Emulating Low-Dose PCCT Image Pairs With Independent Noise for Self-Supervised Spectral Image Denoising

Photon counting CT (PCCT) acquires spectral measurements and enables generation of material decomposition (MD) images that provide distinct advantages in various clinical situations. However, noise amplification is observed in MD images, and denoising is typically applied. Clean or high-quality references are rare in clinical scans, often making supervised learning (Noise2Clean) impractical. Noise2Noise is a self-supervised counterpart, using noisy images and corresponding noisy references with zero-mean, independent noise. PCCT counts transmitted photons separately, and raw measurements are assumed to follow a Poisson distribution in each energy bin, providing the possibility to create noise-independent pairs. The approach is to use binomial selection to split the counts into two low-dose scans with independent noise. We prove that the reconstructed spectral images inherit the noise independence from counts domain through noise propagation analysis and also validated it in numerical simulation and experimental phantom scans. The method offers the flexibility to split measurements into desired dose levels while ensuring the reconstructed images share identical underlying features, thereby strengthening the model's robustness for input dose levels and capability of preserving fine details. In both numerical simulation and experimental phantom scans, we demonstrated that Noise2Noise with binomial selection outperforms other common self-supervised learning methods based on different presumptive conditions.

Iodine Density Threshold to Distinguish Between Enhancing and Nonenhancing Renal Lesions With Dual-Layer Dual-Energy CT

PURPOSE

This study aimed to determine the optimal threshold iodine density to distinguish enhancing and nonenhancing renal masses with dual-layer dual-energy CT (dlDECT).

METHODS

In this retrospective, HIPAA-compliant, institutional review board-approved study, 383 consecutive renal mass dlDECT studies from September 5, 2018, through December 15, 2022, were reviewed for enhancing solid renal masses with ≥∆20 HU. Studies with simple cysts in the same interval served as controls. Lesion ROI HU measurements on unenhanced and nephrographic phases and ROI iodine density measurements of each lesion and the abdominal aorta for normalization were recorded. The mean lesion HU values and absolute and normalized iodine densities were compared with enhancing and nonenhancing renal lesions using a two-sample t test. The diagnostic accuracy of iodine thresholds was assessed by calculating sensitivity and specificity, with receiver operating characteristic curve and AUC analysis.

RESULTS

There were 38 enhancing and 39 nonenhancing renal lesions. The mean (standard deviation [SD]) ∆HU was 73.5 (38.7) and 3.9 (5.1) HU for enhancing and nonenhancing lesions, respectively. The mean absolute iodine density was significantly different for enhancing and nonenhancing lesions (3.2 [1.7] mg/mL and 0.20 [0.22] mg/mL, respectively; P < 0.0001). Normalized mean iodine density was significantly different for enhancing and nonenhancing lesions (0.62 [0.33] mg/mL and 0.04 [0.04] mg/mL, respectively; P < 0.0001). The optimal absolute iodine density threshold of 0.70 mg/mL (AUC, 0.999) and normalized iodine density threshold of 0.11 mg/mL (AUC, 0.999) were 100% sensitive and 97.4% specific for discriminating enhancing and nonenhancing renal lesions.

CONCLUSIONS

This study provides absolute and normalized iodine density thresholds to distinguish enhancing and nonenhancing renal lesions with high sensitivity and specificity using dlDECT.

Trauma and 'Whole' Body Computed Tomography: Role, Protocols, Appropriateness, and Evidence to Support its Use and When

Imaging plays a crucial role in the immediate evaluation of the trauma patient, particularly using multi-detector computed tomography (CT), and especially in moderately to severely injured trauma patients. There are specific areas of relative consensus, while other aspects of whole-body computed tomography (WB-CT) use remain controversial and are subject to opinion/debate based on the current literature. Even a few hours of a delayed diagnosis may result in a detrimental outcome for the patient. One must utilize all the tools available to enhance the interpretation of images. It is also important to recognize imaging pitfalls and artifacts to avoid unnecessary intervention.

Standardization of Dual-Energy CT Iodine Uptake of the Abdomen and Pelvis: Defining Reference Values in a Big Data Cohort

Background: To establish dual-energy-derived iodine density reference values in abdominopelvic organs in a large cohort of healthy subjects. Methods: 597 patients who underwent portal venous phase dual-energy CT scans of the abdomen were retrospectively enrolled. Iodine distribution maps were reconstructed, and regions of interest measurements were placed in abdominal and pelvic structures to obtain absolute iodine values. Subsequently, normalization of the abdominal aorta was conducted to obtain normalized iodine ratios. The values obtained were subsequently analyzed and differences were investigated in subgroups defined by sex, age and BMI. Results: Overall mean iodine uptake values and normalized iodine ratios ranged between 0.31 and 6.08 mg/mL and 0.06 and 1.20, respectively. Women exhibited higher absolute iodine concentration across all organs. With increasing age, normalized iodine ratios mostly tend to decrease, being most significant in the uterus, prostate, and kidneys (p < 0.015). BMI was the parameter less responsible for variations in iodine concentrations; normal weighted patients demonstrated higher values of both absolute and normalized iodine. Conclusions: Iodine concentration values and normalized iodine ratios of abdominal and pelvic organs reveal significant gender-, age-, and BMI-related differences, underscoring the necessity to integrate these variables into clinical practice.

Image-domain material decomposition for dual-energy CT using unsupervised learning with data-fidelity loss

BACKGROUND

Dual-energy computed tomography (DECT) and material decomposition play vital roles in quantitative medical imaging. However, the decomposition process may suffer from significant noise amplification, leading to severely degraded image signal-to-noise ratios (SNRs). While existing iterative algorithms perform noise suppression using different image priors, these heuristic image priors cannot accurately represent the features of the target image manifold. Although deep learning-based decomposition methods have been reported, these methods are in the supervised-learning framework requiring paired data for training, which is not readily available in clinical settings.

PURPOSE

This work aims to develop an unsupervised-learning framework with data-measurement consistency for image-domain material decomposition in DECT.

METHODS

The proposed framework combines iterative decomposition and deep learning-based image prior in a generative adversarial network (GAN) architecture. In the generator module, a data-fidelity loss is introduced to enforce the measurement consistency in material decomposition. In the discriminator module, the discriminator is trained to differentiate the low-noise material-specific images from the high-noise images. In this scheme, paired images of DECT and ground-truth material-specific images are not required for the model training. Once trained, the generator can perform image-domain material decomposition with noise suppression in a single step.

RESULTS

In the simulation studies of head and lung digital phantoms, the proposed method reduced the standard deviation (SD) in decomposed images by 97% and 91% from the values in direct inversion results. It also generated decomposed images with structural similarity index measures (SSIMs) greater than 0.95 against the ground truth. In the clinical head and lung patient studies, the proposed method suppressed the SD by 95% and 93% compared to the decomposed images of matrix inversion.

CONCLUSIONS

Since the invention of DECT, noise amplification during material decomposition has been one of the biggest challenges, impeding its quantitative use in clinical practice. The proposed method performs accurate material decomposition with efficient noise suppression. Furthermore, the proposed method is within an unsupervised-learning framework, which does not require paired data for model training and resolves the issue of lack of ground-truth data in clinical scenarios.

Update on Renal Cell Carcinoma Diagnosis with Novel Imaging Approaches

This review highlights recent advances in renal cell carcinoma (RCC) imaging. It begins with dual-energy computed tomography (DECT), which has demonstrated a high diagnostic accuracy in the evaluation of renal masses. Several studies have suggested the potential benefits of iodine quantification, particularly for distinguishing low-attenuation, true enhancing solid masses from hyperdense cysts. By determining whether or not a renal mass is present, DECT could avoid the need for additional imaging studies, thereby reducing healthcare costs. DECT can also provide virtual unenhanced images, helping to reduce radiation exposure. The review then provides an update focusing on the advantages of multiparametric magnetic resonance (MR) imaging performance in the histological subtyping of RCC and in the differentiation of benign from malignant renal masses. A proposed standardized stepwise reading of images helps to identify clear cell RCC and papillary RCC with a high accuracy. Contrast-enhanced ultrasound may represent a promising diagnostic tool for the characterization of solid and cystic renal masses. Several combined pharmaceutical imaging strategies using both sestamibi and PSMA offer new opportunities in the diagnosis and staging of RCC, but their role in risk stratification needs to be evaluated. Although radiomics and tumor texture analysis are hampered by poor reproducibility and need standardization, they show promise in identifying new biomarkers for predicting tumor histology, clinical outcomes, overall survival, and the response to therapy. They have a wide range of potential applications but are still in the research phase. Artificial intelligence (AI) has shown encouraging results in tumor classification, grade, and prognosis. It is expected to play an important role in assessing the treatment response and advancing personalized medicine. The review then focuses on recently updated algorithms and guidelines. The Bosniak classification version 2019 incorporates MRI, precisely defines previously vague imaging terms, and allows a greater proportion of masses to be placed in lower-risk classes. Recent studies have reported an improved specificity of the higher-risk categories and better inter-reader agreement. The clear cell likelihood score, which adds standardization to the characterization of solid renal masses on MRI, has been validated in recent studies with high interobserver agreement. Finally, the review discusses the key imaging implications of the 2017 AUA guidelines for renal masses and localized renal cancer.

Image Domain Multi-Material Decomposition Noise Suppression Through Basis Transformation and Selective Filtering

Spectral CT can provide material characterization ability to offer more precise material information for diagnosis purposes. However, the material decomposition process generally leads to amplification of noise which significantly limits the utility of the material basis images. To mitigate such problem, an image domain noise suppression method was proposed in this work. The method performs basis transformation of the material basis images based on a singular value decomposition. The noise variances of the original spectral CT images were incorporated in the matrix to be decomposed to ensure that the transformed basis images are statistically uncorrelated. Due to the difference in noise amplitudes in the transformed basis images, a selective filtering method was proposed with the low-noise transformed basis image as guidance. The method was evaluated using both numerical simulation and real clinical dual-energy CT data. Results demonstrated that compared with existing methods, the proposed method performs better in preserving the spatial resolution and the soft tissue contrast while suppressing the image noise. The proposed method is also computationally efficient and can realize real-time noise suppression for clinical spectral CT images.

Iodine quantification of renal lesions: Preliminary results using spectral-based material extraction on photon-counting CT

Objectives

To assess the range of quantitative iodine values in renal cysts (RC) (with a few renal neoplasms [RNs] as a comparison) to develop an expected range of values for RC that can be used in future studies for their differentiation.

Material and Methods

Consecutive patients (n = 140) with renal lesions who had undergone abdominal examination on a clinical photon-counting computed tomography (PCCT) were retrospectively included. Automated iodine quantification maps were reconstructed, and region of interest (ROI) measurements of iodine concentration (IC) (mg/cm3) were performed on whole renal lesions. In addition, for heterogeneous lesions, a secondary ROI was placed on the area most suspicious for malignancy. The discriminatory values of minimum, maximum, mean, and standard deviation for IC were compared using simple logistic regression and receiver operating characteristic curves (area under the curve [AUC]).

Results

A total of 259 renal lesions (243 RC and 16 RN) were analyzed. There were significant differences between RC and RN for all IC measures with the best-performing metrics being mean and maximum IC of the entire lesion ROI (AUC 0.912 and 0.917, respectively) but also mean and minimum IC of the most suspicious area in heterogeneous lesions (AUC 0.983 and 0.992, respectively). Most RC fell within a range of low measured iodine values although a few had higher values.

Conclusion

Automated iodine quantification maps reconstructed from clinical PCCT have a high diagnostic ability to differentiate RCs and neoplasms. The data from this pilot study can be used to help establish quantitative values for clinical differentiation of renal lesions.

Spectral CT imaging versus conventional CT post-processing technique in differentiating malignant and benign renal tumors

OBJECTIVE

Quantitative comparison of the diagnostic value of spectral CT imaging and conventional CT post-processing technique in differentiating malignant and benign renal tumors.

METHODS

A total of 209 patients with renal tumors who had undergone CT enhancement were assigned to three groups-clear cell renal cell carcinoma (ccRCC, n = 106), non-ccRCC (n = 60), and benign renal tumor (n = 43) groups. Parametric CT enhancement of each tumor based on spectral CT and conventional CT was performed using in-house software, and the iodine concentration, water content, slope, and density values among the three groups were compared. The receiver operating characteristic (ROC) curve analysis was performed to determine the optimum diagnostic thresholds, the area under the ROC curve (AUC), sensitivity, specificity, and accuracy of the above parameters.

RESULTS

The iodine concentration, slope, and density values were higher in the ccRCCs group compared to the non-ccRCCs and benign renal tumor groups (p < 0.05). Moreover, the iodine concentration, slope, and density values were higher in benign renal tumors compared to non-ccRCCs (p < 0.05). According to the ROC curve analysis, iodine concentration presented the highest diagnostic efficacy in differentiating ccRCCs/non-ccRCCs from benign renal tumors. The pairwise comparisons of the ROC curves and the diagnostic efficacies revealed that ROI-based CT enhancement was worse than the spectral CT imaging analysis in terms of density (p < 0.05).

CONCLUSION

Iodine concentration presented the highest diagnostic efficacy in differentiating ccRCCs/non-ccRCCs from benign renal tumors.

ADVANCES IN KNOWLEDGE

1. The iodine concentration, slope, and density values were higher for the ccRCCs compared to non-ccRCCs and benign renal tumors.2. Iodine concentration presented the highest diagnostic efficacy in differentiating ccRCCs/non-ccRCCs from benign renal tumors.3. Spectral CT imaging analysis performed better than conventional CT in differentiating malignant and benign renal tumors.

Dual-Energy Computed Tomography Applications in the Genitourinary Tract

By virtue of material differentiation capabilities afforded through dedicated postprocessing algorithms, dual-energy CT (DECT) has been shown to provide benefit in the evaluation of various diseases. In this article, we review the diagnostic use of DECT in the assessment of genitourinary diseases, with emphasis on its role in renal stone characterization, incidental renal and adrenal lesion characterization, retroperitoneal trauma, reduction of radiation, and contrast dose and cost-effectiveness potential. We also discuss future perspectives of the DECT scanning mode, including the use of novel contrast injection strategies and photon-counting detector computed tomography.

Leveraging Dual-Energy Computed Tomography to Improve Emergency Radiology Practice

Dual-energy computed tomography affords emergency radiologists with important tools to aid in the detection and discrimination of commonly encountered ED pathologies. In doing so, it can increase the speed of diagnosis and diagnostic certainty while sparing patients potentially unnecessary downsteam workups and radiation exposure. This article demonstrates these clinical benefits through a case-based approach.

Spectral performance of two split-filter dual-energy CT systems: A phantom study

BACKGROUND

Recently, a second generation of split filter dual-energy CT (SFCT) platform has been developed. The thicknesses of the gold and tin filters used to obtain both low- and high-energy spectra have been changed. These differences in filter thickness may affect the spectral separation between the two spectra and thus the quality of spectral images.

PURPOSE

To compare the spectral performance of two Split-Filter Dual-Energy CT systems (SFCT-1st and SFCT-2nd ) on virtual monoenergetic images (VMIs) and iodine map.

METHODS

A Multi-Energy CT phantom was scanned on two SFCT with a tube voltage of 120 kVp for both systems (SFCT-1st -120 and SFCT-2nd -120) and 140 kVp only for the second generation (SFCT-2nd -140). Acquisitions were performed on the phantom with a CTDIvol close to 11 mGy. Noise power spectrum (NPS) and task-based transfer function (TTF) were evaluated on VMIs from 40 to 70 keV. A detectability index (d') was computed to assess the detection of two contrast-enhanced lesions on VMIs. Hounsfield Unit (HU) accuracy was assessed on VMIs and the accuracy of iodine concentration was assessed on iodine maps.

RESULTS

For all keV, noise magnitude values were lower with the SFCT-2nd -120 than with the SFCT-1st -120 (on average: -22.5 ± 2.9%) and higher with the SFCT-2nd -140 than with the SFCT-2nd -120 (on average: 25.0 ± 6.2%). Average NPS spatial frequencies (fav ) were lower with the SFCT-1st -120 than with the SFCT-2nd -120 (-6.0 ± 0.5%) and the SFCT-2nd -140 (-3.6 ± 1.6%). Similar TTF50% values were found for both systems and both kVp for blood and iodine inserts at 2 mg/mL (0.29 ± 0.01 mm-1 ) and at 4 mg/mL (0.31 ± 0.01 mm-1 ). d' values peaked at 40 keV for the SFCT-2nd and at 70 keV for the SFCT-1st . Highest d' values were found for the SFCT-2nd -120 for both simulated lesions. Accuracy of HU values and iodine concentration was higher with the SFCT-2nd than with the SFCT 1st .

CONCLUSION

Compared to the SFCT-1st , with similar spatial resolution and noise texture values, the SFCT-2nd -120 exhibited the lowest values for noise magnitude, the highest detectability index values, and more accurate HU values and iodine concentrations.

Incidental Indeterminate Renal Lesions: Distinguishing Non-Enhancing from Potential Enhancing Renal Lesions Using Iodine Quantification on Portal Venous Dual-Layer Spectral CT

The purpose of our study is to determine a threshold for iodine quantification to distinguish definitely non-enhancing benign renal lesions from potential enhancing masses on portal venous dual-layer spectral computed tomography (CT) to reduce the need for additional multiphase CT. In this single-center retrospective study, patients (≥18 years) scanned between April 2021 and January 2023 following the local renal CT protocol were included. Exclusion criteria were patients without renal lesions, lesions smaller than 10 mm, only fat-containing lesions, abscesses or infarction, follow-up after radiofrequent ablation, wrong scan protocol, or artefacts. Scans were performed on a dual layer detector-based spectral CT (CT 7500, Philips Healthcare, Best, The Netherlands). Iodine concentration (mgI/mL) in renal lesions was determined using spectral data. Analyses were performed for all lesions and for lesions of >30 HU on portal venous CT. Enhancement on multiphase CT (≥20 ΔHU from true unenhanced (TUE) to portal venous phase (PVP) CT) was used as reference standard. To determine thresholds for iodine concentration receiver operating characteristic (ROC) curves, area under the curve (AUC) and 95% confidence intervals were calculated. To obtain thresholds for definite (non-)enhancement, 100% sensitivity with maximum specificity and 100% specificity with maximum sensitivity were noted. Data were measured using one reader. To assess interobserver agreement, a second reader performed measurements on the PVP CT scans. A total of 103 patients (62 years ± 14, 68 men) were included. We measured 328 renal lesions, 56 enhancing lesions (17%) in 38 patients and 272 non-enhancing lesions (83%) in 86 patients. The threshold for non-enhancing lesions was 0.76 mgI/mL or lower (100% sensitivity, 76% specificity). The threshold for a definite enhancing mass was 1.69 mgI/mL or higher (100% specificity, 78% sensitivity). A total of 77% of indeterminate lesions (>30 HU on PVP CT) in our study could be definitely characterized. Renal lesions can be definitively classified as non-enhancing or enhancing on PVP spectral CT using thresholds of 0.76 mgI/mL or 1.69 mgI/mL, respectively, eliminating the need for multiphase imaging.

Comparison of radiomics models and dual-energy material decomposition to decipher abdominal lymphoma in contrast-enhanced CT

PURPOSE

The radiologists' workload is increasing, and computational imaging techniques may have the potential to identify visually unequivocal lesions, so that the radiologist can focus on equivocal and critical cases. The purpose of this study was to assess radiomics versus dual-energy CT (DECT) material decomposition to objectively distinguish visually unequivocal abdominal lymphoma and benign lymph nodes.

METHODS

Retrospectively, 72 patients [m, 47; age, 63.5 (27-87) years] with nodal lymphoma (n = 27) or benign abdominal lymph nodes (n = 45) who had contrast-enhanced abdominal DECT between 06/2015 and 07/2019 were included. Three lymph nodes per patient were manually segmented to extract radiomics features and DECT material decomposition values. We used intra-class correlation analysis, Pearson correlation and LASSO to stratify a robust and non-redundant feature subset. Independent train and test data were applied on a pool of four machine learning models. Performance and permutation-based feature importance was assessed to increase the interpretability and allow for comparison of the models. Top performing models were compared by the DeLong test.

RESULTS

About 38% (19/50) and 36% (8/22) of the train and test set patients had abdominal lymphoma. Clearer entity clusters were seen in t-SNE plots using a combination of DECT and radiomics features compared to DECT features only. Top model performances of AUC = 0.763 (CI = 0.435-0.923) were achieved for the DECT cohort and AUC = 1.000 (CI = 1.000-1.000) for the radiomics feature cohort to stratify visually unequivocal lymphomatous lymph nodes. The performance of the radiomics model was significantly (p = 0.011, DeLong) superior to the DECT model.

CONCLUSIONS

Radiomics may have the potential to objectively stratify visually unequivocal nodal lymphoma versus benign lymph nodes. Radiomics seems superior to spectral DECT material decomposition in this use case. Therefore, artificial intelligence methodologies may not be restricted to centers with DECT equipment.

Impact of Phantom Size on Low-Energy Virtual Monoenergetic Images of Three Dual-Energy CT Platforms

The purpose of this study was to compare the quality of low-energy virtual monoenergetic images (VMIs) obtained with three Dual-Energy CT (DECT) platforms according to the phantom diameter. Three sections of the Mercury Phantom 4.0 were scanned on two generations of split-filter CTs (SFCT-1st and SFCT-2nd) and on one Dual-source CT (DSCT). The noise power spectrum (NPS), task-based transfer function (TTF), and detectability index (d') were assessed on VMIs from 40 to 70 keV. The highest noise magnitude values were found with SFCT-1st and noise magnitude was higher with DSCT than with SFCT-2nd for 26 cm (10.2% ± 1.3%) and 31 cm (7.0% ± 2.5%), and the opposite for 36 cm (-4.2% ± 2.5%). The highest average NPS spatial frequencies and TTF values at 50% (f50) values were found with DSCT. For all energy levels, the f50 values were higher with SFCT-2nd than SFCT-1st for 26 cm (3.2% ± 0.4%) and the opposite for 31 cm (-6.9% ± 0.5%) and 36 cm (-5.6% ± 0.7%). The lowest d' values were found with SFCT-1st. For all energy levels, the d' values were lower with DSCT than with SFCT-2nd for 26 cm (-6.2% ± 0.7%), similar for 31 cm (-0.3% ± 1.9%) and higher for 36 cm (5.4% ± 2.7%). In conclusion, compared to SFCT-1st, SFCT-2nd exhibited a lower noise magnitude and higher detectability values. Compared with DSCT, SFCT-2nd had a lower noise magnitude and higher detectability for the 26 cm, but the opposite was true for the 36 cm.

The Value of Dual-Energy Computed Tomography Angiography-Derived Parameters in the Evaluation of Clot Composition

OBJECTIVES

We aimed to assess the value of dual-energy computed tomography angiography (DE-CTA) derived parameters as a quantitative biomarker of thrombus composition in acute ischemic stroke (AIS).

METHODS

AIS patients who underwent DE-CTA before thrombectomy between August 2016 and September 2022 were included in this study. We assessed the relative proportion of red blood cells (RBCs) and the fibrin/platelet ratio (F/P) of the retrieved clots and categorized the clots as RBC-dominant (RBCs > F/P) or F/P-dominant (F/P > RBCs). The thrombus based parameters were measured on polyenergetic images (PEI), virtual monoenergetic (VM), virtual non-contrast (VNC), iodine concentration (IC), and effective atomic number (Zeff) images respectively, and the slope of the spectral Hounsfield unit curve (λHU) was calculated. These parameters were compared in the DE-CTA images of RBC- and F/P-dominant thrombi. The diagnostic performance of the parameters was analyzed using the ROC curve. Correlations between thrombus composition and DE-CTA-derived parameters were assessed.

RESULTS

The retrieved clots in 54 of 88 patients (61.36%) were RBC-dominant. The RBC-dominant thrombi showed significantly higher VNC values and lower IC, λHU, and Zeff values than the F/P-dominant thrombi (p < 0.05). The CT density measured on IC images showed the largest AUC value (AUC, 0.94; sensitivity, 77.78%; specificity, 100.00%). The Spearman rank-order correlation coefficient values showed that CT density measured on IC images of the thrombus showed the strongest association with the proportion of RBCs (r = -0.64, p < 0.001) and F/P (r = 0.65, p < 0.001).

CONCLUSIONS

DE-CTA-derived parameters, especially the CT density measured on IC images, could be associated with thrombus composition and allow for personalized thrombectomy strategies.

New Frontiers in Oncological Imaging With Computed Tomography: From Morphology to Function

The latest evolutions in Computed Tomography (CT) technology have several applications in oncological imaging. The innovations in hardware and software allow for the optimization of the oncological protocol. Low-kV acquisitions are possible thanks to the new powerful tubes. Iterative reconstruction algorithms and artificial intelligence are helpful for the management of image noise during image reconstruction. Functional information is provided by spectral CT (dual-energy and photon counting CT) and perfusion CT.

Dual-Energy CT Evaluation of Gastrointestinal Bleeding

Gastrointestinal (GI) bleeding is a potentially life-threatening condition accounting for more than 300 000 annual hospitalizations. Multidetector abdominopelvic CT angiography is commonly used in the evaluation of patients with GI bleeding. Given that many patients with severe overt GI bleeding are unlikely to tolerate bowel preparation, and inpatient colonoscopy is frequently limited by suboptimal preparation obscuring mucosal visibility, CT angiography is recommended as a first-line diagnostic test in patients with severe hematochezia to localize a source of bleeding. Assessment of these patients with conventional single-energy CT systems typically requires the performance of a noncontrast series followed by imaging during multiple postcontrast phases. Dual-energy CT (DECT) offers several potential advantages for performing these examinations. DECT may eliminate the need for a noncontrast acquisition by allowing the creation of virtual noncontrast (VNC) images from contrast-enhanced data, affording significant radiation dose reduction while maintaining diagnostic accuracy. VNC images can help radiologists to differentiate active bleeding, hyperattenuating enteric contents, hematomas, and enhancing masses. Additional postprocessing techniques such as low-kiloelectron voltage virtual monoenergetic images, iodine maps, and iodine overlay images can increase the conspicuity of contrast material extravasation and improve the visibility of subtle causes of GI bleeding, thereby increasing diagnostic confidence and assisting with problem solving. GI bleeding can also be diagnosed with routine single-phase DECT scans by constructing VNC images and iodine maps. Radiologists should also be aware of the potential pitfalls and limitations of DECT. ^©RSNA, 2023 Quiz questions for this article are available through the Online Learning Center.

Phantom-Less Nonlinear Magnetic Resonance Imaging Calibration With Multiple Input Blood Flow Model

PURPOSE

Previous work used phantoms to calibrate the nonlinear relationship between the gadolinium contrast concentration and the intensity of the magnetic resonance imaging signal. This work proposes a new nonlinear calibration procedure without phantoms and considers the variation of contrast agent mass minimum combined with the multiple input blood flow system. This also proposes a new single-input method with meaningful variables that is not influenced by reperfusion or noise generated by aliasing. The reperfusion in the lung is usually neglected and is not considered by the indicator dilution method. However, in cases of lung cancer, reperfusion cannot be neglected. A new multiple input method is formulated, and the contribution of the pulmonary artery and bronchial artery to lung perfusion can be considered and evaluated separately.

METHODS

The calibration procedure applies the minimum variation of contrast agent mass in 3 different regions: (1) pulmonary artery, (2) left atrium, and (3) aorta. It was compared with four dimensional computerized tomography with iodine, which has a very high proportional relationship between contrast agent concentration and signal intensity.

RESULTS

Nonlinear calibration was performed without phantoms, and it is in the range of phantom calibration. It successfully separated the contributions of the pulmonary and bronchial arteries. The proposed multiple input method was verified in 6 subjects with lung cancer, and perfusion from the bronchial artery, rich in oxygen, was identified as very high in the cancer region.

CONCLUSIONS

Nonlinear calibration of the contrast agent without phantoms is possible. Separate contributions of the pulmonary artery and aorta can be determined.

Imaging biomarkers to stratify lymph node metastases in abdominal CT - Is radiomics superior to dual-energy material decomposition?

PURPOSE

To assess the potential of radiomic features in comparison to dual-energy CT (DECT) material decomposition to objectively stratify abdominal lymph node metastases.

MATERIALS AND METHODS

In this retrospective study, we included 81 patients (m, 57; median age, 65 (interquartile range, 58.7-73.3) years) with either lymph node metastases (n = 36) or benign lymph nodes (n = 45) who underwent contrast-enhanced abdominal DECT between 06/2015-07/2019. All malignant lymph nodes were classified as unequivocal according to RECIST criteria and confirmed by histopathology, PET-CT or follow-up imaging. Three investigators segmented lymph nodes to extract DECT and radiomics features. Intra-class correlation analysis was applied to stratify a robust feature subset with further feature reduction by Pearson correlation analysis and LASSO. Independent training and testing datasets were applied on four different machine learning models. We calculated the performance metrics and permutation-based feature importance values to increase interpretability of the models. DeLong test was used to compare the top performing models.

RESULTS

Distance matrices and t-SNE plots revealed clearer clusters using a combination of DECT and radiomic features compared to DECT features only. Feature reduction by LASSO excluded all DECT features of the combined feature cohort. The top performing radiomic features model (AUC = 1.000; F1 = 1.000; precision = 1.000; Random Forest) was significantly superior to the top performing DECT features model (AUC = 0.942; F1 = 0.762; precision = 0.800; Stochastic Gradient Boosting) (DeLong < 0.001).

CONCLUSION

Imaging biomarkers have the potential to stratify unequivocal lymph node metastases. Radiomics models were superior to DECT material decomposition and may serve as a support tool to facilitate stratification of abdominal lymph node metastases.

Accurate Image Reconstruction in Dual-Energy CT with Limited-Angular-Range Data Using a Two-Step Method

Dual-energy CT (DECT) with scans over limited-angular ranges (LARs) may allow reductions in scan time and radiation dose and avoidance of possible collision between the moving parts of a scanner and the imaged object. The beam-hardening (BH) and LAR effects are two sources of image artifacts in DECT with LAR data. In this work, we investigate a two-step method to correct for both BH and LAR artifacts in order to yield accurate image reconstruction in DECT with LAR data. From low- and high-kVp LAR data in DECT, we first use a data-domain decomposition (DDD) algorithm to obtain LAR basis data with the non-linear BH effect corrected for. We then develop and tailor a directional-total-variation (DTV) algorithm to reconstruct from the LAR basis data obtained basis images with the LAR effect compensated for. Finally, using the basis images reconstructed, we create virtual monochromatic images (VMIs), and estimate physical quantities such as iodine concentrations and effective atomic numbers within the object imaged. We conduct numerical studies using two digital phantoms of different complexity levels and types of structures. LAR data of low- and high-kVp are generated from the phantoms over both single-arc (SA) and two-orthogonal-arc (TOA) LARs ranging from 14∘ to 180∘. Visual inspection and quantitative assessment of VMIs obtained reveal that the two-step method proposed can yield VMIs in which both BH and LAR artifacts are reduced, and estimation accuracy of physical quantities is improved. In addition, concerning SA and TOA scans with the same total LAR, the latter is shown to yield more accurate images and physical quantity estimations than the former. We investigate a two-step method that combines the DDD and DTV algorithms to correct for both BH and LAR artifacts in image reconstruction, yielding accurate VMIs and estimations of physical quantities, from low- and high-kVp LAR data in DECT. The results and knowledge acquired in the work on accurate image reconstruction in LAR DECT may give rise to further understanding and insights into the practical design of LAR scan configurations and reconstruction procedures for DECT applications.

Contrast thresholds for detection of various iodine concentrations in subtraction CT and dual-energy CT systems

OBJECTIVE

To estimate the minimum iodine concentrations detectable in simulated vessels of various diameters for both subtraction computed tomography (CT) and dual-energy CT systems.

METHODS

Fillable tubes (diameters: 1, 3, and 5 mm) were filled with a variety of iodine concentrations (range: 0-20 mg/ml), placed in the center of 28-mm cylindrical rods and surrounded with water. Rods with and without fillable tubes were placed in a 20-cm cylindrical solid-water phantom to simulate administration of iodine in blood vessels. The phantom was scanned with clinical subtraction CT (SCT) and dual-energy CT (DECT) head protocols to assess the detection of minimum iodine concentrations in both systems. The SCT and DECT images were evaluated quantitatively with a MATLAB script to extract regions of interest (ROIs) of each simulated vessel. ROI measurements were used to calculate the limit of detectability (LOD) and signal-to-noise ratio of Rose criteria for the assessment of the contrast thresholds.

RESULTS

Both SNR_Rose and LOD methods agreed and determined the minimum detectable iodine concentration to be 0.4 mg/ml in the 5-mm diameter vessel for SCT. However, the minimum detectable concentration in the 5-mm vessel with DECT was 1 mg/ml. The 3-mm vessel had a minimum detectable concentration of 0.8 mg/ml for SCT and 2 mg/ml for DECT. Lastly, the minimum detectable iodine concentration for the 1-mm vessel was 10 mg/ml for SCT and 10 mg/ml for DECT.

CONCLUSION

In this phantom study, SCT showed the capability to detect lower iodine concentrations compared to DECT. Contrast thresholds varied for vessels of different diameters and the smaller vessels required a higher iodine concentration for detection. Based on this knowledge, radiologists can modify their protocols to increase contrast enhancement.

Dual-energy CT based material decomposition to differentiate intrahepatic cholangiocarcinoma from hepatocellular carcinoma

PURPOSE

To assess the potential of material decomposition in dual-energy CT (DECT) to differentiate intrahepatic cholangiocarcinoma (iCCA) from hepatocellular carcinoma (HCC).

METHOD

In this retrospective study, we included 94 patients (26 female (27.7 %), median age 64.5 (interquartile range 55.5-74.5) years) with either iCCA or HCC who underwent abdominal contrast-enhanced DECT in arterial phase. To test for differences between iCCA (n = 47) and HCC (n = 47), we evaluated mean attenuation and DECT material density values including iodine density (ID), normalized iodine uptake (NIU), fat fraction, and lesion-to-liver parenchyma ratio. Histopathology served as reference standard for all lesions. We used univariate logistic regression models for the outcome iCCA versus HCC. ROC curve analysis was applied to assess discriminative ability of the model. Model accuracy was evaluated by calculating the Brier score. Youden index was applied to establish thresholds to differentiate between iCCA and HCC.

RESULTS

Comparison of quantitative image parameters revealed significant differences between iCCA and HCC for ID (1.6 ± 0.5 mg/ml vs 2.8 ± 0.8 mg/ml, p < 0.001), NIU (14.5 ± 4.8 vs 24.8 ± 10.3, p < 0.001), attenuation (41.9 ± 10.1 HU vs 47.9 ± 8.9 HU, p = 0.003), and fat fraction (12.0 ± 7.8 % vs 9.0 ± 6.4 %, p = 0.045). ROC curve analysis revealed highest ability to differentiate iCCA from HCC for ID (AUC = 0.93, 95 % CI 0.89-0.98). For ID, an optimal threshold of 2.33 mg/dl was determined to discriminate between iCCA and HCC (sensitivity 89.4 %, specificity 76.6 %).

CONCLUSIONS

DECT-based iodine quantification can serve as a tool for the differentiation of iCCA and HCC in contrast-enhanced CT. ID yielded the highest diagnostic performance and may assist in clinical routine CT diagnostics.

Renal Lesion Characterization by Dual-Layer Dual-Energy CT: Comparison of Virtual and True Unenhanced Images

BACKGROUND. Prior studies have provided mixed results for the ability to replace true unenhanced (TUE) images with virtual unenhanced (VUE) images when characterizing renal lesions by dual-energy CT (DECT). Detector-based dual-layer DECT (dlDECT) systems may optimize performance of VUE images for this purpose. OBJECTIVE. The purpose of this article was to compare dual-phase dlDECT examinations evaluated using VUE and TUE images in differentiating cystic and solid renal masses. METHODS. This retrospective study included 110 patients (mean age, 64.3 ± 11.8 years; 46 women, 64 men) who underwent renal-mass protocol dlDECT between July 2018 and February 2022. TUE, VUE, and nephrographic phase image sets were reconstructed. Lesions were diagnosed as solid masses by histopathology or MRI. Lesions were diagnosed as cysts by composite criteria reflecting findings from MRI, ultrasound, and the TUE and nephrographic phase images of the dlDECT examinations. One radiologist measured lesions' attenuation on all dlDECT image sets. Lesion characterization was compared between use of VUE and TUE images, including when considering enhancement of 20 HU or greater to indicate presence of a solid mass. RESULTS. The analysis included 219 lesions (33 solid masses; 186 cysts [132 simple, 20 septate, 34 hyperattenuating]). TUE and VUE attenuation were significantly different for solid masses (33.4 ± 7.1 HU vs 35.4 ± 8.6 HU, p = .002), simple cysts (10.8 ± 5.6 HU vs 7.1 ± 8.1 HU, p < .001), and hyperattenuating cysts (56.3 ± 21.0 HU vs 47.6 ± 16.3 HU, p < .001), but not septate cysts (13.6 ± 8.1 HU vs 14.0 ± 6.8 HU, p = .79). Frequency of enhancement 20 HU or greater when using TUE and VUE images was 90.9% and 90.9% in solid masses, 0.0% and 9.1% in simple cysts, 15.0% and 10.0% in septate cysts, and 11.8% and 38.2% in hyperattenuating cysts. All solid lesions were concordant in terms of enhancement 20 HU or greater when using TUE and VUE images. Twelve simple cysts and nine hyperattenuating cysts showed enhancement of 20 HU or greater when using VUE but not TUE images. CONCLUSION. Use of VUE images reliably detected enhancement in solid masses. However, VUE images underestimated attenuation of simple and hyperattenuating cysts, leading to false-positive findings of enhancement by such lesions. CLINICAL IMPACT. The findings do not support replacement of TUE acquisitions with VUE images when characterizing renal lesions by dlDECT.

Quantitative dual-energy CT techniques in the abdomen

Advances in dual-energy CT (DECT) technology and spectral techniques are catalyzing the widespread implementation of this technology across multiple radiology subspecialties. The inclusion of energy- and material-specific datasets has ushered overall improvements in CT image contrast and noise as well as artifacts reduction, leading to considerable progress in radiologists' ability to detect and characterize pathologies in the abdomen. The scope of this article is to provide an overview of various quantitative clinical DECT applications in the abdomen and pelvis. Several of the reviewed applications have not reached mainstream clinical use and are considered investigational. Nonetheless awareness of such applications is critical to having a fully comprehensive knowledge base to DECT and fostering future clinical implementation.

Iodine Images in Dual-energy CT: Detection of Hepatic Steatosis by Quantitative Iodine Concentration Values

Hepatic steatosis is a common condition and an early manifestation of a systemic metabolic syndrome. As of today, there is no broadly accepted method for the diagnosis of hepatic steatosis in contrast-enhanced CT images. This retrospective study evaluates the potential of quantitative iodine values in portal venous phase iodine images in dual-energy CT (DECT) by measuring iodine concentrations in regions of interest (ROI) and analyzing the absolute iodine concentration of the liver parenchyma as well as three different blood-normalized iodine concentrations in a study cohort of 251 patients. An independent two sample t-test (p < 0.05) was used to compare the iodine concentrations of healthy and fatty liver. Diagnostic performance was assessed by ROC (receiver operating characteristic) curve analysis. The results showed significant differences between the average iodine concentration of healthy and fatty liver parenchyma for the absolute and for the blood-normalized iodine concentrations. The study concludes that the iodine uptake of the liver parenchyma is impaired by hepatic steatosis, and that the measurement of iodine concentration can provide a suitable method for the detection of hepatic steatosis in quantitative iodine images. Suitable thresholds of quantitative iodine concentration values for the diagnosis of hepatic steatosis are provided.

Iodine images in dual energy CT: A monocentric study benchmarking quantitative iodine concentration values of the healthy liver

Dual energy computed tomography (DECT) allows the quantification of specific materials such as iodine contrast agent in human body tissue, potentially providing additional diagnostic data. Yet full diagnostic value can only be achieved if physiological normal values for iodine concentrations are known. We retrospectively evaluated abdominal DECT scans of 105 patients with healthy liver between March and August 2018 (age 17 to 86 years, 43 female and 62 male). The iodine concentrations within ROIs of the liver parenchyma as well as of the abdominal aorta and main portal vein were obtained. We evaluated the absolute iodine concentration and blood-normalized iodine concentrations relating the measured iodine concentration of the liver parenchyma to those of the supplying vessels. The influence of age and gender on the iodine uptake was assessed. The absolute iodine concentration was significantly different for the male and female cohort, but the difference was eliminated by the blood-normalized values. The average blood-normalized iodine concentrations were 2.107 mg/ml (+/- 0.322 mg/ml), 2.125 mg/ml (+/- 0.426 mg/ml) and 2.103 mg/ml (+/- 0.317 mg/ml) for the portal vein normalized, aorta normalized and mixed blood normalized iodine concentrations, respectively. A significant negative correlation between the patients’ age and the iodine concentration was detected only for the blood-normalized values. A physiological range for iodine concentration in portal venous phase contrast enhanced DECT images can be defined for absolute and blood-normalized values. Deviations of blood-normalized iodine concentration values might be a robust biomarker for diagnostic evaluation. Patient age but not the gender influences the blood-normalized iodine concentrations in healthy liver parenchyma.

Recent advances in imaging techniques of renal masses

Multiphasic multidetector computed tomography (CT) forms the mainstay for the characterization of renal masses whereas magnetic resonance imaging (MRI) acts as a problem-solving tool in some cases. However, a few of the renal masses remain indeterminate even after evaluation by conventional imaging methods. To overcome the deficiency in current imaging techniques, advanced imaging methods have been devised and are being tested. This review will cover the role of contrast-enhanced ultrasonography, shear wave elastography, dual-energy CT, perfusion CT, MR perfusion, diffusion-weighted MRI, blood oxygen level-dependent MRI, MR spectroscopy, positron emission tomography (PET)/prostate-specific membrane antigen-PET in the characterization of renal masses.

Can radiomic analysis of a single-phase dual-energy CT improve the diagnostic accuracy of differentiating enhancing from non-enhancing small renal lesions?

BACKGROUND

The value of dual-energy computed tomography (DECT)-based radiomics in renal lesions is unknown.

PURPOSE

To develop DECT-based radiomic models and assess their incremental values in comparison to conventional measurements for differentiating enhancing from non-enhancing small renal lesions.

MATERIAL AND METHODS

A total of 349 patients with 519 small renal lesions (390 non-enhancing, 129 enhancing) who underwent contrast-enhanced nephrographic phase DECT examinations between June 2013 and January 2020 on multiple DECT platforms were retrospectively recruited. Cohort A included all lesions, while cohort B included Bosniak II-IV and solid enhancing renal lesions. Radiomic models were built with features selected by the least absolute shrinkage and selection operator regression (LASSO). ROC analyses were performed to compare the diagnostic accuracy among conventional and radiomic models for predicting enhancing renal lesions.

RESULTS

The individual iodine concentration (IC), normalized IC, mean attenuation on 75-keV images, radiomic model of iodine images, 75-keV images and a combined model integrating all the above-mentioned features all demonstrated high AUCs for predicting renal lesion enhancement in cohort A (AUCs = 0.934-0.979) as well as in the test dataset (AUCs = 0.892-0.962) of cohort B (P values with Bonferroni correction >0.003). The AUC (0.864) of mean attenuation on 75-keV images was significantly lower than those of other models (all P values ≤0.001) except the radiomic model of 75-keV images (P = 0.038) in the training dataset of cohort B.

CONCLUSION

No incremental value was found by adding radiomic and machine learning analyses to iodine images for differentiating enhancing from non-enhancing renal lesions.

Accuracy of iodine quantification using dual-energy computed tomography with focus on low concentrations

BACKGROUND

Iodine quantification using dual-energy computed tomography (DECT) is helpful in characterizing, and follow-up after treatment of tumors. Some malignant masses, for instance papillary renal cell carcinomas (p-RCC), are hard to differentiate from benign lesions because of very low contrast enhancement. In these cases, iodine concentrations might be very low, and it is therefore important that iodine quantification is reliable even at low concentrations if this technique is used.

PURPOSE

To examine the accuracy of iodine quantification and to determine whether it is also accurate for low iodine concentrations.

MATERIAL AND METHODS

Twenty-six syringes with different iodine concentrations (0-30 mg I/mL) were scanned in a phantom model using a DECT scanner with two different kilovoltage and image reconstruction settings. Iodine concentrations were measured and compared to known concentration. Absolute and relative errors were calculated.

RESULTS

For concentrations of 1 mg I/mL or higher, there was an excellent correlation between true and measured iodine concentrations for all settings (R = 0.999-1.000; P < 0.001). For concentrations <1.0 mg I/mL, the relative error was greater. Absolute and relative errors were smaller using tube voltages of 80/Sn140 kV than 100/Sn140 kV (P < 0.01). Reconstructions using a 3.0-mm slice thickness had less variance between repeated acquisitions versus 0.6 mm (P < 0.001).

CONCLUSION

Iodine quantification using DECT was in general very accurate, but for concentrations < 1.0 mg I/mL the technique was less reliable. Using a tube voltage with larger spectral separation was more accurate and the result was more reproducible using thicker image reconstructions.

Value of energy spectrum CT parameters in the differential diagnosis of high-grade clear cell renal cell carcinoma and type II papillary renal cell carcinoma

BACKGROUND

Energy spectrum computed tomography (CT) has become a promising approach for the differential diagnosis of tumor subtypes.

PURPOSE

To explore the value of energy spectrum CT parameters in the differential diagnosis of high-grade clear cell renal cell carcinoma (ccRCC) and type II papillary renal cell carcinoma (pRCC).

MATERIAL AND METHODS

Forty-two cases of high-grade ccRCC and 28 cases of type II pRCC were retrospectively reviewed. All region of interest (ROI) measurements were maintained consistently between the two-phase contrast-enhanced examinations. The ROIs encompassed as much of the enhancing areas of the lesions as possible. Energy spectrum CT parameters of all cases, including the 70 keV (HU) value, normalized iodine concentration (NIC), and energy spectrum curve slope were recorded by two radiologists with over 10 years of experience in abdominal CT diagnosis.

RESULTS

In the cortical phase (CP) and parenchymal phase (PP), the 70 keV (HU) value, NIC, and slope value of the energy spectrum curve of high-grade ccRCC were significantly higher than those of type II pRCC. In the CP, NIC showed the highest differential diagnosis efficiency for the two group tumors, with a sensitivity of 78.9% and a specificity of 77.0%. There was no statistical difference in tumor hemorrhage, tumor envelope, tumor morphology, tumor border, lymph node metastasis, embolism, renal pelvis invasion, or tumor calcification between the two tumor types. However, there was significant difference in the number of tumors (P = 0.019).

CONCLUSION

Energy spectrum CT parameters are valuable for the differential diagnosis of high-grade ccRCC and type II pRCC.

Comparison of virtual monoenergetic imaging between a rapid kilovoltage switching dual-energy computed tomography with deep-learning and four dual-energy CTs with iterative reconstruction

Background

To assess the spectral performance of rapid kV switching dual-energy CT (KVSCT-Canon) equipped with a Deep-Learning spectral reconstruction algorithm on virtual-monoenergetic images at low-energy levels and to compare its performances with four other dual-energy CT (DECT) platforms equipped with iterative reconstruction algorithms.

Methods

Two CT phantoms were scanned on five DECT platforms: KVSCT-Canon, fast kV-switching CT (KVSCT-GE), split filter CT, dual-source CT (DSCT), and dual-layer CT (DLCT). The classical parameters of abdomen-pelvic examinations were used for all phantom acquisitions, and a CTDI_vol close to 10 mGy. For KVSCT-Canon, virtual-monoenergetic images were reconstructed with a clinical slice thickness of 0.5 and 1.5 mm to be close to other platforms. Noise power spectrum (NPS) and task-based transfer function (TTF) were evaluated from 40 to 80 keV of virtual-monoenergetic images. A detectability index (d') was computed to model the detection task of two contrast-enhanced lesions as function of keV.

Results

For KVSCT-Canon, the noise magnitude and average NPS spatial frequency (f_av) decreased from 40 to 70 keV and increased thereafter. Similar noise magnitude outcomes were found for KVSCT-GE but the opposite for f_av. For the other DECT platforms, the noise magnitude decreased as the keV increased. For split filter CT, DSCT and DLCT, the f_av values increased from 40 to 80 keV. For all DECT platforms, TTF at 50% (f₅₀) decreased as the keV increased, decreasing spatial resolution. For KVSCT-Canon, d' values peaked at 60 and 70 keV for both simulated lesions and from 50 to 70 keV for KVSCT-GE. d' decreased between 40 and 70 keV for DSCT, DLCT and split filter CT. For KVSCT-Canon, the increase in slice thickness decreases noise magnitude, f_av and f₅₀ and increases d' values. The highest d' values were found for DLCT at 40 and 50 keV and for KVSCT-Canon at 1.5 mm for other keV.

Conclusions

For KVSCT-Canon, the detectability of contrast-enhanced lesions was highest at 60 keV. The highest d' values were found for DLCT at 40 and 50 keV and for KVSCT-Canon at 1.5 mm for other keV.

Realistic Kidney Tissue Surrogates for Multienergy Computed Tomography-Feasibility and Estimation of Energy-Dependent Attenuation Thresholds for Renal Lesion Enhancement in Low-kV and Virtual Monoenergetic Imaging

PURPOSE

The aims of this study were to assess if kidney tissue surrogates (KTSs) are superior to distilled water-iodine solutions in the emulation of energy-dependent computed tomography (CT) attenuation characteristics of renal parenchyma and to estimate attenuation thresholds for definite lesion enhancement for low-kV single-energy and low-keV dual-energy virtual monoenergetic imaging.

METHODS

A water-filled phantom (diameter, 30 cm) with multiple vials was imaged on a dual-source dual-energy CT (DS-DE) and a single-source split-filter dual-energy CT (SF-DE), both in single-energy mode at 80, 100, 120, 140 kVp and in dual-energy mode at 80/Sn150, 90/Sn150, and 100/Sn150 kVp for DS-DE and AuSn120 kVp for SF-DE. Single-energy images, linear-blended dual-energy images, and virtual monoenergetic imaging at energy levels from 40 to 190 keV were reconstructed. First, attenuation characteristics of KTS in solid and liquid consistencies were compared. Second, solid KTSs were developed to match the CT attenuation of unenhanced renal parenchyma at 120 kVp as retrospectively measured in 100 patients. Third, CT attenuation of KTS-iodine and water-iodine solutions at 8 different iodine concentrations (0-10 mg I/mL) were compared as a function of tube voltage and of keV level using multiple linear regression models. Energy-dependent attenuation thresholds for definite lesion enhancement were calculated.

RESULTS

Unenhanced renal parenchyma at 120 kVp measured on average 30 HU on both scanners in the patient cohort. Solid KTS with a water content of 80% emulated the attenuation of unenhanced renal parenchyma (30 HU) more accurately compared with water-iodine solutions (0 HU). Attenuation difference between KTS-iodine and water-iodine solutions converged with increasing iodine concentration and decreasing x-ray energy due to beam-hardening effects. A slight attenuation difference of approximately 2 HU was found between the 2 CT scanners. Attenuation thresholds for definite lesion enhancement were dependent on tube voltage and keV level and ranged from 16.6 to 33.2 HU and 3.2 to 68.3 HU for single-energy and dual-energy CT scan modes for DS-DE and from 16.1 to 34.3 HU and 3.3 to 92.2 HU for SF-DE.

CONCLUSIONS

Kidney tissue surrogates more accurately emulate the energy-dependent CT attenuation characteristics of renal parenchyma for multienergy CT compared with conventional water-iodine approaches. Energy-dependent thresholds for definite lesion enhancement could facilitate lesion characterization when imaging at different energies than the traditional 120 kVp.

A Method for Reducing Variability Across Dual-Energy CT Manufacturers in Quantification of Low Iodine Content Levels

Background: Clinical use of the dual-energy CT (DECT) iodine quantification technique is hindered by between-platform (i.e., across different manufacturers) variability in iodine concentration (IC), particularly at low iodine levels. Objective: To develop in an anthropomorphic phantom a method for reducing between-platform variability in quantification of low iodine content levels using DECT and to evaluate the method's performance in patients undergoing serial clinical DECT examinations on different platforms. Methods: An anthropomorphic phantom in three body sizes, incorporating varied lesion types and scanning conditions, was imaged with three distinct DECT implementations from different manufacturers at varying radiation exposures. A cross-platform iodine quantification model for correcting between-platform variability at low iodine content was developed using the phantom data. The model was tested in a retrospective series of 30 patients (20 men, 10 women; median age, 62 years) who each underwent three serial contrast-enhanced DECT examinations of the abdomen and pelvis (90 scans total) for routine oncology surveillance, using the same three DECT platforms as in the phantom. Estimated accuracy of phantom IC values was summarized using rootmean-squared error (RMSE) relative to known IC. Between-platform variability in patients was summarized using root-mean-square-deviation (RMSD). RMSE and RMSD were compared between platform-based IC (ICPB) and cross-platform IC (ICCP). ICPB was normalized to aorta and portal vein. Results: In the phantom study, mean RMSE of ICPB across platforms and other experimental conditions was 0.65 ± 0.18 mgI/mL compared with 0.40 ± 0.075 mgI/mL for ICCP (38% decrease in mean RMSE; P<.05). Intra-patient between-platform variability across serial DECT examinations was lower for ICPB than ICCP (RMSD: 97% vs 88%; P<.001). Between-platform variability was not reduced by normalization of ICPB to aorta (RMSD: 97% vs 101%; P=.12) or portal vein (RMSD: 97% vs 97%; P=.81). Conclusion: The developed cross-platform method significantly decreased between-platform variability occurring at low iodine content with platform-based DECT iodine quantification. Clinical Impact: With further validation, the cross-platform method, which has been implemented as a webbased app, may expand clinical use of DECT iodine quantification, yielding meaningful IC values that reflect tissue biologic viability or treatment response in patients who undergo serial examinations on different platforms.

Dual-Energy CT Vital Iodine Tumor Burden for Response Assessment in Patients With Metastatic GIST Undergoing TKI Therapy: Comparison to Standard CT and FDG PET/CT Criteria

Background: CT-based criteria for assessing gastroinstestinal stromal tumor (GIST) response to tyroskine kinase inhibitor (TKI) therapy are limited partly because tumor attenuation is influenced by treatment-related changes including hemorrhage and calcification. Iodine concentration may be less impacted by such changes. Objective: To determine whether DECT vital iodine tumor burden (TB) provides improved differentiation between responders and non-responders in patients with metastatic GIST undergoing TKI therapy compared to established CT and PET/CT criteria. Methods: An anthropomorphic phantom with spherical inserts mimicking GIST lesions of varying iodine concentrations and having non-enhancing central necrotic cores underwent DECT to determine a threshold iodine concentration. Forty patients (median age 57 years; 25 women, 15 men) treated with TKI for metaststic GIST were retrospectively evaluated. Patients underwent baseline and follow-up DECT and FDG PET/CT. Response assessment was performed using RECIST 1.1, modified Choi (mChoi), vascular tumor burden (VTB), DECT vital iodine TB, and European Organization for Research and Treatment of Cancer (EORTC PET) criteria. DECT vital iodine TB used the same percentage changes as RECIST 1.1 response categories. Progression-free survival (PFS) was compared between responders and non-responders for each response criteria using Cox proportional hazard ratios and Harrell's c-indices. Results: The phantom experiment identified a 0.5 mg/mL threshold to differentiate vital from non-vital tissue. Using DECT vital iodine TB, median PFS was significantly different between non-responders and responders (587 vs 167 days, respectively; p=.02). Hazard ratio for progression for DECT vital iodine TB non-responders versus responders was 6.9, versus 7.6 for EORTC PET, 3.3 for VTB, 2.3 for RECIST 1.1, and 2.1 for mChoi. C-index was 0.74 for EORTC PET, 0.73 for DECT vital iodine TB, 0.67 for VTB, 0.61 for RECIST 1.1, and 0.58 for mChoi. C-index was significantly greater for DECT vital iodine TB than RECIST 1.1 (p=.02) and mChoi (p=.002), but not different than VTB and EORTC PET (p>.05). Conclusion: DECT vital iodine TB criteria showed comparable performance as EORTC PET and outperformed RECIST 1.1 and mChoi for response assessment of metastatic GIST under TKI therapy. Clinical Impact: DECT vital iodine TB could help guide early management decisions in patients on TKI therapy.

Image Quality of Virtual Monochromatic Reconstructions of Noncontrast CT on a Dual-Source CT Scanner in Adult Patients

RATIONALE AND OBJECTIVES

To evaluate the image quality of virtual monochromatic images (VMI) reconstructed from dual-energy dual-source noncontrast head CT with different reconstruction kernels.

MATERIALS AND METHODS

Twenty-five consecutive adult patients underwent noncontrast dual-energy CT. VMI were retrospectively reconstructed at 5-keV increments from 40 to 140 keV using quantitative and head kernels. CT-number, noise levels (SD), signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) in the gray and white matter and artifacts using the posterior fossa artifact index (PFAI) were evaluated.

RESULTS

CT-number increased with decreasing VMI energy levels, and SD was lowest at 85 keV. SNR was maximized at 80 keV and 85 keV for the head and quantitative kernels, respectively. CNR was maximum at 40 keV; PFAI was lowest at 90 (head kernel) and 100 (quantitative kernel) keV. Optimal VMI image quality was significantly better than conventional CT.

CONCLUSION

Optimal image quality of VMI energies can improve brain parenchymal image quality compared to conventional CT but are reconstruction kernel dependent and depend on indication for performing noncontrast CT.

Design and implementation of a practical quality control program for dual-energy CT

A novel routine dual‐energy computed tomography (DECT) quality control (QC) program was developed to address the current deficiency of routine QC for this technology. The dual‐energy quality control (DEQC) program features (1) a practical phantom with clinically relevant materials and concentrations, (2) a clinically relevant acquisition, reconstruction, and postprocessing protocol, and (3) a fully automated analysis software to extract quantitative data for database storage and trend analysis. The phantom, designed for easy set up for standalone or adjacent imaging next to the ACR phantom, was made in collaboration with an industry partner and informed by clinical needs to have four iodine inserts (0.5, 1, 2, and 5 mg/ml) and one calcium insert (100 mg/ml) equally spaced in a cylindrical water‐equivalent background. The imaging protocol was based on a clinical DECT abdominal protocol capable of producing material specific concentration maps, virtual unenhanced images, and virtual monochromatic images. The QC automated analysis software uses open‐source technologies which integrates well with our current automated CT QC database. The QC program was tested on a GE 750 HD scanner and two Siemens SOMATOM FLASH scanners over a 3‐month period. The automated algorithm correctly identified the appropriate region of interest (ROI) locations and stores measured values in a database for monitoring and trend analysis. Slight variations in protocol settings were noted based on manufacturer. Overall, the project proved to provide a convenient and dependable clinical tool for routine oversight of DE CT imaging within the clinic.

Dual-energy CT imaging with limited-angular-range data

In dual-energy computed tomography (DECT), low- and high-kVp data are collected often over a full-angular range (FAR) of 360^○. While there exists strong interest in DECT with low- and high-kVp data acquired over limited-angular ranges (LARs), there remains little investigation of image reconstruction in DECT with LAR data.Objective: We investigate image reconstruction with minimized LAR artifacts from low- and high-kVp data over LARs of ≤180^○by using a directional-total-variation (DTV) algorithm.Methods: Image reconstruction from LAR data is formulated as a convex optimization problem in which data-l₂is minimized with constraints on image's DTVs along orthogonal axes. We then achieve image reconstruction by applying the DTV algorithm to solve the optimization problem. We conduct numerical studies from data generated over arcs of LARs, ranging from 14^○to 180^○, and perform visual inspection and quantitative analysis of images reconstructed.Results: Monochromatic images of interest obtained with the DTV algorithm from LAR data show substantially reduced artifacts that are observed often in images obtained with existing algorithms. The improved image quality also leads to accurate estimation of physical quantities of interest, such as effective atomic number and iodine-contrast concentration.Conclusion: Our study reveals that from LAR data of low- and high-kVp, monochromatic images can be obtained that are visually, and physical quantities can be estimated that are quantitatively, comparable to those obtained in FAR DECT.Significance: As LAR DECT is of high practical application interest, the results acquired in the work may engender insights into the design of DECT with LAR scanning configurations of practical application significance.

Iterative material decomposition for spectral CT using self-supervised Noise2Noise prior

Compared to conventional computed tomography (CT), spectral CT can provide the capability of material decomposition, which can be used in many clinical diagnosis applications. However, the decomposed images can be very noisy due to the dose limit in CT scanning and the noise magnification of the material decomposition process. To alleviate this situation, we proposed an iterative one-step inversion material decomposition algorithm with a Noise2Noise prior. The algorithm estimated material images directly from projection data and used a Noise2Noise prior for denoising. In contrast to supervised deep learning methods, the designed Noise2Noise prior was built based on self-supervised learning and did not need external data for training. In our method, the data consistency term and the Noise2Noise network were alternatively optimized in the iterative framework, respectively, using a separable quadratic surrogate (SQS) and the Adam algorithm. The proposed iterative algorithm was validated and compared to other methods on simulated spectral CT data, preclinical photon-counting CT data and clinical dual-energy CT data. Quantitative analysis showed that our proposed method performs promisingly on noise suppression and structure detail recovery.

Dual-Energy CT-Based Iodine Quantification in Liver Tumors - Impact of Scan-, Patient-, and Position-Related Factors

RATIONALE AND OBJECTIVES

To quantify the contribution of lesion location and patient positioning, dual-energy approach, patient size, and radiation dose to the error of dual-energy CT-based iodine quantification (DECT-IQ) in liver tumors.

MATERIALS AND METHODS

A phantom with four liver lesions (diameter 15 mm; iodine concentration 0-5 mgI/mL) and two sizes was used. One lesion emulated a subdiaphragmatic lesion. Both sizes were imaged in dual-energy mode on (1) a dual-source DECT (DS-DE) at 100/Sn150 kV and (2) a single-source split-filter DECT (SF-DE) at AuSn120 kV at two radiation doses (8 and 12 mGy). Scans were performed at seven different vertical table positions (from -6 to + 6 cm from the gantry isocenter). Iodine concentration was repeatedly measured and absolute errors (error_abs) were calculated. Errors were compared using robust repeated-measures ANOVAs with post-hoc comparisons. A linear mixed effect model was used to determine the factors influencing the error of DECT-IQ.

RESULTS

The linear mixed effect models showed that errors were significantly influenced by DECT approach, phantom size, and lesion location (all p < 0.001). The impact of lesion location on the error was stronger in SF-DE compared to DS-DE. Radiation dose did not significantly influence error (p = 0.22). When averaged across all setups, error_abs was significantly higher for SF-DE (2.08 ± 1.92 mgI/mL) compared to DS-DE (0.37 ± 0.29 mgI/mL) (all p < 0.001). Artefacts were found in the subdiaphragmatic lesion for SF-DE with significantly increased error_abs compared to DS-DE (p < 0.001). Error_abs was significantly higher in the large compared to the medium phantom for DS-DE (0.30 ± 0.23 mgI/mL vs. 0.43 ± 0.33 mgI/mL) and SF-DE (1.68 ± 1.99 vs. 2.36 ± 1.81 mgI/mL) (p < 0.001).

CONCLUSION

The dual-energy approach, patient size, and lesion location modified by patient position significantly impacted DECT-IQ in simulated liver tumors.

Evaluating renal lesions using deep-learning based extension of dual-energy FoV in dual-source CT-A retrospective pilot study

PURPOSE

Dual-source (DS) CT, dual-energy (DE) field of view (FoV) is limited to the size of the smaller detector array. The purpose was to establish a deep learning-based approach to DE extrapolation by estimating missing image data using data from both tubes to evaluate renal lesions.

METHOD

A DE extrapolation deep-learning (DEEDL) algorithm had been trained on DECT data of 50 patients using a DSCT with DE-FoV = 33 cm (Somatom Flash). Data from 128 patients with known renal lesions falling within DE-FoV was retrospectively collected (100/140 kVp; reference dataset 1). A smaller DE-FoV = 20 cm was simulated excluding the renal lesion of interest (dataset 2) and the DEEDL was applied to this dataset. Output from the DEEDL algorithm was evaluated using ReconCT v14.1 and Syngo.via. Mean attenuation values in lesions on mixed images (HU) were compared calculating the root-mean-squared-error (RMSE) between the datasets using MATLAB R2019a.

RESULTS

The DEEDL algorithm performed well reproducing the image data of the kidney lesions (Bosniak 1 and 2: 125, Bosniak 2F: 6, Bosniak 3: 1 and Bosniak 4/(partially) solid: 32) with RSME values of 10.59 HU, 15.7 HU for attenuation, virtual non-contrast, respectively. The measurements performed in dataset 1 and 2 showed strong correlation with linear regression (r2: attenuation = 0.89, VNC = 0.63, iodine = 0.75), lesions were classified as enhancing with an accuracy of 0.91.

CONCLUSION

This DEEDL algorithm can be used to reconstruct a full dual-energy FoV from restricted data, enabling reliable HU value measurements in areas not covered by the smaller FoV and evaluation of renal lesions.

A material decomposition method for dual-energy CT via dual interactive Wasserstein generative adversarial networks

PURPOSE

Dual-energy computed tomography (DECT) is highly promising for material characterization and identification, whereas reconstructed material-specific images are affected by magnified noise and beam-hardening artifacts. Although various DECT material decomposition methods have been proposed to solve this problem, the quality of the decomposed images is still unsatisfactory, particularly in the image edges. In this study, a data-driven approach using dual interactive Wasserstein generative adversarial networks (DIWGAN) is developed to improve DECT decomposition accuracy and perform edge-preserving images.

METHODS

In proposed DIWGAN, two interactive generators are used to synthesize decomposed images of two basis materials by modeling the spatial and spectral correlations from input DECT reconstructed images, and the corresponding discriminators are employed to distinguish the difference between the generated images and labels. The DECT images reconstructed from high- and low-energy bins are sent to two generators separately, and each generator synthesizes one material-specific image, thereby ensuring the specificity of the network modeling. In addition, the information from different energy bins is exploited through the feature sharing of two generators. During decomposition model training, a hybrid loss function including L₁ loss, edge loss, and adversarial loss is incorporated to preserve the texture and edges in the generated images. Additionally, a selector is employed to define the generator that should be trained in each iteration, which can ensure the modeling ability of two different generators and improve the material decomposition accuracy. The performance of the proposed method is evaluated using digital phantom, XCAT phantom, and real data from a mouse.

RESULTS

On the digital phantom, the regions of bone and soft tissue are strictly and accurately separated using the trained decomposition model. The material densities in different bone and soft-tissue regions are near the ground truth, and the error of material densities is lower than 3 mg/ml. The results from XCAT phantom show that the material-specific images generated by directed matrix inversion and iterative decomposition methods have severe noise and artifacts. Regarding to the learning-based methods, the decomposed images of fully convolutional network (FCN) and butterfly network (Butterfly-Net) still contain varying degrees of artifacts, while proposed DIWGAN can yield high quality images. Compared to Butterfly-Net, the root-mean-square error (RMSE) of soft-tissue images generated by the DIWGAN decreased by 0.01 g/ml, whereas the peak-signal-to-noise ratio (PSNR) and structural similarity (SSIM) of the soft-tissue images reached 31.43 dB and 0.9987, respectively. The mass densities of the decomposed materials are nearest to the ground truth when using the DIWGAN method. The noise standard deviation of the decomposition images reduced by 69%, 60%, 33%, and 21% compared with direct matrix inversion, iterative decomposition, FCN, and Butterfly-Net, respectively. Furthermore, the performance of the mouse data indicates the potential of the proposed material decomposition method in real scanned data.

CONCLUSIONS

A DECT material decomposition method based on deep learning is proposed, and the relationship between reconstructed and material-specific images is mapped by training the DIWGAN model. Results from both the simulation phantoms and real data demonstrate the advantages of this method in suppressing noise and beam-hardening artifacts.

Intraindividual Consistency of Iodine Concentration in Dual-Energy Computed Tomography of the Chest and Abdomen

OBJECTIVES

Dual-energy computed tomography (DECT)-derived quantification of iodine concentration (IC) is increasingly used in oncologic imaging to characterize lesions and evaluate treatment response. However, only limited data are available on intraindividual consistency of IC and its variation. This study investigates the longitudinal reproducibility of IC in organs, vessels, and lymph nodes in a large cohort of healthy patients who underwent repetitive DECT imaging.

MATERIALS AND METHODS

A total of 159 patients, who underwent a total of 469 repetitive (range, 2-4), clinically indicated portal-venous phase DECT examinations of the chest and abdomen, were retrospectively included. At time of imaging, macroscopic tumor burden was excluded by follow-up imaging (≥3 months). Iodine concentration was measured region of interest-based (N = 43) in parenchymatous organs, vessels, lymph nodes, and connective tissue. Normalization of IC to the aorta and to the trigger delay as obtained from bolus tracking was performed. For statistical analysis, intraclass correlation coefficient and modified variation coefficient (MVC) were used to assess intraindividual agreement of IC and its variation between different time points, respectively. Furthermore, t tests and analysis of variance with Tukey-Kramer post hoc test were used.

RESULTS

The mean intraclass correlation coefficient over all regions of interest was good to excellent (0.642-0.936), irrespective of application of normalization or the normalization technique. Overall, MVC ranged from 1.8% to 25.4%, with significantly lower MVC in data normalized to the aorta (5.8% [1.8%-15.8%]) in comparison with the MVC of not normalized data and data normalized to the trigger delay (P < 0.01 and P = 0.04, respectively).

CONCLUSIONS

Our study confirms intraindividual, longitudinal variation of DECT-derived IC, which varies among vessels, lymph nodes, organs, and connective tissue, following different perfusion characteristics; normalizing to the aorta seems to improve reproducibility when using a constant contrast media injection protocol.

Performance of four dual-energy CT platforms for abdominal imaging: a task-based image quality assessment based on phantom data

OBJECTIVES

To compare the spectral performance of dual-energy CT (DECT) platforms using task-based image quality assessment based on phantom data.

MATERIALS AND METHODS

Two CT phantoms were scanned on four DECT platforms: fast kV-switching CT (KVSCT), split filter CT (SFCT), dual-source CT (DSCT), and dual-layer CT (DLCT). Acquisitions on each phantom were performed using classical parameters of abdomen-pelvic examination and a CTDI_vol at 10 mGy. Noise power spectrum (NPS) and task-based transfer function (TTF) were evaluated from 40 to 140 keV of virtual monoenergetic images. A detectability index (d') was computed to model the detection task of two contrast-enhanced lesions as function of keV.

RESULTS

The noise magnitude decreased from 40 to 70 keV for all DECT platforms, and the highest noise magnitude values were found for KVSCT and SFCT and the lowest for DSCT and DLCT. The average NPS spatial frequency shifted towards lower frequencies as the energy level increased for all DECT platforms, smoothing the image texture. TTF values decreased with the increase of keV deteriorating the spatial resolution. For both simulated lesions, higher detectability (d' value) was obtained at 40 keV for DLCT, DSCT, and SFCT but at 70 keV for KVSCT. The detectability of both simulated lesions was highest for DLCT and DSCT.

CONCLUSION

Highest detectability was found for DLCT for the lowest energy levels. The task-based image quality assessment used for the first time for DECT acquisitions showed the benefit of using low keV for the detection of contrast-enhanced lesions.

KEY POINTS

• Detectability of both simulated contrast-enhanced lesions was higher for dual-layer CT for the lowest energy levels. • The image noise increased and the image texture changed for the lowest energy levels. • The detectability of both simulated contrast-enhanced lesions was highest at 40 keV for all dual-energy CT platforms except for fast kV-switching platform.

Computed tomography in rheumatology - From DECT to high-resolution peripheral quantitative CT

In this chapter, we discuss current updates and applications of Dual Energy Computed Tomography (DECT), iodine-DECT mapping, and high-resolution peripheral quantitative CT (HR-pQCT) in rheumatology. DECT provides a noninvasive diagnosis of gout and can help to differentiate gout from CPPD. Accuracy of DECT varies in various stages of gout. DECT needs specialized hardware, software, and skilled post-processing and interpretation. Sensitivity reduces significantly with deeper tissues such as hip and shoulder. Iodine map enables to delineate inflammatory lesions such as capsulitis and tenosynovitis by improving iodine contrast. Iodine quantification with an iodine map is a promising objective method to evaluate therapeutic effect of inflammatory arthritis. HR-pQCT allows for highly sensitive and specific measures of bone erosions and osteophytes in inflammatory joint diseases, documenting change over time, e.g. in cohorts undergoing immunosuppressive treatments. However, assessing the images requires trained readers, and (semi)-automated scripts to detect bone damage are still undergoing validation and further development.

Use of dual-energy CT for renal mass assessment

Although dual-energy CT (DECT) may prove useful in a variety of abdominal imaging tasks, renal mass evaluation represents the area where this technology can be most impactful in abdominal imaging compared to routinely performed contrast-enhanced-only single-energy CT exams. DECT post-processing techniques, such as creation of virtual unenhanced and iodine density images, can help in the characterization of incidentally discovered renal masses that would otherwise remain indeterminate based on post-contrast imaging only. The purpose of this article is to review the use of DECT for renal mass assessment, including its benefits and existing limitations. KEY POINTS: • If DECT is selected as the scanning mode for most common abdominal protocols, many incidentally found renal masses can be fully triaged within the same exam. • Virtual unenhanced and iodine density DECT images can provide additional information when renal masses are discovered in the post-contrast-only setting. • For renal mass evaluation, virtual unenhanced and iodine density DECT images should be interpreted side-by-side to troubleshoot pitfalls that can potentially lead to erroneous interpretation.

Iodine concentration of healthy lymph nodes of neck, axilla, and groin in dual-energy computed tomography

BACKGROUND

Lymph nodes (LN) are examined in every computed tomography (CT) scan. Until now, an evaluation is only possible based on morphological criteria. With dual-energy CT (DECT) systems, iodine concentration (IC) can be measured which could conduct in an improved diagnostic evaluation of LNs.

PURPOSE

To define standard values for IC of cervical, axillary, and inguinal LNs in DECT.

MATERIAL AND METHODS

Imaging data of 297 patients who received a DECT scan of the neck, thorax, abdomen-pelvis, or a combination of those in a portal-venous phase were retrospectively collected from the institutional PACS. No present history of malignancy, inflammation, or trauma in the examined region was present. For each examined region, the data of 99 patients were used. The IC of the three largest LNs, the main artery, the main vein, and a local muscle of the examined area was measured, respectively.

RESULTS

Normalization of the IC of LNs to the artery, vein, muscle, or a combination of those did not lead to a decreased value-range. The smallest range and confidence interval (CI) of IC was found when using absolute values of IC for each region. Hereby, mean values (95% CI) for IC of LN were found: 2.09 mg/mL (2.00-2.18 mg/mL) for neck, 1.24 mg/mL (1.16-1.33 mg/mL) for axilla, and 1.11 mg/mL (1.04-1.17 mg/mL) for groin.

CONCLUSION

The present study suggests standard values for IC of LNs in dual-layer CT could be used to differentiate between healthy and pathological lymph nodes, considering the used contrast injection protocol.

Deep-learning-based direct inversion for material decomposition

PURPOSE

To develop a convolutional neural network (CNN) that can directly estimate material density distribution from multi-energy computed tomography (CT) images without performing conventional material decomposition.

METHODS

The proposed CNN (denoted as Incept-net) followed the general framework of encoder-decoder network, with an assumption that local image information was sufficient for modeling the nonlinear physical process of multi-energy CT. Incept-net was implemented with a customized loss function, including an in-house-designed image-gradient-correlation (IGC) regularizer to improve edge preservation. The network consisted of two types of customized multibranch modules exploiting multiscale feature representation to improve the robustness over local image noise and artifacts. Inserts with various densities of different materials [hydroxyapatite (HA), iodine, a blood-iodine mixture, and fat] were scanned using a research photon-counting detector (PCD) CT with two energy thresholds and multiple radiation dose levels. The network was trained using phantom image patches only, and tested with different-configurations of full field-of-view phantom and in vivo porcine images. Furthermore, the nominal mass densities of insert materials were used as the labels in CNN training, which potentially provided an implicit mass conservation constraint. The Incept-net performance was evaluated in terms of image noise, detail preservation, and quantitative accuracy. Its performance was also compared to common material decomposition algorithms including least-square-based material decomposition (LS-MD), total-variation regularized material decomposition (TV-MD), and U-net-based method.

RESULTS

Incept-net improved accuracy of the predicted mass density of basis materials compared with the U-net, TV-MD, and LS-MD: the mean absolute error (MAE) of iodine was 0.66, 1.0, 1.33, and 1.57 mgI/cc for Incept-net, U-net, TV-MD, and LS-MD, respectively, across all iodine-present inserts (2.0-24.0 mgI/cc). With the LS-MD as the baseline, Incept-net and U-net achieved comparable noise reduction (both around 95%), both higher than TV-MD (85%). The proposed IGC regularizer effectively helped both Incept-net and U-net to reduce image artifact. Incept-net closely conserved the total mass densities (i.e., mass conservation constraint) in porcine images, which heuristically validated the quantitative accuracy of its outputs in anatomical background. In general, Incept-net performance was less dependent on radiation dose levels than the two conventional methods; with approximately 40% less parameters, the Incept-net achieved relatively improved performance than the comparator U-net, indicating that performance gain by Incept-net was not achieved by simply increasing network learning capacity.

CONCLUSION

Incept-net demonstrated superior qualitative image appearance, quantitative accuracy, and lower noise than the conventional methods and less sensitive to dose change. Incept-net generalized and performed well with unseen image structures and different material mass densities. This study provided preliminary evidence that the proposed CNN may be used to improve the material decomposition quality in multi-energy CT.