Territory-wide cohort study of Brugada syndrome in Hong Kong: predictors of long-term outcomes using random survival forests and non-negative matrix factorisation

Applying Artificial Intelligence for Phenotyping of Inherited Arrhythmia Syndromes

Inherited arrhythmia disorders account for a significant proportion of sudden cardiac death, particularly among young individuals. Recent advances in our understanding of these syndromes have improved patient diagnosis and care, yet certain clinical gaps remain, particularly within case ascertainment, access to genetic testing and risk stratification. Artificial intelligence (AI), specifically machine learning and its subset deep learning, present promising solutions to these challenges. The capacity of AI to process vast amounts of patient data and identify disease patterns differentiates them from traditional methods, which are time and resource intensive. To date, AI models have demonstrated immense potential in condition detection (including asymptomatic/concealed disease) and genotype and phenotype identification, exceeding expert cardiologists in these tasks. Additionally, they have exhibited applicability for general population screening, improving case ascertainment in a set of conditions that are often asymptomatic such as left ventricular dysfunction. Third, models have displayed ability to improve testing protocols, as through model identification of disease and genotype, specific clinical testing (e.g. drug challenges or further diagnostic imaging) can be avoided, reducing health care expenses, speeding diagnosis, and possibly allowing for more incremental or targeted genetic testing approaches. These significant benefits warrant continued investigation of the field, particularly regarding the development and implementation of clinically applicable screening tools. This review summarizes key developments in the field, including studies in Long QT Syndrome, Brugada Syndrome, Hypertrophic Cardiomyopathy, and Arrhythmogenic Cardiomyopathies, and provides direction for effective future AI implementation in clinical practice.

Risk Prediction for Sudden Cardiac Death in the General Population: A Systematic Review and Meta-Analysis

Objective: Identification of SCD risk is important in the general population from a public health perspective. The objective is to summarize and appraise the available prediction models for the risk of SCD among the general population. Methods: Data were obtained searching six electronic databases and reporting prediction models of SCD risk in the general population. Studies with duplicate cohorts and missing information were excluded from the meta-analysis. Results: Out of 8,407 studies identified, fifteen studies were included in the systematic review, while five studies were included in the meta-analysis. The Cox proportional hazards model was used in thirteen studies (96.67%). Study locations were limited to Europe and the United States. Our pooled meta-analyses included four predictors: diabetes mellitus (ES = 2.69, 95%CI: 1.93, 3.76), QRS duration (ES = 1.16, 95%CI: 1.06, 1.26), spatial QRS-T angle (ES = 1.46, 95%CI: 1.27, 1.69) and factional shortening (ES = 1.37, 95%CI: 1.15, 1.64). Conclusion: Risk prediction model may be useful as an adjunct for risk stratification strategies for SCD in the general population. Further studies among people except for white participants and more accessible factors are necessary to explore.

Healthcare Big Data in Hong Kong: Development and Implementation of Artificial Intelligence-Enhanced Predictive Models for Risk Stratification

Routinely collected electronic health records (EHRs) data contain a vast amount of valuable information for conducting epidemiological studies. With the right tools, we can gain insights into disease processes and development, identify the best treatment and develop accurate models for predicting outcomes. Our recent systematic review has found that the number of big data studies from Hong Kong has rapidly increased since 2015, with an increasingly common application of artificial intelligence (AI). The advantages of big data are that i) the models developed are highly generalisable to the population, ii) multiple outcomes can be determined simultaneously, iii) ease of cross-validation by for model training, development and calibration, iv) huge numbers of useful variables can be analyzed, v) static and dynamic variables can be analyzed, vi) non-linear and latent interactions between variables can be captured, vii) artificial intelligence approaches can enhance the performance of prediction models. In this paper, we will provide several examples (cardiovascular disease, diabetes mellitus, Brugada syndrome, long QT syndrome) to illustrate efforts from a multi-disciplinary team to identify data from different modalities to develop models using territory-wide datasets, with the possibility of real-time risk updates by using new data captured from patients. The benefit is that only routinely collected data are required for developing highly accurate and high-performance models. AI-driven models outperform traditional models in terms of sensitivity, specificity, accuracy, area under the receiver operating characteristic and precision-recall curve, and F1 score. Web and/or mobile versions of the risk models allow clinicians to risk stratify patients quickly in clinical settings, thereby enabling clinical decision-making. Efforts are required to identify the best ways of implementing AI algorithms on the web and mobile apps.

Secular trends of health care resource utilization and costs between Brugada syndrome and congenital long QT syndrome: A territory-wide study

BACKGROUND

Health care resource utilization (HCRU) and costs are important metrics of health care burden, but they have rarely been explored in the setting of cardiac ion channelopathies.

HYPOTHESIS

This study tested the hypothesis that attendance-related HCRUs and costs differed between patients with Brugada syndrome (BrS) and congenital long QT syndrome (LQTS).

METHODS

This was a retrospective cohort study of consecutive BrS and LQTS patients at public hospitals or clinics in Hong Kong, China. HCRUs and costs (in USD) for Accident and Emergency (A&E), inpatient, general outpatient and specialist outpatient attendances were analyzed between 2001 and 2019 at the cohort level. Comparisons were made using incidence rate ratios (IRRs [95% confidence intervals]).

RESULTS

Over the 19-year period, 516 BrS (median age of initial presentation: 51 [interquartile range: 38-61] years, 92% male) and 134 LQTS (median age of initial presentation: 21 [9-44] years, 32% male) patients were included. Compared to LQTS patients, BrS patients had lower total costs (2 008 126 [2 007 622-2 008 629] vs. 2 343 864 [2 342 828-2 344 900]; IRR: 0.857 [0.855-0.858]), higher costs for A&E attendances (83 113 [83 048-83 177] vs. 70 604 [70 487-70 721]; IRR: 1.177 [1.165-1.189]) and general outpatient services (2,176 [2,166-2,187] vs. 921 [908-935]; IRR: 2.363 [2.187-2.552]), but lower costs for inpatient stay (1 391 624 [1 391 359-1 391 889] vs. 1 713 742 [1 713 166-1 714 319]; IRR: 0.812 [0.810-0.814]) and lower costs for specialist outpatient services (531 213 [531 049-531 376] vs. 558 597 [558268-558926]; IRR: 0.951 [0.947-0.9550]).

CONCLUSIONS

Overall, BrS patients consume 14% less health care resources compared to LQTS patients in terms of attendance costs. BrS patients require more A&E and general outpatient services, but less inpatient and specialist outpatient services than LQTS patients.

Ethnic and racial differences in Asian populations with ion channelopathies associated with sudden cardiac death

Cardiovascular diseases are associated with several morbidities and are the most common cause of worldwide disease-related fatalities. Studies show that treatment and outcome-related differences for cardiovascular diseases disproportionately affect minorities in the United States. The emergence of ethnic and racial differences in sudden cardiac death (SCD) and related ion channelopathies complicates cardiovascular disease prevention, diagnosis, management, prognosis, and treatment objectives for patients and physicians alike. This review compiles and synthesizes current research in cardiac ion channelopathies and genetic disorders in Asian populations, an underrepresented population in cardiovascular literature. We first present a brief introduction to SCD, noting relevant observations and statistics from around the world, including Asian populations. We then examined existing differences between Asian and White populations in research, treatment, and outcomes related to cardiac ion channelopathies and SCD, showing progression in thought and research over time for each ion channelopathy. The review also identifies research that explored phenotypic abnormalities, device usage, and risk of death in Asian patients. We touch upon the unique genetic risk factors in Asian populations that lead to cardiac ion channelopathies and SCD while comparing them to White and Western populations, particularly in the United States, where Asians comprise approximately 7% of the total population. We also propose potential solutions such as improving early genetic screening, addressing barriers affecting access to medical care and device utilization, physician training, and patient education on risks.

Unmasking a Silent Killer and Understanding Sudden Cardiac Death in Brugada Syndrome: A Traditional Review

Brugada syndrome (BrS) is an intricate and heterogeneous genetic disorder that engenders a formidable risk of life-threatening ventricular arrhythmias (VAs). While initially regarded as an electrophysiological aberration, emergent studies have illuminated the presence of underlying structural anomalies in select BrS cases. Although mutations in the SCN5A gene encoding the α-subunit of the cardiac sodium channel were originally identified as a primary causative factor; they account for only a fraction of the syndrome's multifaceted complexity pointing at genetic heterogeneity as a contributing factor. Remarkably, BrS has been linked to a higher incidence of fatal arrhythmic incidents and sudden cardiac death (SCD) with about 4% of SCD cases thought to be caused by BrS. Patients who spontaneously exhibit type one Brugada ECGs are more likely to experience cardiac events, emphasizing the importance of early risk stratification. To aid in risk stratification, the Shanghai score; a multifactorial risk stratification scoring system that incorporates ECG, clinical history, family history, and genetic test results; is utilized to identify those most susceptible to SCD. Beyond single ECGs, evaluation of arrhythmic findings from 24-hour Holter monitoring, ECG variables, electrophysiologic study (EPS) status in the temporal domain, and EPS data collected over time are all critical factors in risk classification. Among management options avoidance of triggers, early risk stratification, and implantation of an Implantable Cardioverter-Defibrillator (ICD) are recommended for asymptomatic patients. For symptomatic patients, pharmacotherapy and ICD implantation are available, with the latter being a highly effective choice for treating and preventing lethal arrhythmias in BrS.

Artificial Intelligence in Ventricular Arrhythmias and Sudden Death

Sudden cardiac arrest due to lethal ventricular arrhythmias is a major cause of mortality worldwide and results in more years of potential life lost than any individual cancer. Most of these sudden cardiac arrest events occur unexpectedly in individuals who have not been identified as high-risk due to the inadequacy of current risk stratification tools. Artificial intelligence tools are increasingly being used to solve complex problems and are poised to help with this major unmet need in the field of clinical electrophysiology. By leveraging large and detailed datasets, artificial intelligence-based prediction models have the potential to enhance the risk stratification of lethal ventricular arrhythmias. This review presents a synthesis of the published literature and a discussion of future directions in this field.

Association between sick sinus syndrome and atrial fibrillation: A systematic review and meta-analysis

AIMS

Sick sinus syndrome (SSS) and atrial fibrillation (AF) frequently coexist and show a bidirectional relationship. This systematic review and meta-analysis aimed to decipher the precise relationship between SSS and AF, further exploring and comparing different therapy strategies on the occurrence or progression of AF in patients with SSS.

METHODS AND RESULTS

A systematic literature search was conducted until November 2022. A total of 35 articles with 37,550 patients were included. Patients with SSS were associated with new-onset AF compared to those without SSS. Catheter ablation was associated with a lower risk of AF recurrence, AF progression, all-cause mortality, stroke and hospitalization of heart failure compared to pacemaker therapy. Regarding the different pacing strategies for SSS, VVI/VVIR has higher risk of new-onset AF than DDD/DDDR. No significant difference was found between AAI/AAIR and DDD/DDDR, as well as between DDD/DDDR and minimal ventricular pacing (MVP) for AF recurrence. AAI/AAIR was associated with higher risk of all-cause mortality when compared to DDD/DDDR, but lower risk of cardiac death when compared to DDD/DDDR. Right atrial septum pacing was associated with a similar risk of new-onset AF or AF recurrence compared to right atrial appendage pacing.

CONCLUSION

SSS is associated with a higher risk of AF. For patients with both SSS and AF, catheter ablation should be considered. This meta-analysis re-emphasizes that high percentage of ventricular pacing should be avoided in patients with SSS in order to decrease AF burden and mortality.

Sudden cardiac death multiparametric classification system for Chagas heart disease's patients based on clinical data and 24-hours ECG monitoring

About 6.5 million people are infected with Chagas disease (CD) globally, and WHO estimates that $ > million people worldwide suffer from ChHD. Sudden cardiac death (SCD) represents one of the leading causes of death worldwide and affects approximately 65% of ChHD patients at a rate of 24 per 1000 patient-years, much greater than the SCD rate in the general population. Its occurrence in the specific context of ChHD needs to be better exploited. This paper provides the first evidence supporting the use of machine learning (ML) methods within non-invasive tests: patients' clinical data and cardiac restitution metrics (CRM) features extracted from ECG-Holter recordings as an adjunct in the SCD risk assessment in ChHD. The feature selection (FS) flows evaluated 5 different groups of attributes formed from patients' clinical and physiological data to identify relevant attributes among 57 features reported by 315 patients at HUCFF-UFRJ. The FS flow with FS techniques (variance, ANOVA, and recursive feature elimination) and Naive Bayes (NB) model achieved the best classification performance with 90.63% recall (sensitivity) and 80.55% AUC. The initial feature set is reduced to a subset of 13 features (4 Classification; 1 Treatment; 1 CRM; and 7 Heart Tests). The proposed method represents an intelligent diagnostic support system that predicts the high risk of SCD in ChHD patients and highlights the clinical and CRM data that most strongly impact the final outcome.

Attendance-related Healthcare Resource Utilisation and Costs in Patients With Brugada Syndrome in Hong Kong: A Retrospective Cohort Study

Understanding health care resource utilisation and its associated costs are important for identifying areas of improvement regarding resource allocations. However, there is limited research exploring this issue in the setting of Brugada syndrome (BrS).This was a retrospective territory-wide study of BrS patients from Hong Kong. Healthcare resource utilisation for accident and emergency (A&E), inpatient and specialist outpatient attendances were analyzed over a 19-year period, with their associated costs presented in US dollars. A total of 507 BrS patients with a mean presentation age of 49.9 ± 16.3 years old were included. Of these, 384 patients displayed spontaneous type 1 electrocardiographic (ECG) Brugada pattern and 77 patients had presented with ventricular tachycardia/ventricular fibrillation (VT/VF). At the individual patient level, the median annualized costs were $110 (52-224) at the (A&E) setting, $6812 (1982-32414) at the inpatient setting and $557 (326-1001) for specialist outpatient attendances. Patients with initial VT/VF presentation had overall greater costs in inpatient ($20161 [9147-189215] vs $5290 [1613-24937],P < 0.0001) and specialist outpatient setting ($776 [438-1076] vs $542 [293-972],P = 0.015) compared to those who did not present VT. In addition, patients without Type 1 ECG pattern had greater median costs in the specialist outpatient setting ($7036 [3136-14378] vs $4895 [2409-10554],p=0.019). There is a greater health care demand in the inpatient and specialist outpatient settings for BrS patients. The most expensive attendance type was inpatient setting stay at $6812 per year. The total median annualized cost of BrS patients without VT/VF presentation was 78% lower compared to patients with VT/VF presentation.

Machine learning in sudden cardiac death risk prediction: a systematic review

AIMS

Most patients who receive implantable cardioverter defibrillators (ICDs) for primary prevention do not receive therapy during the lifespan of the ICD, whilst up to 50% of sudden cardiac death (SCD) occur in individuals who are considered low risk by conventional criteria. Machine learning offers a novel approach to risk stratification for ICD assignment.

METHODS AND RESULTS

Systematic search was performed in MEDLINE, Embase, Emcare, CINAHL, Cochrane Library, OpenGrey, MedrXiv, arXiv, Scopus, and Web of Science. Studies modelling SCD risk prediction within days to years using machine learning were eligible for inclusion. Transparency and quality of reporting (TRIPOD) and risk of bias (PROBAST) were assessed. A total of 4356 studies were screened with 11 meeting the inclusion criteria with heterogeneous populations, methods, and outcome measures preventing meta-analysis. The study size ranged from 122 to 124 097 participants. Input data sources included demographic, clinical, electrocardiogram, electrophysiological, imaging, and genetic data ranging from 4 to 72 variables per model. The most common outcome metric reported was the area under the receiver operator characteristic (n = 7) ranging between 0.71 and 0.96. In six studies comparing machine learning models and regression, machine learning improved performance in five. No studies adhered to a reporting standard. Five of the papers were at high risk of bias.

CONCLUSION

Machine learning for SCD prediction has been under-applied and incorrectly implemented but is ripe for future investigation. It may have some incremental utility in predicting SCD over traditional models. The development of reporting standards for machine learning is required to improve the quality of evidence reporting in the field.

Comparing the performance of published risk scores in Brugada syndrome: a multi-center cohort study

INTRODUCTION

The management of Brugada Syndrome (BrS) patients at intermediate risk of arrhythmic events remains controversial. The present study evaluated the predictive performance of different risk scores in an Asian BrS population and its intermediate risk subgroup.

METHODS

This retrospective cohort study included consecutive patients diagnosed with BrS from January 1^st, 1997 to June 20^th, 2020 from Hong Kong. The primary outcome is sustained ventricular tachyarrhythmias. Two novel risk risk scores and seven machine learning-based models (random survival forest, Ada boost classifier, Gaussian naïve Bayes, light gradient boosting machine, random forest classifier, gradient boosting classifier and decision tree classifier) were developed. The area under the receiver operator characteristic (ROC) curve (AUC) [95% confidence intervals] was compared between the different models.

RESULTS

This study included 548 consecutive BrS patients (7% female, age at diagnosis: 50±16 years, follow-up: 84±55 months). For the whole cohort, the score developed by Sieira et al. showed the best performance (AUC: 0.806 [0.747-0.865]). A novel risk score was developed using the Sieira score and additional variables significant on univariable Cox regression (AUC: 0.855 [0.808-0.901]). A simpler score based on non-invasive results only showed a statistically comparable AUC (0.784 [0.724-0.845]), improved using random survival forests (AUC: 0.942 [0.913-0.964]). For the intermediate risk subgroup (N=274), a gradient boosting classifier model showed the best performance (AUC: 0.814 [0.791-0.832]).

CONCLUSION

A simple risk score based on clinical and electrocardiographic variables showed a good performance for predicting VT/VF, improved using machine learning. Abstract: The management of Brugada Syndrome (BrS) patients at intermediate risk of arrhythmic events remains controversial. This study evaluated the predictive performance of published risk scores in a cohort of BrS patients from Hong Kong (N=548) and its intermediate risk subgroup (N=274). A novel risk score developed by modifying the best performing existing score (by. Sieira et al.) showed an area under the curve of 0.855 and 0.760 for the whole BrS cohort and the intermediate risk subgroup, respectively. The performance of the different scores was significantly improved machine learning-based methods, such as random survival forests and gradient boosting classifier.

Development of an Electronic Frailty Index for Predicting Mortality and Complications Analysis in Pulmonary Hypertension Using Random Survival Forest Model

Background

The long-term prognosis of the cardio-metabolic and renal complications, in addition to mortality in patients with newly diagnosed pulmonary hypertension, are unclear. This study aims to develop a scalable predictive model in the form of an electronic frailty index (eFI) to predict different adverse outcomes.

Methods

This was a population-based cohort study of patients diagnosed with pulmonary hypertension between January 1st, 2000 and December 31st, 2017, in Hong Kong public hospitals. The primary outcomes were mortality, cardiovascular complications, renal diseases, and diabetes mellitus. The univariable and multivariable Cox regression analyses were applied to identify the significant risk factors, which were fed into the non-parametric random survival forest (RSF) model to develop an eFI.

Results

A total of 2,560 patients with a mean age of 63.4 years old (interquartile range: 38.0–79.0) were included. Over a follow-up, 1,347 died and 1,878, 437, and 684 patients developed cardiovascular complications, diabetes mellitus, and renal disease, respectively. The RSF-model-identified age, average readmission, anti-hypertensive drugs, cumulative length of stay, and total bilirubin were among the most important risk factors for predicting mortality. Pair-wise interactions of factors including diagnosis age, average readmission interval, and cumulative hospital stay were also crucial for the mortality prediction. Patients who developed all-cause mortality had higher values of the eFI compared to those who survived (P < 0.0001). An eFI ≥ 9.5 was associated with increased risks of mortality [hazard ratio (HR): 1.90; 95% confidence interval [CI]: 1.70–2.12; P < 0.0001]. The cumulative hazards were higher among patients who were 65 years old or above with eFI ≥ 9.5. Using the same cut-off point, the eFI predicted a long-term mortality over 10 years (HR: 1.71; 95% CI: 1.53–1.90; P < 0.0001). Compared to the multivariable Cox regression, the precision, recall, area under the curve (AUC), and C-index were significantly higher for RSF in the prediction of outcomes.

Conclusion

The RSF models identified the novel risk factors and interactions for the development of complications and mortality. The eFI constructed by RSF accurately predicts the complications and mortality of patients with pulmonary hypertension, especially among the elderly.

Machine learning techniques for arrhythmic risk stratification: a review of the literature

Ventricular arrhythmias (VAs) and sudden cardiac death (SCD) are significant adverse events that affect the morbidity and mortality of both the general population and patients with predisposing cardiovascular risk factors. Currently, conventional disease-specific scores are used for risk stratification purposes. However, these risk scores have several limitations, including variations among validation cohorts, the inclusion of a limited number of predictors while omitting important variables, as well as hidden relationships between predictors. Machine learning (ML) techniques are based on algorithms that describe intervariable relationships. Recent studies have implemented ML techniques to construct models for the prediction of fatal VAs. However, the application of ML study findings is limited by the absence of established frameworks for its implementation, in addition to clinicians’ unfamiliarity with ML techniques. This review, therefore, aims to provide an accessible and easy-to-understand summary of the existing evidence about the use of ML techniques in the prediction of VAs. Our findings suggest that ML algorithms improve arrhythmic prediction performance in different clinical settings. However, it should be emphasized that prospective studies comparing ML algorithms to conventional risk models are needed while a regulatory framework is required prior to their implementation in clinical practice.

Pathogenesis and Management of Brugada Syndrome: Recent Advances and Protocol for Umbrella Reviews of Meta-Analyses in Major Arrhythmic Events Risk Stratification

Brugada syndrome (BrS) is a primary electrical disease associated with life-threatening arrhythmias. It is estimated to cause at least 20% of sudden cardiac deaths (SCDs) in patients with normal cardiac anatomy. In this review paper, we discuss recent advances in complex BrS pathogenesis, diagnostics, and current standard approaches to major arrhythmic events (MAEs) risk stratification. Additionally, we describe a protocol for umbrella reviews to systematically investigate clinical, electrocardiographic, electrophysiological study, programmed ventricular stimulation, and genetic factors associated with BrS, and the risk of MAEs. Our evaluation will include MAEs such as sustained ventricular tachycardia, ventricular fibrillation, appropriate implantable cardioverter–defibrillator therapy, sudden cardiac arrest, and SCDs from previous meta-analytical studies. The protocol was written following the Preferred Reporting Items for Systematic review and Meta-Analysis Protocols (PRISMA-P) guidelines. We plan to extensively search PubMed, Embase, and Scopus databases for meta-analyses concerning risk-stratification in BrS. Data will be synthesized integratively with transparency and accuracy. Heterogeneity patterns across studies will be reported. The Joanna Briggs Institute (JBI) methodology, A MeaSurement Tool to Assess systematic Reviews 2 (AMSTAR 2), and the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) are planned to be applied for design and execution of our evidence-based research. To the best of our knowledge, these will be the first umbrella reviews to critically evaluate the current state of knowledge in BrS risk stratification for life-threatening ventricular arrhythmias, and will potentially contribute towards evidence-based guidance to enhance clinical decisions.

Syncope Prediction Scores in the Emergency Department

Syncope is a common clinical presentation defined as a transient loss of consciousness (TLOC) due to cerebral hypoperfusion, characterized by a rapid onset, short duration, and spontaneous complete recovery. Different clinical decision rules (CDRs) and risk stratification scores have been developed to predict short- and long-term risks for adverse outcomes after syncope. The central theme of these prediction systems is consistent with the ESC syncope guidelines. Initial assessment according to the ESC guideline is essential until an optimal and well-validated risk score is available. The focus should be accurate risk stratification to allow prevention of adverse outcomes and optimize the use of limited healthcare resources. In this review article, we summarize and critically appraise the evidence regarding the CDRs for patients presenting with syncope.

Paediatric/young versus adult patients with long QT syndrome

Introduction

Long QT syndrome (LQTS) is a less prevalent cardiac ion channelopathy than Brugada syndrome in Asia. The present study compared the outcomes between paediatric/young and adult LQTS patients.

Methods

This was a population-based retrospective cohort study of consecutive patients diagnosed with LQTS attending public hospitals in Hong Kong. The primary outcome was spontaneous ventricular tachycardia/ventricular fibrillation (VT/VF).

Results

A total of 142 LQTS (mean onset age=27±23 years old) were included. Arrhythmias other than VT/VF (HR 4.67, 95% CI (1.53 to 14.3), p=0.007), initial VT/VF (HR=3.25 (95% CI 1.29 to 8.16), p=0.012) and Schwartz score (HR=1.90 (95% CI 1.11 to 3.26), p=0.020) were predictive of the primary outcome for the overall cohort, while arrhythmias other than VT/VF (HR=5.41 (95% CI 1.36 to 21.4), p=0.016) and Schwartz score (HR=4.67 (95% CI 1.48 to 14.7), p=0.009) were predictive for the adult subgroup (>25 years old; n=58). A random survival forest model identified initial VT/VF, Schwartz score, initial QTc interval, family history of LQTS, initially asymptomatic and arrhythmias other than VT/VF as the most important variables for risk prediction.

Conclusion

Clinical and ECG presentation varies between the paediatric/young and adult LQTS population. Machine learning models achieved more accurate VT/VF prediction.

Supervised machine learning for the assessment of Chronic Kidney Disease advancement

BACKGROUND AND OBJECTIVE

Chronic Kidney Disease (CKD) is a condition characterized by a progressive loss of kidney function over time caused by many diseases. The most effective weapons against CKD are early diagnosis and treatment, which in most of the cases can only postpone the onset of complete kidney failure. The CKD grading system is classified based on the estimated Glomerular Filtration Rate (eGFR), and it helps to stratify patients for risk, follow up and management planning. This study aims to effectively predict how soon a CKD patient will need to be dialyzed, thus allowing personalized care and strategic planning of treatment.

METHODS

To accurately predict the time frame within which a CKD patient will necessarily have to be dialyzed, a computational model based on a supervised machine learning approach is developed. Many techniques, regarding both information extraction and model training phases, are compared in order to understand which approaches are most effective. The different models compared are trained on the data extracted from the Electronic Medical Records of the Vimercate Hospital.

RESULTS

As final model, we propose a set of Extremely Randomized Trees classifiers considering 27 features, including creatinine level, urea, red blood cells count, eGFR trend (which is not even the most important), age and associated comorbidities. In predicting the occurrence of complete renal failure within the next year rather than later, it obtains a test accuracy of 94%, specificity of 91% and sensitivity of 96%. More and shorter time-frame intervals, up to 6 months of granularity, can be specified without relevantly worsening the model performance.

CONCLUSIONS

The developed computational model provides nephrologists with a great support in predicting the patient's clinical pathway. The model promising results, coupled with the knowledge and experience of the clinicians, can effectively lead to better personalized care and strategic planning of both patient's needs and hospital resources.

Predictive scores for identifying patients with type 2 diabetes mellitus at risk of acute myocardial infarction and sudden cardiac death

Introduction

The present study evaluated the application of incorporating non‐linear J/U‐shaped relationships between mean HbA1c and cholesterol levels into risk scores for predicting acute myocardial infarction (AMI) and non‐AMI‐related sudden cardiac death (SCD) respectively, amongst patients with type 2 diabetes mellitus.

Methods

This was a territory‐wide cohort study of patients with type 2 diabetes mellitus above the age 40 and free from prior AMI and SCD, with or without prescriptions of anti‐diabetic agents between January 1st, 2009 to December 31st, 2009 at government‐funded hospitals and clinics in Hong Kong. Patients recruited were followed up until 31 December 2019 or their date of death. Risk scores were developed for predicting incident AMI and non‐AMI‐related SCD. The performance of conditional inference survival forest (CISF) model compared to that of random survival forests (RSF) model and multivariate Cox model.

Results

This study included 261 308 patients (age = 66.0 ± 11.8 years old, male = 47.6%, follow‐up duration = 3552 ± 1201 days, diabetes duration = 4.77 ± 2.29 years). Mean HbA1c and low high‐density lipoprotein‐cholesterol (HDL‐C) were significant predictors of AMI on multivariate Cox regression. Mean HbA1c was linearly associated with AMI, whilst HDL‐C was inversely associated with AMI. Mean HbA1c and total cholesterol were significant multivariate predictors with a J‐shaped relationship with non‐AMI‐related SCD. The AMI and SCD risk scores had an area under the curve (AUC) of 0.666 (95% confidence interval (CI) = [0.662, 0.669]) and 0.677 (95% CI = [0.673, 0.682]), respectively. CISF significantly improves prediction performance of both outcomes compared to RSF and multivariate Cox models.

Conclusion

A holistic combination of demographic, clinical and laboratory indices can be used for the risk stratification of patients with type 2 diabetes mellitus for AMI and SCD.