Machine learning methods offer great promise for fast and accurate detection and prognostication of coronavirus disease 2019 (COVID-19) from standard-of-care chest radiographs (CXR) and chest computed tomography (CT) images. Many articles have been published in 2020 describing new machine learning-based models for both of these tasks, but it is unclear which are of potential clinical utility. In this systematic review, we consider all published papers and preprints, for the period from 1 January 2020 to 3 October 2020, which describe new machine learning models for the diagnosis or prognosis of COVID-19 from CXR or CT images. All manuscripts uploaded to bioRxiv, medRxiv and arXiv along with all entries in EMBASE and MEDLINE in this timeframe are considered. Our search identified 2,212 studies, of which 415 were included after initial screening and, after quality screening, 62 studies were included in this systematic review. Our review finds that none of the models identified are of potential clinical use due to methodological flaws and/or underlying biases. This is a major weakness, given the urgency with which validated COVID-19 models are needed. To address this, we give many recommendations which, if followed, will solve these issues and lead to higher-quality model development and well-documented manuscripts.
BackgroundUnlike most noninvasive imaging modalities, coronary computed tomography angiography can characterize subtypes of atherosclerotic plaque.ObjectivesThe purpose of this study was to investigate the prognostic implications of adverse coronary plaque characteristics in patients with suspected coronary artery disease.MethodsIn this SCOT-HEART (Scottish COmputed Tomography of the HEART Trial) post hoc analysis, the presence of adverse plaque (positive remodeling or low attenuation plaque), obstructive disease, and coronary artery calcification within 15 coronary segments was assessed on coronary computed tomography angiography of 1,769 patients who were followed-up for 5 years.ResultsAmong study participants (mean age 58 ± 10 years; 56% male), 608 (34%) patients had 1 or more adverse plaque features. Coronary heart disease death or nonfatal myocardial infarction was 3 times more frequent in patients with adverse plaque (n = 25 of 608 [4.1%] vs. n = 16 of 1,161 [1.4%]; p < 0.001; hazard ratio [HR]: 3.01; 95% confidence interval (CI): 1.61 to 5.63; p = 0.001) and was twice as frequent in those with obstructive disease (n = 22 of 452 [4.9%] vs. n = 16 of 671 [2.4%]; p = 0.024; HR: 1.99; 95% CI: 1.05 to 3.79; p = 0.036). Patients with both obstructive disease and adverse plaque had the highest event rate, with a 10-fold increase in coronary heart disease death or nonfatal myocardial infarction compared with patients with normal coronary arteries (HR: 11.50; 95% CI: 3.39 to 39.04; p < 0.001). However, these associations were not independent of coronary artery calcium score, a surrogate measure of coronary plaque burden.ConclusionsAdverse coronary plaque characteristics and overall calcified plaque burden confer an increased risk of coronary heart disease death or nonfatal myocardial infarction. (Scottish COmputed Tomography of the HEART Trial [SCOT-HEART]; NCT01149590)
BackgroundRapid access chest pain clinics have facilitated the early diagnosis and treatment of patients with coronary heart disease and angina. Despite this important service provision, coronary heart disease continues to be under-diagnosed and many patients are left untreated and at risk. Recent advances in imaging technology have now led to the widespread use of noninvasive computed tomography, which can be used to measure coronary artery calcium scores and perform coronary angiography in one examination. However, this technology has not been robustly evaluated in its application to the clinic.Methods/designThe SCOT-HEART study is an open parallel group prospective multicentre randomized controlled trial of 4,138 patients attending the rapid access chest pain clinic for evaluation of suspected cardiac chest pain. Following clinical consultation, participants will be approached and randomized 1:1 to receive standard care or standard care plus ≥64-multidetector computed tomography coronary angiography and coronary calcium score. Randomization will be conducted using a web-based system to ensure allocation concealment and will incorporate minimization. The primary endpoint of the study will be the proportion of patients diagnosed with angina pectoris secondary to coronary heart disease at 6 weeks. Secondary endpoints will include the assessment of subsequent symptoms, diagnosis, investigation and treatment. In addition, long-term health outcomes, safety endpoints, such as radiation dose, and health economic endpoints will be assessed. Assuming a clinic rate of 27.0% for the diagnosis of angina pectoris due to coronary heart disease, we will need to recruit 2,069 patients per group to detect an absolute increase of 4.0% in the rate of diagnosis at 80% power and a two-sided P value of 0.05. The SCOT-HEART study is currently recruiting participants and expects to report in 2014.DiscussionThis is the first study to look at the implementation of computed tomography in the patient care pathway that is outcome focused. This study will have major implications for the management of patients with cardiovascular disease.Trial registrationClinicalTrials.gov Identifier: NCT01149590
COPD is the second most common cause of pulmonary hypertension, and is a common complication of severe COPD with significant implications for both quality of life and mortality. However, the use of a rigid diagnostic threshold of a mean pulmonary arterial pressure (mPAP) of ≥25mHg when considering the impact of the pulmonary vasculature on symptoms and disease is misleading. Even minimal exertion causes oxygen desaturation and elevations in mPAP, with right ventricular hypertrophy and dilatation present in patients with mild to moderate COPD with pressures below the threshold for diagnosis of pulmonary hypertension. This has significant implications, with right ventricular dysfunction associated with poorer exercise capability and increased mortality independent of pulmonary function tests.The compliance of the pulmonary artery (PA) is a key component in decoupling the right ventricle from the pulmonary bed, allowing the right ventricle to work at maximum efficiency and protecting the microcirculation from large pressure gradients. PA stiffness increases with the severity of COPD, and correlates well with the presence of exercise induced pulmonary hypertension. A curvilinear relationship exists between PA distensibility and mPAP and pulmonary vascular resistance (PVR) with marked loss of distensibility before a rapid rise in mPAP and PVR occurs with resultant right ventricular failure. This combination of features suggests PA stiffness as a promising biomarker for early detection of pulmonary vascular disease, and to play a role in right ventricular failure in COPD. Early detection would open this up as a potential therapeutic target before end stage arterial remodelling occurs.
BackgroundThere is considerable overlap between left ventricular noncompaction (LVNC) and other cardiomyopathies. LVNC has been reported in up to 40% of the general population, raising questions about whether it is a distinct pathological entity, a remodeling epiphenomenon, or merely an anatomical phenotype.ObjectivesThe authors determined the prevalence and predictors of LVNC in a healthy population using 4 cardiac magnetic resonance imaging diagnostic criteria.MethodsVolunteers >40 years of age (N = 1,651) with no history of cardiovascular disease (CVD), a 10-year risk of CVD < 20%, and a B-type natriuretic peptide level greater than their gender-specific median underwent magnetic resonance imaging scan as part of the TASCFORCE (Tayside Screening for Cardiac Events) study. LVNC ratios were measured on the horizontal and vertical long axis cine sequences. All individuals with a noncompaction ratio of ≥2 underwent short axis systolic and diastolic LVNC ratio measurements, and quantification of noncompacted and compacted myocardial mass ratios. Those who met all 4 criteria were considered to have LVNC.ResultsOf 1,480 participants analyzed, 219 (14.8%) met ≥1 diagnostic criterion for LVNC, 117 (7.9%) met 2 criteria, 63 (4.3%) met 3 criteria, and 19 (1.3%) met all 4 diagnostic criteria. There was no difference in demographic or allometric measures between those with and without LVNC. Long axis noncompaction ratios were the least specific, with current diagnostic criteria positive in 219 (14.8%), whereas the noncompacted to compacted myocardial mass ratio was the most specific, only being met in 61 (4.4%).ConclusionsA significant proportion of an asymptomatic population free from CVD satisfy all currently used cardiac magnetic resonance imaging diagnostic criteria for LVNC, suggesting that those criteria have poor specificity for LVNC, or that LVNC is an anatomical phenotype rather than a distinct cardiomyopathy.
The celiac axis (CA) and its branches are critically important arteries that supply blood to the vital solid and hollow abdominal viscera of the foregut. There are many potential anatomic configurations, with up to half the population having a variation from the classic pattern of the CA bifurcating into the hepatosplenic trunk and left gastric artery. These configurations result from permutations in the fusion of the paired dorsal aortas during the first trimester. Despite the short length of the CA, it is affected by a wide range of pathologic conditions, including mesenteric ischemia due to intrinsic occlusion (secondary to causes such as atherosclerosis or thromboembolic events) and extrinsic compression from masses or the median arcuate ligament. Symptoms of mesenteric ischemia are nonspecific and include postprandial abdominal pain and weight loss; thus, the underlying pathologic condition may be found only when being sought specifically. More unusual pathologic conditions include dissection, aneurysms, and vascular malformations. Awareness of the pathologic conditions that affect the CA is important for both diagnostic and interventional radiologists. Early recognition and treatment of CA disease may prevent catastrophic hemorrhage and bowel infarction. Both endovascular and surgical approaches to treatment are greatly enhanced by correct identification of arterial anatomic variants; catheter angiography, computed tomographic angiography, and magnetic resonance angiography can facilitate detection of these variants. Knowledge of the different anatomic permutations is essential to guide endovascular procedures, such as hemorrhage control, transarterial interventional oncologic therapy, and treatment of visceral artery aneurysms. Online supplemental material is available for this article.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.