Citations

Introduction: Multiple supervised machine learning (ML) models have been used to assess the outcomes in patients who underwent transcatheter aortic valve replacement (TAVR) procedure, however, supervised ML requires significant expert time for annotation and labeling. Unsupervised ML can provide insight into the dataset without this process. Hypothesis: We hypothesized that unsupervised k-means ML can identify distinct phenogroups in TAVR patients that can predict prognosis without a priori knowledge. Methods: A retrospective study of patients who underwent TAVR (Jan 2012-Dec 2017) at Mayo Clinic sites was performed. Baseline demographics, EKG, cardiac computed tomography, and echocardiographic data were abstracted. Variables with ≥ 20% missing data were excluded. Prior to the k-means clustering algorithm, principal component analysis was used for dimensionality reduction. Input variables were analyzed according to k-means determined clusters. Kaplan-Meier survival analysis was used to assess the survival of each cluster. Results: A total of 1071 consecutive patients were included in the final analysis. A total of 141 unique variables were used. The mean age was 80.7±8.6 years, 58.2% were male. Patients were divided into 2 distinct phenogroups. In the two phenogroups, cluster 0 (n=639) has significantly better survival than cluster 1 (n= 432) (log-rank p=0.0002) (Figure 1). Cluster 1 has features of low flow-low gradient (LFLG) hemodynamic profile, including lower systolic blood pressure, mean aortic valve gradient, peak aortic valve velocity, and worse LV ejection fraction. Conclusions: Unsupervised ML can identify distinct phenotype features with prognostic significance in TAVR patients. Our results suggest that patients with LFLG hemodynamic features are intrinsically different from other TAVR patients and require more attention to avoid adverse outcomes.

BackgroundAdvanced diffusion‐based MRI biomarkers may provide insight into microstructural and perfusion changes associated with neurodegeneration and cognitive decline.PurposeTo assess longitudinal microstructural and perfusion changes using apparent diffusion coefficient (ADC) and intravoxel incoherent motion diffusion‐weighted imaging (IVIM‐DWI) parameters in cognitively impaired (CI) and healthy control (HC) groups.Study TypeProspective/longitudinal.PopulationTwelve CI patients (75% female) and 13 HC subjects (69% female).Field Strength/Sequence3 T; Spin‐Echo‐IVIM‐DWI.AssessmentTwo MRI scans were performed with a 12‐month interval. ADC and IVIM‐DWI metrics (diffusion coefficient [D] and perfusion fraction [f]) were generated from monoexponential and biexponential fits, respectively. Additionally, voxel‐based correlations were evaluated between change in Montreal Cognitive Assessment (ΔMoCA) and baseline imaging parameters.Statistical TestsAnalysis of covariance with sex and age as covariates was performed for main effects of group and time (false discovery rate [FDR] corrected) with post hoc comparisons using Bonferroni correction. Partial‐η2 and Hedges' g were used for effect‐size analysis. Spearman's correlations (FDR corrected) were used for the relationship between ΔMoCA score and imaging. P < 0.05 was considered statistically significant.ResultsSignificant differences were found for the main effects of group (HC vs. CI) and time. For group effects, higher ADC, IVIM‐D, and IVIM‐f were observed in the CI group compared to HC (ADC: 1.23 ± 0.08.10−3 vs. 1.09 ± 0.07.10−3 mm2/sec; IVIM‐D: 0.82 ± 0.01.10−3 vs. 0.73 ± 0.01.10−3 mm2/sec; and IVIM‐f: 0.317 ± 0.008 vs. 0.253 ± 0.009). Significantly higher ADC, IVIM‐D, and IVIM‐f values were observed in the CI group after 12 months (ADC: 1.45 ± 0.05.10−3 vs. 1.50 ± 0.07.10−3 mm2/sec; IVIM‐D: 0.87 ± 0.01.10−3 vs. 0.94 ± 0.02.10−3 mm2/sec; and IVIM‐f: 0.303 ± 0.007 vs. 0.332 ± 0.008), but not in the HC group at large effect size. ADC, IVIM‐D, and IVIM‐f negatively correlated with ΔMoCA score (ρ = −0.49, −0.51, and −0.50, respectively).Data ConclusionThese findings demonstrate that longitudinal differences between CI and HC cohorts can be measured using IVIM‐based metrics.Level of Evidence2Technical Efficacy Stage2

Introduction: Conduction disturbance after transcatheter aortic valve replacement (TAVR) remains a common complication. Predictive tools to identify conduction disturbance requiring pacing are limited. Our main objective was to develop a deep learning model to predict the need for a permanent pacemaker. Hypothesis: An electrocardiogram (ECG) based deep learning algorithm can predict the need for permanent pacemaker post TAVR. Methods: We evaluated patients who had TAVR at Mayo Clinic between June 2010 and May 2021. Digital ECGs done on the TAVR date or within 24 hours post procedure were partitioned into training, validation, and testing datasets. We trained a deep learning algorithm on the training data and optimized hyperparameters against the validation data. All model performance metrics are based on the reserved test data that were estimated once model architecture was optimized. Need for permanent pacemaker, defined as pacemaker placement due to development of conduction disturbance post TAVR and at least 1% ventricular pacing on device interrogation done at least 2 weeks post pacemaker implantation, was the training target. Results: A total of 3021 patients were evaluated for inclusion and 2,407 cases were available for modeling (exclusions: prior pacemaker (n=419), unsuccessful TAVR (n=27) and no ECG within the pre-specified period (n=168)). Overall, the median age was 82 (Q1: 76, Q3: 86) and 41% were female. In the withheld test data, 15% had a permanent pacemaker with at least 1% ventricular pacing. The algorithm detected the need for pacemaker in the test set with an AUC of 0.83 (95% CI: 0.73 to 0.92). Accuracy, sensitivity, specificity, positive and negative predictive were 80%, 72%, 81%, 41% and 94%, respectively. Conclusions: This study demonstrates the ability of a deep learning model based on ECG data to assist in making the decision to implant a pacemaker following TAVR with a high negative predictive value. Larger validation studies are warranted.

Background: Transcatheter aortic valve replacement (TAVR) is an approved alternative to surgical aortic valve replacement. However, as the clinical indication and adoption of TAVR expands, it remains important to identify predictors of mortality after TAVR. We aimed to study the impact of aortic valve calcium score on long term mortality following TAVR in a large patient population. Methods: We included consecutive patients who successfully underwent TAVR from June 2010 – June 2019 across all Mayo Clinic sites. Patients were followed through February 2021. Aortic valve calcium scores were obtained from preoperative computed tomography angiography of the chest. Additional data was abstracted from medical records. Cox-proportional hazard regression models were employed to evaluate the independent effect of aortic valve calcium score on mortality. Results: A total of 1988 patients were evaluated in the final analysis. Forty-two percent were women, median age was 83 years (Q1: 77, Q3: 87), 16% received a permanent pacemaker following TAVR and 83% received an Edwards Sapien valve. We demonstrate an incremental increase in mortality risk as aortic valve calcium score values increased after adjusting for age, gender, pacemaker implantation, left bundle branch block, and Charlson Comorbidity Index. Selected calcium score cut-points with adjusted hazard ratios (aHR) are displayed: 1200 AU (aHR: 1.13, 95% CI 1.06, 1.20); 2000 AU (aHR: 1.22, 95% CI 1.09, 1.35); 3000 AU (aHR: 1.34, 95% CI: 1.14, 1.58). Conclusion: This study demonstrates an association between aortic valve calcium score and long-term mortality in a large patient population. This highlights the potential predictive value of the calcium score in preoperative planning and may provide clinicians with an additional measure for risk stratification prior to TAVR.

Introduction: Unsupervised machine learning (ML) can provide insight into a dataset without significant expert time for annotation and labeling, but was never used to assess the heterogeneous MitraClip population. Hypothesis: We hypothesized that unsupervised k-means ML could identify prognostically-distinct phenogroups in MitraClip patients without a priori knowledge of the dataset. Methods: Patients who underwent MitraClip (June 2014-September 2020) at Mayo Clinic sites were identified from the Mayo institutional NCDR database for baseline and follow-up data. Variables with ≥ 50% missing data were excluded. A k-means algorithm was used for clustering analysis. Input variables were grouped according to 4 distinct k-means determined clusters. Kaplan-Meier survival analysis was used to assess the survival of each cluster. Results: A total of 389 consecutive patients were included in the final analysis, and 95 unique variables were used. The mean age was 80.3±8.7 years; 256 (65.8%) were male. The mean STS MV Replace score was 9.1±5.7. Fifty-five (14.5%) patients died during the mean follow duration (185 days). Kaplan-Meier analysis showed significant survival differences among clusters (Figure 1). The cluster (cluster 2, n=120) with best survival performance (all-cause mortality 3.3%) had features of non-calcified mitral leaflets, nonischemic cardiomyopathy, and less than moderate tricuspid regurgitation. Cluster 3 (n=66) had the worst survival performance (all-cause mortality 27.3%) and opposite features than cluster 2. Conclusions: Unsupervised ML can identify distinct phenotype clusters with prognostic significance in MitraClip patients. Patients with non-calcified mitral leaflets, nonischemic cardiomyopathy, and less than moderate tricuspid regurgitation are intrinsically different from other MitraClip patients and should be considered the potential optimal candidates.

Background/Introduction: More than moderate tricuspid regurgitation (TR) at baseline is associated with worse outcome in MitraClip patients. The presence of iatrogenic ASD (iASD) may cause right-sided volume overload and lead to adverse events. However, the real effects of iASD are not well-investigated in these patients who are vulnerable to right-sided volume overload. Methods: Patients with more than moderate TR who underwent a successful MitraClip procedure in the period from June 2014 to September 2020 were identified from the Mayo institutional NCDR database. Follow-up transthoracic echocardiography (TTE) at about 1 year after the procedure was used to determine the presence of iASD, and speckle-tracking strain analysis was done. Patients were grouped according to the presence of iASD (iASD+ group vs. iASD- group) for statistical analysis. Results: A total of 141 consecutive patients were included for the final analysis (iASD+ group n=44 and iASD- group n=97). The mean age was 79.3 ± 8.8 years, and 72.3% were male. Follow-up echocardiography and speckle-tracking strain data did not show a significant difference in either left or right ventricular function in conventional echocardiographic parameters and strain. LA reservoir strain (13.7 ± 6.9 vs 11.5±7.5, P=0.0418) was the only parameter with significant difference among the two groups. Details are summarized in Table 1 . Conclusion: The presence of iASD can decompress the left atrium, but does not affect left or right ventricular strain function in MitraClip patients with more than moderate baseline TR.

AbstractBackgroundFrontotemporal lobar degeneration (FTLD) refers to a group of neurodegenerative conditions, affecting the frontal and/or temporal lobes. Ongoing research has provided insight into developing clinical trials for FTLD and key clinical measures such as structural MRI. To inform clinical trial design and optimize participation, it is imperative to explore facilitators and barriers for potential candidates.ObjectiveThe objective of this study is to explore facilitators and barriers to participating in future clinical trials for FTLD.MethodsAdvancing Research and Treatment for Frontotemporal Lobar Degeneration (ARTFL) and Longitudinal Evaluation of Familial Frontotemporal Dementia Subjects (LEFFTDS) are observational studies focused on characterizing FTLD syndromes in preparation for clinical trials. The 584 participants enrolled across 18 research sites in the United States and Canada completed a survey assessing interest in clinical trial participation.Results29% of respondents self‐reported as patients (63±10 years), 26% self‐reported as caregivers answering on behalf of patients (65±10 years), and 45% self‐reported as healthy but at risk for FTLD (48±14 years). Travel reimbursement was the most common factor reported to positively influence participation (≧66%), with the healthy but at risk group showing the strongest endorsement (83%). Cost and time involved in travel were possible barriers for about half of the patients (48%) and healthy but at risk respondents (53%). The respondents value receiving feedback on the study findings (≧80%) and being informed of their individual disease progression (≧75%). Particularly, keeping participation confidential was very important for the healthy but at risk group (62%). In regard to research assessments, most participants demonstrated a high interest in physical and neurological exams at a research center (≧87%) whereas only half were interested in doing more invasive procedures such as the lumbar puncture (≧52%). Overall, respondents showed a positive attitude and support for research participation (≧77%) and trusted that their health information would remain confidential in a clinical trial (≧53%).ConclusionsFavorable attitudes and interest towards medical research exist among participants. To optimize participation, clinical trials should allocate funding for travel and involve participants in feedback about study results and their disease progression. Alternatives to invasive assessments may increase participation.

AbstractObjectivesTo assess the impact of pulmonary hypertension (PH) on outcomes of patients with severe mitral annular calcification (MAC) undergoing transcatheter mitral valve replacement (TMVR).BackgroundPH is associated with poor outcomes after mitral valve surgery. Whether the presence of PH in patients with MAC undergoing (TMVR) is associated with poor outcomes, is unknown.MethodsRetrospective evaluation of 116 patients from 51 centers in 11 countries who underwent TMVR with valve in mitral annular calcification (ViMAC) using balloon‐expandable aortic transcatheter valves (THVs) from September 2012 to March 2017. Pulmonary artery systolic blood pressure (PASP) by echocardiogram was available in 90 patients. The subjects were stratified based on PASP: No PH = PASP ≤35 mmHg (n = 11); mild to moderate PH = PASP 36–49 mmHg (n = 21) and severe PH = PASP ≥50 mmHg (n = 58). Clinical, procedural, and echocardiographic outcomes were assessed.ResultsMean age was 72.7 (±12.8) years, 59 (65.6%) were female, Society of Thoracic Surgeons score was 15.8 + 11.8% and 90.0% where in New York Heart Association (NYHA) class III–IV. There was no significant difference in all‐cause mortality at 30 days (no PH = 27.3%, mild–moderate PH = 19.0%, severe PH = 31.6%; p = 0.55) or at 1 year (no PH = 54.5%, mild–moderate PH = 38.1%, severe PH = 56.1%; p = 0.36). No difference in adverse events, NYHA class or amount of residual mitral regurgitation at 1 year were observed between the groups.ConclusionThis study suggests that the presence of PH in patients with predominantly mitral stenosis with MAC undergoing TMVR does not impact mortality or adverse events. Further studies are needed to fully understand the effect of PH in this group of patients.

Background: Several studies have evaluated preprocedural imaging predictors of left ventricular outflow tract obstruction (LVOTO) after transcatheter mitral valve replacement. The patient cohorts in these studies were heterogeneous and included patients with transcatheter mitral valve replacement in failed bioprostheses, annuloplasty rings, and severe mitral annular calcification (MAC). The goal of this study was to evaluate predictors of LVOTO specific to patients undergoing valve-in-MAC. Methods: This study included patients with severe MAC who underwent valve-in-MAC and had optimal quality preprocedural multidetector row computed tomography scans eligible for retrospective analysis. Baseline demographic, echocardiographic, and procedural data on these patients were collected. multidetector row computed tomography parameters were analyzed for association with LVOTO, defined as increase in mean LVOT gradient by ≥10 mm Hg with accompanying hemodynamic instability. Results: Seventy-one patients with optimal preprocedural computed tomography scans were included in this study (mean age, 72.5±13.5 years), 9 of which developed LVOTO (all female). Baseline mean LVOT area, neo-LVOT area (145.3 versus 270.9 mm 2 ; P =0.006), indexed neo-LVOT area (90.1 versus 157.4; P =0.05), and virtual transcatheter heart valve to septum distance (3.1 versus 6.9 mm; P =0.002) were lower in the LVOTO group. Expected % LVOT area reduction was higher in the latter group (58.3 versus 42.7%; P =0.008). In the univariable analysis, the baseline mean LVOT area, neo-LVOT area, indexed neo-LVOT area, and valve to septum distance were all significantly associated with LVOTO. Conclusions: The systolic mean LVOT area, neo-LVOT area, indexed neo-LVOT, expected percentage LVOT area reduction, and the valve to septum distance were associated with LVOTO after valve-in-MAC.