Montenegrin Digital Academic Innovation Hub: Experience in Supporting Digital Health Innovations.

Montenegrin Digital Academic Innovation Hub established within Erasmus+ project DigNEST is essential institutional support for developing innovations in the field of health in academic-business cooperation and partnership. Experience of 18 months in running Hub service provides preliminary results in analysis received innovation ideas, provided support and potentials/capacities in medical informatics advancements at national, regional and global level.

Demystifying XAI: Requirements for Understandable XAI Explanations.

This paper establishes requirements for assessing the usability of Explainable Artificial Intelligence (XAI) methods, focusing on non-AI experts like healthcare professionals. Through a synthesis of literature and empirical findings, it emphasizes achieving optimal cognitive load, task performance, and task time in XAI explanations. Key components include tailoring explanations to user expertise, integrating domain knowledge, and using non-propositional representations for comprehension. The paper highlights the critical role of relevance, accuracy, and truthfulness in fostering user trust. Practical guidelines are provided for designing transparent and user-friendly XAI explanations, especially in high-stakes contexts like healthcare. Overall, the paper's primary contribution lies in delineating clear requirements for effective XAI explanations, facilitating human-AI collaboration across diverse domains.

Interpretation of elevated baseline concentrations and serial changes of high-sensitivity cardiac troponin T in confirmed muscular dystrophies.

AIMS: Concentrations of high-sensitivity cardiac troponin T (hs-cTnT) are frequently elevated in stable patients with confirmed muscle dystrophies. However, sparse information is available on the interpretation of serial concentration changes. METHODS: Hs-cTnT was collected in 35 stable outpatients with confirmed skeletal muscle dystrophies at 0 and 1 h and after 6-12 months during scheduled outpatient visits. We simulated the effectiveness of the European Society of Cardiology (ESC) 0/1 h algorithm and assessed biological variation at 6-12 months using two established methods: reference change value (RCV) and minimal important difference (MID). RESULTS: Median baseline hs-cTnT concentrations were 34.4 ng/L [inter-quartile range (IQR): 17.5-46.2], and values > 99th percentile upper limit of normal were present in 34 of 35 patients. All patients were stable without cardiovascular adverse events during a follow-up of 6.6 months (IQR: 6-7). Median concentration change was 1.9 ng/L (IQR: 0.7-3.2) and 0.8 ng/L (IQR: 0-7.0) at 60 min and 6-9 months, respectively. Applying the criteria of the ESC 0/1 h algorithm for triage of suspected acute coronary syndrome (ACS) showed poor overall effectiveness of baseline hs-cTnT values. No patient would qualify for rule-out based on hs-cTnT less than the limit of detection, whereas five cases would qualify for rule-in based on hs-cTnT ≥ 52 ng/L. Biological variabilities at 6-12 months per MID and RCV were 1.2 ng/L [95% confidence interval (CI): 0.7-2.1] and 28.6% (95% CI: 27.9-29.6), respectively. A total of 8 (22.9%) and 25 (71.4%) cases exceeded the biological variation range, suggesting some additional myocardial damage. CONCLUSIONS: The high prevalence of elevated hs-cTnT could negatively impact the effectiveness of rule-out and rule-in strategies based on a single hs-cTnT value. Knowledge of the physiological and biological variation of hs-cTnT after 6-12 months is helpful to detect the progression of cardiac involvement or to search for cardiac complications including but not limited to arrhythmias that may trigger acute or chronic myocardial damage.

Personalized care in dilated cardiomyopathy: Rationale and study design of the activeDCM trial.

BACKGROUND: Dilated cardiomyopathy (DCM) is a leading cause of heart failure, particularly in younger individuals. Low physical strength is a global risk factor for cardiovascular mortality, and physical activity and a healthy lifestyle have been shown to improve outcomes in patients with heart failure. However, inappropriate exercise may increase the risk of arrhythmias in certain individuals with DCM. The determinants for predicting individual risks in this setting are poorly understood, and clinicians are hesitant to recommend sports for cardiomyopathy patients. The activeDCM trial aims to assess the safety and efficacy of a personalized exercise and activity programme for individuals with DCM. STUDY DESIGN: The activeDCM trial is a prospective, randomized, interventional trial with a 12 month follow-up. Three hundred patients, aged 18-75 years with DCM, left ventricular ejection fraction (LVEF) ≤ 50% and New York Heart Association (NYHA) classes I-III, will be enrolled. The intervention includes a personalized exercise and activity programme. The primary outcome is the increase in peak oxygen uptake (VO2max, mL/kg/min) from baseline to 12 months. Secondary endpoints include adherence to personalized activity programmes, freedom from clinically relevant arrhythmia, unplanned hospitalization for heart failure and changes in NYHA class, quality of life scores, 6 min walk distance, muscular strength, N-terminal pro-B-type natriuretic peptide (NT-proBNP) and high-sensitivity troponin T (hsTnT) levels and cardiac function. Advanced research questions include high-density phenome and omics analysis combined with digital biomarkers derived from Apple Watch devices. DISCUSSION: The activeDCM trial will provide valuable insights into the safety and efficacy of personalized exercise training in DCM patients, inform clinical practice and contribute to the development of heart failure management programmes. The study will generate data on the impact of exercise on various aspects of cardiovascular disease, including genetic, metabolic, phenotypic and longitudinal aspects, facilitating the development of future digital tools and strategies, including the incorporation of smart wearable devices.

Assessing the Applicability of Cardiac Myosin Inhibitors for Hypertrophic Cardiomyopathy Management in a Large Single Center Cohort.

Background: Cardiac myosin inhibitors (CMIs), including Mavacamten and Aficamten, have emerged as a groundbreaking treatment for hypertrophic cardiomyopathy (HCM). The results from phase 2 and 3 randomized clinical trials for both drugs have showed promising outcomes. However, the highly selective patient recruitment for these trials raises questions about the generalizability of the observed positive effects across broader patient populations suffering from HCM. Methods: A retrospective cohort study at University Hospital Heidelberg included 404 HCM patients. Baseline assessments included family history, electrocardiograms (ECGs), and advanced cardiac imaging, to ensure the exclusion of secondary causes of left ventricular hypertrophy. Results: Among the HCM patients evaluated, only a small percentage met the inclusion criteria for recent CMI trials: 10.4% for EXPLORER-HCM and 4.7% for SEQUOIA-HCM. The predominant exclusion factor was the stringent left ventricular outflow tract (LVOT) gradient requirement. Conclusions: This study highlights a significant discrepancy between patient demographics in clinical trials and those encountered in routine HCM clinical practice. Despite promising results from the initial randomized clinical trials that led to the approval of Mavacamten, the selected patient population only represents a small part of the HCM patient cohort seen in routine clinics. This study advocates for further expanded randomized clinical trials with broader inclusion criteria to represent diverse primary HCM patient populations.

Prediction of diagnosis and diastolic filling pressure by AI-enhanced cardiac MRI: a modelling study of hospital data.

BACKGROUND: With increasing numbers of patients and novel drugs for distinct causes of systolic and diastolic heart failure, automated assessment of cardiac function is important. We aimed to provide a non-invasive method to predict diagnosis of patients undergoing cardiac MRI (cMRI) and to obtain left ventricular end-diastolic pressure (LVEDP). METHODS: For this modelling study, patients who had undergone cardiac catheterisation at University Hospital Heidelberg (Heidelberg, Germany) between July 15, 2004 and March 16, 2023, were identified, as were individual left ventricular pressure measurements. We used existing patient data from routine cardiac diagnostics. From this initial group, we extracted patients who had been diagnosed with ischaemic cardiomyopathy, dilated cardiomyopathy, hypertrophic cardiomyopathy, or amyloidosis, as well as control individuals with no structural phenotype. Data were pseudonymised and only processed within the university hospital's AI infrastructure. We used the data to build different models to predict either demographic (ie, AI-age and AI-sex), diagnostic (ie, AI-coronary artery disease and AI-cardiomyopathy [AI-CMP]), or functional parameters (ie, AI-LVEDP). We randomly divided our datasets via computer into training, validation, and test datasets. AI-CMP was not compared with other models, but was validated in a prospective setting. Benchmarking was also done. FINDINGS: 66 936 patients who had undergone cardiac catheterisation at University Hospital Heidelberg were identified, with more than 183 772 individual left ventricular pressure measurements. We extracted 4390 patients from this initial group, of whom 1131 (25·8%) had been diagnosed with ischaemic cardiomyopathy, 1064 (24·2%) had been diagnosed with dilated cardiomyopathy, 816 (18·6%) had been diagnosed with hypertrophic cardiomyopathy, 202 (4·6%) had been diagnosed with amyloidosis, and 1177 (26·7%) were control individuals with no structural phenotype. The core cohort only included patients with cardiac catherisation and cMRI within 30 days, and emergency cases were excluded. AI-sex was able to predict patient sex with areas under the receiver operating characteristic curves (AUCs) of 0·78 (95% CI 0·77-0·78) and AI-age was able to predict patient age with a mean absolute error of 7·86 years (7·77-7·95), with a Pearson correlation of 0·57 (95% CI 0·56-0·57). The AUCs for the classification tasks ranged between 0·82 (95% CI 0·79-0·84) for ischaemic cardiomyopathy and 0·92 (0·91-0·94) for hypertrophic cardiomyopathy. INTERPRETATION: Our AI models could be easily integrated into clinical practice and provide added value to the information content of cMRI, allowing for disease classification and prediction of diastolic function. FUNDING: Informatics for Life initiative of the Klaus-Tschira Foundation, German Center for Cardiovascular Research, eCardiology section of the German Cardiac Society, and AI Health Innovation Cluster Heidelberg.

Comparison of the clinical chemistry score to other biomarker algorithms for rapid rule-out of acute myocardial infarction and risk stratification in patients with suspected acute coronary syndrome.

BACKGROUND: The clinical chemistry score (CCS) comprising high-sensitivity cardiac troponins (hs-cTn), glucose and estimated glomerular filtration rate has been previously validated with superior accuracy for detection and risk stratification of acute myocardial infarction (AMI) compared to hs-cTn alone. METHODS: The CCS was compared to other biomarker-based algorithms for rapid rule-out and prognostication of AMI including the hs-cTnT limit-of-blank (LOB, <3 ng/L) or limit-of-detection (LOD, <5 ng/L) and a dual marker strategy (DMS) (copeptin <10 pmol/L and hs-cTnT ≤14 ng/L) in 1506 emergency department (ED) patients with symptoms suggestive of acute coronary syndrome. Negative predictive values (NPV) and sensitivities for AMI rule-out, and 12-month combined endpoint rates encompassing mortality, myocardial re-infarction, as well as stroke were assessed. RESULTS: NPVs of 100% (95% CI: 98.3-100%) were observed for CCS = 0, hs-cTnT LoB and hs-cTnT LoD with rule-out efficacies of 11.1%, 7.6% and 18.3% as well as specificities of 13.0% (95% CI: 9.9-16.6%), 8.8% (95% CI: 7.3-10.5%) and 21.4% (95% CI: 19.2-23.8%), respectively. A CCS ≤ 1 achieved a rule-out in 32.2% of all patients with a NPV of 99.6% (95% CI: 98.4-99.9%) and specificity of 37.4% (95% CI: 34.2-40.5%) compared to a rule-out efficacy of 51.2%, NPV of 99.0 (95% CI: 98.0-99.5) and specificity of 59.7% (95% CI: 57.0-62.4%) for the DMS. Rates of the combined end-point of death/AMI within 30 days ranged between 0.0% and 0.7% for all fast-rule-out protocols. CONCLUSIONS: The CCS ensures reliable AMI rule-out with low short and long-term outcome rates for a specific ED patient subset. However, compared to a single or dual biomarker strategy, the CCS displays reduced efficacy and specificity, limiting its clinical utility.