Cooperative AI training for cardiothoracic segmentation in computed tomography: An iterative multi-center annotation approach.

PURPOSE: Radiological reporting is transitioning to quantitative analysis, requiring large-scale multi-center validation of biomarkers. A major prerequisite and bottleneck for this task is the voxelwise annotation of image data, which is time-consuming for large cohorts. In this study, we propose an iterative training workflow to support and facilitate such segmentation tasks, specifically for high-resolution thoracic CT data. METHODS: Our study included 132 thoracic CT scans from clinical practice, annotated by 13 radiologists. In three iterative training experiments, we aimed to improve and accelerate segmentation of the heart and mediastinum. Each experiment started with manual segmentation of 5-25 CT scans, which served as training data for a nnU-Net. Further iterations incorporated AI pre-segmentation and human correction to improve accuracy, accelerate the annotation process, and reduce human involvement over time. RESULTS: Results showed consistent improvement in AI model quality with each iteration. Resampled datasets improved the Dice similarity coefficients for both the heart (DCS 0.91 [0.88; 0.92]) and the mediastinum (DCS 0.95 [0.94; 0.95]). Our AI models reduced human interaction time by 50 % for heart and 70 % for mediastinum segmentation in the most potent iteration. A model trained on only five datasets achieved satisfactory results (DCS > 0.90). CONCLUSIONS: The iterative training workflow provides an efficient method for training AI-based segmentation models in multi-center studies, improving accuracy over time and simultaneously reducing human intervention. Future work will explore the use of fewer initial datasets and additional pre-processing methods to enhance model quality.

Lightweight Techniques to Improve Generalization and Robustness of U-Net Based Networks for Pulmonary Lobe Segmentation.

Lung lobe segmentation in chest CT is relevant to a wide range of clinical applications. However, existing segmentation pipelines often exhibit vulnerabilities and performance degradations when applied to external datasets. This is usually attributed to the size of the available dataset or model. We show that it is possible to enhance generalizability without huge resources by carefully curating the dataset and combining machine learning with medical expertise. Multiple machine learning techniques (self-supervision (SSL), attention (A), and data augmentation (DA)) are used to train a fast and fully-automated lung lobe segmentation model based on 2D U-Net. Our study involved evaluating these techniques on a diverse dataset collected under the RACOON project, encompassing 100 CT chest scans from patients with bacterial, viral, or SARS-CoV2 infections. We compare our model to a baseline U-Net trained on the same dataset. Our approach significantly improved segmentation accuracy (Dice score of 92.8% vs. 82.3%, p < 0.001). Moreover, our model achieved state-of-the-art performance (Dice score of 92.8% vs. 90.8% for the literature's state-of-the-art, p = 0.102) with reduced training examples (69 vs. 231 CT Scans). Among the techniques, data augmentation with expert knowledge displayed the most significant impact, enhancing the Dice score by +0.056. Notably, these enhancements are not limited to lobe segmentation but can be seamlessly integrated into various medical imaging segmentation tasks, demonstrating their versatility and potential for broader applications.

Generative Adversarial Networks: A Primer for Radiologists.

Artificial intelligence techniques involving the use of artificial neural networks-that is, deep learning techniques-are expected to have a major effect on radiology. Some of the most exciting applications of deep learning in radiology make use of generative adversarial networks (GANs). GANs consist of two artificial neural networks that are jointly optimized but with opposing goals. One neural network, the generator, aims to synthesize images that cannot be distinguished from real images. The second neural network, the discriminator, aims to distinguish these synthetic images from real images. These deep learning models allow, among other applications, the synthesis of new images, acceleration of image acquisitions, reduction of imaging artifacts, efficient and accurate conversion between medical images acquired with different modalities, and identification of abnormalities depicted on images. The authors provide an introduction to GANs and adversarial deep learning methods. In addition, the different ways in which GANs can be used for image synthesis and image-to-image translation tasks, as well as the principles underlying conditional GANs and cycle-consistent GANs, are described. Illustrated examples of GAN applications in radiologic image analysis for different imaging modalities and different tasks are provided. The clinical potential of GANs, future clinical GAN applications, and potential pitfalls and caveats that radiologists should be aware of also are discussed in this review. The online slide presentation from the RSNA Annual Meeting is available for this article. ©RSNA, 2021.

[Artificial intelligence in oncological radiology : A (p)review]. / Künstliche Intelligenz in der onkologischen Radiologie : Ein (P)review.

BACKGROUND: Artificial intelligence (AI) has the potential to fundamentally change medicine within the coming decades. Radiological imaging is one of the primary fields of its clinical application. OBJECTIVES: In this article, we summarize previous AI developments with a focus on oncological radiology. Based on selected examples, we derive scenarios for developments in the next 10 years. MATERIALS AND METHODS: This work is based on a review of various literature and product databases, publications by regulatory authorities, reports, and press releases. CONCLUSIONS: The clinical use of AI applications is still in an early stage of development. The large number of research publications shows the potential of the field. Several certified products have already become available to users. However, for a widespread adoption of AI applications in clinical routine, several fundamental prerequisites are still awaited. These include the generation of evidence justifying the use of algorithms through representative clinical studies, adjustments to the framework for approval processes and dedicated education and teaching resources for its users. It is expected that use of AI methods will increase, thus, creating new opportunities for improved diagnostics, therapy, and more efficient workflows.

Value of minimum intensity projections for chest CT in COVID-19 patients.

PURPOSE: To investigate whether minimum intensity projection (MinIP) reconstructions enable more accurate depiction of pulmonary ground-glass opacity (GGO) compared to standard transverse sections and multiplanar reformat (MPR) series in patients with suspected coronavirus disease 2019 (COVID-19). METHOD: In this multinational study, chest CT scans of 185 patients were retrospectively analyzed. Diagnostic accuracy, diagnostic confidence, image quality regarding the assessment of GGO, as well as subjective time-efficiency of MinIP and standard MPR series were analyzed based on the assessment of six radiologists. In addition, the suitability for COVID-19 evaluation, image quality regarding GGO and subjective time-efficiency in clinical routine was assessed by five clinicians. RESULTS: The reference standard revealed a total of 149 CT scans with pulmonary GGO. MinIP reconstructions yielded significantly higher sensitivity (99.9 % vs 95.6 %), specificity (95.8 % vs 86.1 %) and accuracy (99.1 % vs 93.8 %) for assessing of GGO compared with standard MPR series. MinIP reconstructions achieved significantly higher ratings by radiologists concerning diagnostic confidence (medians, 5.00 vs 4.00), image quality (medians, 4.00 vs 4.00), contrast between GGO and unaffected lung parenchyma (medians, 5.00 vs 4.00) as well as subjective time-efficiency (medians, 5.00 vs 4.00) compared with MPR-series (all P < .001). Clinicians preferred MinIP reconstructions for COVID-19 assessment (medians, 5.00 vs 3.00), image quality regarding GGO (medians, 5.00 vs 3.00) and subjective time-efficiency in clinical routine (medians, 5.00 vs 3.00). CONCLUSIONS: MinIP reconstructions improve the assessment of COVID-19 in chest CT compared to standard images and may be suitable for routine application.

DIALAPP: a prospective validation of a new diagnostic algorithm for acute appendicitis.

PURPOSE: The management of patients with suspected appendicitis remains a challenge in daily clinical practice, and the optimal management algorithm is still being debated. Negative appendectomy rates (NAR) continue to range between 10 and 15%. This prospective study evaluated the accuracy of a diagnostic pathway in acute appendicitis using clinical risk stratification (Alvarado score), routine ultrasonography, gynecology consult for females, and selected CT after clinical reassessment. METHODS: Patients presenting with suspected appendicitis between November 2015 and September 2017 from age 18 years and above were included. Decision-making followed a clear management pathway. Patients were followed up for 6 months after discharge. The hypothesis was that the algorithm can reduce the NAR to a value of under 10%. RESULTS: A total of 183 patients were included. In 65 of 69 appendectomies, acute appendicitis was confirmed by histopathology, corresponding to a NAR of 5.8%. Notably, all 4 NAR appendectomies had other pathologies of the appendix. The perforation rate was 24.6%. Only 36 patients (19.7%) received a CT scan. The follow-up rate after 30 days achieved 69%, including no patients with missed appendicitis. The sensitivity and specificity of the diagnostic pathway was 100% and 96.6%, respectively. The potential saving in costs can be as much as 19.8 million €/100,000 cases presenting with the suspicion of appendicitis. CONCLUSION: The risk-stratified diagnostic algorithm yields a high diagnostic accuracy for patients with suspicion of appendicitis. Its implementation can safely reduce the NAR, simultaneously minimizing the use of CT scans and optimizing healthcare-related costs in the treatment of acute appendicitis. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02627781 (December 2015).

Joint Imaging Platform for Federated Clinical Data Analytics.

PURPOSE: Image analysis is one of the most promising applications of artificial intelligence (AI) in health care, potentially improving prediction, diagnosis, and treatment of diseases. Although scientific advances in this area critically depend on the accessibility of large-volume and high-quality data, sharing data between institutions faces various ethical and legal constraints as well as organizational and technical obstacles. METHODS: The Joint Imaging Platform (JIP) of the German Cancer Consortium (DKTK) addresses these issues by providing federated data analysis technology in a secure and compliant way. Using the JIP, medical image data remain in the originator institutions, but analysis and AI algorithms are shared and jointly used. Common standards and interfaces to local systems ensure permanent data sovereignty of participating institutions. RESULTS: The JIP is established in the radiology and nuclear medicine departments of 10 university hospitals in Germany (DKTK partner sites). In multiple complementary use cases, we show that the platform fulfills all relevant requirements to serve as a foundation for multicenter medical imaging trials and research on large cohorts, including the harmonization and integration of data, interactive analysis, automatic analysis, federated machine learning, and extensibility and maintenance processes, which are elementary for the sustainability of such a platform. CONCLUSION: The results demonstrate the feasibility of using the JIP as a federated data analytics platform in heterogeneous clinical information technology and software landscapes, solving an important bottleneck for the application of AI to large-scale clinical imaging data.

Carotid and cerebrovascular dual-energy computed tomography angiography: Optimization of window settings for virtual monoenergetic imaging reconstruction.

PURPOSE: Dedicated post-processing of dual-energy computed tomography angiography (DE-CTA) datasets has been shown to allow for increased vascular contrast. The goal of our study was to define optimal window settings for displaying virtual monoenergetic images (VMI) reconstructed from dual-energy carotid and cerebrovascular DE-CTA. METHODS: Fifty-seven patients who underwent clinically-indicated carotid and cerebrovascular third-generation dual-source DE-CTA were retrospectively evaluated. Standard linearly-blended (M_0.6), 70-keV traditional VMI (M70), and 40-keV noise-optimized VMI (M40+) reconstructions were analyzed. For M70 and M40+ datasets, the subjectively best window setting (width and level, B-W/L) was independently determined by two observers and subsequently related with aortic arch attenuation to calculate optimized values (O-W/L) using linear regression. Subjective evaluation of image quality (IQ) between W/L settings were assessed by two additional readers. Repeated measures analysis of variance were performed to compare W/L settings and IQ indices between M_0.6, M70, and M40 + . RESULTS: B-W/L and O-W/L for M70 were 580/210 and 560/200, and for M40+ were 1630/570 and 1560/550, respectively, higher than standard DE-CTA W/L settings (450/100). Highest subjective scores were observed for M40+ regarding overall IQ (all p < 0.001). CONCLUSION: Application of O-W/L settings is mandatory to optimize subjective IQ of VMI reconstructions of DE-CTA.

Diagnostic accuracy of color-coded virtual noncalcium dual-energy CT for the assessment of bone marrow edema in sacral insufficiency fracture in comparison to MRI.

PURPOSE: To evaluate the diagnostic accuracy of color-coded dual-energy CT virtual noncalcium (VNCa) reconstructions for the assessment of traumatic bone marrow edema in sacral insufficiency fracture (SIF). METHOD: Data from 52 consecutive patients (28 women, 24 men; mean age, 61 ±â€¯13 years; range, 49-94 years) who had undergone third-generation dual-source CT and 3-Tesla (T) MRI due to low back pain without adequate trauma were retrospectively evaluated. Five radiologists, blinded to MRI and clinical information, independently analyzed conventional grayscale dual-energy CT series for sacral fractures according to the Denis classification. Eight weeks later, readers re-assessed all scans using color-coded VNCa reconstructions for sacral bone marrow edema. CT numbers on VNCa reconstructions were measured by a sixth radiologist. One experienced radiologist (33 years of experience in musculoskeletal [MSK] imaging), blinded to CT and clinical information, defined the reference standard by analyzing the MRI scans. The primary indices for diagnostic accuracy were sensitivity, specificity, and the area under the curve (AUC). RESULTS: MRI revealed a total of 39 zones with SIF-associated bone marrow edema in 27 patients. In the qualitative analysis, VNCa showed high overall sensitivity (93 %) and specificity (95 %) for assessing SIF-associated bone marrow edema. The quantitative analysis of color-coded VNCa reconstructions revealed an overall AUC of 0.976. A cut-off value of -43 Hounsfield units provided a sensitivity of 85 % and a specificity of 95 % for differentiating bone marrow edema. CONCLUSIONS: Color-coded dual-energy CT VNCa reconstructions yield excellent diagnostic accuracy in the analysis of SIF-associated bone marrow edema compared to MRI.

Evaluation of Radiation Dose and Image Quality using High-Pitch 70-kV Chest CT in Immunosuppressed Patients. / Evaluation der Strahlendosis und Bildqualität von High-Pitch-70-kV-CT-Untersuchungen des Thorax bei immunsupprimierten Patienten.

PURPOSE: The aim of the study was to evaluate high-pitch 70-kV CT examinations of the thorax in immunosuppressed patients regarding radiation dose and image quality in comparison with 120-kV acquisition. MATERIALS AND METHODS: The image data from 40 patients (14 women and 26 men; mean age: 40.9 ±â€Š15.4 years) who received high-pitch 70-kV CT chest examinations were retrospectively included in this study. A control group (n = 40), matched by age, gender, BMI, and clinical inclusion criteria, had undergone standard 120-kV chest CT imaging. All CT scans were performed on a third-generation dual-source CT unit. For an evaluation of the radiation dose, the CT dose index (CTDIvol), dose-length product (DLP), effective dose (ED), and size-specific dose estimates (SSDE) were analyzed in each group. The objective image quality was evaluated using signal-to-noise (SNR) and contrast-to-noise ratios (CNR). Three blinded and independent radiologists evaluated subjective image quality and diagnostic confidence using 5-point Likert scales. RESULTS: The mean dose parameters were significantly lower for high-pitch 70-kV CT examinations (CTDIvol, 2.9 ±â€Š0.9 mGy; DLP, 99.9 ±â€Š31.0 mGyxcm; ED, 1.5 ±â€Š0.6 mSv; SSDE, 3.8 ±â€Š1.2 mGy) compared to standard 120-kV CT imaging (CTDIvol, 8.8 ±â€Š3.7mGy; DLP, 296.6 ±â€Š119.3 mGyxcm; ED, 4.4 ±â€Š2.1 mSv; SSDE, 11.6 ±â€Š4.4 mGy) (P≤ 0.001). The objective image parameters (SNR: 7.8 ±â€Š2.1 vs. 8.4 ±â€Š1.8; CNR: 7.7 ±â€Š2.4 vs. 8.3 ±â€Š2.8) (P≥ 0.065) and the cumulative subjective image quality (4.5 ±â€Š0.4 vs. 4.7 ±â€Š0.3) (p = 0.052) showed no significant differences between the two protocols. CONCLUSION: High-pitch 70-kV thoracic CT examinations in immunosuppressed patients resulted in a significantly reduced radiation exposure compared to standard 120-kV CT acquisition without a decrease in image quality. KEY POINTS: · Third-generation dual-source CT units enable high-pitch 70-kV CT examinations of the chest.. · High-pitch 70-kV CT examinations show a significantly reduced radiation dose compared to standard 120-kV CT examinations.. · High-pitch 70-kV CT examinations of the chest show comparable objective and subjective image quality.. · Subjectively deteriorated image noise and sharpness of 70-kV CT did not impact diagnostic confidence.. CITATION FORMAT: · Yel I, Martin SS, Wichmann JL et al. Evaluation of Radiation Dose and Image Quality using High-Pitch 70-kV Chest CT in Immunosuppressed Patients . Fortschr Röntgenstr 2019; 191: 122 - 129.

Optimisation of window settings for traditional and noise-optimised virtual monoenergetic imaging in dual-energy computed tomography pulmonary angiography.

OBJECTIVES: To define optimal window settings for displaying virtual monoenergetic images (VMI) of dual-energy CT pulmonary angiography (DE-CTPA). METHODS: Forty-five patients who underwent clinically-indicated third-generation dual-source DE-CTPA were retrospectively evaluated. Standard linearly-blended (M_0.6), 70-keV traditional VMI (M70), and 40-keV noise-optimised VMI (M40+) reconstructions were analysed. For M70 and M40+ datasets, the subjectively best window setting (width and level, B-W/L) was independently determined by two observers and subsequently related with pulmonary artery attenuation to calculate separate optimised values (O-W/L) using linear regression. Subjective evaluation of image quality (IQ) between W/L settings were assessed by two additional readers. Repeated measures of variance were performed to compare W/L settings and IQ indices between M_0.6, M70, and M40+. RESULTS: B-W/L and O-W/L for M70 were 460/140 and 450/140, and were 1100/380 and 1070/380 for M40+, respectively, differing from standard DE-CTPA W/L settings (450/100). Highest subjective scores were observed for M40+ regarding vascular contrast, embolism demarcation, and overall IQ (all p<0.001). CONCLUSIONS: Application of O-W/L settings is beneficial to optimise subjective IQ of VMI reconstructions of DE-CTPA. A width slightly less than two times the pulmonary trunk attenuation and a level approximately of overall pulmonary vessel attenuation are recommended. KEY POINTS: • Application of standard window settings for VMI results in inferior image perception. • No significant differences between B-W/L and O-W/L for M70/M40+ were observed. • O-W/L for M70 were 450/140 and were 1070/380 for M40+. • Improved subjective IQ characteristics were observed for VMI displayed with O-W/L.

Dynamic Liver Magnetic Resonance Imaging in Free-Breathing: Feasibility of a Cartesian T1-Weighted Acquisition Technique With Compressed Sensing and Additional Self-Navigation Signal for Hard-Gated and Motion-Resolved Reconstruction.

OBJECTIVES: The aim of this study was to assess the feasibility of a free-breathing dynamic liver imaging technique using a prototype Cartesian T1-weighted volumetric interpolated breathhold examination (VIBE) sequence with compressed sensing and simultaneous acquisition of a navigation signal for hard-gated and motion state-resolved reconstruction. MATERIALS AND METHODS: A total of 43 consecutive oncologic patients (mean age, 66 ± 11 years; 44% female) underwent free-breathing dynamic liver imaging for the evaluation of liver metastases from colorectal cancer using a prototype Cartesian VIBE sequence (field of view, 380 × 345 mm; image matrix, 320 × 218; echo time/repetition time, 1.8/3.76 milliseconds; flip angle, 10 degrees; slice thickness, 3.0 mm; acquisition time, 188 seconds) with continuous data sampling and additionally acquired self-navigation signal. Data were iteratively reconstructed using 2 different approaches: first, a hard-gated reconstruction only using data associated to the dominating motion state (CS VIBE, Compressed Sensing VIBE), and second, a motion-resolved reconstruction with 6 different motion states as additional image dimension (XD VIBE, eXtended dimension VIBE). Continuous acquired data were grouped in 16 subsequent time increments with 11.57 seconds each to resolve arterial and venous contrast phases. For image quality assessment, both CS VIBE and XD VIBE were compared with the patient's last staging dynamic liver magnetic resonance imaging including a breathhold (BH) VIBE as reference standard 4.5 ± 1.2 months before. Representative quality parameters including respiratory artifacts were evaluated for arterial and venous phase images independently, retrospectively and blindly by 3 experienced radiologists, with higher scores indicating better examination quality. To assess diagnostic accuracy, same readers evaluated the presence of metastatic lesions for XD VIBE and CS VIBE compared with reference BH examination in a second session. RESULTS: Compared with CS VIBE, XD VIBE showed significantly higher overall image quality for both arterial phase (4.2 ± 0.6 vs 3.8 ± 0.7, P = 0.008) and venous phase (4.7 ± 0.4 vs 4.3 ± 0.7, P < 0.001) imaging. There was no significant difference between XD VIBE and BH VIBE for overall image quality in the venous phase (4.7 ± 0.4 vs 4.8 ± 0.4, P = 0.834), whereas arterial phase images were scored slightly lower for XD VIBE (4.5 ± 0.6 vs 4.2 ± 0.6, P = 0.024). Both XD VIBE and BH VIBE were characterized by a very low level of respiratory artifacts with no significant difference between BH and motion-resolved free-breathing strategy (P = 0.505 for arterial phase; P = 0.496 for venous phase). Compared with CS VIBE, obvious quality improvement could be achieved for the extended XD VIBE reconstruction with significantly reduced motion artifacts for venous phase images (P = 0.007). Generally, arterial phase images were scored slightly lower compared with venous phase images when using the free-breathing protocol. Overall, 98% of all metastatic lesions were identified on XD VIBE images and 92% of all metastases were found on CS VIBE. CONCLUSIONS: Dynamic liver imaging using the proposed free-breathing Cartesian strategy is feasible in oncologic patients with excellent image quality, high respiratory motion robustness, and accurate lesion detection. Overall, XD VIBE was superior to CS VIBE in our study.

Automated Attenuation Based Tube Potential Selection of the Lower Extremity Runoff: A Comparison to Fixed Kilovolt with Automated Tube Current Modulation.

PURPOSE: The aim of this study was to evaluate the impact of automated attenuation-based tube potential selection (ATPS) on image quality and radiation dose exposure parameters at a computed tomography angiography (CTA) lower-extremity runoff. MATERIALS AND METHODS: Two hundred forty patients (156 men, 84 women) underwent CTA examinations of the lower-extremity runoff on a second-generation dual-source computed tomography system: 120 patients at a fixed tube potential of 120 kV and a tube current of 180 reference mAs, another 120 patients using automated ATPS. Volume computed tomography dose index (CTDIvol), dose-length-product (DLP), body diameters, noise, signal-to-noise ratio, and subjective image quality were compared. RESULTS: In the ATPS group, 80 kV was automatically selected in 102 patients, 100 kV in 15 patients, and 120 kV in 3 patients; 140 kV was not chosen in any of the cases. The median CTDIvol of 4.81 mGy (2.2-10.6 mGy) and DLP of 568 mGy⋅cm (203-1324 mGy⋅cm) in the ATPS group were significantly lower compared with the CTDIvol of 8.1 mGy (4.4-14.4 mGy) and DLP of 1027.5 mGy⋅cm (509-1806 mGy⋅cm) in the fixed 120-kV group (P < 0.01). Image quality was comparable (P > 0.05). CONCLUSION: Automated ATPS allows for significant dose savings in lower-extremity runoff CTA, whereas image quality remains constant at a high level.

Improved visual delineation of the intimal flap in Stanford type A and B dissections at 3rd generation dual-source high-pitch CT angiography.

OBJECTIVE: Evaluation of the intimal flap visibility comparing 2nd and 3rd generation dual-source high-pitch CT. METHODS: Twenty-five consecutive patients with aortic dissection underwent CT angiography on a second and third generation dual-source CT scanner using prospective ECG-gated high-pitch dual-source CT acquisition mode. Contrast material, saline flush and flow rate were kept equal for optimum comparability. The visibility of the intimal flap as well as the delineation of the different vascular structures was evaluated. RESULTS: In 3rd generation dual-source high-pitch CT we could show a significant improvement of intimal flap visibility in aortic dissection. Especially, the far end of the dissection membrane could be better evaluated in 3rd generation high-pitch CT, reaching statistical significance (P < 0.01). CONCLUSION: 3rd Generation high-pitch CT angiography shows a better delineation of the aortic intimal flap in a small patient cohort, especially in the far ends of the dissection membrane. This might be due to higher tube power in this CT generation. However, to generalise these findings larger trials are needed.

Systematic Comparison of Reduced Tube Current Protocols for High-pitch and Standard-pitch Pulmonary CT Angiography in a Large Single-center Population.

RATIONALE AND OBJECTIVES: Benefits of iterative reconstruction (IR) algorithms combined with dose-reduction techniques have been shown at computed tomography pulmonary angiography (CTPA) in several medium to small patient collectives. In this study, we performed a systematic comparison of image quality to combinations of reduced tube current (RC) and IR for both standard-pitch (SP) single-source and high-pitch (HP) dual-source CTPA in a large, single-center population. MATERIALS AND METHODS: Three hundred eighty-two consecutive patients (October 2010 through December 2012) received clinically indicated CTPA with one of four consecutively changed protocols: (1) HPSC: 180 mAs, weighted filtered back projection, pitch = 3; (2) HPRC: 90 mAs, IR, pitch = 3; (3) SPSC: 180 mAs, weighted filtered back projection, pitch = 1.2; and (4) HPRC: 90 mAs, IR, pitch = 1.2. Tube potential was 100 kV. Vascular attenuation and standardized signal-to-noise ratio (sSNR) were measured in the pulmonary trunk (sSNRPT) and on segmental artery level (sSNRS1, sSNRS10). Dose-length-product was recorded per series. Two independent investigators rated image quality. Kolmogorov-Smirnov test, Kruskal-Wallis test, and kappa statistics were used for statistical analysis. Median values are presented per group. RESULTS: Image quality was consistent between all groups (observer 1: P = 0.118; observer 2: P = 0.122). Inter-reader consistency was very good (κ = 0.866, P < 0.001). Dose-length-product was significantly reduced in HP and RC groups (P < 0.001 for each; SPSC: 139.5 mGycm; HPRC: 92 mGycm; SPSC: 211 mGycm; HPRC: 137 mGycm). sSNR was comparable (sSNRPT overall: P = 0.052; sSNRS1 overall: P = 0.161; and sSNRS10 overall: P = 0.259). CONCLUSIONS: Substantial dose reduction can be within a routine clinical setting without quantifiable loss of image quality either by HP pulmonary angiography or by a combination of IR and RC in either HP or SP acquisition.

Advanced image-based virtual monoenergetic dual-energy CT angiography of the abdomen: optimization of kiloelectron volt settings to improve image contrast.

OBJECTIVES: To compare quantitative image quality parameters in abdominal dual-energy computed tomography angiography (DE-CTA) using an advanced image-based (Mono+) reconstruction algorithm for virtual monoenergetic imaging and standard DE-CTA. METHODS: Fifty-five patients (36 men; mean age, 64.2 ± 12.7 years) who underwent abdominal DE-CTA were retrospectively included. Mono + images were reconstructed at 40, 50, 60, 70, 80, 90 and 100 keV levels and as standard linearly blended M_0.6 images (60 % 100 kV, 40 % 140 kV). The contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) of the common hepatic (CHA), splenic (SA), superior mesenteric (SMA) and left renal arteries (LRA) were objectively measured. RESULTS: Mono+ DE-CTA series showed a statistically superior CNR for 40, 50, 60, 70 and 80 keV (P < 0.031) compared to M_0.6 images for all investigated arteries except SMA at 80 keV (P = 0.08). CNR at 40 keV revealed a mean relative increase of 287.7 % compared to linearly blended images among all assessed arteries (P < 0.001). SNR of Mono+ images was consistently significantly higher at 40, 50, 60 and 70 keV compared to M_0.6 for CHA and SA (P < 0.009). CONCLUSIONS: Compared to linearly blended images, Mono+ reconstructions at low keV levels of abdominal DE-CTA datasets significantly improve quantitative image quality. KEY POINTS: • Mono+ combines increased attenuation with reduced image noise compared to standard DE-CTA. • Mono+ shows superior contrast-to-noise ratios at low keV compared to linearly-blended images. • Contrast-to-noise ratio in monoenergetic DE-CTA peaks at 40 keV. • Mono+ reconstructions significantly improve quantitative image quality at low keV levels.

Quantitative evaluation of beam-hardening artefact correction in dual-energy CT myocardial perfusion imaging.

OBJECTIVES: To assess quantitatively the impact of a novel reconstruction algorithm ("kernel") with beam-hardening correction (BHC) on beam-hardening artefacts of the myocardium at dual-energy CT myocardial perfusion imaging (DE-CTMPI). METHODS: Rest-series of DE-CTMPI examinations from 14 patients were retrospectively analyzed. Six image series were reconstructed for each patient: a) 100 kV, b) 140 kV, and c) linearly blended MIX0.5, each with BHC (D33f kernel) and without (D30f kernel). Seven hundred and fifty-six myocardial regions were assessed. Seven equal regions of interest divided the myocardium in the axial section. Three subdivisions were created within these regions in areas prone to BHA. Reports of SPECT studies performed within 30 days of CT examination were used to confirm the presence and location of true perfusion defects. Paired student t-test was used for statistical evaluation. RESULTS: Overall mean myocardial attenuation was lower using BHC (D30f: 87.3 ± 24.1 HU; D33f: 85.5 ± 21.5 HU; p = 0.009). Overall relative difference from average myocardial attenuation (RDMA) was more homogeneous using BHC (D30f: -0.3 ± 11.4 %; D33f: 0.1 ± 10.1 %; p < 0.001). Changes in RDMA were greatest in the posterobasal myocardium (D30f: -16.2 ± 10.0 %; D33f: 3.4 ± 10.7 %; p < 0.001). CONCLUSIONS: A dedicated reconstruction algorithm with BHC can significantly reduce beam-hardening artefacts in DE-CTMPI. KEY POINTS: • Beam-hardening artefacts (BHA) cause interference with attenuation-based CT myocardial perfusion assessment (CTMPI). • BHA occur mostly in the posterobasal left ventricular wall. • Beam-hardening correction homogenized and decreased mean myocardial attenuation. • BHC can help avoid false-positive findings and increase specificity of static CTMPI.

ECG-gated Versus Non-ECG-gated High-pitch Dual-source CT for Whole Body CT Angiography (CTA).

RATIONALE AND OBJECTIVES: To investigate motion artifacts, image quality, and practical differences in electrocardiographic (ECG)-gated versus non-ECG-gated high-pitch dual-source computed tomography angiography (CTA) of the whole aorta. MATERIALS AND METHODS: Two groups, each including 40 patients, underwent either ECG-gated or non-ECG-gated high-pitch dual-source CTA of the whole aorta. The aortic annulus, aortic valve, coronary ostia, and the presence of motion artifacts of the thoracic aorta as well as vascular contrast down to the femoral arteries were independently assessed by two readers. Additional objective parameters including image noise and signal-to-noise ratio were analyzed. RESULTS: Subjective and objective scoring revealed no presence of motional artifacts regardless of whether the ECG-gated or the non-ECG-gated protocol was used (P > 0.1). Image acquisition parameters (examination length, examination duration, radiation dose) were comparable between the two groups without significant differences. The aortic annulus, aortic valve, and coronary ostia were reliably evaluable in all patients. Vascular contrast was rated excellent in both groups. CONCLUSIONS: High-pitch dual-source CTA of the whole aorta is a robust and dose-efficient examination strategy for the evaluation of aortic pathologies whether or not ECG gating is used.

Influence of technical parameters on epicardial fat volume quantification at cardiac CT.

OBJECTIVES: To systematically analyze the influence of technical parameters on quantification of epicardial fat volume (EATV) at cardiac CT. METHODS: 153 routine cardiac CT data sets were analyzed using three-dimensional pericardial border delineation. Three image series were reconstructed per patient: (a) CTAD: coronary CT angiography (CTA), diastolic phase; (b) CTAS: coronary CTA, systolic phase; (c) CaScD: non-contrast CT, diastolic phase. EATV was calculated using three different upper thresholds (-15HU, -30 HU, -45HU). Repeated measures ANOVA, Spearman's rho, and Bland Altman plots were used. RESULTS: Mean EATV differed between all three image series at a -30HU threshold (CTAD 87.2 ± 38.5 ml, CTAS 90.9 ± 37.7 ml, CaScD 130.7 ± 49.5 ml, P<0.001). EATV of diastolic and systolic CTA reconstructions did not differ significantly (P=0.225). Mean EATV for contrast enhanced CTA at a -15HU threshold (CTAD15 102.4 ± 43.6 ml, CTAS15 105.3 ± 42.3 ml) could be approximated most closely by non-contrast CT at -45HU threshold (CaScD45 105.3 ± 40.8 ml). The correlation was excellent: CTAS15-CTAD15, rho=0.943; CTAD15-CaScD45, rho=0.905; CTAS15-CaScD45, rho=0.924; each P<0.001). Bias values from Bland Altman Analysis were: CTAS15-CTAD15, 4.9%; CTAD15-CaScD45, -4.3%; CTAS15-CaScD45, 0.6%. CONCLUSIONS: Measured EATV can differ substantially between contrast enhanced and non-contrast CT studies, which can be reconciled by threshold modification. Heart cycle phase does not significantly influence EATV measurements.

Assessment of an Advanced Monoenergetic Reconstruction Technique in Dual-Energy Computed Tomography of Head and Neck Cancer.

OBJECTIVES: To define optimal keV settings for advanced monoenergetic (Mono+) dual-energy computed tomography (DECT) in patients with head and neck squamous cell carcinoma (SCC). METHODS: DECT data of 44 patients (34 men, mean age 55.5 ± 16.0 years) with histopathologically confirmed SCC were reconstructed as 40, 55, 70 keV Mono + and M_0.3 (30 % 80 kV) linearly blended series. Attenuation of tumour, sternocleidomastoid muscle, internal jugular vein, submandibular gland, and noise were measured. Three radiologists with >3 years of experience subjectively assessed image quality, lesion delineation, image sharpness, and noise. RESULTS: The highest lesion attenuation was shown for 40 keV series (248.1 ± 94.1 HU), followed by 55 keV (150.2 ± 55.5 HU; P = 0.001). Contrast-to-noise ratio (CNR) at 40 keV (19.09 ± 13.84) was significantly superior to all other reconstructions (55 keV, 10.25 ± 9.11; 70 keV, 7.68 ± 6.31; M_0.3, 5.49 ± 3.28; all P < 0.005). Subjective image quality was highest for 55 keV images (4.53; κ = 0.38, P = 0.003), followed by 40 keV (4.14; κ = 0.43, P < 0.001) and 70 keV reconstructions (4.06; κ = 0.32, P = 0.005), all superior (P < 0.004) to linear blending M_0.3 (3.81; κ = 0.280, P = 0.056). CONCLUSIONS: Mono + DECT at low keV levels significantly improves CNR and subjective image quality in patients with head and neck SCC, as tumour CNR peaks at 40 keV, and 55 keV images are preferred by observers. KEY POINTS: • Mono + DECT combines increased contrast with reduced image noise, unlike linearly blended images. • Mono + DECT imaging allows for superior CNR and subjective image quality. • Head and neck tumour contrast-to-noise ratio peaks at 40 keV. • 55 keV images are preferred over all other series by observers.