Machine Learning Prediction Models for Mitral Valve Repairability and Mitral Regurgitation Recurrence in Patients Undergoing Surgical Mitral Valve Repair.

BACKGROUND: Mitral valve regurgitation (MR) is the most common valvular heart disease and current variables associated with MR recurrence are still controversial. We aim to develop a machine learning-based prognostic model to predict causes of mitral valve (MV) repair failure and MR recurrence. METHODS: 1000 patients who underwent MV repair at our institution between 2008 and 2018 were enrolled. Patients were followed longitudinally for up to three years. Clinical and echocardiographic data were included in the analysis. Endpoints were MV repair surgical failure with consequent MV replacement or moderate/severe MR (>2+) recurrence at one-month and moderate/severe MR recurrence after three years. RESULTS: 817 patients (DS1) had an echocardiographic examination at one-month while 295 (DS2) also had one at three years. Data were randomly divided into training (DS1: n = 654; DS2: n = 206) and validation (DS1: n = 164; DS2 n = 89) cohorts. For intra-operative or early MV repair failure assessment, the best area under the curve (AUC) was 0.75 and the complexity of mitral valve prolapse was the main predictor. In predicting moderate/severe recurrent MR at three years, the best AUC was 0.92 and residual MR at six months was the most important predictor. CONCLUSIONS: Machine learning algorithms may improve prognosis after MV repair procedure, thus improving indications for correct candidate selection for MV surgical repair.

Head to Head Comparison between Different 3-Dimensional Echocardiographic Rendering Tools in the Imaging of Percutaneous Edge-to-Edge Mitral Valve Repair.

MitraClip (MC) is the most common percutaneous treatment for severe mitral regurgitation (MR). An accurate two-dimensional and three-dimensional echocardiographic (3DTEE) imaging is mandatory for the optimal procedural result. Recently transillumination 3DTEE rendering (3DTr) has been introduced integrating a virtual light source into the dataset and with the addition of glass effect (3DGl) allows to adjust tissue transparency improving depth perception and anatomical structure delineation in comparison with the standard 3DTEE (3DSt). The aim of this retrospective study in 30 patients undergoing MC, was to compare 3DSt, 3DTr, and 3DGl in mitral valve (MV) evaluation and procedural result assessment. 3DTEE acquisitions obtained before and after MC were processed with 3DSt, 3DTr, and 3DGl rendering. Each reconstruction was scored for quality and for ability to recognize MV anatomy, MR origin, clip position, dimension and grasping. Imaging quality was judged good or optimal in 52%, 76%, and 96% in 3DSt, 3DTr, and 3DGl reconstructions respectively. In 26/30 patients a diagnostic incremental value was found with 3DTr vs. 3DSt and in 15/26 with 3DGl vs. 3DTr and 3DSt. Only 3DGl with perpendicular cropping of the clip allowed to visualize and measure the grasped portion of each mitral leaflets. 3DTEE imaging during MC may be improved by 3DTr and 3DGl providing a better evaluation of MV, of leaflet grasping and of residual MR jets after MC.

Predicting Long-Term Mortality in TAVI Patients Using Machine Learning Techniques.

BACKGROUND: Whereas transcatheter aortic valve implantation (TAVI) has become the gold standard for aortic valve stenosis treatment in high-risk patients, it has recently been extended to include intermediate risk patients. However, the mortality rate at 5 years is still elevated. The aim of the present study was to develop a novel machine learning (ML) approach able to identify the best predictors of 5-year mortality after TAVI among several clinical and echocardiographic variables, which may improve the long-term prognosis. METHODS: We retrospectively enrolled 471 patients undergoing TAVI. More than 80 pre-TAVI variables were collected and analyzed through different feature selection processes, which allowed for the identification of several variables with the highest predictive value of mortality. Different ML models were compared. RESULTS: Multilayer perceptron resulted in the best performance in predicting mortality at 5 years after TAVI, with an area under the curve, positive predictive value, and sensitivity of 0.79, 0.73, and 0.71, respectively. CONCLUSIONS: We presented an ML approach for the assessment of risk factors for long-term mortality after TAVI to improve clinical prognosis. Fourteen potential predictors were identified with the organic mitral regurgitation (myxomatous or calcific degeneration of the leaflets and/or annulus) which showed the highest impact on 5 years mortality.

Predictive Value of Pre-Operative 2D and 3D Transthoracic Echocardiography in Patients Undergoing Mitral Valve Repair: Long Term Follow Up of Mitral Valve Regurgitation Recurrence and Heart Chamber Remodeling.

The "ideal" management of asymptomatic severe mitral regurgitation (MR) in valve prolapse (MVP) is still debated. The aims of this study were to identify pre-operatory parameters predictive of residual MR and of early and long-term favorable remodeling after MVP repair. We included 295 patients who underwent MV repair for MVP with pre-operatory two- and three-dimensional transthoracic echocardiography (2DTTE and 3DTTE) and 6-months (6M) and 3-years (3Y) follow-up 2DTTE. MVP was classified by 3DTTE as simple or complex and surgical procedures as simple or complex. Pre-operative echo parameters were compared to post-operative values at 6M and 3Y. Patients were divided into Group 1 (6M-MR < 2) and Group 2 (6M-MR ≥ 2), and predictors of MR 2 were investigated. MVP was simple in 178/295 pts, and 94% underwent simple procedures, while in only 42/117 (36%) of complex MVP a simple procedure was performed. A significant relation among prolapse anatomy, surgical procedures and residual MR was found. Post-operative MR ≥ 2 was present in 9.8%: complex MVP undergoing complex procedures had twice the percentage of MR ≥ 2 vs. simple MVP and simple procedures. MVP complexity resulted independent predictor of 6M-MR ≥ 2. Favorable cardiac remodeling, initially found in all cases, was maintained only in MR < 2 at 3Y. Pre-operative 3DTTE MVP morphology identifies pts undergoing simple or complex procedures predicting MR recurrence and favorable cardiac remodeling.

Multi-parametric "on board" evaluation of right ventricular function using three-dimensional echocardiography: feasibility and comparison to traditional two-and three dimensional echocardiographic measurements.

Three-dimensional echocardiographic (3DE) of right ventricle (RV) has been validated in many clinical settings. However, the necessity of complicated and off-line dedicated software has reduced its diffusion. A new simplified "on board" 3DE software (OB) has been developed to obtain RV volumes and ejection fraction (EF) together with several conventional parameters automatically derived from 3DE: tricuspid annular plane systolic excursion (TAPSE), fractional area change (FAC), longitudinal strain (LS). Aims of this study were to evaluate feasibility and accuracy of OB RV analysis. A complete 2DE and 3DE with OB 3DRV evaluation was obtained in 35 normal subjects and 105 patients with different pathologies. Results were compared with the conventional off-line software (OFL) and with the 2D-derived corresponding values. A subgroup of 22 patients underwent also cardiac CMR. OB 3DRV was feasible in 133/140 cases (95%) in a mean time of 97.5 ± 33 s lower than OFL analysis (129 ± 52 s plus dataset loading 80 ± 24 s). Imaging quality was good in 84%. OB and OFL 3DE RV volumes and EF were similar. 3DE derived FSA and LS (but not TAPSE) were similar to 2DE values and correlated with tissue Doppler systolic peak velocity, dP/dt, systolic pulmonary pressure and myocardial performance index. OB RV volumes and EF well correlated with CMR. (bias + SD: - 21.5 ± 20 mL for EDV; - 8.2 ± 12.4 mL for ESV; - 1 ± 5.9% for EF). OB 3DE method is feasible, simple, time saving. It easily provides 3DE RV volumes and multiple functional parameters. Off-line operator border adjustment may improve accuracy of 3DE TAPSE.

Effects of blood transfusion on exercise capacity in thalassemia major patients.

Anemia has an important role in exercise performance. However, the direct link between rapid changes of hemoglobin and exercise performance is still unknown.To find out more on this topic, we studied 18 beta-thalassemia major patients free of relevant cardiac dysfunction (age 33.5±7.2 years,males = 10). Patients performed a maximal cardiopulmolmonary exercise test (cycloergometer, personalized ramp protocol, breath-by-breath measurements of expired gases) before and the day after blood transfusion (500 cc of red cell concentrates). After blood transfusion, hemoglobin increased from 10.5±0.8 g/dL to 12.1±1.2 (p<0.001), peak VO2 from 1408 to 1546mL/min (p<0.05), and VO2 at anaerobic threshold from 965 to 1024mL/min (p<0.05). No major changes were observed as regards heart and respiratory rates either at peak exercise or at anaerobic threshold. Similarly, no relevant changes were observed in ventilation efficiency, as evaluated by the ventilation vs. carbon dioxide production relationship, or in O2 delivery to the periphery as analyzed by the VO2 vs. workload relationship. The relationship between hemoglobin and VO2 changes showed, for each g/dL of hemoglobin increase, a VO2 increase = 82.5 mL/min and 35 mL/min, at peak exercise and at anaerobic threshold, respectively. In beta-thalassemia major patients, an acute albeit partial anemia correction by blood transfusion determinates a relevant increase of exercise performance, observed both at peak exercise and at anaerobic threshold.

Incidence and severity of atherosclerotic cardiovascular artery disease in patients undergoing TAVI.

Transcatheter aortic valve implantation (TAVI) has extended the treatment options for severe, symptomatic aortic valve stenosis (AS). Risk factors for AS have been shown to be similar to atherosclerosis. Consequently, coronary artery disease (CAD), peripheral vascular and carotid artery diseases are often found concurrently with diagnostic, procedural and prognostic implications. This study sought to describe comprehensive vascular assessment in terms of prevalence, severity and correlations in TAVI candidates. A total of 323 patients (81 ± 6 years) undergoing TAVI were enrolled. Vascular pathologies were evaluated by invasive coronary angiography, computer-tomography (abdominal aorta, renal, iliac and femoral arteries), echo-color Doppler ultrasound (carotid artery), and transoesophageal echocardiography (thoracic aorta). CAD was found in 173 (54%) patients, of which 65 (38%) had 1-vessel, 45 (26%) 2-vessel and 59 (34%) 3-vessel disease. Carotid artery disease was present in 33.6% patients, of which 23.6% unilateral and 10.0% bilateral. Iliac, femoral and renal stenosis were found in 29.2%, 22.0%, and 4.7%, respectively. Cardiovascular risk factor and gender correlated with CAD. CAD patients presented more frequently with significant stenosis (luminal narrowing ≥ 50%) of at least one other district. Multi-districts significant stenosis stratified patients on long-term survival and the coexistence of 3-districts involvement with CAD negatively impacts on mortality. Multimodality imaging assessment shows that coronary, carotid, and peripheral artery disease are often found concurrently in patients undergoing TAVI. Several risk factors and gender correlate with the presence and severity of CAD and peripheral pathologies. Long-term mortality is increased in patients with a more compromised vascular situation.

Reliability and feasibility of longitudinal AFI global and segmental strain compared with 2D left ventricular volumes and ejection fraction: intra- and inter-operator, test-retest, and inter-cycle reproducibility.

AIMS: Echocardiographic evaluation of 2D longitudinal peak systolic strain (LPSS) can detect initial impairment of left ventricular (LV) function in heart disease. Global LPSS (GLPSS) variability has been assessed in small groups and segmental LPSS has not been determined. We compared variability of GLPSS and segmental LPSS with that of 2D LV volumes and ejection fraction (EF) in patients with and without heart diseases. METHODS AND RESULTS: 2D speckle tracking analysis was performed on LV apical views using automated function imaging (AFI) software (GE Healthcare). Intra-operator, inter-cycle, and test-retest variability (bias and CR, coefficient of reproducibility; MPE, mean percent error; CV, coefficient of variation) was assessed for GLPSS, 18 segments of LPSS, and LV volumes and EF in 40 patients (720 segments), and inter-operator variability in 250 patients (4500 segments). Feasibility of segmental tracking was 93.1%. Variability of GLPSS increased from a minimum intra-operator CV = -2.6% to a maximum test-retest CV = -5.4% and was lower than that assessed for volumes and EF. Segmental intra-operator LPSS CV ranged from -5.6 to -14.7%, and test-retest from -8 to -22%, and was at worst similar to variability of end-systolic volume. In the 8.3% of segments with the highest variability, this was related to suboptimal imaging, minor changes in scan angulation, and insufficient ROI width. CONCLUSION: Overall, reproducibility of GLPSS is excellent and superior to that of 2D EF, whereas segmental LPSS reproducibility is good and similar to that of LV volumes. Both are suitable for diagnosis and follow-up of LV global and regional systolic function.

Feasibility and accuracy of three-dimensional transthoracic echocardiography vs. multidetector computed tomography in the evaluation of aortic valve annulus in patient candidates to transcatheter aortic valve implantation.

AIMS: Proper measurement of the aortic annulus (AoA) is crucial for the success of transcatheter aortic valve implantation (TAVI). Transthoracic echocardiography (TTE) is the first step to assess AoA diameter, but a two-dimensional TTE (2DTTE) measurement is no longer accepted as the sole determinant of prosthetic size. The aims of the study were to evaluate feasibility and accuracy of three-dimensional TTE (3DTTE) estimation of AoA dimensions in comparison with multidetector computer tomography (MDCT). METHODS AND RESULTS: We enrolled 100 consecutive patients referred for TAVI. Feasibility of AoA evaluation was 91% for 3DTTE and in 90% for MDCT. In 81 of 100 patients, AoA maximum diameter (max-D), minimum diameter (min-D), and area were measured and compared using 2DTTE, 3DTTE, and MDCT. Image quality of 3DTTE was sufficient in 47, good in 46 and optimal in 7%. High correlations (P < 0.001) were found between MDCT and 3DTTE (max-D: r = 0.89; min-D r = 0.86; area: r = 0.93), and between MDCT and 2DTTE (min-D: r = 0.81; area 0.78). The 3DTTE measurements were found to be highly reproducible on intra- and interobserver variability analyses. Regarding the choice of prosthesis size, agreement between 3DTTE and MDCT was very good (k = 0.84, P < 0.001) while it was poor between 2DTTE and MDCT (k = 0.36, P < 0.001). CONCLUSIONS: 3DTTE may be a valid imaging alternative in patients unsuitable for MDCT during the preoperative evaluation for TAVI. Evaluation of AoA through 3DTTE is feasible, and measurements closely approximate those of MDCT thus improving TTE accuracy in identifying the correct prosthesis size.

Prevalence of calcification of the mitral valve annulus in patients undergoing surgical repair of mitral valve prolapse.

Factors correlating to mitral annulus calcification (MAC) include risk factors predisposing to atherosclerosis. In patients with mitral valve (MV) prolapse (MVP), other anatomic or mechanical factors have been supposed to facilitate MAC. The aims of this study were, in patients with MVP undergoing MV repair, (1) to describe the prevalence and characteristics of MAC, (2) to correlate MAC with clinical risk factors, coronary involvement, and aortic valve disease, and (3) to describe prevalence, site, and extension of MAC in fibroelastic deficiency (FED) versus Barlow's disease (BD) and correlate MAC to surgical outcomes (repair vs replacement). In 410 consecutive patients with MVP suitable for surgical MV repair, detailed clinical and echocardiographic data were collected to characterize MAC in BD and FED. MAC was found in 99 patients (24%). Age, female gender, coronary artery disease, and cardiovascular risk factors were correlated with MAC. MAC was equally distributed in FED and BD groups despite patients with FED being older with more cardiovascular risk factors. The most common localization of MAC was annular involvement adjacent to P2 (75%), P1 (31%), and P3 (35%). The presence of MAC affected surgical outcomes in both groups (8% patients with MAC underwent replacement after a first attempt of repair vs 3% without MAC). MAC is a common finding in patients undergoing MV repair, and several clinical characteristics correlate with MAC either in FED or BD. In conclusion, despite very high percentage of repairability, MAC influences surgical outcomes and very detailed echo evaluation is advocated.

Role of Intracardiac echocardiography in Atrial Fibrillation Ablation.

In the recent years, several new evidences support catheter-based ablation as a treatment modality of atrial fibrillation (AF). Based on a plenty of different applications, intracardiac echocardiography (ICE) is now a well-established technology in complex electrophysiological procedures, in particular in AF ablation. ICE contributes to improve the efficacy and safety of such procedures defining the anatomical structures involved in ablation procedures and monitoring in real time possible complications. In particular ICE allows: a correct identification of the endocardial structures; a guidance of transseptal puncture; an assessment of accurate placement of the circular mapping catheter; an indirect evaluation of evolving lesions during radiofrequency (RF) energy delivery via visualization of micro and macrobubbles tissue heating; assessment of catheter contact with cardiac tissues. Recently, also the feasibility of the integration of electroanatomical mapping (EAM) and intracardiac echocardiography has been demonstrated, combining accurate real time anatomical information with electroanatomical data. As a matter of fact, different techniques and ablation strategies have been developed throughout the years. In the setting of balloon-based ablation systems, recently adopted by an increasing number of centers, ICE might have a role in the choice of appropriate balloon size and to confirm accurate occlusion of pulmonary veins. Furthermore, in the era of minimally fluoroscopic ablation, ICE has successfully provided a contribute in reducing fluoroscopy time. The purpose of this review is to summarize the current applications of ICE in catheter based ablation strategies of atrial fibrillation, focusing-on electronically phased-array ICE.

Intraoperative 2D and 3D transoesophageal echocardiographic predictors of aortic regurgitation after transcatheter aortic valve implantation.

BACKGROUND: Post-procedural aortic regurgitation (AR) has been described in a large number of patients receiving transcatheter aortic valve implantation (TAVI). OBJECTIVE: The aim of this study was to examine the intraoperative 2-dimensional (2D) and 3-dimensional (3D) echocardiographic features of the aortic valve associated with significant post-procedural paravalvular AR. METHODS: A total of 135 patients (81±7 years) with severe symptomatic aortic stenosis, who underwent TAVI, were imaged with comprehensive 2D and 3D transoesophageal echocardiography before the procedure and peri-procedure. Various baseline and peri-procedural echocardiographic characteristics were tested to predict paravalvular AR post-TAVI: calcifications at the aortic valve commissures and leaflets, 'aortic annulus eccentricity index', 'area cover index', overlap between aortic prosthesis and anterior mitral leaflet. Post-procedural paravalvular AR≥2 was considered significant. RESULTS: Successful TAVI was achieved in all patients. The incidence of paravalvular AR≥2 immediately after the procedure was 21% (28 patients). Commissural calcifications and, particularly, the calcification of the commissure between the right coronary and non-coronary cusps was significantly more frequent in presence of paravalvular AR; the area cover index pre-TAVI was significantly lower among patients with AR (11.1±11.8% vs 20.8±12.5%, p=0.0004). Multivariate analysis revealed that calcification of the commissure between the right coronary and non-coronary cusps (OR=2.66, 95% CI 1.39 to 5.12, p=0.001), and the area cover index pre-TAVI (OR=0.95, 95% CI 0.91 to 0.99, p=0.006) were the only independent predictors of significant paravalvular AR after TAVI. CONCLUSIONS: Intraoperative 2D and 3D transoesophageal echocardiography identified calcification of the commissure between the right coronary and non-coronary cusps and the area cover index as independent predictors of significant paravalvular AR following TAVI.

Feasibility and accuracy of 3DTEE versus CT for the evaluation of aortic valve annulus to left main ostium distance before transcatheter aortic valve implantation.

OBJECTIVES: The aims of this study were to analyze in a large series of patients undergoing transcatheter aortic valve implantation (TAVI): 1) the accuracy of 3-dimensional transesophageal echocardiographic (3DTEE) measurement of left coronary cusp (LCC) length and of the distances from left main coronary ostium (LM) to the aortic annulus (AA) pre-operatively and to the aortic prosthesis post-operatively; and 2) the role of the 3DTEE measurements in predicting the prosthetic deployment and the association between prosthesis position and aortic regurgitation (AR) and/or prosthesis-patient mismatch (PPM). BACKGROUND: Coronary ostia occlusion is a possible complication in TAVI; therefore, the careful pre-operative evaluation of AA-LM and LCC length, and the post-operative analysis of the relationship between the prosthesis and LM, may influence the procedural outcomes. Even though multidetector computed tomography (MDCT) is the gold standard pre-operatively, sometimes it cannot be performed and it is rarely repeated post-operatively. METHODS: In 122 patients undergoing TAVI, pre-operative AA-LM and LCC measurements obtained by 3DTEE and MDCT were compared. Post-operatively, the feasibility of 3DTEE evaluation of the prosthesis-LM distance was performed. The relationship between 3DTEE overlap of the prosthesis with the anterior mitral leaflet and AR/PPM was assessed. RESULTS: Pre-operatively, 3DTEE AA-LM (r = 0.83) and LCC (r = 0.69) significantly correlated with MDCT. Post-operatively, 3DTEE prosthesis-LM distance was 2.1 ± 1.9 mm. The prosthesis reached or exceeded LM in 6 and 10 cases, respectively. Prosthesis overlap with mitral leaflet was 4.7 ± 1.8 mm. Significant correlation between the 3DTEE computed and nominal length of the prosthesis was found (r = 0.61). No correlations were found between prosthesis-mitral leaflet overlap and aortic regurgitation or PPM. CONCLUSIONS: AA-LM distance and LCC length may be accurately estimated by 3DTEE, which may represent a valid alternative to MDCT. Pre- and post-3DTEE data concerning the aortic root, such as LM, aortic valve, and prosthetic morphology, give new insights into TAVI and its complications.

Effects of carvedilol on oxygen uptake and heart rate kinetics in patients with chronic heart failure at simulated altitude.

BACKGROUND: The response to moderate exercise at altitude in heart failure (HF) is unknown. METHODS AND RESULTS: We evaluated 30 HF patients, (NYHA I-III, 25 M/5 F; 59 ± 10 years; LVEF = 39.6 ± 7.1%), in stable clinical conditions, treated with carvedilol at the maximal tolerated dose. We performed a maximal cardiopulmonary exercise test (CPET) with ramp protocol at sea level to evaluate patients' performance and two moderate intensity constant workload CPETs (50% of peak workload) at sea level (normoxia) and simulated altitude (hypoxia). Oxygen uptake (.VO2 ) and heart rate (HR) on-kinetics at constant workload were assessed calculating the time constant (τ) with a monoexponential equation. .VO2 and HR were higher in hypoxia (0.944 ± 0.233 vs 1.031 ± 0.264 l/min; 100 ± 23 vs 108 ± 22 bpm; p < 0.001). On-kinetics showed a different behavior of τ being .VO2 faster in hypoxia (67.1 ± 23.0 vs. 56.3 ± 19.7 s; p = 0.026) and HR faster in normoxia (49.3 ± 19.4 vs. 62.2 ± 22.5 s; p = 0.018). Ten patients, who lowered oxygen kinetics in hypoxia, had greater HR increase during maximal CPET suggesting lower functional beta-blockade. The higher τ of .VO2 in hypoxia is likely to be due to a peripheral effect of carvedilol mediated either by ß- or α-receptor. CONCLUSION: HF patients performing moderate exercise at 2000 m simulated altitude have 20% .VO2 increase without trouble at the beginning of exercise when treated with carvedilol.