TH‐C‐144‐03: Tissue Decomposition From Dual Energy CT Data to Reduce Range Uncertainties in Proton and Carbon Radiotherapy

To compare the radiation exposure from single-energy CT (SECT) against rapid kV-switching dual-energy CT (DECT) imaging in both adults and children when resulting image data offer equivalent lesion identification power. Lesions in an adult and a 10-year-old-child body phantom were imitated using iodine solutions of different concentrations. Phantoms were subjected to several SECT and DECT thoracic and abdominal scans using a rapid kV-switching DECT scanner. The contrast-to-noise ratio (CNR) of each lesion was measured on resulting SECT images and virtual monoenergetic images (VMI) available from DECT. The SECT scans that resulted in CNR values similar to the maximum CNR observed in VMIs derived from corresponding DECT scans were identified. SECT and DECT scans with equivalent lesion-discriminating power were compared regarding the associated radiation dose burden. Doses to the lung, breast, and esophagus from thoracic imaging and doses to the liver, kidneys, and stomach from abdominal imaging were determined through Monte Carlo simulations of SECT and DECT exposures. Compared to SECT imaging of the adult body phantom, organ doses from DECT were found to be 5-11% lower in thoracic imaging and 44-45% lower in abdominal imaging. Compared to SECT imaging of the 10-year-old body phantom, organ doses from DECT were found to be 2.8-3.4 times higher in thoracic imaging and 1.5-1.6 times higher in abdominal imaging. The use of rapid kV-switching DECT instead of SECT imaging may be associated with a similar or lower dose burden in adults but a noticeably higher dose burden in children. Question How does the radiation exposure from single-energy and dual-energy CT imaging compare when both techniques provide equivalent lesion identification power? Findings Rapid kV-switching dual-energy CT compared to single-energy CT may result in a similar or lower radiation dose in adults, but higher radiation dose in children. Clinical relevance Rapid kV-switching dual-energy CT imaging in children should be preferred over single-energy CT imaging only in cases where the additional information provided is crucial for an effective diagnosis.

Purpose:Dual energy CT can predict stopping power ratios (SPR) for ion therapy treatment planning. Several approaches have been proposed recently, however accuracy and practicability in a clinical workflow are unaddressed. The aim of this work is to provide a fair comparison of available approaches in a human‐like phantom to find the optimal method for tissue characterization in a clinical situation.Methods:The SPR determination accuracy is investigated using simulated DECT images. A virtual human‐like phantom is created containing 14 different standard human tissues. SECT (120 kV) and DECT images (100 kV and 140 kV Sn) are simulated using the software ImaSim. The single energy CT (SECT) stoichiometric calibration method and four recently published calibration‐based DECT methods are implemented and used to predict the SPRs from simulated images. The difference between SPR predictions and theoretical SPR are compared pixelwize. Mean, standard deviation and skewness of the SPR difference distributions are used as measures for bias, dispersion and symmetry.Results:The average SPR differences and standard deviations are (0.22 ± 1.27)% for SECT, and A) (−0.26 ± 1.30)%, B) (0.08 ± 1.12)%, C) (0.06 ± 1.15)% and D) (−0.05 ± 1.05)% for the four DECT methods. While SPR prediction using SECT is showing a systematic error on SPR, the DECT methods B, C and D are unbiased. The skewness of the SECT distribution is 0.57%, and A) −0.19%, B) −0.56%, C) −0.29% and D) −0.07% for DECT methods respectively.Conclusion:The here presented DECT methods B, C and D outperform the commonly used SECT stoichiometric calibration. These methods predict SPR accurately without a bias and within ± 1.2% (68th percentile). This indicates that DECT potentially improves accuracy of range predictions in proton therapy. A validation of these findings using clinical CT images of real tissues is necessary.

Dual energy CT (DECT) has been shown to estimate stopping power ratio (SPR) map with a higher accuracy than conventional single energy CT (SECT) by obtaining the energy dependence of photon interactions. This work presents a learning-based method to synthesize DECT images from SECT image for proton radiotherapy. The proposed method uses a residual attention generative adversarial network. Residual blocks with attention gates were used to force the model to focus on the difference between DECT images and SECT images. To evaluate the accuracy of the method, we retrospectively investigated 70 head-and-neck cancer patients whose DECT and SECT scans were acquired simultaneously. The model was trained to generate both a high and low energy DECT image based on a SECT image. The generated synthetic low and high DECT images were evaluated against the true DECT images using leave-one-out cross-validation. To evaluate our method in the context of a practical application, we generated SPR maps from synthetic DECT (sDECT) using a dual-energy based stoichiometric method and compared the SPR maps to those generated from DECT. A dosimetric comparison for dose obtained from DECT was performed against that derived from sDECT. The mean of mean absolute error, peak signal-to-noise ratio and normalized cross-correlation for the synthetic high and low energy CT images was 36.9 HU, 29.3 dB, 0.96 and 35.8 HU, 29.2 dB, and 0.96, respectively. The corresponding SPR maps generated from synthetic DECT showed an average normalized mean square deviation of about 1% with reduced noise level and artifacts than those from original DECT. Dose-volume histogram (DVH) metrics for the clinical target volume agree within 1% between the DECT and sDECT calculated dose. Our method synthesized accurate DECT images and showed a potential feasibility for proton SPR map generation. This study investigated a learning-based method to synthesize DECT images from SECT image for proton radiotherapy.

PurposeTo investigate and quantify the potential benefits associated with the use of stopping-power-ratio (SPR) images created from dual-energy computed tomography (DECT) images for proton dose calculation in a clinical proton treatment planning system (TPS).Materials and MethodsThe DECT and single-energy computed tomography (SECT) scans obtained for 26 plastic tissue surrogate plugs were placed individually in a tissue-equivalent plastic phantom. Relative-electron density (ρe) and effective atomic number (Zeff) images were reconstructed from the DECT scans and used to create an SPR image set for each plug. Next, the SPR for each plug was measured in a clinical proton beam for comparison of the calculated values in the SPR images. The SPR images and SECTs were then imported into a clinical TPS, and treatment plans were developed consisting of a single field delivering a 10 × 10 × 10-cm3 spread-out Bragg peak to a clinical target volume that contained the plugs. To verify the accuracy of the TPS dose calculated from the SPR images and SECTs, treatment plans were delivered to the phantom containing each plug, and comparisons of point-dose measurements and 2-dimensional γ-analysis were performed.ResultsFor all 26 plugs considered in this study, SPR values for each plug from the SPR images were within 2% agreement with measurements. Additionally, treatment plans developed with the SPR images agreed with the measured point dose to within 2%, whereas a 3% agreement was observed for SECT-based plans. γ-Index pass rates were > 90% for all SECT plans and > 97% for all SPR image–based plans.ConclusionTreatment plans created in a TPS with SPR images obtained from DECT scans are accurate to within guidelines set for validation of clinical treatment plans at our center. The calculated doses from the SPR image–based treatment plans showed better agreement to measured doses than identical plans created with standard SECT scans.

Metal artefacts pose a common problem in single energy computed tomography (SECT) images used for radiotherapy. Virtual monoenergetic (VME) images constructed with dual energy computed tomography (DECT) scans can be used to reduce beam hardening artefacts. Dual energy metal artefact reduction is compared and combined with iterative metal artefact reduction (iMAR) to determine optimal imaging strategies for patients with metal prostheses. SECT and DECT scans were performed on a Siemens Somatom AS-64 Slice CT scanner. Images were acquired of a modified CIRS pelvis phantom with 6, 12, 20mm diameter stainless steel rods and VME images reconstructed at 100, 120, 140 and 190keV. These were post-reconstructed with and without the iMAR algorithm. Artefact reduction was measured using: (1) the change in Hounsfield Unit (HU) with and without metal artefact reduction (MAR) for 4 regions of interest; (2) the total number of artefact pixels, defined as pixels with a difference (between images with metal rod and without) exceeding a threshold; (3) the difference in the mean pixel intensity of the artefact pixels. DECT, SECT + iMAR and DECT + iMAR were compared. Both SECT + iMAR and DECT + iMAR offer successful MAR for phantom simulating unilateral hip prosthesis. DECT gives minimal artefact reduction over iMAR alone. Quantitative metrics are advantageous for MAR analysis but have limitations that leave room for metric development.

Real-time diagnosis of intracerebral hemorrhage by generating dual-energy CT from single-energy CT

ID: 114 The impact of dual-energy CT tissue segmentation for low-dose rate prostate brachytherapy Monte Carlo dose calculations

Diagnostic value of quantitative parameters for myocardial perfusion assessment in patients with suspected coronary artery disease by single- and dual-energy computed tomography myocardial perfusion imaging.

While dual-energy computed tomography (DECT) images provide clinically useful information than single-energy CT (SECT), SECT remains the most widely used CT system globally, and only a few institutions can use DECT. This study aimed to establish an artificial intelligence (AI)-based image-domain material decomposition technique using multiple keV-output learning of virtual monochromatic images (VMIs) to create DECT-equivalent images from SECT images. This study involved 82 patients with head and neck cancer. Of these, the AI model was built with data from the 67 patients with only DECT scans, while 15 patients with both SECT and DECT scans were used for SECT testing. Our AI model generated VMI50keV and VMI100keV from VMI70keV equivalent to 120-kVp SECT images. We introduced a loss function for material density images (MDIs) in addition to the loss for VMIs. For comparison, we trained the same model with the loss for VMIs only. DECT-equivalent images were generated from SECT images and compared with the true DECT images. The prediction time was 5.4s per patient. The proposed method with the MDI loss function quantitatively provided more accurate DECT-equivalent images than the model trained with the loss for VMIs only. Using real 120-kVp SECT images, the trained model produced precise DECT images of excellent quality. In this study, we developed an AI-based material decomposition approach for head and neck cancer patients by introducing the loss function for MDIs via multiple keV-output learning. Our results suggest the feasibility of AI-based image-domain material decomposition in a conventional SECT system without a DECT scanner.

Improving image quality of triple-low-protocol renal artery CT angiography with deep-learning image reconstruction: a comparative study with standard-dose single-energy and dual-energy CT with adaptive statistical iterative reconstruction

To investigate in an anthropomorphic phantom study the accuracy of dual-energy computed tomography (DECT) techniques for fat quantification in comparison with magnetic resonance (MR) proton density fat fraction (PDFF) and single-energy computed tomography (SECT), using known fat content as reference standard. Between August 2018 and November 2020, organic material-based cylinders, composed of mixtures of lean and fat tissues mimics, iodine, and iron, were constructed to simulate varying fat content levels (0%, 10%, 15%, 25%, 50%, 75%, and 100%) in a parenchymal organ and were embedded into an anthropomorphic phantom simulating 3 patient sizes (circumference, 91, 126, and 161 cm). The phantom was imaged with multiecho MR, DECT, and SECT. Magnetic resonance PDFF, DECT fat fraction, and computed tomography (CT) numbers (SECT polychromatic and DECT monochromatic data, virtual unenhanced images) were estimated. Performances of MR PDFF and CT techniques to detect differences in fat content were measured using the area under the curve (AUC). Noninferiority of each CT technique relative to MR PDFF was tested using a noninferiority margin of -0.1. MR PDFF, DECT 140 keV monochromatic data, and fat fraction most closely correlated with known fat content (R2 = 0.98, 0.98, and 0.96, respectively). Unlike SECT and all other DECT techniques, DECT fat fraction was not affected by presence of iodine (mean difference, 0.3%; 95% confidence interval [CI], -0.9% to 1.5%). Dual-energy computed tomography fat fraction showed noninferiority to MR PDFF in detecting differences of 5% in fat content in medium-sized phantoms (ΔAUC, -0.05; 95% CI, -0.08 to -0.01), and 7% in large (ΔAUC, -0.04; 95% CI, -0.0 to 0.00) or extralarge sized phantoms (ΔAUC, -0.02; 95% CI, -0.07 to 0.00). Dual-energy computed tomography fat fraction shows linear correlation with true fat content in the range up to 50% fat fraction. Dual-energy computed tomography fat fraction has comparable estimation error and shows noninferiority to MR PDFF in detecting small differences in fat content across different body sizes.

Ion range estimation by using dual energy computed tomography

PurposeRelative stopping powers (RSPs) for proton therapy are estimated using single‐energy computed tomography (SECT), calibrated with standardized tissues of the adult male. It is assumed that those tissues are representative of tissues of all age and sex. Female, male, and pediatric tissues differ from one another in density and composition. In this study, we use tabulated pediatric tissues and computational phantoms to investigate the impact of this assumption on pediatric proton therapy. The potential of dual‐energy CT (DECT) to improve the accuracy of these calculations is explored.MethodsWe study 51 human body tissues, categorized into male/female for the age groups newborn, 1‐, 5‐, 10‐, and 15‐year‐old children, and adult, with given compositions and densities. CT numbers are simulated and RSPs are estimated using SECT and DECT methods. Estimated tissue RSPs from each method are compared to theoretical RSPs. The dose and range errors of each approach are evaluated on three computational phantoms (Ewing's sarcoma, salivary sarcoma, and glioma) derived from pediatric proton therapy patients.ResultsWith SECT, soft tissues have mean estimation errors and standard deviation up to (1.96 ± 4.18)% observed in newborns, compared to (0.20 ± 1.15)% in adult males. Mean estimation errors for bones are up to (−3.35 ± 4.76)% in pediatrics as opposed to (0.10 ± 0.66)% in adult males. With DECT, mean errors reduce to (0.17 ± 0.13)% and (0.23 ± 0.22)% in newborns (soft tissues/bones). With SECT, dose errors in a Ewing's sarcoma phantom are exceeding 5 Gy (10% of prescribed dose) at the distal end of the treatment field, with volumes of dose errors >5 Gy of Vdiff>5=4630.7 mm3. Similar observations are made in the head and neck phantoms, with overdoses to healthy tissue exceeding 2 Gy (4%). A systematic Bragg peak shift resulting in either over‐ or underdosage of healthy tissues and target volumes depending on the crossed tissues RSP prediction errors is observed. Water equivalent range errors of single beams are between −1.53 and 5.50 mm (min, max) (Ewing's sarcoma phantom), −0.78 and 3.62 mm (salivary sarcoma phantom), and −0.43 and 1.41 mm (glioma phantom). DECT can reduce dose errors to <1 Gy and range errors to <1 mm.ConclusionSingle‐energy computed tomography estimates RSPs for pediatric tissues with systematic shifts. DECT improves the accuracy of RSPs and dose distributions in pediatric tissues compared to the SECT calibration curve based on adult male tissues.

Organ delineation is crucial to diagnosis and therapy, while it is also labor-intensive and observer-dependent. Dual energy CT (DECT) provides additional image contrast than conventional single energy CT (SECT), which may facilitate automatic organ segmentation. This work aims to develop an automatic multi-organ segmentation approach using deep learning for head-and-neck region on DECT. We proposed a mask scoring regional convolutional neural network (R-CNN) where comprehensive features are firstly learnt from two independent pyramid networks and are then combined via deep attention strategy to highlight the informative ones extracted from both two channels of low and high energy CT. To perform multi-organ segmentation and avoid misclassification, a mask scoring subnetwork was integrated into the Mask R-CNN framework to build the correlation between the class of potential detected organ’s region-of-interest (ROI) and the shape of that organ’s segmentation within that ROI. We evaluated our model on DECT images from 127 head-and-neck cancer patients (66 training, 61 testing) with manual contours of 19 organs as training target and ground truth. For large- and mid-sized organs such as brain and parotid, the proposed method successfully achieved average Dice similarity coefficient (DSC) larger than 0.8. For small-sized organs with very low contrast such as chiasm, cochlea, lens and optic nerves, the DSCs ranged between around 0.5 and 0.8. With the proposed method, using DECT images outperforms using SECT in almost all 19 organs with statistical significance in DSC (p<0.05). Meanwhile, by using the DECT, the proposed method is also significantly superior to a recently developed FCN-based method in most of organs in terms of DSC and the 95th percentile Hausdorff distance. Quantitative results demonstrated the feasibility of the proposed method, the superiority of using DECT to SECT, and the advantage of the proposed R-CNN over FCN on the head-and-neck patient study. The proposed method has the potential to facilitate the current head-and-neck cancer radiation therapy workflow in treatment planning.

Proton radiotherapy treatment plans are currently restricted by the range uncertainties originating from the stopping power ratio (SPR) prediction based on single-energy computed tomography (SECT). Various studies have shown that multi-energy CT (MECT) can reduce the range uncertainties due to medical implant materials and age-related variations in tissue composition. None of these has directly applied the basis material density (MD) images produced by projection-based MECT systems for SPR prediction. To present and evaluate a novel proton SPR prediction method based on MD images from dual-energy CT (DECT), which could reduce the range uncertainties currently associated with proton radiotherapy. A theoretical basis material decomposition into water and iodine material densities was performed for various pediatric and adult human reference tissues, as well as other non-tissue materials, by minimizing the root-mean-square relative attenuation error in the energy interval from 40 to 140keV. A model (here called MD-SPR) mapping predicted MDs to theoretically calculated reference SPRs was created with locally weighted scatterplot smoothing (LOWESS) data-fitting. The goodness of fit of the MD-SPR model was evaluated for the included reference tissues. MD images of two electron density phantoms, combined to form a head- and an abdomen-sized phantom setup, were acquired with a clinical projection-based fast-kV switching DECT scanner. The MD images were compared to the theoretically predicted MDs of the tissue surrogates and other non-tissue materials in the phantoms, as well as used for input to the MD-SPR model for generation of SPR images. The SPR images were subsequently compared to theoretical reference SPRs of the materials in the phantoms, as well as to SPR images from a commercial algorithm (DirectSPR, Siemens Healthineers, Forchheim, Germany) using image-based consecutive scan DECT for the same phantom setups. The predicted SPRs of the tissue surrogates were similar for MD-SPR and DirectSPR, where the adipose and bone tissue surrogates were within 1% difference to the reference SPRs, while other non-adipose soft tissue surrogates (breast, brain, liver, muscle) were all underestimated by between -0.7% and -1.8%. The SPRs of the non-tissue materials (polymethyl methacrylate (PMMA), polyether ether ketone (PEEK), graphite and Teflon) were within 2.8% for MD-SPR images, compared to 6.8% for DirectSPR. The MD-SPR model performed similar compared to other published methods for the human reference tissues. The SPR prediction for tissue surrogates was similar to DirectSPR and showed potential to improve SPR prediction for non-tissue materials.