Harnessing feature extraction capacities from a pre-trained convolutional neural network (VGG-16) for the unsupervised distinction of aortic outflow velocity profiles in patients with severe aortic stenosis.

Aims: Hypothesizing that aortic outflow velocity profiles contain more valuable information about aortic valve obstruction and left ventricular contractility than can be captured by the human eye, features of the complex geometry of Doppler tracings from patients with severe aortic stenosis (AS) were extracted by a convolutional neural network (CNN). Methods and results: After pre-training a CNN (VGG-16) on a large data set (ImageNet data set; 14 million images belonging to 1000 classes), the convolutional part was employed to transform Doppler tracings to 1D arrays. Among 366 eligible patients [age: 79.8 ± 6.77 years; 146 (39.9%) women] with pre-procedural echocardiography and right heart catheterization prior to transcatheter aortic valve replacement (TAVR), good quality Doppler tracings from 101 patients were analysed. The convolutional part of the pre-trained VGG-16 model in conjunction with principal component analysis and k-means clustering distinguished two shapes of aortic outflow velocity profiles. Kaplan-Meier analysis revealed that mortality in patients from Cluster 2 (n = 40, 39.6%) was significantly increased [hazard ratio (HR) for 2-year mortality: 3; 95% confidence interval (CI): 1-8.9]. Apart from reduced cardiac output and mean aortic valve gradient, patients from Cluster 2 were also characterized by signs of pulmonary hypertension, impaired right ventricular function, and right atrial enlargement. After training an extreme gradient boosting algorithm on these 101 patients, validation on the remaining 265 patients confirmed that patients assigned to Cluster 2 show increased mortality (HR for 2-year mortality: 2.6; 95% CI: 1.4-5.1, P-value: 0.004). Conclusion: Transfer learning enables sophisticated pattern recognition even in clinical data sets of limited size. Importantly, it is the left ventricular compensation capacity in the face of increased afterload, and not so much the actual obstruction of the aortic valve, that determines fate after TAVR.

Subphenotyping of Patients With Aortic Stenosis by Unsupervised Agglomerative Clustering of Echocardiographic and Hemodynamic Data.

OBJECTIVES: The aim of this retrospective analysis was to categorize patients with severe aortic stenosis (AS) according to clinical presentation by applying unsupervised machine learning. BACKGROUND: Patients with severe AS present with heterogeneous clinical phenotypes, depending on disease progression and comorbidities. METHODS: Unsupervised agglomerative clustering was applied to preprocedural data from echocardiography and right heart catheterization from 366 consecutively enrolled patients undergoing transcatheter aortic valve replacement for severe AS. RESULTS: Cluster analysis revealed 4 distinct phenotypes. Patients in cluster 1 (n = 164 [44.8%]), serving as a reference, presented with regular cardiac function and without pulmonary hypertension (PH). Accordingly, estimated 2-year survival was 90.6% (95% CI: 85.8%-95.6%). Clusters 2 (n = 66 [18.0%]) and 4 (n = 91 [24.9%]) both comprised patients with postcapillary PH. Yet patients in cluster 2 with preserved left and right ventricular structure and function showed a similar survival as those in cluster 1 (2-year survival 85.8%; 95% CI: 76.9%-95.6%), whereas patients in cluster 4 with dilatation of all heart chambers and a high prevalence of mitral and tricuspid regurgitation (12.5% and 14.8%, respectively) died more often (2-year survival 74.9% [95% CI: 65.9%-85.2%]; HR for 2-year mortality: 2.8 [95% CI: 1.4-5.5]). Patients in cluster 3, the smallest (n = 45 [12.3%]), displayed the most extensive disease characteristics (ie, left and right heart dysfunction together with combined pre- and postcapillary PH), and 2-year survival was accordingly reduced (77.3% [95% CI: 65.2%-91.6%]; HR for 2-year mortality: 2.6 [95% CI: 1.1-6.2]). CONCLUSIONS: Unsupervised machine learning aids in capturing complex clinical presentations as observed in patients with severe AS. Importantly, structural alterations in left and right heart morphology, possibly due to genetic predisposition, constitute an equally sensitive indicator of poor prognosis compared with high-grade PH.

Myocardial hypoxic stress mediates functional cardiac extracellular vesicle release.

AIMS: Increased shedding of extracellular vesicles (EVs)-small, lipid bilayer-delimited particles with a role in paracrine signalling-has been associated with human pathologies, e.g. atherosclerosis, but whether this is true for cardiac diseases is unknown. METHODS AND RESULTS: Here, we used the surface antigen CD172a as a specific marker of cardiomyocyte (CM)-derived EVs; the CM origin of CD172a+ EVs was supported by their content of cardiac-specific proteins and heart-enriched microRNAs. We found that patients with aortic stenosis, ischaemic heart disease, or cardiomyopathy had higher circulating CD172a+ cardiac EV counts than did healthy subjects. Cellular stress was a major determinant of EV release from CMs, with hypoxia increasing shedding in in vitro and in vivo experiments. At the functional level, EVs isolated from the supernatant of CMs derived from human-induced pluripotent stem cells and cultured in a hypoxic atmosphere elicited a positive inotropic response in unstressed CMs, an effect we found to be dependent on an increase in the number of EVs expressing ceramide on their surface. Of potential clinical relevance, aortic stenosis patients with the highest counts of circulating cardiac CD172a+ EVs had a more favourable prognosis for transcatheter aortic valve replacement than those with lower counts. CONCLUSION: We identified circulating CD172a+ EVs as cardiac derived, showing their release and function and providing evidence for their prognostic potential in aortic stenosis patients.

Changes of Right Ventricular Function After Transcatheter Aortic Valve Replacement and Association With Outcomes.

BACKGROUND: Baseline right ventricular (RV) dysfunction represents a predictor for poor outcome in patients undergoing transcatheter aortic valve replacement (TAVR). However, RV function may improve after TAVR, which could have important implications on outcomes. The aim of the present study was to assess changes in RV function after TAVR and its prognostic value regarding clinical outcome. METHODS AND RESULTS: Patients undergoing TAVR at our institution were consecutively enrolled and categorized into 4 groups according to changes in RV function during echocardiographic follow-up at 6 months. A total of 188 patients were included. Of those showing normal function at baseline, 87% (130/149) had preserved RV function at follow-up (group 1), whereas 13% (19/149) developed new RV dysfunction (group 2). Of those with RV dysfunction at baseline (39 patients), RV function normalized in 46% (18/39) (group 3) and remained impaired in 54% (21/39) (group 4). The Kaplan-Meier estimated survival at 3 years was highest in patients in group 1 (83%), intermediate in group 2 (65%) and 3 (69%), whereas group 4 had the worst survival (37%; P < .001). Furthermore, new or persistent RV dysfunction was identified to be independently associated with mortality during follow-up (hazard ratio 2.55; interquartile range 1.03-6.47, P = .004). CONCLUSIONS: Patients with preserved RV function have a high 3-year survival. Normalization of RV function showed improved survival compared with patients with persistent RV dysfunction, who had a dismal prognosis despite TAVR.

Impact of sinus rhythm versus atrial fibrillation on left ventricular remodeling after transcatheter aortic valve replacement.

AIMS: Atrial fibrillation (AF) is associated with increased mortality after transcatheter aortic valve replacement (TAVR). Cerebrovascular complications and bleeding events associated with anticoagulation therapy are discussed to be possible causes for this increased mortality. The present study sought to assess whether AF is associated with impaired left ventricular (LV) reverse remodeling representing another possible mechanism for poor outcome. METHODS: All patients who underwent TAVR in our institution and had 1-year echocardiography follow-up were included. LV mass index (LVMI) at baseline and follow-up as well as LVMI change at 1 year were assessed with respect to the presence of AF (either at baseline or during hospitalization after TAVR) and sinus rhythm (SR). RESULTS: A total of 213 patients (n = 95 in AF; n = 118 in SR) were enrolled in the present study. Patients with AF had higher LVMI at 1 year compared to those with SR (173 ± 61 g/m2 vs. 154 ± 55 g/m2; p = 0.02) and they showed lower relative LVMI change at 1 year (- 2 ± 28% vs. - 9 ± 29%; p = 0.04). In linear regression analysis, AF was independently associated with relative LVMI change (regression coefficient ß 0.076 [95% CI 0.001-0.150]; p = 0.04). With respect to clinical outcome depending on AF and LVMI regression, the Kaplan-Meier estimated event-free of death or cardiac rehospitalization at 3 years was lowest among patients with AF and no LVMI regression. CONCLUSIONS: The present study identified a significant association of AF with changes in LVMI after TAVR, which was also shown to be associated with clinical outcome.

Dose-Dependent Effect of Renin-Angiotensin System Blockade Following Transcatheter Aortic Valve Replacement.

BACKGROUND: There is growing body of evidence from retrospective studies that renin-angiotensin system (RAS) blockade is associated with improved outcome after transcatheter aortic valve replacement (TAVR). However, it remains unknown whether the effect of RAS blockade is dose dependent. The current study sought to assess the dose-dependent effect of RAS blockade on survival and left-ventricular (LV) remodelling after TAVR. METHODS: Patients who were enrolled into our observational TAVR study at our institution were retrospectively assessed according to different doses of RAS blockade: group 1 (no RAS blockade), group 2 (25% of maximum daily dose), group 3 (50% of maximum daily dose), and group 4 (full daily dose). RESULTS: A total of 323 patients between January 2015 and September 2019 were included. Patients with higher doses of RAS blockade showed a trend toward higher overall survival at 3-year follow-up (56% with no RAS blockade vs 66% with the 25% dose vs 79% with the 50% dose vs 78% with the full dose; P = 0.063). After adjustment for baseline characteristics, the difference in survival was significant (P = 0.042). Besides New York Heart Association class and left-ventricular ejection fraction (LVEF), RAS blockade dose was identified as independent predictor for all-cause mortality (hazard ratio [HR] 0.72; 95% confidence interval [CI], 0.54-0.97; P = 0.03). With respect to LV remodelling, a significantly larger reduction of LV mass index was observed during the follow-up with higher doses of RAS blockade. CONCLUSIONS: The current study showed that the impact of RAS blockade treatment on clinical outcome and LV remodelling after TAVR is dose dependent.

Impact of transcatheter mitral valve repair using MitraClip on right ventricular remodeling.

The potential of the MitraClip to prevent from right heart failure or to restore right ventricular (RV) function is still unclear. The aim of the present study was to analyze the impact of the MitraClip implantation on RV function and its association with clinical outcome. After MitraClip implantation patients underwent echocardiography follow-up scheduled between 3 and 6 months after the procedure in the present single-center registry. A total of 93 patients were included. Compared to baseline, RV function declined in 20%, was unchanged in 25% and improved in 55% of the patients. Factors associated with decline in RV performance were atrial fibrillation, decrease in left ventricular function and lack of reduction in pulmonary artery pressure. Patients who experienced worsening in RV function had a significantly lower survival after mean follow-up of 11 ± 7 months compared to those with preserved or improved RV function (15% vs. 83% vs. 83%; p log rank = 0.001). Furthermore, changes in TAPSE were found to be an independent predictor for all-cause mortality [HR 0.88 (0.77-0.99); p = 0.04]. The majority of patients suffering from severe MR benefited from MitraClip with respect to RV remodeling. However, 20% of the patients experienced a decline in RV function, which was associated with poor prognosis. Importantly, changes in RV function after MitraClip were identified as independent predictor for survival in contrast to baseline RV function and, therefore, should be implemented in follow-up routine for better outcome prediction.

Evaluation of Hybrid Surgical Access Approaches for Pulmonary Valve Implantation in an Acute Swine Model.

Percutaneous implantation of the pulmonary valve through peripheral vascular access can be limited due to poor venous access, low patient weight, hemodynamic or rhythmic instability, and size constraints related to the valve. In such cases, hybrid procedures may provide alternatives. Because the most commonly used median sternotomy is unsuitable for chronic trials in large animals, we evaluated several hybrid approaches for pulmonary valve replacement in a swine model. We tested the feasibility of hybrid pulmonary valve implantation in pigs by using inhouse-generated valves containing bare-metal or nitinol stents. Valves consisted of bovine jugular veins, bovine pericardial valves, or sprayed polyurethane valves. Access was achieved through median sternotomy, lower partial sternotomy, transverse sternotomy, or right lateral thoracotomy. The delivery device was introduced in a transventricular manner. Implantation took place under fluoroscopic and epicardial echocardiographic guidance. We achieved implantation of the stented valve in 12 (92.3%) pigs, of which 5 (41.7%) of the implanted valves were in an optimal position. Paravalvular leakage occurred in 2 trials (16.7%). Lower partial sternotomy provided the best trade-off between feasibility and minimized trauma for long-term animal trials. Here we describe our experience with hybrid pulmonary valve implantation in an acute large-animal (swine) model. We demonstrate the feasibility of the procedure in terms of surgical technique and the perioperative management and preparation of the field for a chronic trial.

Prognostic Impact of Left Atrial Function Following Transcatheter Mitral Valve Repair.

Background Left atrial ( LA ) function predicts clinical outcome in a variety of cardiovascular diseases. However, limited data are available in the setting of mitral regurgitation. The aim of the present study was to assess potential changes in LA ejection fraction (LAEF) and its prognostic value in patients following transcatheter mitral valve repair using the MitraClip. Methods and Results A total of 88 consecutive patients undergoing MitraClip implantation with complete echocardiography at baseline and follow-up between 3 and 6 months postprocedure were enrolled. LAEF improved in 58% of the population. Change in LAEF was associated with residual mitral regurgitation, residual transmitral gradient and left ventricular ejection fraction changes. Compared with their counterparts, patients with residual mitral regurgitation ≥grade 2 (change in LAEF, -6% [Interquartile [IQR], -9-1%] versus 4% [IQR, -5-15%]; P=0.05) and with residual transmitral gradient ≥5 mm Hg (change in LAEF, -2% [IQR, -9-9%] versus 5% [IQR, -4-16%]; P=0.03) showed a decline in LAEF , respectively. Furthermore, LAEF significantly correlated with changes in left ventricular ejection fraction ( r=0.40; P=0.001). With regards to clinical outcome, heart failure symptoms as assessed by New York Heart Association class were more severe in patients with worsened LAEF at follow-up. Finally, LAEF change was identified as an independent predictor of all-cause mortality (hazard ratio, 0.94; 95% CI, 0.90-0.98 [ P=0.008]). Conclusions The present analysis showed that changes in LA function in patients undergoing MitraClip implantation are associated with important measures including residual mitral regurgitation, elevated transmitral gradient, and left ventricular function. Importantly, LA function alterations represent a strong predictor for all-cause mortality.