Comparing Radiomics features of tumour and healthy liver tissue in a limited CT dataset: A machine learning study

IntroductionLiver cancer lesions on Computed Tomography (CT) withholds a great amount of data, which is not visible to the radiologists and radiographer. Radiomics features can be extracted from the lesions and used to train Machine Learning (ML) algorithms to predict between tumour and liver tissue. The purpose of this study was to investigate and classify Radiomics features extracted from liver tumours and normal liver tissue in a limited CT dataset.MethodsThe Liver Tumour Segmentation Benchmark (LiTS) dataset consisting of 131 CT scans of the liver with segmentations of tumour tissue and healthy liver was used to extract Radiomic features. Extracted Radiomic features included size, shape, and location extracted with morphological and statistical techniques according to the International Symposium on Biomedical Imaging manual. Relevant features was selected with chi2 correlation and principal component analysis (PCA) with tumour and healthy liver tissue as outcome according to a consensus between three experienced radiologists. Logistic regression, random forest and support vector machine was used to train and validate the dataset with a 10-fold cross-validation method and the Grid Search as hyper-parameter tuning. Performance was evaluated with sensitivity, specificity and accuracy.ResultsThe performance of the ML algorithms achieved sensitivities, specificities and accuracy ranging from 96.30% (95% CI: 81.03%–99.91%) to 100.00% (95% CI: 86.77%–100.00%), 91.30% (95% CI: 71.96%–98.93%) to 100.00% (95% CI: 83.89%–100.00%)and 94.00% (95% CI: 83.45%–98.75%) to 100.00% (95% CI: 92.45%–100.00%), respectively.ConclusionML algorithms classifies Radiomics features extracted from healthy liver and tumour tissue with perfect accuracy. The Radiomics signature allows for a prognostic biomarker for hepatic tumour screening on liver CT.Implications for practiceDifferentiation between tumour and liver tissue with Radiomics ML algorithms have the potential to increase the diagnostic accuracy, assist in the decision-making of supplementary multiphasic enhanced medical imaging, as well as for developing novel prognostic biomarkers for liver cancer patients.     

Deep learning with convolutional neural network for estimation of the characterisation of coronary plaques: Validation using IB-IVUS

IntroductionDeep learning approaches have shown high diagnostic performance in image classifications, such as differentiation of malignant tumors and calcified coronary plaque. However, it is unknown whether deep learning is useful for characterizing coronary plaques without the presence of calcification using coronary computed tomography angiography (CCTA). The purpose of this study was to compare the diagnostic performance of deep learning with a convolutional neural network (CNN) with that of radiologists in the estimation of coronary plaques.MethodsWe retrospectively enrolled 178 patients (191 coronary plaques) who had undergone CCTA and integrated backscatter intravascular ultrasonography (IB-IVUS) studies. IB-IVUS diagnosed 81 fibrous and 110 fatty or fibro-fatty plaques. We manually captured vascular short-axis images of the coronary plaques as Portable Network Graphics (PNG) images (150 × 150 pixels). The display window level and width were 100 and 700 Hounsfield units (HU), respectively. The deep-learning system (CNN; GoogleNet Inception v3) was trained on 153 plaques; its performance was tested on 38 plaques. The area under the curve (AUC) obtained by receiver operating characteristic analysis of the deep learning system and by two board-certified radiologists was compared.ResultsWith the CNN, the AUC and the 95% confidence interval were 0.83 and 0.69–0.96, respectively; for radiologist 1 they were 0.61 and 0.42–0.80; for radiologist 2 they were 0.68 and 0.51–0.86, respectively. The AUC for CNN was significantly higher than for radiologists 1 (p = 0.04); for radiologist 2 it was not significantly different (p = 0.22).ConclusionDL-CNN performed comparably to radiologists for discrimination between fatty and fibro-fatty plaque on CCTA images.Implications for practiceThe diagnostic performance of the CNN and of two radiologists in the assessment of 191 ROIs on CT images of coronary plaques whose type corresponded with their IB-IVUS characterization was comparable.     

Feasibility of dynamic myocardial CT perfusion using single-source 64-row CT

BackgroundDynamic myocardial computed tomography perfusion (CTP) is an emerging technique to diagnose significant coronary stenosis. However, this procedure has not been reported using single-source 64-row CT.ObjectiveTo investigate the radiation dose and the diagnostic performance of dynamic CTP to diagnose significant stenosis by catheter exam.MethodsWe prospectively included 165 patients who underwent CTP exam under adenosine stress using a single-source 64-row CT. MBF was calculated using the deconvolution technique. Quantitative perfusion ratio (QPR) was defined as the myocardial blood flow (MBF) of the myocardium with coronary stenosis divided by the MBF of the myocardium without significant stenosis or infarct. Of the 44 patients who underwent subsequent coronary angiography, we assessed the diagnostic performance to diagnose ≥50% stenosis by quantitative coronary analysis (QCA).ResultsThe average effective dose of dynamic CTP and the entire scans were 2.5 ± 0.7 and 7.3 ± 1.8 mSv, respectively. The MBF of the myocardium without significant stenosis was 1.20 ± 0.32 ml/min/g, which significantly decreased to 0.98 ± 0.24 ml/min/g (p < 0.01) in the area with ≥50% stenosis by CT angiography. The QPR of the myocardium with QCA ≥50% stenosis was significantly lower than 1 (0.84 ± 0.32, 95% confidence interval (CI), 0.77–0.90, p < 0.001). The accuracy to detect QCA ≥50% stenosis was 82% (95%CI, 74–88%) using CT angiography alone and significantly increased to 87% (95%CI, 80–92%, p < 0.05) including QPR.ConclusionDynamic myocardial CTP could be performed using 64-row CT with a low radiation dose and would improve the diagnostic performance to detect QCA ≥50% stenosis than CT angiography alone.     

The Value of Preprocedural MR Imaging in Genicular Artery Embolization for Patients with Osteoarthritic Knee Pain

PurposeTo determine the value of preprocedural MR imaging in genicular artery embolization (GAE) for patients with osteoarthritic knee pain.Materials and MethodsThis single-center study retrospectively analyzed 28 knees in 18 patients who underwent GAE for intractable knee pain < 1 month after MR imaging. The pain experienced in each knee was evaluated on a 100-mm visual analog scale (VAS) at baseline and 1- and 3-month after GAE. “GAE responders” were defined as knees that exhibited greater than 30% reduction of VAS pain scores from baseline at both follow-up visits. Musculoskeletal radiologists evaluated MR images of the affected knee compartment regarding cartilage defects, osteophytes, subchondral cysts, bone marrow lesions (BMLs), meniscal injury, and joint effusion. The performances of Kellgren–Lawrence (KL) grading and MR findings in predicting GAE responders was estimated based on receiver operating characteristic curves.ResultsThe mean VAS pain score was 84.3 mm. BML (area under the curve [AUC], 0.860; P < .001), meniscal injury (AUC, 0.811; P = .003), and KL grading (AUC, 0.898; P < .001) were significantly associated with GAE outcome. To predict GAE responders, KL grade ≤ 2 yielded a sensitivity of 87.5% and a specificity of 60.9%, BML grade ≤ 1 yielded a sensitivity of 75.0% and a specificity of 69.6%, and meniscal injury grade ≤ 2 yielded a sensitivity of 83.3% and a specificity of 72.7%.ConclusionsLarge BMLs and severe meniscal injuries on MR imaging, as well as high KL grades, indicated poor responses to GAE.     

Deep Learning Based on MR Imaging for Predicting Outcome of Uterine Fibroid Embolization

PurposeTo develop and validate a deep learning model based on routine magnetic resonance (MR) imaging obtained before uterine fibroid embolization to predict procedure outcome.Materials and MethodsClinical data were collected on patients treated with uterine fibroid embolization at the Hospital of the University of Pennsylvania from 2007 to 2018. Fibroids for each patient were manually segmented by an abdominal radiologist on a T1-weighted contrast-enhanced (T1C) sequence and a T2-weighted sequence of MR imaging obtained before and after embolization. A residual convolutional neural network (ResNet) model to predict clinical outcome was trained using MR imaging obtained before the procedure.ResultsInclusion criteria were met by 727 fibroids in 409 patients. At clinical follow-up, 85.6% (n = 350) of 409 patients (590 of 727 fibroids; 81.1%) experienced symptom resolution or improvement, and 14.4% (n = 59) of 409 patients (137 of 727 fibroids; 18.9%) had no improvement or worsening symptoms. The T1C trained model achieved a test accuracy of 0.847 (95% confidence interval [CI], 0.745–0.914), sensitivity of 0.932 (95% CI, 0.833–0.978), and specificity of 0.462 (95% CI, 0.232–0.709). In comparison, the average of 4 radiologists achieved a test accuracy of 0.722 (95% CI, 0.609–0.813), sensitivity of 0.852 (95% CI, 0.737–0.923), and specificity of 0.135 (95% CI, 0.021–0.415).ConclusionsThis study demonstrates that deep learning based on a ResNet model achieves good accuracy in predicting outcome of uterine fibroid embolization. If further validated, the model may help clinicians better identify patients who can most benefit from this therapy and aid clinical decision making.     

Management of patients with suspected or confirmed COVID-19, in the radiology department

ObjectivesFrom December 2019, a novel coronavirus disease named COVID-19 was reported in China. Within 3 months, the World Health Organization defined COVID-19 as a pandemic, with more than 370,000 cases and 16,000 deaths worldwide. In consideration of the crucial role of diagnostic testing during COVID-19, the aim of this technical note was to provide a complete synthesis of approaches implemented for the management of suspected or confirmed COVID-19 patients.Key findingsThe planning of a robust plan to prevent the transmission of the virus to patients and department staff members should be fundamental in each radiology service. Moreover, the speed of spread and the incidence of the pandemic make it necessary to optimize the use of personal protective devices and dedicated COVID-19 equipment, given the limited availability of supplies.ConclusionIn the management of radiographic and CT imaging, staff should take special precautions to limit contamination between patients and other patients or professionals.Implications for practiceAn isolated imaging room should be dedicated to suspected or confirmed COVID-19 cases, including radiography and CT scanners. This paper will provide guidance concerning disposable protective gear to be utilized, as well as on the cleaning and sanitation of radiology room and equipment.     

Validation of clinical photography and a laser fluorescence device for assessment of enamel demineralization during multi-bracketed fixed orthodontic treatment

AimTo compare the validity of DIAGNOdent pen™ 2190 and clinical photographs in the diagnosis (presence or absence) and assessment of white spot lesions (WSLs) during multi-bracketed fixed orthodontic treatment.MethodsA total of 1607 readings from 99 patients were obtained using DIAGNOdent pen™ 2190, clinical photographs, and visual assessment criteria to assess WSLs at 6 -, 12- and 18 months after bonding of brackets. Receiver operating characteristic curves were constructed, and areas under the curve (AUC) were calculated to assess the presence or absence of WSLs and distinguish the severity of WSLs as slight or severe. Sensitivities, specificities, and AUC were calculated for outcomes at an alpha of 0.05 and 95% confidence intervals (CIs).ResultsThe values of sensitivity and specificity of DIAGNOdent pen™ 2190 were found to be 0.78 and 0.83, respectively, to distinguish presence/absence of WSLs with good accuracy (AUC= 0.855; 95% CI: 0.820–0.889) and for digital photographs, accuracy was excellent (AUC= 0.973, 95% CI: 0.956–0.991) with sensitivity and specificity of 0.96 and 0.99, respectively. However, the accuracy of DIAGNOdent pen™ 2190 was poor (AUC= 0.543, 95% CI: 0.418–0.668) in delineating slight and severe WSLs with a sensitivity of 0.31 and specificity of 0.83; while the values of sensitivity and specificity for distinguishing slight or severe lesions for clinical photographs were 0.96 and 0.99, respectively with excellent accuracy (AUC=0.978, 95% CI: 0.936–1.000).ConclusionsClinical photographs had better validity than DIAGNOdent pen™ 2190 in diagnosis and assessment of severity of demineralized lesions during multi-bracketed fixed orthodontic treatment.     

Liver Tumor Ablation Procedure Duration and Estimated Patient Radiation Dose: Comparing Positron Emission Tomography/CT and CT Guidance

PurposeTo compare procedure duration and patient radiation dose in positron emission tomography/computed tomography (PET/CT) and CT-guided liver tumor ablation procedures.Materials and MethodsIn this retrospective, case-control study, 275 patients underwent 368 image-guided ablation procedures to treat 537 tumors. Radiologists used PET/CT guidance for 117 procedures and CT guidance for 251 procedures. PET/CT-guided procedures were performed by one radiologist (C: P.B.S.). All 3 radiologists (A: J.G.S., B: a radiologist who is not an author on this article, and C: P.B.S.) performed CT-guided procedures. Potential confounders included patient demographics, clinical and tumor characteristics, and procedural variables.ResultsThe mean duration and estimated patient radiation dose of PET/CT-guided procedures performed by radiologist C were 21.5 ± 4.9 minutes longer and 0.7 ± 2.8 mSv higher than CT-guided procedures performed by all radiologists in an unadjusted comparison. Adjusting for confounding, mean duration and estimated dose of PET/CT-guided procedures performed by radiologist C were 28.3 ± 3.8 minutes longer (P < .0001) and 6.2 ± 2.9 mSv higher (P = .03) than CT-guided procedures performed by the same radiologist. Comparing CT-guided procedures performed by all 3 radiologists, adjusted mean durations and estimated patient doses of procedures by the least experienced radiologist, radiologist A, and the second most experienced radiologist, radiologist B, were 24.2 ± 5.1 (P < .0001) and 18.1 ± 8.9 (P = .04) minutes longer and 13.1 ± 3.7 (P < .001) and 14.5 ± 6.4 (P = .02) mSv higher, respectively, than procedures performed by the most experienced radiologist, radiologist C.ConclusionsPET/CT-guided liver ablations had a slightly longer duration with slightly higher estimated patient radiation dose than similar CT-guided liver ablations. Procedure duration and patient dose do not appear to be major impediments to the emerging field of PET/CT-guided tumor ablation.     

A Real-Time Blood Flow Measurement Device for Patients with Peripheral Artery Disease

PurposeTo evaluate the feasibility of a new optical device that measures peripheral blood flow as a diagnostic and monitoring tool for patients with peripheral artery disease (PAD).Materials and MethodsIn this prospective study, 167 limbs of 90 patients (mean age, 76 y; 53% men) with suspected PAD were evaluated with the FlowMet device, which uses a new type of dynamic light-scattering technology to assess blood flow in real time. Measurements of magnitude and phasicity of blood flow were combined into a single-value flow–waveform score and compared vs ankle–brachial index (ABI), toe–brachial index (TBI), and clinical presentation of patients per Rutherford category (RC). Receiver operating characteristic curves were constructed to predict RC. Area under the curve (AUC), sensitivity, and specificity were compared among flow–waveform score, ABI, and TBI.ResultsQualitatively, the FlowMet waveforms were analogous to Doppler velocity measurements, and degradation of waveform phasicity and amplitude were observed with increasing PAD severity. Quantitatively, the flow, waveform, and composite flow–waveform scores decreased significantly with decreasing TBI. In predicting RC ≥ 4, the flow–waveform score (AUC = 0.83) showed a linear decrease with worsening patient symptoms and power comparable to that of TBI (AUC = 0.82) and better than that of ABI (AUC = 0.71). Optimal sensitivity and specificity pairs were found to be 56%/83%, 72%/81%, and 89%/74% for ABI, TBI, and flow–waveform score, respectively.ConclusionsThe technology tested in this pilot study showed a high predictive value for diagnosis of critical limb ischemia. The device showed promise as a diagnostic tool capable of providing clinical feedback in real time.     

Comparison of coronary CT angiography-based and invasive coronary angiography-based quantitative flow ratio for functional assessment of coronary stenosis: A multicenter retrospective analysis

BackgroundThe aim of this study was to evaluate the diagnostic performance of coronary CT angiography (CTA)-based quantitative flow ratio (QFR), namely CT-QFR, and compare it with invasive coronary angiography (ICA)-based Murray law QFR (μQFR), using fractional flow reserve (FFR) as the reference standard.MethodsPatients who underwent coronary CTA, ICA and pressure wire-based FFR assessment within two months were retrospectively analyzed. CT-QFR and μQFR were computed in blinded fashion and compared with FFR, all applying the same cut-off value of ≤0.80 to identify hemodynamically significant stenosis.ResultsPaired comparison between CT-QFR and μQFR was performed in 191 vessels from 167 patients. Average FFR was 0.81 ?± ?0.10 and 42.4% vessels had an FFR ≤0.80. CT-QFR had a slightly lower correlation with FFR compared with μQFR, although statistically non-significant (r ?= ?0.87 versus 0.90, p ?= ?0.110). The vessel-level diagnostic performance of CT-QFR was slightly lower but without statistical significance than μQFR (AUC ?= ?0.94 versus 0.97, difference: ?0.03 [95%CI: ?0.00-0.06], p ?= ?0.095), and substantially higher than diameter stenosis by CTA (AUC difference: 0.17 [95%CI: ?0.10-0.23], p ?< ?0.001). The patient-level diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio and negative likelihood ratio for CT-QFR to identify FFR value ?≤ ?0.80 was 88%, 90%, 86%, 86%, 91%, 6.59 and 0.12, respectively. The diagnostic accuracy of CT-QFR was 84% in extensively calcified lesions, while in vessels with no or less calcification, CT-QFR showed a comparable diagnostic accuracy with μQFR (91% versus 92%, p ?= ?0.595). Intra- and inter-observer variability in CT-QFR analysis was ?0.00 ?± ?0.04 and 0.00 ?± ?0.04, respectively.ConclusionsPerformance in diagnosis of hemodynamically significant coronary stenosis by CT-QFR was slightly lower but without statistical significance than μQFR, and substantially higher than CTA-derived diameter stenosis. Extensively calcified lesions reduced the diagnostic accuracy of CT-QFR.     

An investigation of lumbar spine magnetic resonance referrals in two Irish university teaching centres: Radiology clinical judgement versus iRefer guideline compliance

IntroductionWith the annual increase in medical imaging demand, the appropriateness of the lumbar spine magnetic resonance imaging LSMRI referrals is worldwide gaining attention. This study aims to determine the appropriateness of LSMRI referrals and compare radiology clinical decisions to iRefer compliance based solely on referral text content.MethodsReferral text was extracted from 1021 LSMRI referrals. Two review panels were recruited: three expert radiologists and three MRI radiographers. Radiologists classified cases as indicated or not indicated for scanning based on their clinical judgement. The radiographers classified based solely on iRefer guidelines. Majority voting for each case was applied to both review panels and reviewer agreement was tested using Kappa analysis. Logistic regression models were developed to identify medical disciplines associated with high rates of indicated referrals.Results21.7% and 11.9% of the cases were found not indicated for MRI for radiologists and radiographers, respectively. Radiology review identified 18% of the GPs referrals as not indicated and 17% in the radiographers’ review. Panel agreement was fair: Kappa values of 0.23 and0.26 for the radiologists and radiographers respectively. Neurosurgery was associated with the highest rate of indicated referrals across both review panels: oncology referrals raised the highest number of open comments.ConclusionThe study identified a lower number of not indicated referrals compared to previous research. Findings indicate the importance of both guidelines compliance and clinical judgement to optimise practice.Implications for practiceFindings in this study found that even when strict instructions were given to the MR radiographers to vet referrals using the iRefer guidelines, ambiguity within the guidelines resulted in variations in decision-making. This suggests that detailed protocols are required to support radiographers in the vetting process to ensure a standardised approach.     

Transarterial Yttrium-90 Radioembolization for Unresectable Intrahepatic Cholangiocarcinoma: A Systematic Review and Meta-Analysis

PurposeTo investigate the overall efficacy and survival profile of yttrium-90 (⁹⁰Y) radioembolization for unresectable intrahepatic cholangiocarcinoma (ICC).Materials and MethodsA systematic literature review and meta-analysis was completed using a random-effects model. Studies describing the use of ⁹⁰Y for unresectable ICC were included. The disease control rate (DCR), downstaged-to-resectable rate, cancer antigen 19-9 (CA19-9) response rate, pooled median overall survival (OS), pooled median progression-free survival (PFS), and mean reported survival rates ranging from 3 to 36 months were evaluated.ResultsTwenty-one studies detailing a total of 921 patients were included. The overall DCR was 82.3% (95% confidence interval [CI], 76.7%–87.8%; I² = 81%). In 11% of the cases, patients were downstaged to being surgically resectable (95% CI, 6.1%–15.9%; I² = 78%). The CA19-9 response rate was 67.2% (95% CI, 54.5%–79.8%; I² = 60%). From the time of radioembolization, PFS was 7.8 months (95% CI, 4.2–11.3 months; I² = 94%) and median OS was 12.7 months (95% CI, 10.6–14.8 months; I² = 62%). Lastly, the mean overall reported survival proportions were 84% at 3 months (standard deviation [SD], 10%), 69% at 6 months (SD, 16%), 47% at 12 months (SD, 19%), 31% at 18 months (SD, 21%), 30% at 24 months (SD, 19%), 21% at 30 months (SD, 27%), and 5% at 36 months (SD, 7%).ConclusionsRadioembolization with ⁹⁰Y for unresectable ICC results in substantial downstaging, disease control, and survival.     

Applying patient characteristics,stent-graft selection,and pre-operative computed tomographic angiography data to a machine learning algorithm: Is endoleak prediction possible?

IntroductionThis study aims to predict endoleak after endovascular aneurysm repair (EVAR) using machine learning (ML) integration of patient characteristics, stent-graft configuration, and a selection of vessel lengths, diameters and angles measured using pre-operative computed tomography angiography (CTA).MethodsWe evaluated 1-year follow-up CT scans (arterial and delayed phases) in patients who underwent EVAR for the presence or absence of an endoleak. We also obtained data on the patient characteristics, stent-graft selection, and preoperative CT vessel morphology (diameter, length, and angle). The extreme gradient boosting (XGBoost) for the ML system was trained on 30 patients with endoleaks and 81 patients without. We evaluated 5217 items in 111 patients with abdominal aortic aneurysms, including the patient characteristics, stent-graft configuration and vascular morphology acquired using pre-EVAR abdominal CTA. We calculated the area under the curve (AUC) of our receiver operating characteristic analysis using the ML method.ResultsThe AUC, accuracy, 95% confidence interval (CI), sensitivity, and specificity were 0.88, 0.88, 0.79–0.97, 0.85, and 0.91 for ML applying XGBoost, respectively.ConclusionsThe diagnostic performance of the ML method was useful when factors such as the patient characteristics, stent-graft configuration and vessel length, diameter and angle of the vessels were considered from pre-EVAR CTA.Implications for practiceBased on our findings, we suggest that this is a potential application of ML for the interpretation of abdominal CTA scans in patients with abdominal aortic aneurysms scheduled for EVAR.     

Value of SPECT/CT in the diagnosis of avascular necrosis of the head of femur: A meta-analysis

IntroductionTo evaluate the diagnostic value of single photon emission computed tomography/computed tomography (SPECT/CT) in patients with avascular necrosis (AVN) of the femoral head.MethodsA search of the databases PubMed, ScienceDirect, and Google Scholar was undertaken using targeted index word. Studies satisfying inclusions and exclusions criteria were identified and additional related literature sources were considered. Analysis included pooled sensitivity and specificity, odds ratio (OR), and receiver operating characteristic (ROC) analysis.ResultsSeven studies were identified that specifically related to the diagnostic accuracy of SPECT/CT in the detection or characterisation of AVN of the femoral head. These were included in the meta-analysis. The pooled sensitivity and specificity were 94% (95% confidence interval of 87–97%) and 75% (95% confidence interval of 68–81%) respectively, the OR was 44.5 (with a 95% confidence interval of 18.2–108.6), and the area under the SROC was 93.73% (with a 95% confidence interval of 89.2%–98.3%).ConclusionThis meta-analysis provides strong evidence of the high diagnostic efficacy of SPECT/CT in the evaluation of AVN of the femoral head.Implications for practiceIn the absence of or contraindication to MRI, SPECT/CT is able to provide a suitable alternative for diagnosis of AVN of the head of femur.     

Características en tomografía computarizada de la neumonía por coronavirus-2019 (COVID-19) en tres grupos de pacientes iraníes: estudio de un solo centro

Introduction and ObjectivesThe pivotal role of chest computed tomographic (CT) to diagnosis and prognosis coronavirus disease-2019 (COVID-19) is still an open field to be explored. This study was conducted to assess the CT features in confirmed cases with COVID-19.Materials and MethodsRetrospectively, initial chest CT data of 363 confirmed cases with COVID-19 were reviewed. All subjects were stratified into three groups based on patients’ clinical outcomes; non-critical group (n=194), critical group (n=65), and death group (n=104). The detailed of CT findings were collected from patients’ medical records and then evaluated for each group. In addition, multinomial logistic regression was used to analyze risk factors according to CT findings in three groups of patients with COVID-19.ResultsCompared with the non-critical group, mixed ground-glass opacities (GGO) and consolidation lesion, pleural effusion lesion, presence of diffuse opacity in cases, more than 2 lobes involved and opacity scores were significantly higher in the critical and death groups (P<0.05). Having more mixed GGO with consolidation, pleural effusion, lack of pure GGO, more diffuse opacity, involvement of more than 2 lobes and high opacity score identified as independent risk factors of critical and death groups.ConclusionCT images of non-critical, critical and death groups with COVID-19 had definite characteristics. CT examination plays a vital role in managing the current COVID-19 outbreak, for early detection of COVID-19 pneumonia. In addition, initial CT findings may be useful to stratify patients, which have a potentially important utility in the current global medical situation.     

Retinal microvascular changes in patients recovered from COVID-19 compared to healthy controls: A meta-analysis

PurposeTo investigate changes in retinal microcirculation in patients recovered from COVID-19 infection compared to healthy controls, using optical coherence tomography-angiography.MethodsMeta-analysis of eligible studies comparing retinal microcirculation between patients recovered from COVID-19 infection and healthy controls up to 7th of September 2022 was performed, according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2009 guidelines. The following search algorithm was used: (COVID-19 OR coronavirus) AND (retina OR optical coherence tomography OR optical coherence tomography angiography OR vessel density OR foveal avascular zone). Standardized Mean Difference (SMD) with 95% confidence interval (CI) was calculated to compare continuous variables. Revman 5.3 was used for the analysis.Results12 studies were included in our analysis. Foveal avascular zone (FAZ) area was larger in patients recovered from COVID-19 infection compared to healthy controls, while there was no statistically significant difference in FAZ perimeter between the two groups. The foveal, parafoveal and whole image vessel density in the superficial capillary plexus showed no significant difference between the two groups. The foveal, parafoveal and whole image vessel density in the deep capillary plexus was statistically lower in patients recovered from COVID-19 compared to healthy controls.ConclusionFAZ area was enlarged and foveal, parafoveal and whole image vessel density in deep capillary plexus were reduced in patients recovered from COVID-19 infection compared to healthy controls, suggesting that COVID-19 infection may induce long-term retinal microvascular changes in patients recovered from the virus infection.     

A novel density-volume calcium score by non-contrast CT predicts coronary plaque burden on coronary CT angiography: Results from the MACS (Multicenter AIDS cohort study)

BackgroundThe purpose of this study is to determine if a new score calculated with coronary artery calcium (CAC) density and volume is associated with total coronary artery plaque burden and composition on coronary CT angiography (CCTA) compared to the Agatston score (AS).MethodsWe identified 347 men enrolled in the Multicenter AIDS cohort study who underwent contrast and non-contrast CCTs, and had CAC>0. CAC densities (mean Hounsfield Units [HU]) per plaque) and volumes on non-contrast CCT were measured. A Density-Volume Calcium score was calculated by multiplying the plaque volume by a factor based on the mean HU of the plaque (4, 3, 2 and 1 for 130–199, 200–299, 300–399, and ≥400HU). Total Density-Volume Calcium score was determined by the sum of these individual scores. The semi-quantitative partially calcified and total plaque scores (PCPS and TPS) on CCTA were calculated. The associations between Density-Volume Calcium score, PCPS and TPS were examined.ResultsOverall, 2879 CAC plaques were assessed. Multivariable linear regression models demonstrated a stronger association between the log Density-Volume Calcium score and both the PCPS (β 0.99, 95%CI 0.80–1.19) and TPS (β 2.15, 95%CI 1.88–2.42) compared to the log of AS (PCPS: β 0.77, 95%CI 0.61–0.94; TPS: β 1.70, 95%CI 1.48–1.94). Similar results were observed for numbers of PC or TP segments.ConclusionThe new CAC score weighted towards lower density demonstrated improved correlation with semi-quantitative PC and TP burden on CCTA compared to the traditional AS, which suggests it has utility as an alternative measure of atherosclerotic burden.     

Quantification of extracellular volume with cardiac computed tomography in patients with dilated cardiomyopathy

BackgroundCardiac computed tomography (CCT) was recently validated to measure extracellular volume (ECV) in the setting of cardiac amyloidosis, showing good agreement with cardiovascular magnetic resonance (CMR). However, no evidence is available with a whole-heart single source, single energy CT scanner in the clinical context of newly diagnosed left ventricular dysfunction. Therefore, the aim of this study was to test the diagnostic accuracy of ECV_CCT in patients with a recent diagnosis of dilated cardiomyopathy, having ECV_CMR as the reference technique.Methods39 consecutive patients with newly diagnosed dilated cardiomyopathy (LVEF <50%) scheduled for clinically indicated CMR were prospectively enrolled. Myocardial segment evaluability assessment with each technique, agreement between ECV_CMR and ECV_CCT, regression analysis, Bland-Altman analysis and interclass correlation coefficient (ICC) were performed.ResultsMean age of enrolled patients was 62 ​± ​11 years, and mean LVEF at CMR was 35.4 ​± ​10.7%. Overall radiation exposure for ECV estimation was 2.1 ​± ​1.1 ​mSv. Out of 624 myocardial segments available for analysis, 624 (100%) segments were assessable by CCT while 608 (97.4%) were evaluable at CMR. ECV_CCT demonstrated slightly lower values compared to ECV_CMR (all segments, 31.8 ​± ​6.5% vs 33.9 ​± ​8.0%, p ​< ​0.001). At regression analysis, strong correlations were described (all segments, r ​= ​0.819, 95% CI: 0.791 to 0.844). On Bland-Altman analysis, bias between ECV_CMR and ECV_CCT for global analysis was 2.1 (95% CI: −6.8 to 11.1). ICC analysis showed both high intra-observer and inter-observer agreement for ECV_CCT calculation (0.986, 95%CI: 0.983 to 0.988 and 0.966, 95%CI: 0.960 to 0.971, respectively).ConclusionsECV estimation with a whole-heart single source, single energy CT scanner is feasible and accurate. Integration of ECV measurement in a comprehensive CCT evaluation of patients with newly diagnosed dilated cardiomyopathy can be performed with a small increase in overall radiation exposure.     

Nationwide Outcomes following Percutaneous Cholecystostomy for Acute Calculous Cholecystitis and the Impact of Coronavirus Disease 2019: Results of the Multicentre Audit of Cholecystostomy and Further Interventions (MACAFI study)

PurposeTo assess the mortality, readmission rates, and practice variation of percutaneous cholecystostomy (PC) in patients with acute calculous cholecystitis in the United Kingdom (UK).Materials and MethodsA total of 1,186 consecutive patients (636 men [53.6%]; median age, 75 years; range, 24–102 years) who underwent PC for acute calculous cholecystitis between January 1, 2019, and December 31, 2020, were included from 36 UK hospitals. The exclusion criteria were diagnostic aspirations, absence of acute calculous cholecystitis, and age less than 16 years. The coronavirus disease 2019 (COVID-19) lockdown was declared on March 26, 2020, in the UK, which served to distinguish among groups.ResultsMost patients (66.3%) underwent PC as definitive treatment, whereas 31.3% underwent PC as a bridge to surgery. The overall 30-day readmission rate was 42.2% (500/1,186), and the 30-day mortality was 9.1% (108/1,186). Centers performing fewer than 30 PCs per year had higher 90-day mortality than those performing more than 60 (19.3% vs 11.0%, respectively; P = .006). A greater proportion of patients presented with complicated acute calculous cholecystitis during the COVID-19 pandemic compared to prior (49.9% vs 40.9%, respectively; P = .007), resulting in more PCs (61.3 vs 37.9 per month, respectively; P < .001). More PCs were performed in tertiary hospitals than in district general hospitals (9 vs 3 per 100 beds, respectively; P < .001), with a greater proportion performed as a bridge to surgery (50.5% vs 22.8%, respectively; P < .001).ConclusionsThe practice of PC is highly variable throughout the UK. The readmission rates are high, and there is significant correlation between mortality and PC case volume.     

Detection of metallic objects on digital radiographs with convolutional neural networks: A MRI screening tool

IntroductionScreening for metallic implants and foreign bodies before magnetic resonance imaging (MRI) examinations, are crucial for patient safety. History of health are supplied by the patient, a family member, screening of electronic health records or the picture and archive systems (PACS). PACS securely store and transmits digital radiographs (DR) and related reports with patient information. Convolutional neural networks (CNN) can be used to detect metallic objects in DRs stored in PACS. This study evaluates the accuracy of CNNs in the detection of metallic objects on DRs as an MRI screening tool.MethodsThe musculoskeletal radiographs (MURA) dataset consisting of 14.863 upper extremity studies were stratified into datasets with and without metal. For each anatomical region: Elbow, finger, hand, humerus, forearm, shoulder and wrist we trained and validated CNN algorithms to classify radiographs with and without metal. Algorithm performance was evaluated with area under the receiver-operating curve (AUC), sensitivity, specificity, predictive values and accuracies compared with a reference standard of manually labelling.ResultsSensitivities, specificities and area under the ROC-curves (AUC) for the six anatomic regions ranged from 85.33% (95% CI: 78.64%–90.57%) to 100.00% (95% CI: 98.16%–100.00%), 75.44% (95% CI: 62.24%–85.87%) to 93.57% (95% CI: 88.78%–96.75%) and 0.95 to 0.99, respectively.ConclusionCNN algorithms classify DRs with metallic objects for six different anatomic regions with near-perfect accuracy. The rapid and iterative capability of the algorithms allows for scalable expansion and as a substitute MRI screening tool for metallic objects.Implications for practiceAll CNNs would be able to assist in metal detection of digital radiographs prior to MRI, an substantially decrease screening time.