Non-contrast CT Images Research Articles

MP54-01 A SYSTEMATIC REVIEW OF CLINICAL SIGNS AND SYMPTOMS, PREDICTION RULES AND IMAGING MODALITIES IN PATIENTS WITH SUSPECTED RENAL COLIC

You have accessJournal of UrologyStone Disease: Epidemiology & Evaluation II (MP54)1 Sep 2021MP54-01 A SYSTEMATIC REVIEW OF CLINICAL SIGNS AND SYMPTOMS, PREDICTION RULES AND IMAGING MODALITIES IN PATIENTS WITH SUSPECTED RENAL COLIC Philipp Dahm, Alex Koziarz, Charles J. Gerardo, Simranjeet Benipal, Daniel K. Nishijima, Jae Hung Jung, and Ali S. Raja Philipp DahmPhilipp Dahm More articles by this author , Alex KoziarzAlex Koziarz More articles by this author , Charles J. GerardoCharles J. Gerardo More articles by this author , Simranjeet BenipalSimranjeet Benipal More articles by this author , Daniel K. NishijimaDaniel K. Nishijima More articles by this author , Jae Hung JungJae Hung Jung More articles by this author , and Ali S. RajaAli S. Raja More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002084.01AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Suspected renal colic frequently prompts computed tomography (CT) imaging, which is resource-intense, exposes patients to ionizing radiation, and may not always be necessary. We performed this systematic review (SR) and meta-analysis (MTA) to assess the diagnostic accuracy of clinical signs and symptoms, laboratory investigations, prediction rules and imaging modalities commonly used in patients with clinically suspected renal colic. METHODS: We conducted a SR and MTA according to an a priori, registered protocol (PROSPERO CRD42017055153). Literature search was performed across several databases from inception to July 2, 2020. Two independent reviewers performed study screening, selection, quality assessment and data abstraction using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-2 tool. RESULTS: Among 7641 references screened, 53 were included in the MTA. The overall pooled prevalence for ureteral stones among patients with clinically suspected renal colic was 63% (95% CI: 58-67%). No individual demographic feature, symptom, or sign when present had an LR+≥2.0. The presence of any RBCs on the UA increased the likelihood of a ureteral stone (sensitivity 0.83; specificity 0.59; LR 2.2, 95% CI 1.3-3.6). A score of ≥10 for the Sex, Timing, Origin, Nausea and Erythrocytes (STONE) clinical prediction rule (range: 0–13) had a sensitivity of 0.49; specificity of 0.91; LR+ of 5.3, 95% CI, 4.1-6.7) whereas a score ≥6 had a sensitivity of 0.94, specificity of 0.43; and was able to rule out stones accurately when <6 (LR of 0.15, 95% CI 0.10-0.22). Low dose CT scanning was the most useful imaging technique for identifying patients with or without ureteral stones (sensitivity 0.93; specificity 0.94; LR+ 17, 95% CI, 8.8-31; LR-0.08; 95% CI, 0.03-0.19), ultrasound (sensitivity 0.74; specificity 0.89; LR+ 7.0, 95% CI, 3.4-14; LR- 0.31; 95% CI, 0.16-0.49), and plain x-ray radiographs (sensitivity 0.55; specificity 0.79; LR+ 2.7, 95% CI, 1.9-3.7; LR- 0.57; 95% CI, 0.46-0.68). CONCLUSIONS: Individual signs, symptoms, or the presence of microscopic hematuria do not substantially impact the likelihood of ureteral stones in patients with clinically suspected renal colic. The STONE score may sufficiently guide physicians’ decision to obtain imaging, however further research is required. Low dose non-contrast CT-imaging provides superior diagnostic accuracy compared to all other imaging index tests. Source of Funding: Departmental © 2021 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 206Issue Supplement 3September 2021Page: e948-e948 Advertisement Copyright & Permissions© 2021 by American Urological Association Education and Research, Inc.MetricsAuthor Information Philipp Dahm More articles by this author Alex Koziarz More articles by this author Charles J. Gerardo More articles by this author Simranjeet Benipal More articles by this author Daniel K. Nishijima More articles by this author Jae Hung Jung More articles by this author Ali S. Raja More articles by this author Expand All Advertisement Loading ...

An automated ASPECTS method with atlas-based segmentation

Background and PurposeAs a simple and reliable systematic method to evaluate the early ischemic changes in the blood supply region of the middle cerebral artery of patients with ischemic stroke, the Alberta Stroke Program Early CT score (ASPECTS) can be used for rapid semi-quantitative evaluation of ischemic lesions, which is helpful to select potential candidates for intravenous and intra-arterial therapies, determine the thrombolytic effect and long-term prognosis. This method mainly relies on doctors’ visual observation. However, due to different levels of doctor's experience, the poor inter-reader agreement may result in errors in the final ASPECTS. The purpose of this work was to propose an automated semi-quantitative method for the diagnosis of acute ischemic stroke based on non-contrast computed tomography (NCCT), to provide a reference for doctors in the diagnosis and evaluation. MethodsNCCT data from a total of 90 patients were included for auto-ASPECTS training and testing. After preprocessing CT images, the regions of interest (ROI) for ASPECTS were labeled using atlas-based segmentation. The mean difference, mean ratio and brain density shifts (BDS) of the corresponding regions of the contralateral brain were used as the standard for quantitative analysis. The auto-ASPECTS method was developed and validated to predict early ischemic changes whose performance was evaluated by the agreement (accuracy) of predictions and consensus scores of two observers. ResultsA comparison was made among the results on mean difference, mean ratio, BDS and the combination of multiple parameters as the standard. The result of using BDS alone was relatively better than the result of using any other parameter alone or any combination of multiple parameters, and accuracy in the test set was 0.80. In the test set, accuracy with using different BDS thresholds increased by 6.67% compared with using the consistent BDS threshold. After dichotomy of auto-ASPECTS and consensus scores with the threshold of 7, the agreement of them was 83.3% and there was no significant difference between the two distributions (p = 0.344) in McNemar test. ConclusionsThe proposed auto-ASPECTS method for NCCT images can provide useful information for early diagnosis and evaluation of patients with acute ischemic stroke (AIS).

88. Generation of CT myelograms using noncontrast CT with machine learning

<h3>BACKGROUND CONTEXT</h3> Preoperative imaging is essential in surgical planning for spinal deformity. While MRIs have reduced the need for invasive imaging in most cases, some patients, especially patients with prior hardware, still require CT myelograms, which involve injection of contrast into the thecal sac. Creating an algorithm that can generate a "predicted CT myelogram" may help reduce the risks associated with invasive imaging while providing valuable clinical information. <h3>PURPOSE</h3> Our study aims to develop an algorithm that utilizes the recent advances in computer vision, GANs in particular, to generate predicted CT myelograms from noncontrast CT images of the spine to reduce the need for unnecessary invasive imaging. <h3>STUDY DESIGN/SETTING</h3> Cross-sectional database review study. <h3>PATIENT SAMPLE</h3> Routine imaging data from 32 patients aged 18 or older from 01/01/2005 to 11/01/2020 that underwent both noncontrast CT and CT myelogram at a major academic center. <h3>OUTCOME MEASURES</h3> Qualitative assessment of predicted CT myelograms compared to actual CT myelograms on the 10-patient hold-out dataset. <h3>METHODS</h3> Using routine imaging data from 22 patients, we trained a GAN to generate predicted CT myelograms from noncontrast CTs. The GAN's effectiveness was then evaluated on a hold-out dataset of 10 additional patients. <h3>RESULTS</h3> We successfully trained a GAN using 1875 slices of noncontrast CT and CT myelogram data from 22 patients. When evaluated on data from a 10 patient hold-out dataset, the GAN was able to enhance the thecal sac on some of the patient images. <h3>CONCLUSIONS</h3> GANs can enhance presence of the thecal sac using information present in noncontrast CT images, but may require additional optimizations to generate complete contrast. Further optimizations and training with more data may allow GANs to reduce resource utilization and patient harms by decreasing the need for invasive imaging in the future. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.

Non-Contrast CT-Based Radiomics Score for Predicting Hematoma Enlargement in Spontaneous Intracerebral Hemorrhage.

To develop anon-contrast computed tomography-(CT)-based radiomics score for predicting the risk of hematoma early enlargement in spontaneous intracerebral hemorrhage. A total of 258 patients from asingle-center database with acute spontaneous intracerebral parenchymal hemorrhage were collected. Radiomics software was explored to segment hematomas on baseline non-contrast CT images, and the texture features were extracted. Minimal Redundancy and Maximal Relevance (mRMR) and Least Absolute Shrinkage and Selection Operator (LASSO), were used to select optimized subset of features and radiomics score was calculated. The radiomics model (radiomics score-based), radiomics nomogram (radiomics score combined with clinical factors-based) and clinical model (clinical factors-based) were built in a training cohort and validated in a test cohort. The discrimination, calibration, and clinical usefulness of the models were evaluated. Finally, asubgroup analysis was performed to assess the predictive value of radiomics score in specific hemorrhage location. Radiomics score was composed of 12 radiomics features. The radiomics model and radiomics nomogram both showed good performance in predicting hematoma enlargement (area under the curve, AUC 0.83 [0.71-0.95], AUC 0.82 [0.72, 0.93]), and were both better than clinical model (AUC 0.66 [0.54-0.79]). The radiomics model and radiomics nomogram showed satisfactory calibration and clinical usefulness for detecting hematoma enlargement. For subgroup analysis, radiomics score also showed good predictive value for hematoma enlargement in different locations (AUC were 0.828, 0.940, 0.836 and 0.904, respectively, for supratentorial, subtentorial, deep and lobes). Aradiomics score based on non-contrast CT may be considered as apotential biomarker for prediction of hematoma enlargement in patients with spontaneous intracerebral hemorrhage (SICH), and it presented ahigh incremental value to clinical factors for hematoma enlargement prediction.

Posttreatment Ischemic Lesion Evolution Is Associated With Reduced Favorable Functional Outcome in Patients With Stroke

Ischemic lesion volume can increase even 24 hours after onset of an acute ischemic stroke. In this study, we investigated the association of lesion evolution with functional outcome and the influence of successful recanalization on this association. We included patients from the MR CLEAN trial (Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands) who received good quality noncontrast CT images 24 hours and 1 week after stroke onset. The ischemic lesion delineations included infarct, edema, and hemorrhagic transformation. Lesion evolution was defined as the difference between the volumes measured on the 1-week and 24-hour noncontrast CTs. The association of lesion evolution with functional outcome was evaluated using unadjusted and adjusted logistic regression. Adjustments were made for baseline, clinical, and imaging parameters that were associated P<0.10) in univariate analysis with favorable functional outcome, defined as modified Rankin Scale score of ≤2. Interaction analysis was performed to evaluate the influence of successful recanalization, defined as modified Arterial Occlusion Lesion score of 3 points, on this association. Of the 226 patients who were included, 69 (31%) patients achieved the favorable functional outcome. Median lesion evolution was 22 (interquartile range, 10–45) mL. Lesion evolution was significantly inversely correlated with favourable functional outcome: unadjusted odds ratio, 0.76 (95% CI, 0.66–0.86; per 10 mL of lesion evolution; P<0.01) and adjusted odds ratio: 0.85 (95% CI, 0.72–0.97; per 10 mL of lesion evolution; P=0.03). There was no significant interaction of successful recanalization on the association of lesion evolution and favorable functional outcome (odds ratio, 1.01 [95% CI, 0.77–1.36]; P=0.94). In our population, subacute ischemic lesion evolution is associated with unfavorable functional outcome. This study suggests that even 24 hours after onset of stroke, deterioration of the brain continues, which has a negative effect on functional outcome. This finding may warrant additional treatment in the subacute phase.

Fully automated waist circumference measurement on abdominal CT: Comparison with manual measurements and potential value for identifying overweight and obesity as an adjunct output of CT scan.

ObjectiveWaist circumference (WC) is a widely accepted anthropometric parameter of central obesity. We investigated a fully automated body segmentation algorithm for measuring WC on abdominal computed tomography (CT) in comparison to manual WC measurements (WC-manual) and evaluated the performance of CT-measured WC for identifying overweight/obesity.Materials and methodsThis retrospective study included consecutive adults who underwent both abdominal CT scans and manual WC measurements at a health check-up between January 2013 and November 2019. Mid-waist WCs were automatically measured on noncontrast axial CT images using a deep learning-based body segmentation algorithm. The associations between CT-measured WC and WC-manual was assessed by Pearson correlation analysis and their agreement was assessed through Bland-Altman analysis. The performance of these WC measurements for identifying overweight/obesity (i.e., body mass index [BMI] ≥25 kg/m2) was evaluated using receiver operating characteristics (ROC) curve analysis.ResultsAmong 763 subjects whose abdominal CT scans were analyzed using a fully automated body segmentation algorithm, CT-measured WCs were successfully obtained in 757 adults (326 women; mean age, 54.3 years; 64 women and 182 men with overweight/obesity). CT-measured WC was strongly correlated with WC-manual (r = 0.919, p < 0.001), and showed a mean difference of 6.1 cm with limits of agreement between -1.8 cm and 14.0 cm in comparison to WC-manual. For identifying overweight/obesity, CT-measured WC showed excellent performance, with areas under the ROC curve (AUCs) of 0.960 (95% CI, 0.933–0.979) in women and 0.909 (95% CI, 0.878–0.935) in men, which were comparable to WC-manual (AUCs of 0.965 [95% CI, 0.938–0.982] and 0.916 [95% CI, 0.886–0.941]; p = 0.735 and 0.437, respectively).ConclusionCT-measured WC using a fully automated body segmentation algorithm was closely correlated with manually-measured WC. While radiation issue may limit its general use, it can serve as an adjunctive output of abdominal CT scans to identify overweight/obesity.

Contrast-enhanced to noncontrast CT transformation via an adjacency content-transfer-based deep subtraction residual neural network

To reduce overall patient radiation exposure in some clinical scenarios (since cancer patients need frequent follow-ups), noncontrast CT is not used in some institutions. However, although less desirable, noncontrast CT could provide additional important information. In this article, we propose a deep subtraction residual network based on adjacency content transfer to reconstruct noncontrast CT from contrast CT and maintain image quality comparable to that of a CT scan originally acquired without contrast. To address the slight structural dissimilarity of the paired CT images (noncontrast CT and contrast CT) due to involuntary physiological motion, we introduce a contrastive loss network derived from the adjacency content-transfer strategy. We evaluate the results of various similarity metrics (MSE, SSIM, NRMSE, PSNR, MAE) and the fitting curve (HU distribution) of the output mapping to estimate the reconstruction performance of the algorithm. To build the model, we randomly select a total of 15,405 CT paired images (noncontrast CT and contrast-enhanced CT) for training and 10,270 CT paired images for testing. The proposed algorithm preserves the robust structures from the contrast-enhanced CT scans and learns the noncontrast attenuation pattern from the noncontrast CT scans. During the evaluation, the deep subtraction residual network achieves higher MSE, MAE, NRMSE, and PSNR scores (by 30%) than those of the baseline models (BEGAN, CycleGAN, Pixel2Pixel) and better simulates the HU curve of noncontrast CT attenuation. After validation based on an analysis of the experimental results, we can report that the noncontrast CT images reconstructed by our proposed algorithm not only preserve the high-quality structures from the contrast-enhanced CT images, but also mimic the CT attenuation of the originally acquired noncontrast CT images.

Understanding the Clinical Implications of Intracranial Arterial Calcification Using Brain CT and Vessel Wall Imaging.

Background and Purpose: Intracranial arterial calcification (IAC) has been the focus of much attention by clinicians and researchers as an indicator of intracranial atherosclerosis, but correlations of IAC patterns (intimal or medial) with the presence of atherosclerotic plaques and plaque stability are still a matter of debate. Our study aimed to assess the associations of IAC patterns identified on computed tomography (CT) with the presence of plaque detected on vessel wall magnetic resonance imaging and plaque stability.Materials and Methods: Patients with stroke or transient ischemic attack and intracranial artery stenosis were recruited. IAC was detected and localized (intima or media) on non-contrast CT images. Intracranial atherosclerotic plaques were identified using vessel wall magnetic resonance imaging and matched to corresponding CT images. Associations between IAC patterns and culprit atherosclerotic plaques were assessed by using multivariate regression.Results: Seventy-five patients (mean age, 63.4 ± 11.6 years; males, 46) were included. Two hundred and twenty-one segments with IAC were identified on CT in 66 patients, including 86 (38.9%) predominantly intimal calcifications and 135 (61.1%) predominantly medial calcifications. A total of 72.0% of intimal calcifications coexisted with atherosclerotic plaques, whereas only 10.2% of medial calcifications coexisted with plaques. Intimal calcification was more commonly shown in non-culprit plaques than culprit plaques (25.9 vs. 9.4%, P = 0.008). The multivariate mixed logistic regression adjusted for the degree of stenosis showed that intimal calcification was significantly associated with non-culprit plaques (OR, 2.971; 95% CI, 1.036–8.517; P = 0.043).Conclusion: Our findings suggest that intimal calcification may indicate the existence of a stable form of atherosclerotic plaque, but plaques can exist in the absence of intimal calcification especially in the middle cerebral artery.

Predictive Radiomic Models for the Chemotherapy Response in Non-Small-Cell Lung Cancer based on Computerized-Tomography Images.

The heterogeneity and complexity of non-small cell lung cancer (NSCLC) tumors mean that NSCLC patients at the same stage can have different chemotherapy prognoses. Accurate predictive models could recognize NSCLC patients likely to respond to chemotherapy so that they can be given personalized and effective treatment. We propose to identify predictive imaging biomarkers from pre-treatment CT images and construct a radiomic model that can predict the chemotherapy response in NSCLC. This single-center cohort study included 280 NSCLC patients who received first-line chemotherapy treatment. Non-contrast CT images were taken before and after the chemotherapy, and clinical information were collected. Based on the Response Evaluation Criteria in Solid Tumors and clinical criteria, the responses were classified into two categories: response (n = 145) and progression (n = 135), then all data were divided into two cohorts: training cohort (224 patients) and independent test cohort (56 patients). In total, 1629 features characterizing the tumor phenotype were extracted from a cube containing the tumor lesion cropped from the pre-chemotherapy CT images. After dimensionality reduction, predictive models of the chemotherapy response of NSCLC with different feature selection methods and different machine-learning classifiers (support vector machine, random forest, and logistic regression) were constructed. For the independent test cohort, the predictive model based on a random-forest classifier with 20 radiomic features achieved the best performance, with an accuracy of 85.7% and an area under the receiver operating characteristic curve of 0.941 (95% confidence interval, 0.898–0.982). Of the 20 selected features, four were first-order statistics of image intensity and the others were texture features. For nine features, there were significant differences between the response and progression groups (p < 0.001). In the response group, three features, indicating heterogeneity, were overrepresented and one feature indicating homogeneity was underrepresented. The proposed radiomic model with pre-chemotherapy CT features can predict the chemotherapy response of patients with non-small cell lung cancer. This radiomic model can help to stratify patients with NSCLC, thereby offering the prospect of better treatment.

Total and regional ASPECT score for non-contrast CT, CT angiography, and CT perfusion: inter-rater agreement and its association with the final infarction in acute ischemic stroke patients.

Alberta Stroke Program Early Computed Tomography Score (ASPECTS) is a grading system to assess the extent and distribution of early ischemic changes. To assess inter-rater agreement for total and regional ASPECTS on non-contrast computed tomography (NCCT) images, CT angiography source images (CTA-SI), and CT-perfusion cerebral blood volume (CTP-CBV) maps, and their association with final infarction in patients with acute ischemic stroke (AIS). A total of 96 consecutive patients with AIS who underwent pre-treatment NCCT and CTP were retrospectively enrolled. CTA-SI was reconstructed using the raw data of CTP. Total and regional ASPECTS were assessed on baseline NCCT, CTA-SI, and CTP-CBV, and on follow-up NCCT or diffusion-weighted imaging. Follow-up ASPECTS served as the reference standard for final infarction. CTP-CBV demonstrated higher concordance for total ASPECTS (interclass correlation coefficient, 0.895 vs. 0.771 vs. 0.777) and regional ASPECTS in internal capsule, lentiform, caudate nuclei, M5 and M6, compared with NCCT and CTA-SI. CTP-CBV showed a trend of stronger correlation with final ASPECTS than NCCT and CTA-SI (0.717 vs. 0.711 vs. 0.565; P > 0.05). ASPECTS in the internal capsule (ρ, 0.756 vs. 0.556; P = 0.016) and caudate nucleus (ρ, 0.717 vs. 0.476; P = 0.010) on CTP-CBV were more strongly correlated with follow-up ASPECTS than NCCT. CTP-CBV showed higher accuracy for predicting final infarction in the internal capsule (92.5% vs. 90.3% and 87.1%; P > 1.000, P = 0.125, respectively) and caudate nucleus (87.1% vs. 79.6% and 77.4%; P = 0.453, P = 0.039, respectively) than CTA-SI and NCCT. CTP-CBV ASPECTS might be more reliable for delineating early ischemic changes and predicting final infarction.

A prediction of hematoma expansion in hemorrhagic patients using a novel dual-modal machine learning strategy

Objective. Hematoma expansion is closely associated with adverse functional outcomes in patients with intracerebral hemorrhage (ICH). Prediction of hematoma expansion would therefore be of great clinical significance. We therefore attempted to predict hematoma expansion using a dual-modal machine learning (ML) strategy which combines information from non-contrast computed tomography (NCCT) images and multiple clinical variables. Approach. We retrospectively identified 140 ICH patients (57 with hematoma expansion) with 5616 NCCT images of hematoma (2635 with hematoma expansion) and 10 clinical variables. The dual-modal ML strategy consists of two steps. The first step is to derive a mono-modal predictor from a deep convolutional neural network using solely NCCT images. The second step is to achieve a dual-modal predictor by combining the mono-modal predictor with 10 clinical variables to predict hematoma growth using a multi-layer perception network. Main results. For the mono-modal predictor, the best performance was merely 69.5% in accuracy with solely the NCCT images, whereas the dual-modal predictor could boost the accuracy greatly to be 86.5% by combining clinical variables. Significance. To our knowledge, this is the best performance from using ML to predict hematoma expansion. It could be potentially useful as a screening tool for high-risk patients with ICH, though further clinical tests would be necessary to show its performance on a larger cohort of patients.

Detection of Perfusion Deficits in Multiphase Computed Tomography Angiography-A Stroke Imaging Technique Based on Iodine Mapping on Spectral Computed Tomography: Initial Findings.

The purpose of this study was to explore a novel method for brain tissue differentiation using quantitative analysis of multiphase computed tomography (CT) angiography (MP-CTA) on spectral CT, to assess whether it can distinguish underperfused from normal tissue, using CT perfusion (CTP) as reference. Noncontrast CT and MP-CTA images from 10 patients were analyzed in vascular regions through measurements of Hounsfield unit (HU) at 120 kV, HU at 40 keV, and iodine density. Regions were categorized as normal or ischemic according to CTP. Hounsfield unit and iodine density were compared regarding ability to separate normal and ischemic tissue, the difference in maximum time derivative of the right over left hemisphere ratio. Iodine density had the highest maximum time derivatives and generated the largest mean separation between normal and ischemic tissue. The method can be used to categorize tissue as normal or underperfused. Using iodine quantification seems to give a more distinct differentiation of perfusion defects compared with conventional HU.

Comparison of Radiomic Models Based on Different Machine Learning Methods for Predicting Intracerebral Hemorrhage Expansion.

The objective of this study was to predict hematoma expansion (HE) by radiomic models based on different machine learning methods and determine the best radiomic model through the comparison. A total of 108 patients with intracerebral hemorrhage were retrospectively evaluated. Images of baseline non-contrast computed tomography (NCCT) and follow-up NCCT scan within 24 h were retrospectively reviewed. An HE was defined as avolume increase of more than 33% or an increase greater than 12.5 mL from the volume of the baseline NCCT. Texture parameters of the baseline NCCT images were selected by the least absolute shrinkage and selection operator (LASSO) regression. We used support vector machine (SVM), decision tree (DT), conditional inference trees (CIT), random forest (RF), k‑nearest neighbors (KNN), back-propagation neural network (BPNet) and Bayes to build models. Receiver operating characteristic (ROC) analysis and decision curve analysis (DCA) was performed and compared among models. Every model had arelatively high AUC (all > 0.75), SVM and KNN had the highest AUC of 0.91. There were significant differences between SVM and CIT (Z > 2.266, p = 0.02345), KNN and CIT (Z = 2.4834, p = 0.01301), RF and CIT (Z = 2.6956, p = 0.007027), KNN and BPNet (Z = 2.0122, p = 0.0442), RF and BPNet (Z = 1.9793, p = 0.04778). There was no significant difference among SVM, DT, RF, KNN and Bayes (p> 0.05). The SVM obtained the largest net benefit when the threshold probability was less than 0.33, while KNN obtained the largest net benefit when the threshold probability was greater than 0.33. Combined with ROC and DCA, SVM and KNN performed better in all the models for predicting HE. Radiomic models based on different machine learning methods can be used to predict HE and the models generated by SVM and KNN performed best.

Prognostic Value of Coronary Artery Calcification Identified by the Semi-quantitative Weston Method in the Emergency Room or Other Hospitalized Patients.

Background: Coronary artery calcification (CAC) may provide insight to the patients' coronary artery disease (CAD) risks and influence early intervention. With increasing use of non-gated CT scans in clinical practice, the visual coronary artery scoring system (Weston Method) could quickly provide clinicians with important information of CAC for patient triage and management.Methods: We retrospectively studied the available CT imaging data and estimated CAC burden using the Weston method in 493 emergency room or other hospitalized patients. The Weston scores were calculated by the sum of the score for each vessel including the left main, left anterior descending, left circumflex artery and right coronary artery (range 0–12). The primary endpoint was a composite of the major adverse cardiac events (MACEs), including cardiac death, myocardial infarction, stroke, and coronary revascularization.Results: During a median follow-up of 85 months, a total of 25 (5.1%) MACE were recorded and 57 (11.2%) patients died from any causes. Detectable CAC was most common (96%) in the left anterior descending coronary arteries. Multivariable analysis showed that CAC total scores were independent predictors for MACE and all-cause mortality. Receiver operating characteristic analysis showed that CAC total score ≥5 was the optimal cutoff value for predicting MACEs.Conclusions: In the emergency room and hospitalized patients, the semi-quantitation of CAC burden using the Weston score system was related to the long-term cardiovascular outcomes including mortality. Clinicians and radiologists should maximize the value of non-contrast chest CT images by reporting CAC details.

The diagnostic value of a non-contrast computed tomography scan-based radiomics model for acute aortic dissection.

To investigate the diagnostic value of a computed tomography (CT) scan-based radiomics model for acute aortic dissection.For the dissection group, we retrospectively selected 50 patients clinically diagnosed with acute aortic dissection between October 2018 and November 2019, for whom non-contrast CT and CT angiography images were available. Fifty individuals with available non-contrast CT and CT angiography images for other causes were selected for inclusion in the non-dissection group. Based on the aortic dissection locations on the CT angiography images, we marked the corresponding regions-of-interest on the non-contrast CT images of both groups. We collected 1203 characteristic parameters from these regions by extracting radiomics features. Subsequently, we used a random number table to include 70 individuals in the training group and 30 in the validation group. Finally, we used the Lasso regression for dimension reduction and predictive model construction. The diagnostic performance of the model was evaluated by a receiver operating characteristic (ROC) curve.Fourteen characteristic parameters with non-zero coefficients were selected after dimension reduction. The accuracy, sensitivity, specificity, and area under the ROC curve of the prediction model for the training group were 94.3% (66/70), 91.2% (31/34), 97.2% (35/36), and 0.988 (95% confidence interval [CI]: 0.970–0.998), respectively. The respective values for the validation group were 90.0% (27/30), 94.1% (16/17), 84.6% (11/13), and 0.952 (95% CI: 0.883–0.986).Our non-contrast CT scan-based radiomics model accurately facilitated acute aortic dissection diagnosis.

Appropriate iMAR presets for metal artifact reduction from surgical clips and titanium burr hole covers on postoperative non-contrast brain CT

PurposeTo evaluate suitable iterative metal artifact reduction (iMAR) presets for titanium neurosurgical clips and burr hole covers (BHCs) on postoperative non-contrast computed tomography (NCCT). MethodTwenty-two patients who underwent NCCT after intracranial aneurysmal clipping were included. NCCT images were postprocessed using eight currently available iMAR presets. In each image, a circular region of interest (ROI) was placed around clip, BHC, and on parietal lobe as reference. Standard deviation (SD) and attenuation value (HU) were measured in each ROI to obtain artifact index (AI) and contrast-to-noise ratio (CNR). For each iMAR preset, SD, AI, HU, and CNR were compared with those without iMAR for clips and BHCs. Visual assessment around each clip and BHC was performed by two neuroradiologists using three-point visual score (VS) (1 = no apparent, 2 = minor, and 3 = severe artifacts). ResultsAmong the presets, the neuro-coils preset (iMAR-NC) showed the lowest SD, AI, and VS for clips (P < 0.001). For BHCs, HU, CNR, and VS with iMAR-NC were significantly higher than those without iMAR (P < 0.001). SD, AI, and VS with the shoulder implants preset (iMAR-ShI) were significantly lower than those without iMAR for clips (P = 0.002, 0.002, and P < 0.001, respectively). For BHCs, VS with iMAR-ShI was lowest among the presets (P = 0.004). ConclusionsAlthough iMAR-NC reduces metal artifacts from clips, it strengthens artifacts from BHCs. For postoperative NCCT, iMAR-ShI most effectively reduces metal artifacts from both clips and BHCs in a single preset.

Ensembled Deep Neural Network for Intracranial Hemorrhage Detection and Subtype Classification on Noncontrast CT Images

Rapid and accurate diagnosis of intracranial hemorrhage is clinically significant to ensure timely treatment. In this study, we developed an ensembled deep neural network for the detection and subtype classification of intracranial hemorrhage. The model consisted of two parallel network pathways, one using three different window level/width settings to enhance the image contrast of brain, blood, and soft tissue. The other extracted spatial information of adjacent image slices to the target slice. Both pathways exploited the EfficientNet-B0 as the basic architecture and were ensembled to generate the final prediction. Class activation mapping was applied in both pathways to highlight the regions of detected hemorrhage and the associated subtypes. The model was trained and tested using Intracranial Hemorrhage Detection Challenge (IHDC) dataset launched by the Radiological Society of North America (RSNA) in 2019, which contained 674,258 head noncontrasts computer tomography images acquired from 19,530 patients. An independent dataset (CQ500) acquired from another institution was used to test the generalizability of the trained model. The overall accuracy, sensitivity, and F1 score for intracranial hemorrhage detection were 95.7%, 85.9%, and 86.7% on IHDC testing dataset and 92.4%, 92.6%, and 93.4% on external CQ500 testing dataset. The heatmaps by class activation mapping successfully demonstrated discriminative feature regions of the predicted hemorrhage locations and subtypes, providing visual guidance for radiologists to assist in rapid diagnosis of intracranial hemorrhage.

Artificial Intelligence Can Effectively Predict Early Hematoma Expansion of Intracerebral Hemorrhage Analyzing Noncontrast Computed Tomography Image.

This study aims to develop and validate an artificial intelligence model based on deep learning to predict early hematoma enlargement (HE) in patients with intracerebral hemorrhage. A total of 1,899 noncontrast computed tomography (NCCT) images of cerebral hemorrhage patients were retrospectively analyzed to establish a predicting model and 1,117 to validate the model. And a total of 118 patients with intracerebral hemorrhage were selected based on inclusion and exclusion criteria so as to validate the value of the model for clinical prediction. The baseline noncontrast computed tomography images within 6 h of intracerebral hemorrhage onset and the second noncontrast computed tomography performed at 24 ± 3 h from the onset were used to evaluate the prediction of intracerebral hemorrhage growth. In validation dataset 1, the AUC was 0.778 (95% CI, 0.768–0.786), the sensitivity was 0.818 (95% CI, 0.790–0.843), and the specificity was 0.601 (95% CI, 0.565–0.632). In validation dataset 2, the AUC was 0.780 (95% CI, 0.761–0.798), the sensitivity was 0.732 (95% CI, 0.682–0.788), and the specificity was 0.709 (95% CI, 0.658–0.759). The sensitivity of intracerebral hemorrhage hematoma expansion as predicted by an artificial intelligence imaging system was 89.3%, with a specificity of 77.8%, a positive predictive value of 55.6%, a negative predictive value of 95.9%, and a Yoden index of 0.671, which were much higher than those based on the manually labeled noncontrast computed tomography signs. Compared with the existing prediction methods through computed tomographic angiography (CTA) image features and noncontrast computed tomography image features analysis, the artificial intelligence model has higher specificity and sensitivity in the prediction of early hematoma enlargement in patients with intracerebral hemorrhage.

Automatic pulmonary vessel segmentation on noncontrast chest CT: deep learning algorithm developed using spatiotemporally matched virtual noncontrast images and low-keV contrast-enhanced vessel maps.

ObjectivesTo develop a deep learning–based pulmonary vessel segmentation algorithm (DLVS) from noncontrast chest CT and to investigate its clinical implications in assessing vascular remodeling of chronic obstructive lung disease (COPD) patients.MethodsFor development, 104 pulmonary CT angiography scans (49,054 slices) using a dual-source CT were collected, and spatiotemporally matched virtual noncontrast and 50-keV images were generated. Vessel maps were extracted from the 50-keV images. The 3-dimensional U-Net-based DLVS was trained to segment pulmonary vessels (with a vessel map as the output) from virtual noncontrast images (as the input). For external validation, vendor-independent noncontrast CT images (n = 14) and the VESSEL 12 challenge open dataset (n = 3) were used. For each case, 200 points were selected including 20 intra-lesional points, and the probability value for each point was extracted. For clinical validation, we included 281 COPD patients with low-dose noncontrast CTs. The DLVS-calculated volume of vessels with a cross-sectional area < 5 mm2 (PVV5) and the PVV5 divided by total vessel volume (%PVV5) were measured.ResultsDLVS correctly segmented 99.1% of the intravascular points (1,387/1,400) and 93.1% of the extravascular points (1,309/1,400). The areas-under-the receiver-operating characteristic curve (AUROCs) were 0.977 and 0.969 for the two external validation datasets. For the COPD patients, both PPV5 and %PPV5 successfully differentiated severe patients whose FEV1 < 50 (AUROCs; 0.715 and 0.804) and were significantly correlated with the emphysema index (Ps < .05).ConclusionsDLVS successfully segmented pulmonary vessels on noncontrast chest CT by utilizing spatiotemporally matched 50-keV images from a dual-source CT scanner and showed promising clinical applicability in COPD.Key Points • We developed a deep learning pulmonary vessel segmentation algorithm using virtual noncontrast images and 50-keV enhanced images produced by a dual-source CT scanner. • Our algorithm successfully segmented vessels on diseased lungs. • Our algorithm showed promising results in assessing the loss of small vessel density in COPD patients. Supplementary InformationThe online version contains supplementary material available at 10.1007/s00330-021-08036-z.

Optimizing Deep Learning Algorithms for Segmentation of Acute Infarcts on Non-Contrast Material-enhanced CT Scans of the Brain Using Simulated Lesions.

To test the efficacy of lesion segmentation using a deep learning algorithm on non-contrast material-enhanced CT (NCCT) images with synthetic lesions resembling acute infarcts. In this retrospective study, 40 diffusion-weighted imaging (DWI) lesions in patients with acute stroke (median age, 69 years; range, 62-76 years; 17 women; screened between 2011 and 2017) were coregistered to 40 normal NCCT scans (median age, 70 years; range, 55-76 years; 25 women; screened between 2008 and 2011), which produced 640 combinations of DWI-NCCT with and without lesions for training (n = 420), validation (n = 110), and testing (n = 110). The signal intensity on the NCCT scans was depressed by 4 HU (a 13% drop) in the region of the diffusion-weighted lesion. Two U-Net architectures (standard and symmetry aware) were trained with two different training strategies. One was a naive strategy, in which the model started training with random coefficients. The other was a progressive strategy, which started with coefficients derived from a model trained on a dataset with lesions that were depressed by 10 HU. The Dice scores from the two architectures and training strategies were compared from the test dataset. Dice scores of symmetry-aware U-Nets were 25% higher than those of standard U-Nets (median, 0.49 vs 0.65; P < .001). Use of a progressive training strategy had no clear effect on model performance. Symmetry-aware U-Nets offer promise for segmentation of acute stroke lesions on NCCT scans.Keywords: Adults, CT-Quantitative, StrokeSupplemental material is available for this article.© RSNA, 2021.