What do we mean by complex percutaneous coronary intervention? An assessment of agreement amongst interventional cardiologists for defining complexity.

BACKGROUND: In the last decade, percutaneous coronary intervention (PCI) has evolved toward the treatment of complex disease in patients with multiple comorbidities. Whilst there are several definitions of complexity, it is unclear whether there is agreement between cardiologists in classifying complexity of cases. Inconsistent identification of complex PCI can lead to significant variation in clinical decision-making. AIM: This study aimed to determine the inter-rater agreement in rating the complexity and risk of PCI procedures. METHOD: An online survey was designed and disseminated amongst interventional cardiologists by the European Association of Percutaneous Cardiovascular Intervention (EAPCI) board. The survey presented four patient vignettes, with study participants assessing these cases to classify their complexity. RESULTS: From 215 respondents, there was poor inter-rater agreement in classifying the complexity level (k = 0.1) and a fair agreement (k = 0.31) in classifying the risk level. The experience level of participants did not show any significant impact on the inter-rater agreement of rating the complexity level and the risk level. There was good level of agreement between participants in terms of rating 26 factors for classifying complex PCI. The top five factors were (1) impaired left ventricular function, (2) concomitant severe aortic stenosis, (3) last remaining vessel PCI, (4) requirement fort calcium modification and (5) significant renal impairment. CONCLUSION: Agreement among cardiologists in classifying complexity of PCI is poor, which may lead to suboptimal clinical decision-making, procedural planning as well as long-term management. Consensus is needed to define complex PCI, and this requires clear criteria incorporating both lesion and patient characteristics.

Exploring decision making 'noise' when interpreting the electrocardiogram in the context of cardiac cath lab activation.

In this commentary paper, we discuss the use of the electrocardiogram to help clinicians make diagnostic and patient referral decisions in acute care settings. The paper discusses the factors that are likely to contribute to the variability and noise in the clinical decision making process for catheterization lab activation. These factors include the variable competence in reading ECGs, the intra/inter rater reliability, the lack of standard ECG training, the various ECG machine and filter settings, cognitive biases (such as automation bias which is the tendency to agree with the computer-aided diagnosis or AI diagnosis), the order of the information being received, tiredness or decision fatigue as well as ECG artefacts such as the signal noise or lead misplacement. We also discuss potential research questions and tools that could be used to mitigate this 'noise' and improve the quality of ECG based decision making.

New models of atherosclerosis and multi-drug therapeutic interventions.

MOTIVATION: Atherosclerosis is amongst the leading causes of death globally. However, it is challenging to study in vivo or in vitro and no detailed, openly-available computational models exist. Clinical studies hint that pharmaceutical therapy may be possible. Here, we develop the first detailed, computational model of atherosclerosis and use it to develop multi-drug therapeutic hypotheses. RESULTS: We assembled a network describing atheroma development from the literature. Maps and mathematical models were produced using the Systems Biology Graphical Notation and Systems Biology Markup Language, respectively. The model was constrained against clinical and laboratory data. We identified five drugs that together potentially reverse advanced atheroma formation. AVAILABILITY AND IMPLEMENTATION: The map is available in the Supplementary Material in SBGN-ML format. The model is available in the Supplementary Material and from BioModels, a repository of SBML models, containing CellDesigner markup. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Machine learning techniques for detecting electrode misplacement and interchanges when recording ECGs: A systematic review and meta-analysis.

INTRODUCTION: Electrode misplacement and interchange errors are known problems when recording the 12­lead electrocardiogram (ECG). Automatic detection of these errors could play an important role for improving clinical decision making and outcomes in cardiac care. The objectives of this systematic review and meta-analysis is to 1) study the impact of electrode misplacement on ECG signals and ECG interpretation, 2) to determine the most challenging electrode misplacements to detect using machine learning (ML), 3) to analyse the ML performance of algorithms that detect electrode misplacement or interchange according to sensitivity and specificity and 4) to identify the most commonly used ML technique for detecting electrode misplacement/interchange. This review analysed the current literature regarding electrode misplacement/interchange recognition accuracy using machine learning techniques. METHOD: A search of three online databases including IEEE, PubMed and ScienceDirect identified 228 articles, while 3 articles were included from additional sources from co-authors. According to the eligibility criteria, 14 articles were selected. The selected articles were considered for qualitative analysis and meta-analysis. RESULTS: The articles showed the effect of lead interchange on ECG morphology and as a consequence on patient diagnoses. Statistical analysis of the included articles found that machine learning performance is high in detecting electrode misplacement/interchange except left arm/left leg interchange. CONCLUSION: This review emphasises the importance of detecting electrode misplacement detection in ECG diagnosis and the effects on decision making. Machine learning shows promise in detecting lead misplacement/interchange and highlights an opportunity for developing and operationalising deep learning algorithms such as convolutional neural network (CNN) to detect electrode misplacement/interchange.

A novel role for CRIM1 in the corneal response to UV and pterygium development.

Pterygium is a pathological proliferative condition of the ocular surface, characterised by formation of a highly vascularised, fibrous tissue arising from the limbus that invades the central cornea leading to visual disturbance and, if untreated, blindness. Whilst chronic ultraviolet (UV) light exposure plays a major role in its pathogenesis, higher susceptibility to pterygium is observed in some families, suggesting a genetic component. In this study, a Northern Irish family affected by pterygium but reporting little direct exposure to UV was identified carrying a missense variant in CRIM1 NM_016441.2: c.1235 A > C (H412P) through whole-exome sequencing and subsequent analysis. CRIM1 is expressed in the developing eye, adult cornea and conjunctiva, having a role in cell differentiation and migration but also in angiogenesis, all processes involved in pterygium formation. We demonstrate elevated CRIM1 expression in pterygium tissue from additional individual Northern Irish patients compared to unaffected conjunctival controls. UV irradiation of HCE-S cells resulted in an increase in ERK phosphorylation and CRIM1 expression, the latter further elevated by the addition of the MEK1/2 inhibitor, U0126. Conversely, siRNA knockdown of CRIM1 led to decreased UV-induced ERK phosphorylation and increased BCL2 expression. Transient expression of the mutant H412P CRIM1 in corneal epithelial HCE-S cells showed that, unlike wild-type CRIM1, it was unable to reduce the cell proliferation, increased ERK phosphorylation and apoptosis induced through a decrease of BCL2 expression levels. We propose here a series of intracellular events where CRIM1 regulation of the ERK pathway prevents UV-induced cell proliferation and may play an important role in the in the pathogenesis of pterygium.

Exploring the Involvement of NLRP3 and IL-1ß in Osteoarthritis of the Hand: Results from a Pilot Study.

Hand osteoarthritis (HOA) includes different subsets; a particular and uncommon form is erosive HOA (EHOA). Interleukin- (IL-) 1ß plays a crucial role in the pathogenesis of osteoarthritis (OA); it is synthesized as an inactive precursor which requires the intervention of a cytosolic multiprotein complex, named inflammasome, for its activation. The aim of this study was to investigate the involvement of IL-1ß and the NOD-like receptor pyrin domain containing 3 (NLRP3) inflammasome in patients with EHOA and nonerosive HOA (NEHOA) compared to healthy controls. In particular, we evaluated the gene expression of IL-1ß and NLRP3, the serum levels of IL-1ß, IL-6, IL-17, and tumor necrosis factor- (TNF-) α, and the protein levels of IL-1ß and NLRP3. We also assessed the relationships between IL-1ß and NLRP3 and clinical, laboratory, and radiological findings. Fifty-four patients with HOA (25 EHOA and 29 NEHOA) and 20 healthy subjects were included in the study. Peripheral blood mononuclear cell (PBMC) gene and protein expressions of IL-1ß and NLRP3 were quantified by quantitative real-time PCR and western blot. IL-1ß, IL-6, IL-17, and TNF-α serum levels were determined by ELISA. IL-1ß gene expression was significantly reduced (p = 0.0208) in EHOA compared to healthy controls. NLRP3 protein levels were significantly increased in the NEHOA group versus the control (p = 0.0063) and EHOA groups (p = 0.0038). IL-1ß serum levels were not significantly different across the groups; IL-6, IL-17, and TNF-α were not detectable in any sample. IL-1ß concentrations were negatively correlated with the Kellgren-Lawrence score in the whole population (r = -0.446; p = 0.0008) and in NEHOA (r = -0.608; p = 0.004), while IL-1ß gene expression was positively correlated with the number of joint swellings in the EHOA group (r = 0.512; p = 0.011). Taken together, our results, showing poorly detectable IL-1ß concentrations and minimal inflammasome activity in the PBMCs of HOA patients, suggest a low grade of systemic inflammation in HOA. This evidence does not preclude a possible involvement of these factors at the local level.

Data driven feature selection and machine learning to detect misplaced V1 and V2 chest electrodes when recording the 12­lead electrocardiogram.

BACKGROUND: Electrocardiogram (ECG) lead misplacement can adversely affect ECG diagnosis and subsequent clinical decisions. V1 and V2 are commonly placed superior of their correct position. The aim of the current study was to use machine learning approaches to detect V1 and V2 lead misplacement to enhance ECG data quality. METHOD: ECGs for 453 patients, (normal n = 151, Left Ventricular Hypertrophy (LVH) n = 151, Myocardial Infarction n = 151) were extracted from body surface potential maps. These were used to extract both the correct and incorrectly placed V1 and V2 leads. The prevalence for correct and incorrect leads were 50%. Sixteen features were extracted in three different domains: time-based, statistical and time-frequency features using a wavelet transform. A hybrid feature selection approach was applied to select an optimal set of features. To ensure optimal model selection, five classifiers were used and compared. The aforementioned feature selection approach and classifiers were applied for V1 and V2 misplacement in three different positions: first, second and third intercostal spaces (ICS). RESULTS: The accuracy for V1 misplacement detection was 93.9%, 89.3%, 72.8% in the first, second and third ICS respectively. In V2, the accuracy was 93.6%, 86.6% and 68.1% in the first, second and third ICS respectively. There is a noticeable decline in accuracy when detecting misplacement in the third ICS which is expected.

An exploratory analysis investigating blood protein biomarkers to augment ECG diagnosis of ACS.

BACKGROUND: Acute Coronary Syndrome (ACS) is currently diagnosed using a 12­lead Electrocardiogram (ECG). Our recent work however has shown that interpretation of the 12­lead ECG is complex and that clinicians can be sub-optimal in their interpretation. Additionally, ECG does not always identify acute total occlusions in certain patients. PURPOSE: The aim of the present study was to undertake an exploratory analysis to compare protein expression profiles of ACS patients that may in the future augment ECG diagnosis. METHODS: Patients were recruited consecutively at the cardiac catheterization laboratory at Altnagelvin Hospital over a period of 6 months. A low risk control group was recruited by advertisement. Blood samples were analysed using the multiplex proximity extension assays by OLINK proteomics. Support vector machine (SVM) learning was used as a classifier to distinguish between patient groups on training data. The ST segment elevation level was extracted from each ECG for a subset of patients and combined with the protein markers. Quadratic SVM (QSVM) learning was then used as a classifier to distinguish STEMI from NSTEMI on training and test data. RESULTS: Of the 344 participants recruited, 77 were initially diagnosed with NSTEMI, 7 with STEMI, and 81 were low risk controls. The other participants were those diagnosed with angina (stable and unstable) or non-cardiac patients. Of the 368 proteins analysed, 20 proteins together could significantly differentiate between patients with ACS and patients with stable angina (ROC-AUC = 0.96). Six proteins discriminated significantly between the stable angina and the low risk control groups (ROC-AUC = 1.0). Additionally, 16 proteins together perfectly discriminated between the STEMI and NSTEMI patients (ROC-AUC = 1). ECG comparisons with the protein biomarker data for a subset of patients (STEMI n = 6 and NSTEMI n = 6), demonstrated that 21 features (20 proteins + ST elevation) resulted in the highest classification accuracy 91.7% (ROC-AUC = 0.94). The 20 proteins without the ST elevation feature gave an accuracy of 80.6% (ROC-AUC 0.91), while the ST elevation feature without the protein biomarkers resulted in an accuracy of 69.3% (ROC-AUC = 0.81). CONCLUSIONS: This preliminary study identifies panels of proteins that should be further explored in prospective studies as potential biomarkers that may augment ECG diagnosis and stratification of ACS. This work also highlights the importance for future studies to be designed to allow a comparison of blood biomarkers not only with ECG's but also with cardio angiograms.

Computational modelling of atherosclerosis.

Atherosclerosis is one of the principle pathologies of cardiovascular disease with blood cholesterol a significant risk factor. The World Health Organization estimates that approximately 2.5 million deaths occur annually because of the risk from elevated cholesterol, with 39% of adults worldwide at future risk. Atherosclerosis emerges from the combination of many dynamical factors, including haemodynamics, endothelial damage, innate immunity and sterol biochemistry. Despite its significance to public health, the dynamics that drive atherosclerosis remain poorly understood. As a disease that depends on multiple factors operating on different length scales, the natural framework to apply to atherosclerosis is mathematical and computational modelling. A computational model provides an integrated description of the disease and serves as an in silico experimental system from which we can learn about the disease and develop therapeutic hypotheses. Although the work completed in this area to date has been limited, there are clear signs that interest is growing and that a nascent field is establishing itself. This article discusses the current state of modelling in this area, bringing together many recent results for the first time. We review the work that has been done, discuss its scope and highlight the gaps in our understanding that could yield future opportunities.

Automation bias in medicine: The influence of automated diagnoses on interpreter accuracy and uncertainty when reading electrocardiograms.

INTRODUCTION: Interpretation of the 12­lead Electrocardiogram (ECG) is normally assisted with an automated diagnosis (AD), which can facilitate an 'automation bias' where interpreters can be anchored. In this paper, we studied, 1) the effect of an incorrect AD on interpretation accuracy and interpreter confidence (a proxy for uncertainty), and 2) whether confidence and other interpreter features can predict interpretation accuracy using machine learning. METHODS: This study analysed 9000 ECG interpretations from cardiology and non-cardiology fellows (CFs and non-CFs). One third of the ECGs involved no ADs, one third with ADs (half as incorrect) and one third had multiple ADs. Interpretations were scored and interpreter confidence was recorded for each interpretation and subsequently standardised using sigma scaling. Spearman coefficients were used for correlation analysis and C5.0 decision trees were used for predicting interpretation accuracy using basic interpreter features such as confidence, age, experience and designation. RESULTS: Interpretation accuracies achieved by CFs and non-CFs dropped by 43.20% and 58.95% respectively when an incorrect AD was presented (p < 0.001). Overall correlation between scaled confidence and interpretation accuracy was higher amongst CFs. However, correlation between confidence and interpretation accuracy decreased for both groups when an incorrect AD was presented. We found that an incorrect AD disturbs the reliability of interpreter confidence in predicting accuracy. An incorrect AD has a greater effect on the confidence of non-CFs (although this is not statistically significant it is close to the threshold, p = 0.065). The best C5.0 decision tree achieved an accuracy rate of 64.67% (p < 0.001), however this is only 6.56% greater than the no-information-rate. CONCLUSION: Incorrect ADs reduce the interpreter's diagnostic accuracy indicating an automation bias. Non-CFs tend to agree more with the ADs in comparison to CFs, hence less expert physicians are more effected by automation bias. Incorrect ADs reduce the interpreter's confidence and also reduces the predictive power of confidence for predicting accuracy (even more so for non-CFs). Whilst a statistically significant model was developed, it is difficult to predict interpretation accuracy using machine learning on basic features such as interpreter confidence, age, reader experience and designation.

The genetics and screening of familial hypercholesterolaemia.

Familial Hypercholesterolaemia is an autosomal, dominant genetic disorder that leads to elevated blood cholesterol and a dramatically increased risk of atherosclerosis. It is perceived as a rare condition. However it affects 1 in 250 of the population globally, making it an important public health concern. In communities with founder effects, higher disease prevalences are observed.We discuss the genetic basis of familial hypercholesterolaemia, examining the distribution of variants known to be associated with the condition across the exons of the genes LDLR, ApoB, PCSK9 and LDLRAP1. We also discuss screening programmes for familial hypercholesterolaemia and their cost-effectiveness. Diagnosis typically occurs using one of the Dutch Lipid Clinic Network (DCLN), Simon Broome Register (SBR) or Make Early Diagnosis to Prevent Early Death (MEDPED) criteria, each of which requires a different set of patient data. New cases can be identified by screening the family members of an index case that has been identified as a result of referral to a lipid clinic in a process called cascade screening. Alternatively, universal screening may be used whereby a population is systematically screened.It is currently significantly more cost effective to identify familial hypercholesterolaemia cases through cascade screening than universal screening. However, the cost of sequencing patient DNA has fallen dramatically in recent years and if the rate of progress continues, this may change.

Predicting the presence of coronary plaques featuring high-risk characteristics using polygenic risk scores and targeted proteomics in patients with suspected coronary artery disease.

BACKGROUND: The presence of coronary plaques with high-risk characteristics is strongly associated with adverse cardiac events beyond the identification of coronary stenosis. Testing by coronary computed tomography angiography (CCTA) enables the identification of high-risk plaques (HRP). Referral for CCTA is presently based on pre-test probability estimates including clinical risk factors (CRFs); however, proteomics and/or genetic information could potentially improve patient selection for CCTA and, hence, identification of HRP. We aimed to (1) identify proteomic and genetic features associated with HRP presence and (2) investigate the effect of combining CRFs, proteomics, and genetics to predict HRP presence. METHODS: Consecutive chest pain patients (n = 1462) undergoing CCTA to diagnose obstructive coronary artery disease (CAD) were included. Coronary plaques were assessed using a semi-automatic plaque analysis tool. Measurements of 368 circulating proteins were obtained with targeted Olink panels, and DNA genotyping was performed in all patients. Imputed genetic variants were used to compute a multi-trait multi-ancestry genome-wide polygenic score (GPSMult). HRP presence was defined as plaques with two or more high-risk characteristics (low attenuation, spotty calcification, positive remodeling, and napkin ring sign). Prediction of HRP presence was performed using the glmnet algorithm with repeated fivefold cross-validation, using CRFs, proteomics, and GPSMult as input features. RESULTS: HRPs were detected in 165 (11%) patients, and 15 input features were associated with HRP presence. Prediction of HRP presence based on CRFs yielded a mean area under the receiver operating curve (AUC) ± standard error of 73.2 ± 0.1, versus 69.0 ± 0.1 for proteomics and 60.1 ± 0.1 for GPSMult. Combining CRFs with GPSMult increased prediction accuracy (AUC 74.8 ± 0.1 (P = 0.004)), while the inclusion of proteomics provided no significant improvement to either the CRF (AUC 73.2 ± 0.1, P = 1.00) or the CRF + GPSMult (AUC 74.6 ± 0.1, P = 1.00) models, respectively. CONCLUSIONS: In patients with suspected CAD, incorporating genetic data with either clinical or proteomic data improves the prediction of high-risk plaque presence. TRIAL REGISTRATION: https://clinicaltrials.gov/ct2/show/NCT02264717 (September 2014).

Potential Plasma Proteins (LGALS9, LAMP3, PRSS8 and AGRN) as Predictors of Hospitalisation Risk in COVID-19 Patients.

Background: The COVID-19 pandemic, caused by the novel coronavirus SARS-CoV-2, has posed unprecedented challenges to healthcare systems worldwide. Here, we have identified proteomic and genetic signatures for improved prognosis which is vital for COVID-19 research. Methods: We investigated the proteomic and genomic profile of COVID-19-positive patients (n = 400 for proteomics, n = 483 for genomics), focusing on differential regulation between hospitalised and non-hospitalised COVID-19 patients. Signatures had their predictive capabilities tested using independent machine learning models such as Support Vector Machine (SVM), Random Forest (RF) and Logistic Regression (LR). Results: This study has identified 224 differentially expressed proteins involved in various inflammatory and immunological pathways in hospitalised COVID-19 patients compared to non-hospitalised COVID-19 patients. LGALS9 (p-value < 0.001), LAMP3 (p-value < 0.001), PRSS8 (p-value < 0.001) and AGRN (p-value < 0.001) were identified as the most statistically significant proteins. Several hundred rsIDs were queried across the top 10 significant signatures, identifying three significant SNPs on the FSTL3 gene showing a correlation with hospitalisation status. Conclusions: Our study has not only identified key signatures of COVID-19 patients with worsened health but has also demonstrated their predictive capabilities as potential biomarkers, which suggests a staple role in the worsened health effects caused by COVID-19.

Genomic, Proteomic, and Phenotypic Biomarkers of COVID-19 Severity: Protocol for a Retrospective Observational Study.

BACKGROUND: Health organizations and countries around the world have found it difficult to control the spread of COVID-19. To minimize the future impact on the UK National Health Service and improve patient care, there is a pressing need to identify individuals who are at a higher risk of being hospitalized because of severe COVID-19. Early targeted work was successful in identifying angiotensin-converting enzyme-2 receptors and type II transmembrane serine protease dependency as drivers of severe infection. Although a targeted approach highlights key pathways, a multiomics approach will provide a clearer and more comprehensive picture of severe COVID-19 etiology and progression. OBJECTIVE: The COVID-19 Response Study aims to carry out an integrated multiomics analysis to identify biomarkers in blood and saliva that could contribute to host susceptibility to SARS-CoV-2 and the development of severe COVID-19. METHODS: The COVID-19 Response Study aims to recruit 1000 people who recovered from SARS-CoV-2 infection in both community and hospital settings on the island of Ireland. This protocol describes the retrospective observational study component carried out in Northern Ireland (NI; Cohort A); the Republic of Ireland cohort will be described separately. For all NI participants (n=519), SARS-CoV-2 infection has been confirmed by reverse transcription-quantitative polymerase chain reaction. A prospective Cohort B of 40 patients is also being followed up at 1, 3, 6, and 12 months postinfection to assess longitudinal symptom frequency and immune response. Data will be sourced from whole blood, saliva samples, and clinical data from the electronic care records, the general health questionnaire, and a 12-item general health questionnaire mental health survey. Saliva and blood samples were processed to extract DNA and RNA before whole-genome sequencing, RNA sequencing, DNA methylation analysis, microbiome analysis, 16S ribosomal RNA gene sequencing, and proteomic analysis were performed on the plasma. Multiomics data will be combined with clinical data to produce sensitive and specific prognostic models for severity risk. RESULTS: An initial demographic and clinical profile of the NI Cohort A has been completed. A total of 249 hospitalized patients and 270 nonhospitalized patients were recruited, of whom 184 (64.3%) were female, and the mean age was 45.4 (SD 13) years. High levels of comorbidity were evident in the hospitalized cohort, with cardiovascular disease and metabolic and respiratory disorders being the most significant (P<.001), grouped according to the International Classification of Diseases 10 codes. CONCLUSIONS: This study will provide a comprehensive opportunity to study the mechanisms of COVID-19 severity in recontactable participants. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/50733.

Senescence Biomarkers CKAP4 and PTX3 Stratify Severe Kidney Disease Patients.

INTRODUCTION: Cellular senescence is the irreversible growth arrest subsequent to oncogenic mutations, DNA damage, or metabolic insult. Senescence is associated with ageing and chronic age associated diseases such as cardiovascular disease and diabetes. The involvement of cellular senescence in acute kidney injury (AKI) and chronic kidney disease (CKD) is not fully understood. However, recent studies suggest that such patients have a higher-than-normal level of cellular senescence and accelerated ageing. METHODS: This study aimed to discover key biomarkers of senescence in AKI and CKD patients compared to other chronic ageing diseases in controls using OLINK proteomics. RESULTS: We show that senescence proteins CKAP4 (p-value < 0.0001) and PTX3 (p-value < 0.0001) are upregulated in AKI and CKD patients compared with controls with chronic diseases, suggesting the proteins may play a role in overall kidney disease development. CONCLUSIONS: CKAP4 was found to be differentially expressed in both AKI and CKD when compared to UHCs; hence, this biomarker could be a prognostic senescence biomarker of both AKI and CKD.

Combining Polygenic and Proteomic Risk Scores With Clinical Risk Factors to Improve Performance for Diagnosing Absence of Coronary Artery Disease in Patients With de novo Chest Pain.

BACKGROUND: Patients with de novo chest pain, referred for evaluation of possible coronary artery disease (CAD), frequently have an absence of CAD resulting in millions of tests not having any clinical impact. The objective of this study was to investigate whether polygenic risk scores and targeted proteomics improve the prediction of absence of CAD in patients with suspected CAD, when added to the PROMISE (Prospective Multicenter Imaging Study for Evaluation of Chest Pain) minimal risk score (PMRS). METHODS: Genotyping and targeted plasma proteomics (N=368 proteins) were performed in 1440 patients with symptoms suspected to be caused by CAD undergoing coronary computed tomography angiography. Based on individual genotypes, a polygenic risk score for CAD (PRSCAD) was calculated. The prediction was performed using combinations of PRSCAD, proteins, and PMRS as features in models using stability selection and machine learning. RESULTS: Prediction of absence of CAD yielded an area under the curve of PRSCAD-model, 0.64±0.03; proteomic-model, 0.58±0.03; and PMRS model, 0.76±0.02. No significant correlation was found between the genetic and proteomic risk scores (Pearson correlation coefficient, -0.04; P=0.13). Optimal predictive ability was achieved by the full model (PRSCAD+protein+PMRS) yielding an area under the curve of 0.80±0.02 for absence of CAD, significantly better than the PMRS model alone (P<0.001). For reclassification purpose, the full model enabled down-classification of 49% (324 of 661) of the 5% to 15% pretest probability patients and 18% (113 of 611) of >15% pretest probability patients. CONCLUSIONS: For patients with chest pain and low-intermediate CAD risk, incorporating targeted proteomics and polygenic risk scores into the risk assessment substantially improved the ability to predict the absence of CAD. Genetics and proteomics seem to add complementary information to the clinical risk factors and improve risk stratification in this large patient group. REGISTRATION: URL: https://www. CLINICALTRIALS: gov; Unique identifier: NCT02264717.

TACE/ADAM17 substrates associate with ACS (Ep-CAM, HB-EGF) and follow-up MACE (TNFR1 and TNFR2).

Background and aims: TACE/ADAM17 is a membrane bound metalloprotease, which cleaves substrates involved in immune and inflammatory responses and plays a role in coronary artery disease (CAD). We measured TACE and its substrates in CAD patients to identify potential biomarkers within this molecular pathway with potential for acute coronary syndrome (ACS) and major adverse cardiovascular events (MACE) prediction. Methods: Blood samples were obtained from consecutive patients (n = 229) with coronary angiographic evidence of CAD admitted with ACS or electively. MACE were recorded after a median 3-year follow-up. Controls (n = 115) had a <10% CAD risk as per the HeartSCORE. TACE and TIMP3 protein and mRNA levels were measured by ELISA and RT-qPCR respectively. TACE substrates were measured using a multiplex proximity extension assay. Results: TACE mRNA and cell protein levels (p < 0.01) and TACE substrates LDLR (p = 0.006), TRANCE (p = 0.045), LAG-3 (p < 0.001) and ACE2 (p < 0.001) plasma levels were significantly higher in CAD patients versus controls. TACE inhibitor TIMP3 mRNA levels were significantly lower in CAD patients and tended to be lower in the ACS population (p < 0.05). TACE substrates TNFR1 (OR:3.237,CI:1.514-6.923,p = 0.002), HB-EGF (OR:0.484,CI:0.288-0.813,p = 0.006) and Ep-CAM (OR:0.555,CI:0.327-0.829,p = 0.004) accurately classified ACS patients with HB-EGF and Ep-CAM levels being lower compared to electively admitted patients. TNFR1 (OR:2.317,CI:1.377-3.898,p = 0.002) and TNFR2 (OR:1.902,CI:1.072-3.373,p = 0.028) were significantly higher on admission in those patients who developed MACE within 3 years. Conclusions: We demonstrate a possible role of TACE substrates LAG-3, HB-EGF and Ep-CAM in atherosclerotic plaque development and stability. We also underline the importance of measuring TNFR1 and TNFR2 earlier than previously appreciated for MACE prediction. We report an important role of TIMP3 in regulating TACE levels.

Using latent class analysis to identify clinical features of patients with occlusive myocardial infarction: Preangiogram prediction remains difficult.

BACKGROUND: Treatment decisions in myocardial infarction (MI) are currently stratified by ST elevation (ST-elevation myocardial infarction [STEMI]) or lack of ST elevation (non-ST elevation myocardial infarction [NSTEMI]) on the electrocardiogram. This arose from the assumption that ST elevation indicated acute coronary artery occlusion (OMI). However, one-quarter of all NSTEMI cases are an OMI, and have a higher mortality. The purpose of this study was to identify features that could help identify OMI. METHODS: Prospectively collected data from patients undergoing percutaneous coronary intervention (PCI) was analyzed. Data included presentation characteristics, comorbidities, treatments, and outcomes. Latent class analysis was undertaken, to determine patterns of presentation and history associated with OMI. RESULTS: A total of 1412 patients underwent PCI for acute MI, and 263 were diagnosed as OMI. Compared to nonocclusive MI, OMI patients are more likely to have fewer comorbidities but no difference in cerebrovascular disease and increased acute mortality (4.2% vs. 1.1%; p < .001). Of OMI, 29.5% had delays to their treatment such as immediate reperfusion therapy. With latent class analysis, while clusters of similar patients are observed in the data set, the data available did not usefully identify patients with OMI compared to non-OMI. CONCLUSION: Features between OMI and STEMI are broadly very similar. However, there was no difference in age and risk of cerebrovascular disease in the OMI/non-OMI group. There are no reliable characteristics therefore for identifying OMI versus non-OMI. Delays to treatment also suggest that OMI patients are still missing out on optimal treatment. An alternative strategy is required to improve the identification of OMI patients.

Machine learning and the electrocardiogram over two decades: Time series and meta-analysis of the algorithms, evaluation metrics and applications.

BACKGROUND: The application of artificial intelligence to interpret the electrocardiogram (ECG) has predominantly included the use of knowledge engineered rule-based algorithms which have become widely used today in clinical practice. However, over recent decades, there has been a steady increase in the number of research studies that are using machine learning (ML) to read or interrogate ECG data. OBJECTIVE: The aim of this study is to review the use of ML with ECG data using a time series approach. METHODS: Papers that address the subject of ML and the ECG were identified by systematically searching databases that archive papers from January 1995 to October 2019. Time series analysis was used to study the changing popularity of the different types of ML algorithms that have been used with ECG data over the past two decades. Finally, a meta-analysis of how various ML techniques performed for various diagnostic classifications was also undertaken. RESULTS: A total of 757 papers was identified. Based on results, the use of ML with ECG data started to increase sharply (p < 0.001) from 2012. Healthcare applications, especially in heart abnormality classification, were the most common application of ML when using ECG data (p < 0.001). However, many new emerging applications include using ML and the ECG for biometrics and driver drowsiness. The support vector machine was the technique of choice for a decade. However, since 2018, deep learning has been trending upwards and is likely to be the leading technique in the coming few years. Despite the accuracy paradox, accuracy was the most frequently used metric in the studies reviewed, followed by sensitivity, specificity, F1 score and then AUC. CONCLUSION: Applying ML using ECG data has shown promise. Data scientists and physicians should collaborate to ensure that clinical knowledge is being applied appropriately and is informing the design of ML algorithms. Data scientists also need to consider knowledge guided feature engineering and the explicability of the ML algorithm as well as being transparent in the algorithm's performance to appropriately calibrate human-AI trust. Future work is required to enhance ML performance in ECG classification.

Spherical aberration and higher order aberrations with Balafilcon A (PureVision) and Comfilcon A (Biofinity).

BACKGROUND: Advances in the optical design of soft contact lenses have seen certain manufacturers incorporate aspheric optics into soft lenses in an attempt to reduce spherical aberration, to provide superior visual performance. The aim of this study is to determine the on-eye differences in spherical aberration and higher order aberrations (HOA) between the Bausch and Lomb PureVision (Balafilcon A) and the CooperVision Biofinity (Comfilcon A). METHODS: Twenty subjects were recruited in a prospective, randomized, unilateral study. The right eye was dilated and HOA measured with the NIDEK OPD-Scan. Each eye was fitted randomly with a -3.00D PureVision and a -3.00D Biofinity, and HOA were measured with lenses in situ across a 6 mm pupil. Paired t-tests were performed to determine HOA differences with the lenses in situ compared to baseline. RESULTS: Aberrometry was successfully performed on all subjects. Statistical analysis indicated no changes in spherical aberration, but changes in other HOA. With the PureVision, there were increases in Zernike terms Z (3) (1) (from 0.01 µm to -0.11 µm), Z (4) (-2) (from 0.01 µm to 0.13 µm) and Z (5) (-1) (from -0.01 µm to 0.03 µm). With the Biofinity there was an increase in Zernike term Z (3) (3) (from 0.00 µm to 0.09 µm). CONCLUSIONS: No statistically significant changes occurred in spherical aberration. The PureVision caused statistically significant increases in Z (3) (1) , Z (4) (-2) and Z (5) (-1) , and the Biofinity caused an increase in Z (3) (3) . Clinically significant changes (>0.1 µm) occurred with terms Z (3) (1) and Z (4) (-2) with the PureVision only.