Non-enhanced CT Scan Research Articles

Impact of a deep learning-based brain CT interpretation algorithm on clinical decision-making for intracranial hemorrhage in the emergency department

Intracranial hemorrhage is a critical emergency that requires prompt and accurate diagnosis in the emergency department (ED). Deep learning technology can assist in interpreting non-enhanced brain CT scans, but its real-world impact on clinical decision-making is uncertain. This study assessed a deep learning-based intracranial hemorrhage detection algorithm (DLHD) in a simulated clinical environment with ten emergency medical professionals from a tertiary hospital’s ED. The participants reviewed CT scans with clinical information in two steps: without and with DLHD. Diagnostic performance was measured, including sensitivity, specificity, accuracy, and the area under the receiver operating characteristic curve. Consistency in clinical decision-making was evaluated using the kappa statistic. The results demonstrated that DLHD minimally affected experienced participants’ diagnostic performance and decision-making. In contrast, inexperienced participants exhibited significantly increased sensitivity (59.33–72.67%, p < 0.001) and decreased specificity (65.49–53.73%, p < 0.001) with the algorithm. Clinical decision-making consistency was moderate among inexperienced professionals (k = 0.425) and higher among experienced ones (k = 0.738). Inexperienced participants changed their decisions more frequently, mainly due to the algorithm’s false positives. The study highlights the need for thorough evaluation and careful integration of deep learning tools into clinical workflows, especially for less experienced professionals.

The change of epicardial adipose tissue characteristics and vulnerability for atrial fibrillation upon drastic weight loss

ABSTRACT Background Obesity increases the risk of atrial fibrillation (AF). We hypothesize that ‘obese’ epicardial adipose tissue (EAT) is, regardless of comorbidities, associated with markers of AF vulnerability Methods Patients >40y of age undergoing bariatric surgery and using <2 antihypertensive drugs and no insulin were prospectively included. Study investigations were conducted before and 1y after surgery. Heart rhythm and p-wave duration were measured through ECGs and 7-d-holters. EAT-volume and attenuation were determined on non-enhanced CT scans. Serum markers were quantified by ELISA. Results Thirty-seven patients underwent surgery (age: 52.1 ± 5.9y; 27 women; no AF). Increased p-wave duration correlated with higher BMI, larger EAT volumes, and lower EAT attenuations (p < 0.05). Post-surgery, p-wave duration decreased from 109 ± 11 to 102 ± 11ms. Concurrently, EAT volume decreased from 132 ± 49 to 87 ± 52ml, BMI from 43.2 ± 5.2 to 28.9 ± 4.6kg/m2, and EAT attenuation increased from –76.1 ± 4.0 to −71.7 ± 4.4HU (p <0.001). Adiponectin increased from 8.7 ± 0.8 to 14.2 ± 1.0 μg/ml (p <0.001). However, decreased p-wave durations were not related to changed EAT characteristics, BMI or adiponectin. Conclusion In this explorative study, longer p-wave durations related to higher BMIs, larger EAT volume, and lower EAT attenuations. P-wave duration and EAT volume decreased, and EAT attenuation increased upon drastic weightloss. However, there was no relation between decreased p-wave duration and changed BMI or EAT characteristics.

Deep-learning-based method for the segmentation of ureter and renal pelvis on non-enhanced CT scans

This study aimed to develop a deep-learning (DL) based method for three-dimensional (3D) segmentation of the upper urinary tract (UUT), including ureter and renal pelvis, on non-enhanced computed tomography (NECT) scans. A total of 150 NECT scans with normal appearance of the left UUT were chosen for this study. The dataset was divided into training (n = 130) and validation sets (n = 20). The test set contained 29 randomly chosen cases with computed tomography urography (CTU) and NECT scans, all with normal appearance of the left UUT. An experienced radiologist marked out the left renal pelvis and ureter on each scan. Two types of frameworks (entire and sectional) with three types of DL models (basic UNet, UNet3 + and ViT-UNet) were developed, and evaluated. The sectional framework with basic UNet model achieved the highest mean precision (85.5%) and mean recall (71.9%) on the test set compared to all other tested methods. Compared with CTU scans, this method had higher axial UUT recall than CTU (82.5% vs 69.1%, P < 0.01). This method achieved similar or better visualization of UUT than CTU in many cases, however, in some cases, it exhibited a non-ignorable false-positive rate. The proposed DL method demonstrates promising potential in automated 3D UUT segmentation on NECT scans. The proposed DL models could remarkably improve the efficiency of UUT reconstruction, and have the potential to save many patients from invasive examinations such as CTU. DL models could also serve as a valuable complement to CTU.

A 3D Convolutional Neural Network Based on Non-enhanced Brain CT to Identify Patients with Brain Metastases.

Dedicated brain imaging for cancer patients is seldom recommended in the absence of symptoms. There is increasing availability of non-enhanced CT (NE-CT) of the brain, mainly owing to a wider utilization of Positron Emission Tomography-CT (PET-CT) in cancer staging. Brain metastases (BM) are often hard to diagnose on NE-CT. This work aims to develop a 3D Convolutional Neural Network (3D-CNN) based on brain NE-CT to distinguish patients with and without BM. We retrospectively included NE-CT scans for 100 patients with single or multiple BM and 100 patients without brain imaging abnormalities. Patients whose largest lesion was < 5mm were excluded. The largest tumor was manually segmented on a matched contrast-enhanced T1 weighted Magnetic Resonance Imaging (MRI), and shape radiomics were extracted to determine the size and volume of the lesion. The brain was automatically segmented, and masked images were normalized and resampled. The dataset was split into training (70%) and validation (30%) sets. Multiple versions of a 3D-CNN were developed, and the best model was selected based on accuracy (ACC) on the validation set. The median largest tumor Maximum-3D-Diameter was 2.29cm, and its median volume was 2.81cc. Solitary BM were found in 27% of the patients, while 49% had > 5 BMs. The best model consisted of 4 convolutional layers with 3D average pooling layers, dropout layers of 50%, and a sigmoid activation function. Mean validation ACC was 0.983 (SD: 0.020) and mean area under receiver-operating characteristic curve was 0.983 (SD: 0.023). Sensitivity was 0.983 (SD: 0.020). We developed an accurate 3D-CNN based on brain NE-CT to differentiate between patients with and without BM. The model merits further external validation.

A non-enhanced CT-based deep learning diagnostic system for COVID-19 infection at high risk among lung cancer patients.

Pneumonia and lung cancer have a mutually reinforcing relationship. Lung cancer patients are prone to contracting COVID-19, with poorer prognoses. Additionally, COVID-19 infection can impact anticancer treatments for lung cancer patients. Developing an early diagnostic system for COVID-19 pneumonia can help improve the prognosis of lung cancer patients with COVID-19 infection. This study proposes a neural network for COVID-19 diagnosis based on non-enhanced CT scans, consisting of two 3D convolutional neural networks (CNN) connected in series to form two diagnostic modules. The first diagnostic module classifies COVID-19 pneumonia patients from other pneumonia patients, while the second diagnostic module distinguishes severe COVID-19 patients from ordinary COVID-19 patients. We also analyzed the correlation between the deep learning features of the two diagnostic modules and various laboratory parameters, including KL-6. The first diagnostic module achieved an accuracy of 0.9669 on the training set and 0.8884 on the test set, while the second diagnostic module achieved an accuracy of 0.9722 on the training set and 0.9184 on the test set. Strong correlation was observed between the deep learning parameters of the second diagnostic module and KL-6. Our neural network can differentiate between COVID-19 pneumonia and other pneumonias on CT images, while also distinguishing between ordinary COVID-19 patients and those with white lung. Patients with white lung in COVID-19 have greater alveolar damage compared to ordinary COVID-19 patients, and our deep learning features can serve as an imaging biomarker.

Clinical feasibility of deep learning based synthetic contrast enhanced abdominal CT in patients undergoing non enhanced CT scans

Our objective was to develop and evaluate the clinical feasibility of deep-learning-based synthetic contrast-enhanced computed tomography (DL-SynCCT) in patients designated for nonenhanced CT (NECT). We proposed a weakly supervised learning with the utilization of virtual non-contrast CT (VNC) for the development of DL-SynCCT. Training and internal validations were performed with 2202 pairs of retrospectively collected contrast-enhanced CT (CECT) images with the corresponding VNC images acquired from dual-energy CT. Clinical validation was performed using an external validation set including 398 patients designated for true nonenhanced CT (NECT), from multiple vendors at three institutes. Detection of lesions was performed by three radiologists with only NECT in the first session and an additionally provided DL-SynCCT in the second session. The mean peak signal-to-noise ratio (PSNR) and structural similarity index map (SSIM) of the DL-SynCCT compared to CECT were 43.25 ± 0.41 and 0.92 ± 0.01, respectively. With DL-SynCCT, the pooled sensitivity for lesion detection (72.0% to 76.4%, P < 0.001) and level of diagnostic confidence (3.0 to 3.6, P < 0.001) significantly increased. In conclusion, DL-SynCCT generated by weakly supervised learning showed significant benefit in terms of sensitivity in detecting abnormal findings when added to NECT in patients designated for nonenhanced CT scans.

Flexible ureteroscopy in Chinese air force pilots with renal calculi: a cohort study at a single medical centre

IntroductionRenal calculi are the predominant urological ailment in air force pilots. Flexible ureteroscopy (FURS) constitutes a valuable approach for renal calculi treatment. This study presents a decade-long exploration of using...

Building nonenhanced CT based radiomics model in discriminating arteriovenous malformation related hematomas from hypertensive intracerebral hematomas

ObjectiveTo develop and validate radiomics models on non-enhanced CT for discrimination of arteriovenous malformation (AVM) related hematomas from hypertensive intracerebral hematomas.Materials and methodsA total of 571 patients with acute intraparenchymal hematomas and baseline non-enhanced CT scans were retrospectively analyzed, including 297 cases of AVM related hematomas and 274 cases of hypertensive intracerebral hematomas. The patients were divided into training and validation cohorts in a 7:3 ratio with a random seed. A total of 1,688 radiomics features of hematomas were extracted from non-enhanced CT. Then, the least absolute shrinkage and selection operator (LASSO) regression was applied to select features and construct the radiomics models. In this study, a radiomics-based model was constructed that based on the radiomics features only. Furthermore, a combined model was constructed using radiomics features, clinical characteristics and radiological signs by radiologists’ evaluation. In addition, we compared predictive performance of the two models for discrimination of AVM related hematomas from hypertensive intracerebral hematomas.ResultsA total of 67 radiomics features were selected to establish radiomics signature via LASSO regression. The radiomics-based model was constructed with 2 classifiers, support vector machine (SVM) and logistic regression (LR). AUCs of the radiomics-based model in the training set were 0.894 and 0.904, in validation set were 0.774 and 0.782 in SVM classifier and LR classifier, respectively. AUCs of the combined model (combined with radiomics, age and calcification) in the training set were 0.976 and 0.981, in validation set were 0.896 and 0.907 in SVM classifier and LR classifier, respectively. The combined model showed greater AUCs than radiomics-based model in both training set and validation set.ConclusionThe combined model using radiomics, age and calcification showed a satisfactory predictive performance for discrimination of AVM related hematomas from hypertensive intracerebral hematomas and hold great potential for personalized clinical decision.

Longitudinal follow-up of incidental renal calculi on computed tomography.

Measuring small kidney stones on CT is a time-consuming task often neglected. Volumetric assessment provides a better measure of size than linear dimensions. Our objective is to analyze the growth rate and prognosis of incidental kidney stones in asymptomatic patients on CT. This retrospective study included 4266 scans from 2030 asymptomatic patients who underwent two or more nonenhanced CT scans for colorectal screening between 2004 and 2016. The DL software identified and measured the volume, location, and attenuation of 883 stones. The corresponding scans were manually evaluated, and patients without follow-up were excluded. At each follow-up, the stones were categorized as new, growing, persistent, or resolved. Stone size (volume and diameter), attenuation, and location were correlated with the outcome and growth rates of the stones. The stone cohort comprised 407 scans from 189 (M: 124, F: 65, median age: 55.4years) patients. The median number of stones per scan was 1 (IQR: [1, 2]). The median stone volume was 17.1mm3 (IQR: [7.4, 43.6]) and the median peak attenuation was 308 HU (IQR: [204, 532]. The 189 initial scans contained 291stones; 91 (31.3%) resolved, 142 (48.8%) grew, and 58 (19.9) remained persistent at the first follow-up. At the second follow-up (for 27 patients with 2 follow-ups), 14/44 (31.8%) stones had resolved, 19/44 (43.2%) grew and 11/44 (25%) were persistent. The median growth rate of growing stones was 3.3mm3/year, IQR: [1.4,7.4]. Size and attenuation had a moderate correlation (Spearman rho 0.53, P < .001 for volume, and 0.50 P < .001 for peak attenuation) with the growth rate. Growing and persistent stones had significantly greater maximum axial diameter (2.7 vs 2.3mm, P =.047) and peak attenuation (300 vs 258 HU, P =.031) CONCLUSION: We report a 12.7% prevalence of incidental kidney stones in asymptomatic adults, of which about half grew during follow-up with a median growth rate of about 3.3mm3/year.

Atherosclerosis of the iliac arteries for the prediction of erectile dysfunction and epistaxis in men undergoing abdominal CT scan

BackgroundTo investigate the association between erectile dysfunction (ED) as well as epistaxis (ES) in relation to the extent of iliac atherosclerosis.MethodsIn this retrospective cross-sectional study, all consecutive male patients treated at our institution from 01/2016 to 12/2020 undergoing abdominal CT scan were evaluated. Patients (n = 1272) were invited by mail to participate in the study in returning two questionnaires for the evaluation of ED (IIEF-5) and ES. Patients who returned filled-in questionnaires within a 3-month deadline were included in the study. The extent of atherosclerosis in the common iliac artery (CIA) and the internal iliac artery (IIA) was assessed by calcium scoring on unenhanced CT. Stratification of results was performed according to reported IIEF-5 scores and consequential ED groups.ResultsIn total, 437 patients (34.4% of contacted) met the inclusion criteria. Forty-two patients did not fulfill predefined age requirements (< 75 years) and 120 patients had to be excluded as calcium scoring on nonenhanced CT was not feasible. Finally, 275 patients were included in the analysis and stratified into groups of “no-mild” (n = 146) and “moderate-severe” (n = 129) ED. The calcium score (r=-0.28, p < 0.001) and the number of atherosclerotic lesions (r=-0.32, p < 0.001) in the CIA + IIA showed a significant negative correlation to the IIEF-5 score, respectively. Patients differed significantly in CIA + IIA calcium score (difference: 167.4, p < 0.001) and number of atherosclerotic lesions (difference: 5.00, p < 0.001) when belonging to the “no-mild” vs. “moderate-severe” ED group, respectively. A multivariable regression model, after adjusting for relevant baseline characteristics, showed that the number of atherosclerotic CIA + IIA lesions was an independent predictor of ED (OR = 1.05, p = 0.036), whereas CIA + IIA calcium score was not (OR = 1.00031, p = 0.20). No relevant correlation was found between ES episodes and IIEF-5 scores (r=-0.069, p = 0.25), CIA + IIA calcium score (r=-0.10, p = 0.87) or number of atherosclerotic CIA + IIA lesions (r=-0.032, p = 0.60), respectively.ConclusionsThe number of atherosclerotic lesions in the iliac arteries on nonenhanced abdominal CT scans is associated with the severity of ED. This may be used to identify subclinical cardiovascular disease and to quantify the risk for cardiovascular hazards in the future.Trial registrationBASEC-Nr. 2020 − 01637.

A novel fully automated method for measuring ASPECTS to improve stroke diagnosis: Comparison to traditional ASPECTS.

To compare the accuracy of subjective Alberta Stroke Program Early CT Score (sASPECTS) evaluation and that of an automated prototype software (aASPECTS) on nonenhanced CT (NECT) in patients with early anterior territory stroke and controls using side-to-side quantification of hypoattenuated brain areas. We retrospectively analyzed the NECT scans of 42 consecutive patients with ischemic stroke before reperfusion and 42 controls using first sASPECTS and subsequently aASPECTS. We assessed the differences in Alberta Stroke Program Early CT Score (ASPECTS) and calculated the sensitivity and specificity of NECT with CT perfusion, whereas cerebral blood volume (CBV) served as the reference standard for brain infarction. The clot was located in the middle cerebral artery (MCA) in 47.6% of cases and the internal carotid artery (ICA) in 28.6% of cases. Ten cases presented combined ICA and MCA occlusions. The stroke was right sided in 52.4% of cases and left sided in 47.6%. Reader-based NECT analysis yielded a median sASPECTS of 10. The median CBV-based ASPECTS was 7. Compared to the area of decreased CBV, sASPECTS yielded a sensitivity of 12.5% and specificity of 86.8%. The software prototype (aASPECTS) yielded an overall sensitivity of 65.5% and a specificity of 92.2%. The interreader agreement for ASPECTS evaluation of admission NECT and follow-up CT was almost perfect (κ=.93). The interreader agreement of the CBV color map evaluation was substantial (κ=.77). aASPECTS of NECT can outperform sASPECTS for stroke detection.

Textural Analysis of the Hyperdense Artery Sign in Patients with Acute Ischemic Stroke Predicts the Outcome of Thrombectomy.

Textural analysis is pivotal in augmenting the diagnosis and outcomes of endovascular procedures for stroke patients. Due to the detection of changes imperceptible to the human eye, this type of analysis can potentially aid in deciding the optimal type of endovascular treatment. We included 40 patients who suffered from acute ischemic stroke caused by large vessel occlusion, and calculated 130 different textural features based on the non-enhanced CT scan using an open-source software (3D Slicer). Using chi-squared and Mann-Whitney tests and receiver operating characteristics analysis, we identified a total of 21 different textural parameters capable of predicting the outcome of thrombectomy (quantified as the mTICI score), with variable sensitivity (50-97.9%) and specificity (64.6-99.4%) rates. In conclusion, CT-based radiomics features are potential factors that can predict the outcome of thrombectomy in patients suffering from acute ischemic stroke, aiding in the decision between aspiration, mechanical, or combined thrombectomy procedure.

Evaluation of Olfactory Fossa Depth Using Computed Tomography in A Tertiary Center: A Retrospective Study

Introduction:Endoscopic sinus surgery is a widely utilized surgical method for addressing chronic sinonasal diseases. To prevent major postoperative problems, an anatomical understanding of its variants is necessary. Our study aims to determine the occurrence and classification of different Keros types by analyzing paranasal sinus CT scans and retrospectively evaluating the depth of the olfactory fossa in our population.Methods:This is a retrospective single-centre analytical study on 254 patients who had undergone a non-enhanced paranasal sinus CT scan for 1 year after ethical approval from the Institutional Review Committee of Nobel Medical College Teaching Hospital. The statistical program SPSS version 25 was used to conduct the statistical analysis.Results:The average depth of the right olfactory fossa (OF) was 4.4±1.44 mm on the right side and 4.5±1.5 mm on the left side. According to the Keros classification, out of a total of 254 patients, 38 patients (15%) were classified as Keros I, 193 patients (76%) as Keros II, and 23 patients (9%) as Keros III for the right lateral lamella. 44 (17.3%), 180 (70.9%), and 30 (11.8%) patients were classified as Keros I, II, and III respectively, for the left lateral lamella..Conclusions:A preoperative CT scan of the paranasal sinus is useful in assessing the architecture of the anterior skull base and making sure the surgical approach is properly planned.

Diagnostic Accuracy of Dual-Energy CT Material Decomposition Technique for Assessing Bone Status Compared with Quantitative Computed Tomography.

The purpose of this study was to evaluate the diagnostic accuracy when using various base material pairs (BMPs) in dual-energy computed tomography (DECT), and to establish corresponding diagnostic standards for assessing bone status through comparison with quantitative computed tomography (QCT). This prospective study enrolled a total of 469 patients who underwent both non-enhanced chest CT scans under conventional kVp and abdominal DECT. The bone densities of hydroxyapatite (water), hydroxyapatite (fat), hydroxyapatite (blood), calcium (water), and calcium (fat) (DHAP (water), DHAP (fat), DHAP (blood), DCa (water), and DCa (fat)) in the trabecular bone of vertebral bodies (T11-L1) were measured, along with bone mineral density (BMD) via QCT. Intraclass correlation coefficient (ICC) analysis was used to assess the agreement of the measurements. Spearman's correlation test was performed to analyze the relationship between the DECT- and QCT-derived BMD. Receiver operator characteristic (ROC) curves were generated to determine the optimal diagnostic thresholds of various BMPs for diagnosing osteopenia and osteoporosis. A total of 1371 vertebral bodies were measured, and QCT identified 393 with osteoporosis and 442 with osteopenia. Strong correlations were observed between DHAP (water), DHAP (fat), DHAP (blood), DCa (water), and DCa (fat) and the QCT-derived BMD. DHAP (water) showed the best predictive capability for osteopenia and osteoporosis. The area under the ROC curve, sensitivity, and specificity for identifying osteopenia were 0.956, 86.88%, and 88.91% with DHAP (water) ≤ 107.4 mg/cm3, respectively. The corresponding values for identifying osteoporosis were 0.999, 99.24%, and 99.53% with DHAP (water) ≤ 89.62 mg/cm3, respectively. Bone density measurement using various BMPs in DECT enables the quantification of vertebral BMD and the diagnosis of osteoporosis, with DHAP (water) having the highest diagnostic accuracy.

Diagnostic Accuracy of Recognition of Stroke in Emergency Room Score in Patients with Ischemic Stroke

To determine the diagnostic accuracy of ROSIER scale in patients with ischemic stroke taking non-enhanced CT scan brain as gold standard. Materials and Methods: Cross-sectional study conducted at tertiary care hospital in Gujrat from October 2019 to December 2020. Patients of both sex groups aged 30- 75 years presented in emergency within 24 hours of onset of focal neurological deficit which is objectively present and suspected of ischemic stroke. 289 patients were included. Patient history and neurological examination done calculating ROSIER scale and ischemic stroke diagnosed after CT scan brain done 24 hours after onset of symptoms. Sensitivity, specificity, PPV, NPV and diagnostic accuracy of ROSIER scale were calculated. Data analysis was done using SPSS 20.0. Results: In this study group 141(48.8%) were male while 148(51.2%) were female. Mean age was 62.1+11.02 years. 49(17%) patients were aged &lt;50 years while 240(83%) were aged &gt;50 years. 215(74.4%) presented within 12 hours while 74(25.6%) after 12 hours of symptoms onset. Sensitivity of ROSIER scale for diagnosing ischemic stroke was 98.86%(95%CI;95.96%-99.86%), specificity 5.31%(95%CI;1.97%-11.20%), PPV 61.92%(95% CI;60.82%-63.01%) and NPV was 75%(95%CI;38.13%-93.59%). LR+ was 1.04(95%CI; 1.00-1.09) while LR- was 0.21(95%CI; 0.04-1.04). Conclusion: ROSIER score can be used as initial screening test in patients with stroke. It has high sensitivity but low specificity and is a good test to rule in and rule out stroke patients.

Platelet-Neutrophil Association in NETs-Rich Areas in the Retrieved AIS Patient Thrombi.

Histological structure of thrombi is a strong determinant of the outcome of vascular recanalization therapy, the only treatment option for acute ischemic stroke (AIS) patients. A total of 21 AIS patients from this study after undergoing non-enhanced CT scan and multimodal MRI were treated with mechanical stent-based and manual aspiration thrombectomy, and thromboembolic retrieved from a cerebral artery. Complementary histopathological and imaging analyses were performed to understand their composition with a specific focus on fibrin, von Willebrand factor, and neutrophil extracellular traps (NETs). Though distinct RBC-rich and platelet-rich areas were found, AIS patient thrombi were overwhelmingly platelet-rich, with 90% of thrombi containing &lt;40% total RBC-rich contents (1.5 to 37%). Structurally, RBC-rich areas were simple, consisting of tightly packed RBCs in thin fibrin meshwork with sparsely populated nucleated cells and lacked any substantial von Willebrand factor (VWF). Platelet-rich areas were structurally more complex with thick fibrin meshwork associated with VWF. Plenty of leukocytes populated the platelet-rich areas, particularly in the periphery and border areas between platelet-rich and RBC-rich areas. Platelet-rich areas showed abundant activated neutrophils (myeloperoxidase+ and neutrophil-elastase+) containing citrullinated histone-decorated DNA. Citrullinated histone-decorated DNA also accumulated extracellularly, pointing to NETosis by the activated neutrophils. Notably, NETs-containing areas showed strong reactivity to VWF, platelets, and high-mobility group box 1 (HMGB1), signifying a close interplay between these components.

Deep learning-based lesion subtyping and prediction of clinical outcomes in COVID-19 pneumonia using chest CT

The main objective of this work is to develop and evaluate an artificial intelligence system based on deep learning capable of automatically identifying, quantifying, and characterizing COVID-19 pneumonia patterns in order to assess disease severity and predict clinical outcomes, and to compare the prediction performance with respect to human reader severity assessment and whole lung radiomics. We propose a deep learning based scheme to automatically segment the different lesion subtypes in nonenhanced CT scans. The automatic lesion quantification was used to predict clinical outcomes. The proposed technique has been independently tested in a multicentric cohort of 103 patients, retrospectively collected between March and July of 2020. Segmentation of lesion subtypes was evaluated using both overlapping (Dice) and distance-based (Hausdorff and average surface) metrics, while the proposed system to predict clinically relevant outcomes was assessed using the area under the curve (AUC). Additionally, other metrics including sensitivity, specificity, positive predictive value and negative predictive value were estimated. 95% confidence intervals were properly calculated. The agreement between the automatic estimate of parenchymal damage (%) and the radiologists’ severity scoring was strong, with a Spearman correlation coefficient (R) of 0.83. The automatic quantification of lesion subtypes was able to predict patient mortality, admission to the Intensive Care Units (ICU) and need for mechanical ventilation with an AUC of 0.87, 0.73 and 0.68 respectively. The proposed artificial intelligence system enabled a better prediction of those clinically relevant outcomes when compared to the radiologists’ interpretation and to whole lung radiomics. In conclusion, deep learning lesion subtyping in COVID-19 pneumonia from noncontrast chest CT enables quantitative assessment of disease severity and better prediction of clinical outcomes with respect to whole lung radiomics or radiologists’ severity score.

Thin Slice Unenhanced Brain CT Can Detect Aneurysms Larger than 7 mm.

Purpose:To evaluate the detection rate of intracranial aneurysms on reconstructed thin slice non enhanced CT (NECT) scans.Methods:NECT scans from 34 patients with 35 aneurysms and 35 individuals without aneurysms were collected. Thin slice maximum intensity projections of the NECT scans were reconstructed. One observer evaluated the native images twice with a time interval of six month between both passes with respect to the prevalence and location of an aneurysm. The size and location of the aneurysms were evaluated in corresponding CT-Angiography and Time of flight datasets. A logit regression analysis was performed with size and location as dependent variables. 2 × 2 tables were constructed. The sensitivity and false negative rate were calculated for aneurysms with 0–6.9 mm, 7–9.9 mm and 10–20 mm and the under the curve (AUC) was calculated.Results:The overall detection rate of the aneurysms was 63% for the first pass and 66% for the second pass in the reconstructed NECT scans. The detection rate of aneurysms is size dependent. The sensitivity to detect aneurysms with a size of 0–6.9 mm was 0.09 and 0.03, for aneurysms with a size of 7–9.9. mm was 0.8 and 0.7 and for aneurysms with a size of 10–20 mm was 0.92 for both passes.The AUC was 0.77 for the first pass and 0.78 for the second pass.Conclusions:NECT scans can be used to detect a significant proportion of intracranial aneurysms larger than 7 mm if properly displayed and reconstructed. These patients should receive further vascular imaging to prevent future aneurysm related subarachnoid hemorrhage.

Machine-learning-based radiomics identifies atrial fibrillation on the epicardial fat in contrast-enhanced and non-enhanced chest CT.

The purpose is to establish and validate a machine-learning-derived radiomics approach to determine the existence of atrial fibrillation (AF) by analyzing epicardial adipose tissue (EAT) in CT images. Patients with AF based on electrocardiographic tracing who underwent contrast-enhanced (n = 200) or non-enhanced (n = 300) chest CT scans were analyzed retrospectively. After EAT segmentation and radiomics feature extraction, the segmented EAT yielded 1691 radiomics features. The most contributive features to AF were selected by the Boruta algorithm and machine-learning-based random forest algorithm, and combined to construct a radiomics signature (EAT-score). Multivariate logistic regression was used to build clinical factor and nested models. In the test cohort of contrast-enhanced scanning (n = 60/200), the AUC of EAT-score for identifying patients with AF was 0.92 (95%CI: 0.84-1.00), higher than 0.71 (0.58-0.85) of the clinical factor model (total cholesterol and body mass index) (DeLong's p = 0.01), and higher than 0.73 (0.61-0.86) of the EAT volume model (p = 0.01). In the test cohort of non-enhanced scanning (n = 100/300), the AUC of EAT-score was 0.85 (0.77-0.92), higher than that of the CT attenuation model (p < 0.001). The two nested models (EAT-score+volume and EAT-score+volume+clinical factors) for contrast-enhanced scan and one (EAT-score+CT attenuation) for non-enhanced scan showed similar AUCs with that of EAT-score (all p > 0.05). EAT-score generated by machine-learning-based radiomics achieved high performance in identifying patients with AF. A radiomics analysis based on machine learning allows for the identification of AF on the EAT in contrast-enhanced and non-enhanced chest CT.

Correlation Between Intracranial Carotid Artery Calcification and Prognosis of Acute Ischemic Stroke After Intravenous Thrombolysis

ObjectiveTo investigate the correlation between prognosis and intracranial carotid artery calcification (ICAC) in patients with acute ischemic stroke (AIS) who receive intravenous thrombolysis (IVT).MethodsA total of 156 AIS patients who received IVT from March 2019 to March 2020 were enrolled. The modified Woodcock visual score was used to evaluate ICAC in nonenhanced head CT scans. Patients were divided into high calcification burden (HCB; score ≥3) and low calcification burden (LCB; score <3) groups. Demographic, laboratory, imaging and clinical data were compared between the two groups, and whether HCB was a prognostic factor was evaluated.ResultsCompared with the LCB group, the HCB group had a higher incidence of atrial fibrillation (49.2 vs.22.1%, P < 0.001) and coronary heart disease (24.6 vs. 10.0%, P = 0.019) and higher serum homocysteine [15.31 (12.15, 17.50) vs. 14.40 (11.20, 16.20), P = 0.036] and hemoglobin A1c (6.93 ± 1.77 vs. 6.37 ± 0.74, P = 0.023) levels. Binary logistic regression analysis showed that atrial fibrillation (OR = 3.031, 95% CI: 1.312–7.006, P = 0.009) and HbA1c (OR = 1.488, 95% CI: 1.050–2.109, P = 0.026) were independent risk factors for ICAC. After adjusting for other risk factors, symptomatic-side and bilateral ICACs were independent risk factors for poor prognosis (OR = 1.969, 95% CI: 1.220–3.178, P = 0.006), (OR = 1.354, 95% CI: 1.065–1.722, P = 0.013) and mortality (OR = 4.245, 95% CI: 1.114–16.171, P = 0.034), (OR = 2.414, 95% CI = 1.152–5.060, P = 0.020) in patients with AIS who received IVT.ConclusionICAC is closely related to the prognosis of acute ischemic stroke after intravenous thrombolysis.