

PurposeTo study the effect of different reconstruction parameter settings on the performance of a commercially available deep learning based pulmonary nodule CAD system. Materials and methodsWe performed a retrospective analysis of 24 chest CT scans, reconstructed at 16 different reconstruction settings for two different iterative reconstruction algorithms (SAFIRE and ADMIRE) varying in slice thickness, kernel size and iterative reconstruction level strength using a commercially available deep learning pulmonary nodule CAD system. The DL-CAD software was evaluated at 25 different sensitivity threshold settings and nodules detected by the DL-CAD software were matched against a reference standard based on the consensus reading of three radiologists. ResultsA total of 384 CT reconstructions was analysed from 24 patients, resulting in a total of 5786 found nodules. We matched the detected nodules against the reference standard, defined by a team of thoracic radiologists, and showed a gradual drop in recall, and an improvement in precision when the iterative strength levels were increased for a constant kernel size. The optimal DL-CAD threshold setting for use in our clinical workflow was found to be 0.88 with an F2 of 0.73 ± 0.053. ConclusionsThe DL-CAD system behaves differently on IR data than on FBP data, there is a gradual drop in recall, and growth in precision when the iterative strength levels are increased. As a result, caution should be taken when implementing deep learning software in a hospital with multiple CT scanners and different reconstruction protocols. To the best of our knowledge, this is the first study that demonstrates this result from a DL-CAD system on clinical data.

Effect of an iterative reconstruction quantum noise reduction technique on computed tomography radiomic features.

Purpose: The goal of this study was to quantify the effects of iterative reconstruction on radiomics features of normal anatomic structures on head and neck computed tomography (CT) scans. Methods: Regions of interest (ROI) containing five different tissue types and an ROI containing only air were extracted from CT scans of the head and neck from 108 patients. Each scan was reconstructed using three different reconstruction levels (2, 4, and 6) in addition to bone, thin slice (1-mm slice thickness), and thin-bone reconstructions. From each ROI in all reconstructions, 142 radiomic features were calculated. For each of the six ROIs, features were compared between combinations of iDose levels (2v4, 4v6, and 2v6) with a threshold of after correcting for multiple comparisons ( ). Features from reconstructions were also compared to bone, thin slice, and thin-bone reconstructions. Spearman's rank correlation coefficient, , quantified the relative feature value agreement across reconstructions. Results: When comparing radiomics features across the three reconstruction levels, over half of all features reflected significant differences for all tissue types, while no features demonstrated significant differences when extracted from air ROIs. When assessing feature value agreement, at least 97% of features reflected excellent agreement ( ) when comparing the three iDose levels for all ROIs. When comparing to bone, thin slice, and thin-bone reconstructions, more than half of all features demonstrated significant differences for all ROIs and of features reflected excellent agreement for all ROIs. Conclusion: Many radiomics features are dependent on the iterative reconstruction level, and the magnitude of this dependency is affected by the tissue from which features are extracted. For studies using images reconstructed using varying reconstruction levels, features robust to these should be used.

Highly reduced-dose CT of the lumbar spine in a human cadaver model.

PurposeFeasibility of a highly reduced-dose lumbar spine CT protocol using iterative reconstruction (IR) in a human cadaver model.Materials and methodsThe lumbar spine of 20 human cadavers was repeatedly examined using three different reduced-dose protocols (RDCT) with decreasing reference tube current-exposure time products (RDCT-1: 50 mAs; RDCT-2: 30 mAs; RDCT-3: 10 mAs) at a constant tube voltage of 140 kV. A clinical standard-dose protocol (SDCT) served as the reference (reference tube current–exposure time product: 70 mAs; tube voltage: 140 kV). Images were reconstructed using filtered back projection (FBP) and two increasing levels of IR: IRL4 and IRL6. A five-point scale was used by two observers to assess the diagnostic quality of anatomical structures (cortical and trabecular bone, intervertebral foramina, pedicles and intervertebral joints, spinous and transverse processes). Objective image noise (OIN) was measured. Results were interpreted using a linear mixed-effects regression model.ResultsRDCT-2 with IRL6 (1.2 ± 0.5mSv) was the lowest reduced-dose protocol which provided diagnostically acceptable and equivalent image quality compared to the SDCT (2.3 ± 1.1mSV) with FBP (p > 0.05). All RDCT protocols achieved a significant reduction of the mean (±SD) effective radiation doses (RDCT-1: 1.7±0.9mSv; RDCT-2: 1.2±0.5mSv; RDCT-3: 0.4±0.2mSv; p < 0.05) compared to SDCT. OIN was lower in all RDCT protocols with the application of IRL4 and IRL6, compared to the SDCT with FBP (p < 0.05).ConclusionHighly reduced-dose lumbar spine CT providing diagnostically acceptable image quality is feasible using IR in this cadaver model and may be transferred into a clinical setting.