Machine learning to identify a composite indicator to predict cardiac death in ischemic heart disease.

BACKGROUND: Machine learning (ML) employs algorithms that learn from data, building models with the potential to predict events by aggregating a large number of variables and assessing their complex interactions. The aim of this study is to assess ML potential in identifying patients with ischemic heart disease (IHD) at high risk of cardiac death (CD). METHODS: 3987 (mean age 68 ± 11) hospitalized IHD patients were enrolled. We implemented and compared various ML models and their combination into ensembles. Model output constitutes a new ML indicator to be employed for stratification. Primary variable importance was assessed with ablation tests. RESULTS: An ensemble classifier combining three ML models achieved the best performance to predict CD (AUROC of 0.830, F1-macro of 0.726). ML indicator use through Cox survival analysis outperformed the 18 variables individually, producing a better stratification compared to standard multivariate analysis (improvement of ∼20%). Patients in the low risk group defined through ML indicator had a significantly higher survival (88.8% versus 29.1%). The main variables identified were Dyslipidemia, LVEF, Previous CABG, Diabetes, Previous Myocardial Infarction, Smoke, Documented resting or exertional ischemia, with an AUROC of 0.791 and an F1-score of 0.674, lower than that of 18 variables. Both code and clinical data are freely available with this article. CONCLUSION: ML may allow a faster, low-cost and reliable evaluation of IHD patient prognosis by inclusion of more predictors and identification of those more significant, improving outcome prediction towards the development of precision medicine in this clinical field.

Prognostic Role of Sonographic Decongestion in Patients with Acute Heart Failure with Reduced and Preserved Ejection Fraction: A Multicentre Study.

BACKGROUND: We investigated the role of the dynamic changes of pulmonary congestion, as assessed by sonographic B-lines, as a tool to stratify prognosis in patients admitted for acute heart failure with reduced and preserved ejection fraction (HFrEF, HFpEF). METHODS: In this multicenter, prospective study, lung ultrasound was performed at admission and before discharge by trained investigators, blinded to clinical findings. RESULTS: We enrolled 208 consecutive patients (mean age 76 [95% confidence interval, 70-84] years), 125 with HFrEF, 83 with HFpEF (mean ejection fraction 32% and 57%, respectively). The primary composite endpoint of cardiovascular death or HF re-hospitalization occurred in 18% of patients within 6 months. In the overall population, independent predictors of the occurrence of the primary endpoint were the number of B-lines at discharge, NT-proBNP levels, moderate-to-severe mitral regurgitation, and inferior vena cava diameter on admission. B-lines at discharge were the only independent predictor in both HFrEF and HFpEF subgroups. A cut-off of B-lines > 15 at discharge displayed the highest accuracy in predicting the primary endpoint (AUC = 0.80, p < 0.0001). Halving B-lines during hospitalization further improved event classification (continuous net reclassification improvement = 22.8%, p = 0.04). CONCLUSIONS: The presence of residual subclinical sonographic pulmonary congestion at discharge predicts 6-month clinical outcomes across the whole spectrum of acute HF patients, independent of conventional biohumoral and echocardiographic parameters. Achieving effective pulmonary decongestion during hospitalization is associated with better outcomes.

Cardiac troponin T and NT-proBNP for detecting myocardial ischemia in suspected chronic coronary syndrome.

BACKGROUND: Elevated N-terminal pro-B-type natriuretic peptides (NT-proBNP) and cardiac troponin T (cTnT) are associated with poor outcome in patients with chronic coronary syndrome (CCS). The performance of these biomarkers in diagnosing ischemia, and their association with myocardial hypoperfusion and hypokinesis is unclear. METHODS: Patients with suspected CCS (history of angina, estimated cardiovascular risk >15% or a positive stress test) were included in the prospective, multi-center DOPPLER-CIP study. Patients underwent Single Positron Emission Computed Tomography for assessment of ischemia and NT-proBNP and cTnT were measured in venous blood samples. RESULTS: We included 430 patients (25% female) aged 64 ± 8 years. Reversible hypoperfusion and hypokinesis were present in 139 (32%) and 89 (21%), respectively. Concentrations of NT-proBNP and cTnT correlated moderately (rho = 0.50, p < 0.001). NT-proBNP and cTnT concentrations (median [IQR]) were higher in patients with versus without reversible ischemia: 150 (73-294) versus 87 (44-192) ng/L and 10 (6-13) versus 7 (4-11) ng/L, respectively (p < 0.001 for both), and the associations persisted after adjusting for possible confounders. The C-statistics to discriminate ischemia ranged from 63%-73%, were comparable for cTnT and NT-proBNP, and higher for hypokinesis than hypoperfusion, and both were superior to exercise electrocardiography and stress echocardiography. Very low concentrations (≤5 ng/L cTnT and ≤ 60 ng/L NT-proBNP) ruled out reversible hypokinesis with negative predictive value >90%. CONCLUSION: cTnT and NT-proBNP are associated with irreversible and reversible ischemia in patients with suspected CCS, particularly hypokinesis. The diagnostic performance was comparable between the biomarkers, and very low concentrations may reliably rule out ischemia.

The evolving role of cardiac imaging in patients with myocardial infarction and non-obstructive coronary arteries.

Myocardial infarction (MI) with non-obstructive coronary arteries (MINOCA) represents a heterogeneous clinical conundrum accounting for about 6%-8% of all acute MI who are referred for coronary angiography. Current guidelines and consensus documents recommend that these patients are appropriately diagnosed, uncovering the causes of MINOCA, so that specific therapies can be prescribed. Indeed, there are a variety of causes that can result in this clinical condition, and for this reason diagnostic cardiac imaging has an emerging critical role in the assessment of patients with suspected or confirmed MINOCA. In last years, different cardiac imaging techniques have been evaluated in this context, and the comprehension of their strengths and limitations is of the utmost importance for their effective use in clinical practice. Moreover, recent evidence is clearly suggesting that a multimodality cardiac imaging approach, combining different techniques, seems to be crucial for a proper management of MINOCA. However, great variability still exists in clinical practice in the management of patients with suspected MINOCA, also depending on the availability of diagnostic tools and local expertise. Herein, we review the current knowledge supporting the use of different cardiac imaging techniques in patients with MINOCA, underscoring the importance of a comprehensive multimodality cardiac imaging approach and proposing a practical diagnostic algorithm to properly identify and treat the specific causes of MINOCA, in order to improve prognosis and the quality of life in these patients.

A randomized multicenter trial on a lung ultrasound-guided treatment strategy in patients on chronic hemodialysis with high cardiovascular risk.

Lung congestion is a risk factor for all-cause and cardiovascular mortality in patients on chronic hemodialysis, and its estimation by ultrasound may be useful to guide ultrafiltration and drug therapy in this population. In an international, multi-center randomized controlled trial (NCT02310061) we investigated whether a lung ultrasound-guided treatment strategy improved a composite end point (all-cause death, non-fatal myocardial infarction, decompensated heart failure) vs usual care in patients receiving chronic hemodialysis with high cardiovascular risk. Patient-Reported Outcomes (Depression and the Standard Form 36 Quality of Life Questionnaire, SF36) were assessed as secondary outcomes. A total of 367 patients were enrolled: 183 in the active arm and 180 in the control arm. In the active arm, the pre-dialysis lung scan was used to titrate ultrafiltration during dialysis and drug treatment. Three hundred and seven patients completed the study: 152 in the active arm and 155 in the control arm. During a mean follow-up of 1.49 years, lung congestion was significantly more frequently relieved in the active (78%) than in the control (56%) arm and the intervention was safe. The primary composite end point did not significantly differ between the two study arms (Hazard Ratio 0.88; 95% Confidence Interval: 0.63-1.24). The risk for all-cause and cardiovascular hospitalization and the changes of left ventricular mass and function did not differ among the two groups. A post hoc analysis for recurrent episodes of decompensated heart failure (0.37; 0.15-0.93) and cardiovascular events (0.63; 0.41-0.97) showed a risk reduction for these outcomes in the active arm. There were no differences in patient-reported outcomes between groups. Thus, in patients on chronic hemodialysis with high cardiovascular risk, a treatment strategy guided by lung ultrasound effectively relieved lung congestion but was not more effective than usual care in improving the primary or secondary end points of the trial.

Prognostic value of lung ultrasound in patients hospitalized for heart disease irrespective of symptoms and ejection fraction.

AIMS: Lung ultrasound B-lines are the sonographic sign of pulmonary congestion and can be used in the differential diagnosis of dyspnoea to rule in or rule out acute heart failure (AHF). Our aim was to assess the prognostic value of B-lines, integrated with echocardiography, in patients admitted to a cardiology department, independently of the initial clinical presentation, thus in patients with and without AHF, and in AHF with reduced and preserved ejection fraction (HFrEF and HFpEF). METHODS AND RESULTS: We enrolled consecutive patients admitted for various cardiac conditions. Patients were classified into three groups: (i) acute HFrEF; (ii) acute HFpEF; and (iii) non-AHF. All patients underwent an echocardiogram coupled with lung ultrasound at admission, according to standardized protocols. We followed up 1021 consecutive inpatients (69 ± 12 years) for a median of 14.4 months (interquartile range 4.6-24.3) for death and rehospitalization for AHF. During the follow-up, 126 events occurred. Admission B-lines > 30, ejection fraction < 50%, tricuspid regurgitation velocity > 2.8 m/s, and tricuspid annular plane systolic excursion < 17 mm were independent predictors at multivariable analysis. B-lines > 30 had a strong predictive value in HFpEF and non-AHF, but not in HFrEF. CONCLUSIONS: Ultrasound B-lines can detect subclinical pulmonary interstitial oedema in patients thought to be free of congestion and provide useful information not only for the diagnosis but also for the prognosis in different cardiac conditions. Their added prognostic value among standard echocardiographic parameters is more robust in patients with HFpEF compared with HFrEF.

Inflammation is an amplifier of lung congestion by high lv filling pressure in hemodialysis patients: a longitudinal study.

INTRODUCTION: Since inflammation alters vascular permeability, including vascular permeability in the lung, we hypothesized that it can be an amplifier of lung congestion in a category of patients at high risk for pulmonary oedema like end stage kidney disease (ESKD) patients. OBJECTIVE AND METHODS: We investigated the effect modification by systemic inflammation (serum CRP) on the relationship between a surrogate of the filling pressure of the LV [left atrial volume indexed to the body surface area (LAVI)] and lung water in a series of 220 ESKD patients. Lung water was quantified by the number of ultrasound B lines (US-B) on lung US. Six-hundred and three recordings were performed during a 2-year follow up. Longitudinal data analysis was made by the Mixed Linear Model. RESULTS: At baseline, 88 had absent, 101 had mild to moderate lung congestion and 31 severe congestion. The number of US B lines associated with LAVI (r = 0.23, P < 0.001) and serum CRP was a robust modifier of this relationship (P < 0.001). Similarly, in fully adjusted longitudinal analyses US-B lines associated with simultaneous estimates of LAVI (P = 0.002) and again CRP was a strong modifier of this relationship in adjusted analyses (P ≤ 0.01). Overall, at comparable LAVI levels, lung congestion was more pronounced in inflamed than in non-inflamed patients. CONCLUSION: In ESKD systemic inflammation is a modifier of the relationship between LAVI, an integrate measure of LV filling pressure, and lung water. For any given pressure, lung water is increased with higher CRP levels, likely reflecting a higher permeability of the alveolar-capillary barrier.

Anatomical and functional coronary imaging to predict long-term outcome in patients with suspected coronary artery disease: the EVINCI-outcome study.

AIMS: To investigate the prognostic relevance of coronary anatomy, coronary function, and early revascularization in patients with stable coronary artery disease (CAD). METHODS AND RESULTS: From March 2009 to June 2012, 430 patients with suspected CAD (61 ± 9 years, 62% men) underwent coronary anatomical imaging by computed tomography coronary angiography (CTCA) and coronary functional imaging followed by invasive coronary angiography (ICA) if at least one non-invasive test was abnormal. Obstructive CAD was documented by ICA in 119 patients and 90 were revascularized within 90 days of enrolment. Core laboratory analysis showed that 134 patients had obstructive CAD by CTCA (>50% stenosis in major coronary vessels) and 79 significant ischaemia by functional imaging [>10% left ventricular (LV) myocardium]. Over mean follow-up of 4.4 years, major adverse events (AEs) (all-cause death, non-fatal myocardial infarction, or hospital admission for unstable angina or heart failure) or AEs plus late revascularization (LR) occurred in 40 (9.3%) and 58 (13.5%) patients, respectively. Obstructive CAD at CTCA was the only independent imaging predictor of AEs [hazard ratio (HR) 3.2, 95% confidence interval (CI) 1.10-9.30; P = 0.033] and AEs plus LR (HR 4.3, 95% CI 1.56-11.81; P = 0.005). Patients with CAD in whom early revascularization was performed in the presence of ischaemia and deferred in its absence had fewer AEs, similar to patients without CAD (HR 2.0, 95% CI 0.71-5.51; P = 0.195). CONCLUSION: Obstructive CAD imaged by CTCA is an independent predictor of clinical outcome. Early management of CAD targeted to the combined anatomical and functional disease phenotype improves clinical outcome.

Usefulness of dual imaging stress echocardiography for the diagnosis of coronary allograft vasculopathy in heart transplant recipients.

BACKGROUND: Coronary allograft vasculopathy (CAV) is the main factor limiting long-term survival after cardiac transplantation. Dual imaging stress echocardiography with wall motion and Doppler-derived coronary flow reserve (CRF) of the left anterior descending artery (LAD) is a state-of-the-art methodology during dipyridamole stress echocardiography (DiSE). This study involving 74 heart transplanted patients has the purpose to assess the diagnostic value of dipyridamole stress echocardiography with evaluation of wall motion (WM) and Doppler-derived coronary flow reserve for the diagnosis of coronary allograft vasculopathy. METHODS AND RESULTS: All patients underwent DiSE and coronary angiography. Moderate-severe CAV was defined according to International Society of Heart and Lung Transplant (ISHLT) recommended nomenclature for CAV, and CFR < 2 was considered to be impaired. Moderate-severe CAV was present in 11 patients. WM analysis revealed four patients (5%) with rest WM abnormalities. CFR analysis revealed that 40 (54%) individuals had an abnormal result. The combined evaluation of WM analysis and CFR resulted in a sensitivity of 72.7% (95% CI: 39.3 to 92.6%), a specificity of 49.2% (95% CI: 36.5 to 61.9%), a positive predictive value of 20% (95% CI: 9.6 to 36.1%), and negative predictive value of 91.1% (95% CI: 75.1 to 97.6%) for the diagnosis of CAV. CONCLUSIONS: Our results support the inclusion of DiSE performance in Heart transplant follow up protocol. The addition of CFR evaluation offers valuable information to the angiography findings in the detection of CAV and could be helpful in selected patients to adjust the time and indications of coronary angiography.