Evaluation of MR-derived simulated CT-like images and simulated radiographs compared to conventional radiography in patients with shoulder pain: a proof-of-concept study.

BACKGROUND: To evaluate the diagnostic value of MR-derived CT-like images and simulated radiographs compared with conventional radiographs in patients with suspected shoulder pathology. METHODS: 3 T MRI of the shoulder including a 3D T1-weighted gradient echo sequence was performed in 25 patients (mean age 52.4 ± 18 years, 13 women) with suspected shoulder pathology. Subsequently a cone-beam forward projection algorithm was used to obtain intensity-inverted CT-like images and simulated radiographs. Two radiologists evaluated the simulated images separately and independently using the conventional radiographs as the standard of reference, including measurements of the image quality, acromiohumeral distance, critical shoulder angle, degenerative joint changes and the acromial type. Additionally, the CT-like MR images were evaluated for glenoid defects, subcortical cysts and calcifications. Agreement between the MR-derived simulated radiographs and conventional radiographs was calculated using Cohen's Kappa. RESULTS: Measurements on simulated radiographs and conventional radiographs overall showed a substantial to almost perfect inter- and intra-rater agreement (κ = 0.69-1.00 and κ = 0.65-0.85, respectively). Image quality of the simulated radiographs was rated good to excellent (1.6 ± 0.7 and 1.8 ± 0.6, respectively) by the radiologists. A substantial agreement was found regarding diagnostically relevant features, assessed on Y- and anteroposterior projections (κ = 0.84 and κ = 0.69 for the measurement of the CSA; κ = 0.95 and κ = 0.60 for the measurement of the AHD; κ = 0.77 and κ = 0.77 for grading of the Samilson-Prieto classification; κ = 0.83 and κ = 0.67 for the grading of the Bigliani classification, respectively). CONCLUSION: In this proof-of-concept study, clinically relevant features of the shoulder joint were assessed reliably using MR-derived CT-like images and simulated radiographs with an image quality equivalent to conventional radiographs. MR-derived CT-like images and simulated radiographs may provide useful diagnostic information while reducing the amount of radiation exposure.

Towards subject-level cerebral infarction classification of CT scans using convolutional networks.

Automatic evaluation of 3D volumes is a topic of importance in order to speed up clinical decision making. We describe a method to classify computed tomography scans on volume level for the presence of non-acute cerebral infarction. This is not a trivial task, as the lesions are often similar to other areas in the brain regarding shape and intensity. A three stage architecture is used for classification: 1) A cranial cavity segmentation network is developed, trained and applied. 2) Region proposals are generated 3) Connected regions are classified using a multi-resolution, densely connected 3D convolutional network. Mean area under curve values for subject level classification are 0.95 for the unstratified test set, 0.88 for stratification by patient age and 0.93 for stratification by CT scanner model. We use a partly segmented dataset of 555 scans of which 186 scans are used in the unstratified test set. Furthermore we examine possible dataset bias for scanner model and patient age parameters. We show a successful application of the proposed three-stage model for full volume classification. In contrast to black-box approaches, the convolutional network's decision can be further assessed by examination of intermediate segmentation results.

CTPA with a conventional CT at 100 kVp vs. a spectral-detector CT at 120 kVp: Comparison of radiation exposure, diagnostic performance and image quality.

PURPOSE: To compare CT pulmonary angiographies (CTPAs) as well as phantom scans obtained at 100 kVp with a conventional CT (C-CT) to virtual monochromatic images (VMI) obtained with a spectral detector CT (SD-CT) at equivalent dose levels as well as to compare the radiation exposure of both systems. MATERIAL AND METHODS: In total, 2110 patients with suspected pulmonary embolism (PE) were examined with both systems. For each system (C-CT and SD-CT), imaging data of 30 patients with the same mean CT dose index (4.85 mGy) was used for the reader study. C-CT was performed with 100 kVp and SD-CT was performed with 120 kVp; for SD-CT, virtual monochromatic images (VMI) with 40, 60 and 70 keV were calculated. All datasets were evaluated by three blinded radiologists regarding image quality, diagnostic confidence and diagnostic performance (sensitivity, specificity). Contrast-to-noise ratio (CNR) for different iodine concentrations was evaluated in a phantom study. RESULTS: CNR was significantly higher with VMI at 40 keV compared to all other datasets. Subjective image quality as well as sensitivity and specificity showed the highest values with VMI at 60 keV and 70 keV. Hereby, a significant difference to 100 kVp (C-CT) was found for image quality. The highest sensitivity was found using VMI at 60 keV with a sensitivity of more than 97 % for all localizations of PE. For diagnostic confidence and subjective contrast, highest values were found with VMI at 40 keV. CONCLUSION: Higher levels of diagnostic performance and image quality were achieved for CPTAs with SD-CT compared to C-CT given similar dose levels. In the clinical setting SD-CT may be the modality of choice as additional spectral information can be obtained.

Low-dose and sparse sampling MDCT-based femoral bone strength prediction using finite element analysis.

This study aims to evaluate the impact of dose reduction through tube current and sparse sampling on multi-detector computed tomography (MDCT)-based femoral bone strength prediction using finite element (FE) analysis. FE-predicted femoral failure load obtained from MDCT scan data was not significantly affected by 50% dose reductions through sparse sampling. Further decrease in dose through sparse sampling (25% of original projections) and virtually reduced tube current (50% and 25% of the original dose) showed significant effects on the FE-predicted failure load results. PURPOSE: To investigate the effect of virtually reduced tube current and sparse sampling on multi-detector computed tomography (MDCT)-based femoral bone strength prediction using finite element (FE) analysis. METHODS: Routine MDCT data covering the proximal femur of 21 subjects (17 males; 4 females; mean age, 71.0 ± 8.8 years) without any bone diseases aside from osteoporosis were included in this study. Fifty percent and 75% dose reductions were achieved by virtually reducing tube current and by applying a sparse sampling strategy from the raw image data. Images were then reconstructed with a statistically iterative reconstruction algorithm. FE analysis was performed on all reconstructed images and the failure load was calculated. The root mean square coefficient of variation (RMSCV) and coefficient of correlation (R2) were calculated to determine the variation in the FE-predicted failure load data for dose reductions, using original-dose MDCT scan as the standard of reference. RESULTS: Fifty percent dose reduction through sparse sampling showed lower RMSCV and higher correlations when compared with virtually reduced tube current method (RMSCV = 5.70%, R2 = 0.96 vs. RMSCV = 20.78%, R2 = 0.79). Seventy-five percent dose reduction achieved through both methods (RMSCV = 22.38%, R2 = 0.80 for sparse sampling; RMSCV = 24.58%, R2 = 0.73 for reduced tube current) could not predict the failure load accurately. CONCLUSION: Our simulations indicate that up to 50% reduction in radiation dose through sparse sampling can be used for FE-based prediction of femoral failure load. Sparse-sampled MDCT may allow fracture risk prediction and treatment monitoring in osteoporosis with less radiation exposure in the future.

Systematic Evaluation of Low-dose MDCT for Planning Purposes of Lumbosacral Periradicular Infiltrations.

PURPOSE: To evaluate image quality and confidence for planning of periradicular infiltrations using virtually lowered tube currents and in-house developed iterative reconstruction (IR) for multidetector computed tomography (MDCT). METHODS: A total of 20 patients (mean age 54.9 ± 13.1 years) underwent MDCT for planning purposes of periradicular infiltrations at the lumbosacral spine (120 kVp and 100 mAs). Planning scans were simulated as if they were performed at 50% (D50), 10% (D10), 5% (D5), and 1% (D1) of the tube current of original scanning. Image reconstruction was achieved with two levels of IR (A: similar in appearance to clinical reconstructions, B: 10 times stronger noise reduction). Qualitative image evaluation was performed by two readers (R1 and R2) considering overall image quality and artifacts, image contrast, determination of nerve root, and confidence for intervention planning (scoring: 1 high, 2 medium, and 3 low confidence). RESULTS: Level A of IR was favorable regarding overall image quality, artifacts, image contrast, and nerve root depiction according to both readers, with preserved good to excellent scores down to D10 scans. The confidence for intervention planning was not significantly different (p > 0.05) between scans with tube currents virtually lowered down to 10% as compared to the original scans when using level A of IR (R1: 1.2 ± 0.4, R2: 1.1 ± 0.3). Inter-reader agreement for planning confidence was good to excellent (range of weighted Cohen's kappa: 0.62-1.00). CONCLUSION: The use of MDCT for planning purposes of lumbosacral periradicular infiltrations may be possible with tube currents lowered down to 10% of standard dose (equal to 10 mAs) without limitations in planning confidence.

Liver lesion localisation and classification with convolutional neural networks: a comparison between conventional and spectral computed tomography.

PURPOSE: To evaluate the benefit of the additional available information present in spectral CT datasets, as compared to conventional CT datasets, when utilizing convolutional neural networks for fully automatic localisation and classification of liver lesions in CT images. MATERIALS AND METHODS: Conventional and spectral CT images (iodine maps, virtual monochromatic images (VMI)) were obtained from a spectral dual-layer CT system. Patient diagnosis were known from the clinical reports and classified into healthy, cyst and hypodense metastasis. In order to compare the value of spectral versus conventional datasets when being passed as input to machine learning algorithms, we implemented a weakly-supervised convolutional neural network (CNN) that learns liver lesion localisation without pixel-level ground truth annotations. Regions-of-interest are selected automatically based on the localisation results and are used to train a second CNN for liver lesion classification (healthy, cyst, hypodense metastasis). The accuracy of lesion localisation was evaluated using the Euclidian distances between the ground truth centres of mass and the predicted centres of mass. Lesion classification was evaluated by precision, recall, accuracy and F1-Score. RESULTS: Lesion localisation showed the best results for spectral information with distances of 8.22 ± 10.72 mm, 8.78 ± 15.21 mm and 8.29 ± 12.97 mm for iodine maps, 40 keV and 70 keV VMIs, respectively. With conventional data distances of 10.58 ± 17.65 mm were measured. For lesion classification, the 40 keV VMIs achieved the highest overall accuracy of 0.899 compared to 0.854 for conventional data. CONCLUSION: An enhanced localisation and classification is reported for spectral CT data, which demonstrates that combining machine-learning technology with spectral CT information may in the future improve the clinical workflow as well as the diagnostic accuracy.

Bone mineral density measurements derived from dual-layer spectral CT enable opportunistic screening for osteoporosis.

OBJECTIVE: To investigate the in vivo applicability of non-contrast-enhanced hydroxyapatite (HA)-specific bone mineral density (BMD) measurements based on dual-layer CT (DLCT). METHODS: A spine phantom containing three artificial vertebral bodies with known HA densities was measured to obtain spectral data using DLCT and quantitative CT (QCT), simulating different patient positions and grades of obesity. BMD was calculated from virtual monoenergetic images at 50 and 200 keV. HA-specific BMD values of 174 vertebrae in 33 patients (66 ± 18 years; 33% women) were determined in non-contrast routine DLCT and compared with corresponding QCT-based BMD values. RESULTS: Examining the phantom, HA-specific BMD measurements were on a par with QCT measurements. In vivo measurements revealed strong correlations between DLCT and QCT (r = 0.987 [95% confidence interval, 0.963-1.000]; p < 0.001) and substantial agreement in a Bland-Altman plot. CONCLUSION: DLCT-based HA-specific BMD measurements were comparable with QCT measurements in in vivo analyses. This suggests that opportunistic DLCT-based BMD measurements are an alternative to QCT, without requiring phantoms and specific protocols. KEY POINTS: • DLCT-based hydroxyapatite-specific BMD measurements show a substantial agreement with QCT-based BMD measurements in vivo. • DLCT-based hydroxyapatite-specific measurements are on a par with QCT in spine phantom measurements. • Opportunistic DLCT-based BMD measurements may be a feasible alternative for QCT, without requiring dedicated examination protocols or a phantom.

Tube Current Reduction in CT Angiography: How Low Can We Go in Imaging of Patients With Suspected Acute Stroke?

OBJECTIVE. The purpose of this study was to systematically evaluate image quality, detectability of large-vessel occlusion or dissection, and diagnostic confidence in CT angiography (CTA) with virtually lowered tube current and iterative reconstruction in patients with suspected acute stroke. MATERIALS AND METHODS. Thirty patients (15 with large-vessel occlusion or dissection) underwent CTA of the supraaortal up to the intracranial arterial vessels. CTA scans were simulated as if they were made at 50% (D50), 25% (D25), and 10% (D10) of the original tube current. Image reconstruction was achieved with two levels of iterative reconstruction (A, similar to clinical reconstructions; B, two times stronger regularization). Two readers performed qualitative image evaluation considering overall image quality, artifacts, vessel contrast, detection of vessel abnormalities, and diagnostic confidence. RESULTS. Level B of iterative reconstruction was favorable regarding overall image quality and artifacts for D10, whereas level A was favorable for D100 and D50. CTA scans at D25 and both levels of iterative reconstruction still showed good vessel contrast, with even peripheral arterial branches of the anterior, middle, and posterior cerebral arteries being clearly detectable. Furthermore, CTA scans at D25 and level A of iterative reconstruction showed an adequate level of diagnostic confidence without any missed large-vessel occlusion or dissection according to evaluations by both readers. CONCLUSION. CTA with iterative reconstruction and tube currents decreased to 25% of that for original imaging is feasible without limitations in vessel contrast or detection of vessel abnormalities in patients with suspected acute stroke. Thus, the approach evaluated enables substantial reductions in radiation exposure for patients undergoing head and neck CTA.

Sparse sampling computed tomography (SpSCT) for detection of pulmonary embolism: a feasibility study.

OBJECTIVES: Evaluation of sparse sampling computed tomography (SpSCT) regarding subjective and objective image criteria for the detection of pulmonary embolism (PE) at different simulated dose levels. METHODS: Computed tomography pulmonary angiography (CTPA) scans of 20 clinical patients were used to obtain simulated low-dose scans with 100%-50%-25%-12.5%-6.3%-3.1% of the clinical dose, resulting in a total of six dose levels (DL). From these full sampling (FS) data, every second (2-SpSCT) or fourth (4-SpSCT) projection was used to obtain simulated sparse sampling scans. Each image set was evaluated by four blinded radiologists regarding subjective image criteria (artifacts, image quality) and diagnostic performance (confidence, sensitivity, specificity, accuracy, and area under the curve). Additionally, the contrast-to-noise ratio (CNR) was evaluated for objective image quality. RESULTS: Sensitivity was 100% with 2-SpSCT and 4-SpSCT at the 25% DL and the 12.5% DL for all localizations of PE (one subgroup 98.5%). With FS, the sensitivity decreased to 90% at the 12.5% DL. 2-SpSCT and 4-SpSCT showed higher values for sensitivity, specificity, accuracy, and the area under the curve at all DL compared with FS. Subjective image quality was significantly higher for 4-SpSCT compared with FS at each dose level (p < 0.01, paired t test). Only with 4-SpSCT, all examinations were rated as showing diagnostic image quality at the 12.5% DL. CONCLUSIONS: Via SpSCT, a dose reduction down to a 12.5% dose level (corresponding to a mean effective dose of 0.38 mSv in the current study) for CTPA is possible while maintaining high image quality and full diagnostic confidence. KEY POINTS: • With sparse sampling CT, radiation dose could be significantly reduced in clinical routine. • Sparse sampling CT is a novel hardware solution with which less projection images are acquired. • In the current study, a dose reduction of 87.5% (corresponding to a mean effective dose of 0.38 mSv) for CTPA could be achieved while maintaining excellent diagnostic performance.

Perfusion-ventilation CT via three-material differentiation in dual-layer CT: a feasibility study.

Dual-Energy Computed Tomography is of significant clinical interest due to the possibility of material differentiation and quantification. In current clinical routine, primarily two materials are differentiated, e.g., iodine and soft-tissue. A ventilation-perfusion-examination acquired within a single CT scan requires two contrast agents, e.g., xenon and gadolinium, and a three-material differentiation. In the current study, we have developed a solution for three-material differentiation for a ventilation-perfusion-examination. A landrace pig was examined using a dual-layer CT, and three scans were performed: (1) native; (2) xenon ventilation only; (3) xenon ventilation and gadolinium perfusion. An in-house developed algorithm was used to obtain xenon- and gadolinium-density maps. Firstly, lung tissue was segmented from other tissue. Consequently, a two-material decomposition was performed for lung tissue (xenon/soft-tissue) and for remaining tissue (gadolinium/soft-tissue). Results reveal that it was possible to differentiate xenon and gadolinium in a ventilation/perfusion scan of a pig, resulting in xenon and gadolinium density maps. By summation of both density maps, a three-material differentiation (xenon/gadolinium/soft tissue) can be performed and thus, xenon ventilation and gadolinium perfusion can be visualized in a single CT scan. In an additionally performed phantom study, xenon and gadolinium quantification showed very accurate results (r > 0.999 between measured and known concentrations).

Multi-detector CT imaging: impact of virtual tube current reduction and sparse sampling on detection of vertebral fractures.

PURPOSE: To systematically evaluate the effects of virtual tube current reduction and sparse sampling on image quality and vertebral fracture diagnostics in multi-detector computed tomography (MDCT). MATERIALS AND METHODS: In routine MDCT scans of 35 patients (80.0% females, 70.6 ± 14.2 years, 65.7% showing vertebral fractures), reduced radiation doses were retrospectively simulated by virtually lowering tube currents and applying sparse sampling, considering 50%, 25%, and 10% of the original tube current and projections, respectively. Two readers evaluated items of image quality and presence of vertebral fractures. Readout between the evaluations in the original images and those with virtually lowered tube currents or sparse sampling were compared. RESULTS: A significant difference was revealed between the evaluations of image quality between MDCT with virtually lowered tube current and sparse-sampled MDCT (p < 0.001). Sparse-sampled data with only 25% of original projections still showed good to very good overall image quality and contrast of vertebrae as well as minimal artifacts. There were no missed fractures in sparse-sampled MDCT with 50% reduction of projections, and clinically acceptable determination of fracture age was possible in MDCT with 75% reduction of projections, in contrast to MDCT with 50% or 75% virtual tube current reduction, respectively. CONCLUSION: Sparse-sampled MDCT provides adequate image quality and diagnostic accuracy for vertebral fracture detection with 50% of original projections in contrast to corresponding MDCT with lowered tube current. Thus, sparse sampling is a promising technique for dose reductions in MDCT that could be introduced in future generations of scanners. KEY POINTS: • MDCT with a reduction of projection numbers of 50% still showed high diagnostic accuracy without any missed vertebral fractures. • Clinically acceptable determination of vertebral fracture age was possible in MDCT with a reduction of projection numbers of 75%. • With sparse sampling, higher reductions in radiation exposure can be achieved without compromised image or diagnostic quality in routine MDCT of the spine as compared to MDCT with reduced tube currents.

DXA-equivalent quantification of bone mineral density using dual-layer spectral CT scout scans.

OBJECTIVES: To develop and evaluate a method for areal bone mineral density (aBMD) measurement based on dual-layer spectral CT scout scans. METHODS: A post-processing algorithm using a pair of 2D virtual mono-energetic scout images (VMSIs) was established in order to semi-automatically compute the aBMD at the spine similarly to DXA, using manual soft tissue segmentation, semi-automatic segmentation for the vertebrae, and automatic segmentation for the background. The method was assessed based on repetitive measurements of the standardized European Spine Phantom (ESP) using the standard scout scan tube current (30 mA) and other tube currents (10 to 200 mA), as well as using fat-equivalent extension rings simulating different patient habitus, and was compared to dual-energy X-ray absorptiometry (DXA). Moreover, the feasibility of the method was assessed in vivo in female patients. RESULTS: Derived from standard scout scans, aBMD values measured with the proposed method significantly correlated with DXA measurements (r = 0.9925, p < 0.001), and mean accuracy (DXA, 4.12%; scout, 1.60%) and precision (DXA, 2.64%; scout, 2.03%) were comparable between the two methods. Moreover, aBMD values assessed at different tube currents did not differ significantly (p ≥ 0.20 for all), suggesting that the presented method could be applied to scout scans with different settings. Finally, data derived from sample patients were concordant with BMD values from a reference age-matched population. CONCLUSIONS: Based on dual-layer spectral scout scans, aBMD measurements were fast and reliable and significantly correlated with the according DXA measurements in phantoms. Considering the number of CT acquisitions performed worldwide, this method could allow truly opportunistic osteoporosis screening. KEY POINTS: • 2D scout scans (localizer radiographs) from a dual-layer spectral CT scanner, which are mandatory parts of a CT examination, can be used to automatically determine areal bone mineral density (aBMD) at the spine. • The presented method allowed fast (< 25 s/patient), semi-automatic, and reliable DXA-equivalent aBMD measurements for state-of-the-art DXA phantoms at different tube settings and for various patient habitus, as well as for sample patients. • Considering the number of CT scout scan acquisitions performed worldwide on a daily basis, the presented technique could enable truly opportunistic osteoporosis screening with DXA-equivalent metrics, without involving higher radiation exposure since it only processes existing data that is acquired during each CT scan.

Effect of Statistically Iterative Image Reconstruction on Vertebral Bone Strength Prediction Using Bone Mineral Density and Finite Element Modeling: A Preliminary Study.

Statistical iterative reconstruction (SIR) using multidetector computed tomography (MDCT) is a promising alternative to standard filtered back projection (FBP), because of lower noise generation while maintaining image quality. Hence, we investigated the feasibility of SIR in predicting MDCT-based bone mineral density (BMD) and vertebral bone strength from finite element (FE) analysis. The BMD and FE-predicted bone strength derived from MDCT images reconstructed using standard FBP (FFBP) and SIR with (FSIR) and without regularization (FSIRB0) were validated against experimental failure loads (Fexp). Statistical iterative reconstruction produced the best quality images with regard to noise, signal-to-noise ratio, and contrast-to-noise ratio. Fexp significantly correlated with FFBP, FSIR, and FSIRB0. FFBP had a significant correlation with FSIRB0 and FSIR. The BMD derived from FBP, SIRB0, and SIR were significantly correlated. Effects of regularization should be further investigated with FE and BMD analysis to allow for an optimal iterative reconstruction algorithm to be implemented in an in vivo scenario.

MDCT-based Finite Element Analysis of Vertebral Fracture Risk: What Dose is Needed?

PURPOSE: The aim of this study was to compare vertebral failure loads, predicted from finite element (FE) analysis of patients with and without osteoporotic vertebral fractures (OVF) at virtually reduced dose levels, compared to standard-dose exposure from multidetector computed tomography (MDCT) imaging and evaluate whether ultra-low dose derived FE analysis can still differentiate patient groups. MATERIALS AND METHODS: An institutional review board (IRB) approval was obtained for this retrospective study. A total of 16 patients were evaluated at standard-dose MDCT; eight with and eight without OVF. Images were reconstructed at virtually reduced dose levels (i. e. half, quarter and tenth of the standard dose). Failure load was determined at L1-3 from FE analysis and compared between standard, half, quarter, and tenth doses and used to differentiate between fracture and control groups. RESULTS: Failure load derived at standard dose (3254 ± 909 N and 3794 ± 984 N) did not significantly differ from half (3390 ± 890 N and 3860 ± 1063 N) and quarter dose (3375 ± 915 N and 3925 ± 990 N) but was significantly higher for one tenth dose (4513 ± 1762 N and 4766 ± 1628 N) for fracture and control groups, respectively. Failure load differed significantly between the two groups at standard, half and quarter doses, but not at tenth dose. Receiver operating characteristic (ROC) curve analysis also demonstrated that standard, half, and quarter doses can significantly differentiate the fracture from the control group. CONCLUSION: The use of MDCT enables a dose reduction of at least 75% compared to standard-dose for an adequate prediction of vertebral failure load based on non-invasive FE analysis.

Evaluation of MR-derived CT-like images and simulated radiographs compared to conventional radiography in patients with benign and malignant bone tumors.

OBJECTIVES: To evaluate the diagnostic value of MR-derived CT-like images and simulated radiographs compared with conventional radiographs in patients with benign and malignant bone tumors. METHODS: In 32 patients with a benign or malignant bone lesion (mean age 33.9 ± 18.5 years, 17 females), 3-T MR imaging was performed including a 3D T1-weighted gradient echo sequence as the basis for the CT-like images. From these, intensity-inverted MR image volumes were converted into 2D images via a forward projection to obtain simulated radiographs. Two radiologists assessed these images as well as conventional radiographs for the type of periosteal reaction, matrix mineralization and destruction pattern. Agreement between the modalities was calculated using Cohen's κ. RESULTS: The agreement between conventional radiographs and MR-derived CT-like images in combination with simulated radiographs was substantial (periosteal reaction, κ = 0.67; destruction pattern, κ = 0.75), and the sensitivity of both modalities for the final diagnosis of the lesion (aggressive vs. nonaggressive) was high (MR-derived CT-like images, 86.2% vs. conventional radiographs, 90.0%). Additional information on soft tissue extension (MR-derived CT-like images, 21.9% vs. conventional radiographs, 12.5%; p = 0.009) and lobulation (9.4% vs. 0%; p < 0.001) was significantly more often found on MR-derived CT-like images compared with conventional radiographs. CONCLUSIONS: The assessment of the destruction patterns, periosteal reaction and distinction between aggressive and nonaggressive tumors was feasible using MR-derived CT-like images and simulated radiographs and is comparable to that of conventional radiographs. Moreover, MR-derived CT-like images provided additional information on soft tissue extension and tumor architecture. KEY POINTS: • CT-like images and simulated radiographs can be generated from 3D MRI. • Evaluation of bone tumors is feasible with MR-derived images. • CT-like images and simulated radiographs provide additional information on bone tumors.

Evaluation of a preclinical photon-counting CT prototype for pulmonary imaging.

The purpose of this study was to investigate a preclinical spectral photon-counting CT (SPCCT) prototype compared to conventional CT for pulmonary imaging. A custom-made lung phantom, including nodules of different sizes and shapes, was scanned with a preclinical SPCCT and a conventional CT in standard and high-resolution (HR-CT) mode. Volume estimation was evaluated by linear regression. Shape similarity was evaluated with the Dice similarity coefficient. Spatial resolution was investigated via MTF for each imaging system. In-vivo rabbit lung images from the SPCCT system were subjectively reviewed. Evaluating the volume estimation, linear regression showed best results for the SPCCT compared to CT and HR-CT with a root mean squared error of 21.3 mm3, 28.5 mm3 and 26.4 mm3 for SPCCT, CT and HR-CT, respectively. The Dice similarity coefficient was superior for SPCCT throughout nodule shapes and all nodule sizes (mean, SPCCT: 0.90; CT: 0.85; HR-CT: 0.85). 10% MTF improved from 10.1 LP/cm for HR-CT to 21.7 LP/cm for SPCCT. Visual investigation of small pulmonary structures was superior for SPCCT in the animal study. In conclusion, the SPCCT prototype has the potential to improve the assessment of lung structures due to higher resolution compared to conventional CT.

Dual-layer spectral computed tomography: measuring relative electron density.

BACKGROUND: X-ray and particle radiation therapy planning requires accurate estimation of local electron density within the patient body to calculate dose delivery to tumour regions. We evaluate the feasibility and accuracy of electron density measurement using dual-layer computed tomography (DLCT), a recently introduced dual-energy CT technique. METHODS: Two calibration phantoms were scanned with DLCT and virtual monoenergetic images (VMIs) at 50 keV and 200 keV were generated. We investigated two approaches to obtain relative electron densities from these VMIs: to fit an analytic interaction cross-sectional model and to empirically calibrate a conversion function with one of the phantoms. Knowledge of the emitted x-ray spectrum was not required for the presented work. RESULTS: The results from both methods were highly correlated to the nominal values (R > 0.999). Except for the water and lung inserts, the error was within 1.79% (average 1.53%) for the cross-sectional model and 1.61% (average 0.87%) for the calibrated conversion. Different radiation doses did not have a significant influence on the measurement (p = 0.348, 0.167), suggesting that the methods are reproducible. Further, we applied these methods to routine clinical data. CONCLUSIONS: Our study shows a high validity of electron density estimation based on DLCT, which has potential to improve the procedure and accuracy of measuring electron density in clinical practice.

CNN as model observer in a liver lesion detection task for x-ray computed tomography: A phantom study.

PURPOSE: The purpose of this study was the evaluation of anthropomorphic model observers trained with neural networks for the prediction of a human observer's performance. METHODS: To simulate liver lesions, a phantom with contrast targets (acrylic spheres, varying diameters, +30 HU) was repeatedly scanned on a computed tomography scanner. Image data labeled with confidence ratings assessed in a reader study for a detection task of liver lesions were used to build several anthropomorphic model observers. Models were trained with images reconstructed with iterative reconstruction and evaluated with images reconstructed with filtered backprojection. A neural network, based on softmax regression (SR-MO), and convolutional neural networks (CNN-MO) were used to predict the performance of a human observer and compared to a channelized Hotelling observer [with Gabor channels and internal channel noise (CHOi)]. Model observers were evaluated by a receiver operating characteristic curve analysis and compared to the results in the reader study. Two strategies were used to train the SR-MO and CNN-MO: A) building a separate model for each lesion size; B) building one model that was applied to lesions of all sizes. RESULTS: All tested model observers and the human observer were highly correlated at each lesion size and dose level. With strategy A, Pearson's product-moment correlation coefficients r were 0.926 (95% confidence interval (CI): 0.679-0.985) for SR-MO and 0.979 (95% CI: 0.902-0.996) for CNN-MO. With strategy B, r was 0.860 (95% CI: 0.454-0.970) for SR-MO and 0.918 (95% CI: 0.651-0.983) for CNN-MO. For CHOi, r was 0.945 (95% CI: 0.755-0.989). With strategy A, mean absolute percentage differences (MAPD) between the model observers and the human observer were 3.7% for SR-MO and 1.2% for CNN-MO. With strategy B, MAPD were 3.7% for SR-MO and 3.0% for CNN-MO. For the CHOi the MAPD was 2.2%. CONCLUSION: Convolutional neural network model observers can accurately predict the performance of a human observer for all lesion sizes and dose levels in the evaluated signal detection task.

Three-material decomposition with dual-layer spectral CT compared to MRI for the detection of bone marrow edema in patients with acute vertebral fractures.

OBJECTIVES: To assess whether bone marrow edema in patients with acute vertebral fractures can be accurately diagnosed based on three-material decomposition with dual-layer spectral CT (DLCT). MATERIALS AND METHODS: Acute (n = 41) and chronic (n = 18) osteoporotic thoracolumbar vertebral fractures as diagnosed by MRI (hyperintense signal in STIR sequences) in 27 subjects (72 ± 11 years; 17 women) were assessed with DLCT. Spectral data were decomposed into hydroxyapatite, edema-equivalent, and fat-equivalent density maps using an in-house-developed algorithm. Two radiologists, blinded to clinical and MR findings, assessed DLCT and conventional CT independently, using a Likert scale (1 = no edema; 2 = likely no edema; 3 = likely edema; 4 = edema). For DLCT and conventional CT, accuracy, sensitivity, and specificity for identifying acute fractures (Likert scale, 3 and 4) were analyzed separately using MRI as standard of reference. RESULTS: For the identification of acute fractures, conventional CT showed a sensitivity of 0.73-0.76 and specificity of 0.78-0.83, whereas the sensitivity (0.93-0.95) and specificity (0.89) of decomposed DLCT images were substantially higher. Accuracy increased from 0.76 for conventional CT to 0.92-0.93 using DLCT. Interreader agreement for fracture assessment was high in conventional CT (weighted κ [95% confidence interval]; 0.81 [0.70; 0.92]) and DLCT (0.96 [0.92; 1.00]). CONCLUSIONS: Material decomposition of DLCT data substantially improved accuracy for the diagnosis of acute vertebral fractures, with a high interreader agreement. This may spare patients additional examinations and facilitate the diagnosis of vertebral fractures.

Accuracy of iodine quantification in dual-layer spectral CT: Influence of iterative reconstruction, patient habitus and tube parameters.

PURPOSE: Evaluation of the influence of iterative reconstruction, tube settings and patient habitus on the accuracy of iodine quantification with dual-layer spectral CT (DL-CT). MATERIAL AND METHODS: A CT abdomen phantom with different extension rings and four iodine inserts (1, 2, 5 and 10 mg/ml) was scanned on a DL-CT. The phantom was scanned with tube-voltages of 120 and 140 kVp and CTDIvol of 2.5, 5, 10 and 20 mGy. Reconstructions were performed for eight levels of iterative reconstruction (i0-i7). Diagnostic dose levels are classified depending on patient-size and radiation dose. RESULTS: Measurements of iodine concentration showed accurate and reliable results. Taking all CTDIvol-levels into account, the mean absolute percentage difference (MAPD) showed less accuracy for low CTDIvol-levels (2.5 mGy: 34.72%) than for high CTDIvol-levels (20 mGy: 5.89%). At diagnostic dose levels, accurate quantification of iodine was possible (MAPD 3.38%). Level of iterative reconstruction did not significantly influence iodine measurements. Iodine quantification worked more accurately at a tube voltage of 140 kVp. Phantom size had a considerable effect only at low-dose-levels; at diagnostic dose levels the effect of phantom size decreased (MAPD <5% for all phantom sizes). CONCLUSION: With DL-CT, even low iodine concentrations can be accurately quantified. Accuracies are higher when diagnostic radiation doses are employed.