Machine Learning to Predict In-Hospital Mortality in COVID-19 Patients Using Computed Tomography-Derived Pulmonary and Vascular Features.

Pulmonary parenchymal and vascular damage are frequently reported in COVID-19 patients and can be assessed with unenhanced chest computed tomography (CT), widely used as a triaging exam. Integrating clinical data, chest CT features, and CT-derived vascular metrics, we aimed to build a predictive model of in-hospital mortality using univariate analysis (Mann-Whitney U test) and machine learning models (support vectors machines (SVM) and multilayer perceptrons (MLP)). Patients with RT-PCR-confirmed SARS-CoV-2 infection and unenhanced chest CT performed on emergency department admission were included after retrieving their outcome (discharge or death), with an 85/15% training/test dataset split. Out of 897 patients, the 229 (26%) patients who died during hospitalization had higher median pulmonary artery diameter (29.0 mm) than patients who survived (27.0 mm, p < 0.001) and higher median ascending aortic diameter (36.6 mm versus 34.0 mm, p < 0.001). SVM and MLP best models considered the same ten input features, yielding a 0.747 (precision 0.522, recall 0.800) and 0.844 (precision 0.680, recall 0.567) area under the curve, respectively. In this model integrating clinical and radiological data, pulmonary artery diameter was the third most important predictor after age and parenchymal involvement extent, contributing to reliable in-hospital mortality prediction, highlighting the value of vascular metrics in improving patient stratification.

Exploring the Role of Matrix Metalloproteinases as Biomarkers in Sporadic Lymphangioleiomyomatosis and Tuberous Sclerosis Complex. A Pilot Study.

Background: Lymphangioleiomyomatosis can develop in a sporadic form (S-LAM) or in women with tuberous sclerosis complex (TSC). The matrix metalloproteinases (MMPs) are extracellular matrix-degrading enzymes potentially involved in cystic lung destruction, and in the process of migration of LAM cells. The aim of the study was to explore the role of MMP-2 and MMP-7, such as vascular endothelial growth factor (VEGF) -C and -D in women with LAM, including patients with minor pulmonary disease (i.e., <10 lung cysts), and TSC with or without LAM. Methods: We evaluated 50 patients: 13 individuals affected by S-LAM, 20 with TSC-LAM, of whom six with minor pulmonary disease, and 17 with TSC without pulmonary involvement. Sixteen healthy women were used as controls. Results: MMP-2 resulted higher in LAM compared to healthy volunteers, and TSC patients (p = 0.040). MMP-7 was higher in TSC-LAM patient, with even greater values in patients with TSC-LAM minor pulmonary disease, than in S-LAM patients, and in controls (p = 0.001). VEGF-D level was lower than 800 pg/mL in all healthy controls and resulted higher in S-LAM and TSC-LAM than in TSC patients and controls (p < 0.001). VEGF-C values were not statistically different in the study population (p = 0.354). The area under ROC curves (AUCs) of MMP-2, and MMP-7 for predicting LAM diagnosis were of 0.756 ± 0.079 (p = 0.004), and 0.828 ± 0.060 (p < 0.001), respectively. Considering only patients with TSC, the AUCs for MMP-2, and MMP-7 in predicting LAM were 0.694 ± 0.088 (p = 0.044), and 0.713 ± 0.090 (p = 0.027), respectively. Conclusions: Our data suggest that MMP-2 and MMP-7 could be promising biomarkers for LAM diagnosis.

Epicardial fat inflammation response to COVID-19 therapies.

OBJECTIVE: Adipose tissue plays a role in the novel coronavirus disease 2019 (COVID-19). Epicardial adipose tissue (EAT), a unique visceral fat, presents with a high degree of inflammation in severe COVID-19. Whether and how adipose tissue may respond to the COVID-19 therapies is unknown. METHODS: The difference in computed tomography-measured EAT and subcutaneous (SAT) attenuation, defined as mean attenuation expressed in Hounsfield units (HU), was retrospectively analyzed in 72 patients (mean [SD] age was 59.6 [12.4] years, 50 patients [69%] were men) at the hospital admission for COVID-19 and 99 days (interquartile range = 71-129) after discharge. RESULTS: At the admission, EAT-HU was significantly correlated with blood glucose levels, interleukin 6, troponin T levels, and waist circumference. EAT-HU decreased from -87.21 (16.18) to -100.0 (11) (p < 0.001), whereas SAT-HU did not change (-110.21 [12.1] to -111.11 [27.82]; p = 0.78) after therapy. Changes in EAT-HU (expressed as ∆) significantly correlated with dexamethasone therapy (r = -0.46, p = 0.006) and when dexamethasone was combined with tocilizumab (r = -0.24, p = 0.04). CONCLUSIONS: Dexamethasone therapy was associated with significant reduction of EAT inflammation in COVID-19 patients, whereas SAT showed no changes. Anti-inflammatory therapies targeting visceral fat may be helpful in COVID-19.

Computed tomography coronary angiography in patients without known coronary artery disease can demonstrate possible non-cardiovascular causes of non-acute retrosternal chest pain.

OBJECTIVES: To assess the computed tomography coronary angiography (CTCA) accuracy for demonstrating possible non-cardiovascular causes of non-acute retrosternal chest pain in patients without known coronary artery disease (CAD) and to correlate CTCA results with the patient management and relief from pain. METHODS: This prospective observational study was approved by the ethical committee. Consecutive patients suffering non-acute chest pain who underwent CTCA and with not known CAD were enrolled and classified as having coronary diseases (CD) or extracardiac diseases (ECD). Association between age, sex, body mass index (BMI), cardiovascular risk factors, and type of chest pain with CD or ECD was estimated. Correlation between BMI classes and each risk factor was also calculated. RESULTS: A total of 106 patients (60 males; age 62 ± 14 years [mean ± standard deviation]; mean BMI 27) were enrolled. Hypertension was found in 71/106 (67%); smoking was significantly more frequent among males (p = 0.003) and hypercholesterolemia among females (p = 0.017); hypertension and hypercholesterolemia significantly correlated with age, and hypertension also with BMI. Pain was atypical in 70/106 (66%) patients. The kind of pain did not correlate with disease or gender. CTCA showed possible causes of chest pain in 69/106 (65%) patients; 32/69 (47%) having only CD, 23/69 (33%) only ECD, and 14/69 (20%) both CD and ECD. Prevalence was: hiatal hernia 35/106 (33%); significant CAD 24/106 (23%); myocardial bridging 22/106 (21%). At follow-up of 94/106 (89%) patients, 71/94 (76%) were pain-free, 14/17 (82%) significant CAD had been treated, and only one patient with non-significant CAD was treated after CTCA. CONCLUSION: CTCA suggested possible causes of non-acute pain in 65% of patients. MAIN MESSAGES: • CTCA can either rule in or rule out possible causes of chest pain alternative to CAD. • Clinically relevant findings were detected in 65% of patients with non-acute chest pain. • Non-cardiovascular diseases potentially explained symptoms in 35% of patients.

CT-MRI LI-RADS v2017: A Comprehensive Guide for Beginners.

Hepatocellular carcinoma (HCC) is the most common primary liver malignancy and the second leading cause of cancer-related deceases worldwide. Early diagnosis is essential for correct management and improvement of prognosis. Proposed for the first time in 2011 and updated for the last time in 2017, the Liver Imaging-Reporting and Data System (LI-RADS) is a comprehensive system for standardized interpretation and reporting of computed tomography (CT) and magnetic resonance imaging (MRI) liver examinations, endorsed by the American College of Radiology to achieve congruence with HCC diagnostic criteria in at-risk populations. Understanding its algorithm is fundamental to correctly apply LI-RADS in clinical practice. In this pictorial review, we provide a guide for beginners, explaining LI-RADS indications, describing major and ancillary features and eventually elucidating the diagnostic algorithm with the use of some clinical examples.

Use of the ultrasound-based total malignancy score in the management of thyroid nodules.

PURPOSE: The purpose of this study was to validate the role of the total malignancy score (TMS) in identifying thyroid nodules suspicious for malignancy through the sum of their ultrasound features. METHODS: The local ethical committee approved this prospective observational study. We examined 231 nodules in 231 consecutive patients (164 females and 67 males; age range, 20 to 87 years; median age, 59 years; interquartile range, 48 to 70 years) who underwent ultrasound followed by fine-needle aspiration cytology (FNAC). The nodules were further classified using the TMS, which considers ultrasound features (number, echogenicity, structure, halo, margins, Doppler signal, calcifications, and growth), and the Bethesda System for Reporting Thyroid Cytopathology (TBSRTC), which considers cytological features. Patients with non-negative nodules (TBSRTC categories III to VI) underwent histological analysis, repeated FNAC, or 2 years of regular ultrasound follow-up. The associations between the final diagnosis, each of the ultrasound features, and the TMS were estimated using the chi-square test, the Mann-Whitney U test, and multivariate logistic regression. A receiver operating characteristic (ROC) curve was used to evaluate the diagnostic accuracy of the TMS. RESULTS: On ultrasound, 47% of the nodules (108 of 231) had a TMS >3, 18% (42 of 231) had a TMS of 3, and 35% (81 of 231) had a TMS >3. The FNAC results of 85% of the nodules (196 of 231) were benign, while 15% (35 of 231) had non-negative results. Hypoechogenicity, solid structure, the presence of microcalcifications, and the number of nodules were independent predictors of the final diagnosis, and the diagnostic accuracy of the TMS was good (area under the ROC curve, 0.82). CONCLUSION: The TMS system is simple to use, reliable, easily reproducible, and closely reflects malignancy risk. Based on our results, FNAC could be limited to nodules with a TMS ≥3 without missing any cases of carcinoma.

Women with TSC: Relationship between Clinical, Lung Function and Radiological Features in a Genotyped Population Investigated for Lymphangioleiomyomatosis.

The advent of pharmacological therapies for lymphangioleiomyomatosis (LAM) has made early diagnosis important in women with tuberous sclerosis complex (TSC), although the lifelong cumulative radiation exposure caused by chest computer tomography (CT) should not be underestimated. We retrospectively investigated, in a cohort of TSC outpatients of San Paolo Hospital (Milan, Italy) 1) the role of pulmonary function tests (PFTs) for LAM diagnosis, 2) the association between LAM and other features of TSC (e.g. demography, extrapulmonary manifestations, genetic mutations, etc.), and 3) the characteristics of patients with multifocal micronodular pneumocyte hyperplasia (MMPH). Eighty-six women underwent chest CT scan; pulmonary involvement was found in 66 patients (77%; 49% LAM with or without MMPH, and 28% MMPH alone). LAM patients were older, with a higher rate of pneumothorax, presented more frequently with renal and hepatic angiomyolipomas, and tended to have a TSC2 mutation profile. PFTs, assessed in 64% of women unaffected by cognitive impairments, revealed a lower lung diffusion capacity in LAM patients. In multivariate analysis, age, but not PFTs, resulted independently associated with LAM diagnosis. Patients with MMPH alone did not show specific clinical, functional or genetic features. A mild respiratory impairment was most common in LAM-TSC patients: In conclusions, PFTs, even if indicated to assess impairment in lung function, are feasible in a limited number of patients, and are not significantly useful for LAM diagnosis in women with TSC.

Fractional flow reserve based on computed tomography: an overview.

Computed tomography coronary angiography (CTCA) is a technique proved to provide high sensitivity and negative predictive value for the identification of anatomically significant coronary artery disease (CAD) when compared with invasive X-ray coronary angiography. While the CTCA limitation of a ionizing radiation dose delivered to patients is substantially overcome by recent technical innovations, a relevant limitation remains the only anatomical assessment of coronary stenoses in the absence of evaluation of their functional haemodynamic significance. This limitation is highly important for those stenosis graded as intermediate at the anatomical assessment. Recently, non-invasive methods based on computational fluid dynamics were developed to calculate vessel-specific fractional flow reserve (FFR) using data routinely acquired by CTCA [computed tomographic fractional flow reserve (CT-FFR)]. Here we summarize methods for CT-FFR and review the evidence available in the literature up to June 26, 2016, including 16 original articles and one meta-analysis. The perspective of CT-FFR may greatly impact on CAD diagnosis, prognostic evaluation, and treatment decision-making. The aim of this review is to describe technical characteristics and clinical applications of CT-FFR, also in comparison with catheter-based invasive FFR, in order to make a cost-benefit balance in terms of clinical management and patient's health.

Myocardial Fatty Foci in Adult Patients with Tuberous Sclerosis Complex: Association with Gene Mutation and Multiorgan Involvement.

PURPOSE: To estimate the association between myocardial fatty foci (MFF) on chest computed tomographic (CT) images and type of gene mutation or multiorgan involvement in patients with tuberous sclerosis complex (TSC). MATERIALS AND METHODS: This retrospective case-control study was approved by the ethics committee, which waived the need for patient consent. Forty-eight patients with definite TSC (41 women; mean age, 35 years ± 11 [standard deviation]) and 96 age- and sex-matched patients without TSC who had undergone chest CT were evaluated. Two blinded readers independently scored MFF as low-attenuation areas within the myocardium. Patient history, gene mutation, and multiorgan involvement were obtained from clinical records. Cohen κ, Mann-Whitney U, χ(2) or Fisher exact, Kruskal-Wallis, and Spearman statistics were calculated. RESULTS: One or more MFF was detected in 50% (24 of 48) of patients with TSC; however, no MFF was detected in control patients (P < .001). MFFs were oval (62%, 15 of 24) or linear (38%, nine of 24) and involved the left ventricle in 13 patients and both ventricles in 24 patients (mostly the apical or midleft ventricle); median size was 127 mm(2). After four patients with TSC and unknown mutational status (two with MFF) were excluded, MFF was detected in 53% (10 of 19) of patients with TSC1 mutation, 65% (11 of 17) of patients with TSC2 mutation, and 12% (one of eight) of patients with TSC but without an identified mutation (P = .044). MFF presence was associated with brain (P = .011) and multiorgan (P = .008) involvement. The number of MFF per patient correlated with the degree of multiorgan involvement (P = .014). With MFF considered predictive of TSC, 50% (24of 48) sensitivity, 100% (96 of 96) specificity, 100% (24 of 24) positive predictive value, and 80% (96 of 120) negative predictive value were obtained. CONCLUSION: MFF was highly specific for TSC. MFF presence was associated with TSC gene mutations and with brain or multiorgan involvement; their number per patient was correlated with the degree of multiorgan involvement.

Four-year cardiac magnetic resonance (CMR) follow-up of patients treated with percutaneous pulmonary valve stent implantation.

OBJECTIVES: To investigate follow-up after percutaneous pulmonary valve implantation (PPVI). METHODS: Forty patients with pulmonary conduit dysfunction (males/females 24/16; 21 ± 08 years; 12 tetralogy of Fallot, 11 aortic valve disease, 17 other congenital heart disease) were planned for CMR before PPVI and repeated 7 times up to 48 months. CMR prospective results regarded: pressure gradient (PG) and regurgitation fraction (RF); end-diastolic volume, end-systolic volume, and stroke volume indexed to body surface area (EDVI, ESVI, and SVI); ejection fraction (EF) of right and left ventricles (RV, LV). A Friedman test was used for comparisons. RESULTS: Overall, PG (31 ± 06 to 16 ± 4 mmHg), RF (16 ± 17 to 0.3 ± 1 %), RVEDVI (82 ± 38 to 58 ± 12 ml/m(2)), and RVESVI (44 ± 12 to 30 ± 13 ml/m(2)) declined (p < 0.001), RVEF (49 ± 13 to 58 ± 12 %) and RVSVI (from 38 ± 14 to 40 ± 8 ml/m(2)) increased (p < 0.001), LVEDVI (67 ± 17 to 73 ± 18 ml/m(2)) and LVSVI (37 ± 11 to 43 ± 10 ml/m(2)) increased (p = 0.034 and p < 0.001). Two patients had valve fracture at 24 and 36 months and underwent surgery. One patient had stent restenosis at 24 months and underwent percutaneous retreatment. Baseline/follow-up CMR did not predict PPVI failure. CONCLUSIONS: CMR demonstrated restored pulmonary conduit function, reduced RV volumes and increased RV and LV function but did not predict valve fracture/restenosis. KEY POINTS: • A CMR 4-year follow-up after PPVI showed restored pulmonary conduit function • RV volumes were significantly reduced • RV function was significantly better in terms of increased EF and SVI • LV function was significantly better in terms of increased EDVI and SVI • Baseline/follow-up CMR did not predict three cases of PPVI failure.

Contrast enhanced chest-MDCT in oncologic patients. Prospective evaluation of the prevalence of incidental pulmonary embolism and added value of thin reconstructions.

OBJECTIVES: To prospectively assess prevalence/characteristics of clinically unsuspected pulmonary embolism (PE) in cancer patients undergoing follow-up chest MDCT and investigate MDCT protocol. METHODS: We evaluated 1013 oncologic patients. MDCT images at 5 and 1.25 mm thickness were independently evaluated. Pulmonary artery opacification degree was assessed. Presence, level, and site of PE were reported. Type of malignancy and metastases were reported for PE-positive patients. RESULTS: After excluding 1.4% (14/1013) of examinations due to inadequate vessel opacification, 999 patients (572 male; mean age:68 ± 12 years; range:26-93 years) entered the study. Prevalence of PE was 5%. There was significant improvement in the sensitivity for both readers in the evaluation of 1.25 mm compared to 5 mm images (46-50% to 82-92%). 30% (15/51) PE were not described by the radiologist in the prospectively issued report; 53 % (27/51) of PE were segmental, 72.5% (37/51) unilateral. The right lower lobe was the most involved (59%). 27% patients had colon cancer, 18% lung cancer. Among PE-positive patients (25 male; mean age 70 ± 10 years; range:44-87 years), 25% (13/51) had lung cancer, 15% (8/51) colon cancer. CONCLUSIONS: Thin reconstructions are essential for PE diagnosis, regardless of reader experience. Regarding oncologic patients, incidental PE diagnosis influences anticoagulation therapy. KEY POINTS: • CT pulmonary angiography is the gold standard for PE diagnosis. • Cancer and oncological treatments are risk factors for PE. • The prevalence of unsuspected PE was 5%. • Thin reconstructions are essential for PE diagnosis regardless of reader experience. • In oncologic patients, PE diagnosis influences anticoagulation therapy.

Bronchial artery hypertrophy is correlated with coronary artery disease.

BACKGROUND: Bronchial arteries support the systemic pulmonary vasculature and physiologically communicate with pulmonary arteries and coronary arteries. While there is evidence supporting the link between pulmonary diseases and bronchial artery hypertrophy (BAH), few data on the correlation between coronary artery disease (CAD) and BAH have been published. PURPOSE: To evaluate a possible association between BAH and CAD in patients without known pulmonary diseases undergoing computed tomography coronary angiography (CTCA). MATERIAL AND METHODS: This retrospective study was approved by the local ethics committee. One hundred patients with varying degrees of CAD underwent CTCA. Patients were stratified into four groups as follows: group I, 25 patients without CAD or with non-significant CAD; group II, 25 untreated patients with significant CAD; group III, 25 stented patients; group IV, 25 patients with coronary artery bypass grafts. The number and diameter of bronchial arteries were recorded. Correlation between age, CAD, and BAH was estimated. RESULTS: One hundred and ninety-nine bronchial arteries were detected. Approximately 51% were hypertrophic (diameter, >1.5 mm) with a mean diameter of 1.7 ± 0.5 mm. Seventy-six patients showed no pulmonary alterations; 24 were found to have previously undiagnosed pulmonary findings, six of which were severe. Presence and degree of CAD correlated with patients' mean age (60 in group I, 68 in group II, 65 in group III, 69 in group IV; P = 0.023), and mean bronchial artery transverse diameter (1.6 mm, 1.7 mm, 1.8 mm, and 2.0 mm, respectively; P = 0.009). The bronchial artery diameter was not associated with pulmonary findings (P = 0.390). CONCLUSION: There is an association between CAD and BAH. In patients with no pulmonary alterations, BAH could be caused by undiagnosed underlying CAD.

Association of aortic wall thickness on contrast-enhanced chest CT with major cerebro-cardiac events.

BACKGROUND: There is a significant association between aortic atherosclerosis and previous major cardiovascular events. Particularly, thoracic aortic atherosclerosis is closely related to the degree of coronary and carotid artery disease. Thus, there is a rationale for screening the thoracic aorta in patients who undergo a chest computed tomography (CT) for any clinical question, in order to detect patients at increased risk of cerebro-cardiovascular (CCV) events. PURPOSE: To estimate the association between either thoracic aortic wall thickness (AWT) or aortic total calcium score (ATCS) and CCV events. MATERIAL AND METHODS: One hundred and forty-eight non-cardiac patients (78 men; 67 ± 12 years) underwent chest contrast-enhanced multidetector CT (MDCT). The AWT was measured at the level of the left atrium (AWTref) and at the maximum AWT (AWTmax). Correlation with clinical CCV patients' history was estimated. The value of AWTmax and of a semi-quantitative ATCS as a marker for CCV events was assessed using receiver-operating characteristic curve (ROC) analysis and multivariate regression analysis. RESULTS: Out of 148 patients, 59% reported sedentary lifestyle, 44% hypertension, 32% smoking, 23% hypercholesterolemia, 13% family history of cardiac disease, 12% diabetes, and 10% BMI ≥ 30 kg/m(2); 9% reported myocardial infarction, 8% aortic aneurism, 8% myocardial revascularization, and 2% ischemic stroke. Twenty-six percent of patients had a medium-to-high ATCS. Both AWTmax and AWTref correlated with hypertension and age (P < 0.002). At the ROC analysis, a 4.8 mm threshold was associated to a 90% specificity and an odds ratio of 6.3 (AUC = 0.735). Assuming as threshold the AWTmax median value (4.3 mm) of patients who suffered from at least one CCV event in their history, a negative predictive value of 90%, a RR of 3.6 and an OR of 6.3 were found. At the multivariate regression analysis, AWTmax was the only independent variable associated to the frequency of CCV events. CONCLUSION: Patients with increased thoracic AWTmax on chest MDCT could be considered at risk for CCV disease.

Malignant incidental extracardiac findings on cardiac CT: systematic review and meta-analysis.

OBJECTIVE: The objective of our study was to systematically review the evidence on incidental extracardiac findings on cardiac CT with a focus on previously unknown malignancies. MATERIALS AND METHODS: A systematic search was performed (PubMed, EMBASE, Cochrane databases) for studies reporting incidental extracardiac findings on cardiac CT. Among 1099 articles initially found, 15 studies met the inclusion criteria. The references of those articles were hand-searched and 14 additional studies were identified. After review of the full text, 10 articles were excluded. Nineteen studies including 15,877 patients (64% male) were analyzed. A three-level analysis was performed to determine the prevalence of patients with incidental extracardiac findings, the prevalence of patients with major incidental extracardiac findings, and the prevalence of patients with a proven cancer. Heterogeneity was explored for multiple variables. Pooled prevalence and 95% CI were calculated. RESULTS: The prevalence of both incidental extracardiac findings and major incidental extracardiac findings showed a high heterogeneity (I2>95%): The pooled prevalence was 44% (95% CI, 35-54%) and 16% (95% CI, 14-20%), respectively. No significant explanatory variables were found for using or not using contrast material, the size of the FOV, and study design (I2>85%). The pooled cancer prevalence for 10 studies including 5082 patients was 0.7% (95% CI, 0.5-1.0%), with an almost perfect homogeneity (I2<0.1%). Of 29 reported malignancies, 21 (72%) were lung cancers; three, thyroid cancers; two, breast cancers; two, liver cancers; and one, mediastinal lymphoma. CONCLUSION: Although the prevalence of reported incidental extracardiac finding at cardiac CT was highly variable, a homogeneous prevalence of previously unknown malignancies was reported across the studies, for a pooled estimate of 0.7%; more than 70% of these previously unknown malignancies were lung cancers. Extracardiac findings on cardiac CT require careful evaluation and reporting.

Management of thyroid follicular proliferation: an ultrasound-based malignancy score to opt for surgical or conservative treatment.

This study was conducted to evaluate whether ultrasound characteristics of thyroid nodules with a known cytologic diagnosis of "follicular pattern" (indicative of follicular hyperplasia, follicular adenoma or follicular carcinoma) can be used to define a nodule malignancy score to limit surgery to selected, higher-risk cases. In pre-operative ultrasound results of patients diagnosed with a "follicular pattern" on cytology who subsequently underwent surgery, each nodule feature was given a score from 0 (most likely benign) to 2 (most likely malignant), resulting in a total score ranging from 0 to 11. The total nodule score (total malignancy score) was then compared with the definitive histologic diagnosis. According to our results, surgery is advisable in patients with score ≥4 (79% of carcinomas), ultrasound follow-up seems to be appropriate for patients with a score of 3 (21% of carcinomas) and no action is recommended for patients with score <3 (only benign nodules).