The prognostic value of artificial intelligence to predict cardiac amyloidosis in patients with severe aortic stenosis undergoing transcatheter aortic valve replacement.

Aims: Cardiac amyloidosis (CA) is common in patients with severe aortic stenosis (AS) undergoing transcatheter aortic valve replacement (TAVR). Cardiac amyloidosis has poor outcomes, and its assessment in all TAVR patients is costly and challenging. Electrocardiogram (ECG) artificial intelligence (AI) algorithms that screen for CA may be useful to identify at-risk patients. Methods and results: In this retrospective analysis of our institutional National Cardiovascular Disease Registry (NCDR)-TAVR database, patients undergoing TAVR between January 2012 and December 2018 were included. Pre-TAVR CA probability was analysed by an ECG AI predictive model, with >50% risk defined as high probability for CA. Univariable and propensity score covariate adjustment analyses using Cox regression were performed to compare clinical outcomes between patients with high CA probability vs. those with low probability at 1-year follow-up after TAVR. Of 1426 patients who underwent TAVR (mean age 81.0 ± 8.5 years, 57.6% male), 349 (24.4%) had high CA probability on pre-procedure ECG. Only 17 (1.2%) had a clinical diagnosis of CA. After multivariable adjustment, high probability of CA by ECG AI algorithm was significantly associated with increased all-cause mortality [hazard ratio (HR) 1.40, 95% confidence interval (CI) 1.01-1.96, P = 0.046] and higher rates of major adverse cardiovascular events (transient ischaemic attack (TIA)/stroke, myocardial infarction, and heart failure hospitalizations] (HR 1.36, 95% CI 1.01-1.82, P = 0.041), driven primarily by heart failure hospitalizations (HR 1.58, 95% CI 1.13-2.20, P = 0.008) at 1-year follow-up. There were no significant differences in TIA/stroke or myocardial infarction. Conclusion: Artificial intelligence applied to pre-TAVR ECGs identifies a subgroup at higher risk of clinical events. These targeted patients may benefit from further diagnostic evaluation for CA.

Efficacy of Sacubitril-Valsartan on Survival and Cardiac Remodeling in Hypotensive Heart Failure With Reduced Ejection Fraction: A Multicenter Study.

OBJECTIVE: To investigate whether hypotensive patients diagnosed with heart failure and reduced ejection fraction (HFrEF) might benefit from angiotensin receptor-neprilysin inhibitors (ARNis) in real-world practice because patients with baseline systolic blood pressure (SBP) of less than 100 mm Hg have been excluded from landmark trials. PATIENTS AND METHODS: In this multicenter study conducted between January 1, 2013, and December 31, 2021, a total of 7562 symptomatic patients with HFrEF were enrolled and grouped by SBP (hypotension was defined as an SBP of less than 100 mm Hg) and ARNi use as follows: group 1, hypotensive/non-ARNi users (n=484); group 2, hypotensive/ARNi users (n=308); group 3, nonhypotensive/non-ARNi users (n=4560); and group 4, nonhypotensive/ARNi users (n=2210). Inverse probability of treatment weighting was used to balance baseline characteristics for survival analysis. RESULTS: Diverse baseline characteristics and lower rates of medication use were found among non-ARNi users compared with ARNi users. Hypotensive/ARNi users had lower ARNi initiation doses than nonhypotensive/ARNi users. We observed significantly lower mortality, composite heart failure hospitalization, and CV death for hypotensive/ARNi and the other 2 nonhypotensive groups (groups 3 and 4) during a median follow-up of 3.43 years (all P<.05), with a similar effect on reverse remodeling for the hypotensive/ARNi group compared with the hypotensive/non-ARNi group. The event-free survival benefits of ARNi vs renin-angiotensin system inhibitors were consistent with the lower boundary of SBP for clinical benefits found until 88 mm Hg (spline curves) after inverse probability of treatment weighting. CONCLUSION: Patients with HFrEF and hypotension may still benefit from ARNi treatment. Patients with hypotensive HFrEF should not be routinely excluded from ARNi use in a real-world setting.

Echocardiography-Based Deep Learning Model to Differentiate Constrictive Pericarditis and Restrictive Cardiomyopathy.

BACKGROUND: Constrictive pericarditis (CP) is an uncommon but reversible cause of diastolic heart failure if appropriately identified and treated. However, its diagnosis remains a challenge for clinicians. Artificial intelligence may enhance the identification of CP. OBJECTIVES: The authors proposed a deep learning approach based on transthoracic echocardiography to differentiate CP from restrictive cardiomyopathy. METHODS: Patients with a confirmed diagnosis of CP and cardiac amyloidosis (CA) (as the representative disease of restrictive cardiomyopathy) at Mayo Clinic Rochester from January 2003 to December 2021 were identified to extract baseline demographics. The apical 4-chamber view from transthoracic echocardiography studies was used as input data. The patients were split into a 60:20:20 ratio for training, validation, and held-out test sets of the ResNet50 deep learning model. The model performance (differentiating CP and CA) was evaluated in the test set with the area under the curve. GradCAM was used for model interpretation. RESULTS: A total of 381 patients were identified, including 184 (48.3%) CP, and 197 (51.7%) CA cases. The mean age was 68.7 ± 11.4 years, and 72.8% were male. ResNet50 had a performance with an area under the curve of 0.97 to differentiate the 2-class classification task (CP vs CA). The GradCAM heatmap showed activation around the ventricular septal area. CONCLUSIONS: With a standard apical 4-chamber view, our artificial intelligence model provides a platform to facilitate the detection of CP, allowing for improved workflow efficiency and prompt referral for more advanced evaluation and intervention of CP.

Challenges and solutions of echocardiography generalization for deep learning: a study in patients with constrictive pericarditis.

Purpose: The inherent characteristics of transthoracic echocardiography (TTE) images such as low signal-to-noise ratio and acquisition variations can limit the direct use of TTE images in the development and generalization of deep learning models. As such, we propose an innovative automated framework to address the common challenges in the process of echocardiography deep learning model generalization on the challenging task of constrictive pericarditis (CP) and cardiac amyloidosis (CA) differentiation. Approach: Patients with a confirmed diagnosis of CP or CA and normal cases from Mayo Clinic Rochester and Arizona were identified to extract baseline demographics and the apical 4 chamber view from TTE studies. We proposed an innovative preprocessing and image generalization framework to process the images for training the ResNet50, ResNeXt101, and EfficientNetB2 models. Ablation studies were conducted to justify the effect of each proposed processing step in the final classification performance. Results: The models were initially trained and validated on 720 unique TTE studies from Mayo Rochester and further validated on 225 studies from Mayo Arizona. With our proposed generalization framework, EfficientNetB2 generalized the best with an average area under the curve (AUC) of 0.96 (±0.01) and 0.83 (±0.03) on the Rochester and Arizona test sets, respectively. Conclusions: Leveraging the proposed generalization techniques, we successfully developed an echocardiography-based deep learning model that can accurately differentiate CP from CA and normal cases and applied the model to images from two sites. The proposed framework can be further extended for the development of echocardiography-based deep learning models.

Developing an Echocardiography-Based, Automatic Deep Learning Framework for the Differentiation of Increased Left Ventricular Wall Thickness Etiologies.

AIMS: Increased left ventricular (LV) wall thickness is frequently encountered in transthoracic echocardiography (TTE). While accurate and early diagnosis is clinically important, given the differences in available therapeutic options and prognosis, an extensive workup is often required to establish the diagnosis. We propose the first echo-based, automated deep learning model with a fusion architecture to facilitate the evaluation and diagnosis of increased left ventricular (LV) wall thickness. METHODS AND RESULTS: Patients with an established diagnosis of increased LV wall thickness (hypertrophic cardiomyopathy (HCM), cardiac amyloidosis (CA), and hypertensive heart disease (HTN)/others) between 1/2015 and 11/2019 at Mayo Clinic Arizona were identified. The cohort was divided into 80%/10%/10% for training, validation, and testing sets, respectively. Six baseline TTE views were used to optimize a pre-trained InceptionResnetV2 model. Each model output was used to train a meta-learner under a fusion architecture. Model performance was assessed by multiclass area under the receiver operating characteristic curve (AUROC). A total of 586 patients were used for the final analysis (194 HCM, 201 CA, and 191 HTN/others). The mean age was 55.0 years, and 57.8% were male. Among the individual view-dependent models, the apical 4-chamber model had the best performance (AUROC: HCM: 0.94, CA: 0.73, and HTN/other: 0.87). The final fusion model outperformed all the view-dependent models (AUROC: HCM: 0.93, CA: 0.90, and HTN/other: 0.92). CONCLUSION: The echo-based InceptionResnetV2 fusion model can accurately classify the main etiologies of increased LV wall thickness and can facilitate the process of diagnosis and workup.

Real-World Experience of Angiotensin Receptor-Neprilysin Inhibition in Reduced Ejection Fraction Heart Failure Patients With Advanced Kidney Disease.

OBJECTIVE: To investigate the effectiveness and safety of angiotensin receptor-neprilysin inhibitors (ARNIs) in real-world patients with heart failure with reduced ejection fraction (HFrEF) and advanced chronic kidney disease (estimated glomerular filtration rate [eGFR] < 30 mL/min per 1.73 m2), which have been excluded from the landmark trials. PATIENTS AND METHODS: This study examined 3281 patients pooled from two multicenter HFrEF cohorts, and 661 patients with baseline eGFR less than 30 mL/min per 1.73 m2 were further analyzed (the Taiwan Society of Cardiology - Heart Failure with reduced Ejection Fraction (TSOC-HFrEF) registry: May 1, 2013 to October 31, 2014, and the Treatment with Angiotensin Receptor neprilysin inhibitor fOr Taiwan Heart Failure patients (TAROT-HF) study: March 1, 2017, to December 31, 2018). Propensity score matching was performed to adjust for confounders. At 1-year follow-up, all-cause mortality, total heart failure hospitalizations, renal function, and left ventricular ejection fraction (LVEF) were used as the endpoints. RESULTS: After propensity score matching, 510 patients (age, 69.8±13.9 years; male, 61.0%; mean LVEF, 29.8±7.3%; mean eGFR, 19.8±9.0 mL/min per 1.73 m2) were included in the final analysis, including 278 patients receiving ARNI treatment (ARNI group) and 232 patients not on ARNI treatment (non-ARNI group). Baseline characteristics were comparable between the two groups. At 1 year, eGFR and LVEF measurements were significantly higher in the ARNI group than in the non-ARNI group (25.0±17.1 mL/min per 1.73 m2 vs 21.4±17.5 mL/min per 1.73 m2; P=.04; and 40.1±12.9% vs. 33.1±10.8%, P<.001, respectively). The ARNI group had significantly lower risks of 1-year all-cause mortality (19.4 vs 30.9 per 100-person year; P=.02), and total HF rehospitalizations (70.0 vs 110.4 per 100-person year; P=.01) than non-ARNI users. CONCLUSION: Our results show the effectiveness of ARNIs in HFrEF patients with advanced chronic kidney disease in a real-world setting.

A New Insight Into Nonbacterial Thrombotic Endocarditis: A Systematic Review of Cases.

BACKGROUND: Nonbacterial thrombotic endocarditis is characterized by the presence of organized thrombi on cardiac valves, often associated with hypercoagulable states. There is a paucity of data regarding the predictors of mortality in patients with nonbacterial thrombotic endocarditis. Our primary aim was to identify predictors of in-hospital mortality in patients with nonbacterial thrombotic endocarditis. METHODS: A systematic literature review of all published cases and case series was performed until May 2018 according to Preferred Reporting Items for Systematic Review and Meta-analyses statement guidelines. We applied random forest machine learning model to identify predictors of in-patient mortality in patients with nonbacterial thrombotic endocarditis. RESULTS: Our search generated a total of 163 patients (mean age, 46 ± 17 years; women, 69%) with newly diagnosed nonbacterial thrombotic endocarditis. The in-hospital mortality rate in the study cohort was 30%. Among the patients who died in the hospital, initial presentation of pulmonary embolism (12.2 vs. 2.6%), splenic (38.7 vs. 10.5%), and renal (40.8 vs. 9.6%) infarcts were higher compared to patients alive at the time of discharge. Higher rates of malignancy (71.4 vs. 39.4%, P = .0003) and lower rates of antiphospholipid syndrome (8.1 vs. 48.2%, P = .0001) were noted in deceased patients. Random forest machine learning analysis showed that older age, presence of antiphospholipid syndrome, splenic infarct, renal infarct, peripheral thromboembolism, pulmonary embolism, myocardial infarction, and mitral valve regurgitation were significantly associated with increased risk of in-hospital mortality. CONCLUSION: Patients admitted with nonbacterial thrombotic endocarditis have a high rate of in-hospital mortality. Factors including older age, presence of antiphospholipid syndrome, splenic/renal infarct, lower limb thromboembolism, pulmonary embolism, myocardial infarction, and mitral valve regurgitation were significantly associated with increased risk of in-hospital mortality in patients with nonbacterial thrombotic endocarditis.

Global and Regional Variations in Transthyretin Cardiac Amyloidosis: A Comparison of Longitudinal Strain and 99mTc-Pyrophosphate Imaging.

There are limited data on the head-to-head comparison of 99mTc-pyrophosphate (99mTc-PYP) and echocardiographic strain imaging in the assessment of transthyretin (TTR) cardiac amyloidosis. Methods: At Mayo Clinic Arizona, patients who had undergone both a 99mTc-PYP scan and a transthoracic echocardiogram within a 90-d period were retrospectively identified for chart review and strain imaging analysis. Patients were divided into 2 groups according to their 99mTc-PYP results (PYP-positive [PYP+] or PYP-negative [PYP-]) for the comparison. A standard 17-segment model was used for segmental, regional, and global longitudinal strain comparison. A P value of less than 0.05 was deemed significant. Results: In total, 64 patients were included, the mean age was 75.1 ± 13.0 y, and 57 (89.1%) were male. Comparing the PYP+ to the PYP- group, the left ventricular global longitudinal strain was significantly worse in the former (PYP+ vs. PYP-, -10.5 ± 2.6 vs. -13.1 ± 4.1; P = 0.003). PYP+ patients also had worse regional basal strain (-4.6 ± 2.6 vs. -8.8 ± 4.0, P < 0.001) and a trend toward worse midventricular strain (-9.6 ± 4.0 vs. -11.7 ± 4.4, P = 0.07), but there was no statistical difference in the apical region (-17.6 ± 4.73 vs. -19.0 ± 6.46, P = 0.35). This is consistent with an apex-sparing pattern shown by the relative apical longitudinal strain index (1.3 ± 0.5 vs. 1.0 ± 0.3, P = 0.008). Segment-to-segment analysis demonstrated a significant difference in strain between PYP+ and PYP- segments in 4 segments: basal inferior (P = 0.006), basal anterolateral (P = 0.01), apical septal (P = 0.002), and apical inferior (P = 0.001). Left ventricular diastolic dysfunction was significantly different, with 17 (77.3%) patients in the PYP+ group versus 15 (36.6%) in PYP- participants (P = 0.002). Conclusion: Our study suggested that 99mTc-PYP uptake is related to overall worse LV segmental, regional, and global longitudinal strain function, as well as diastolic function, compared with patients without 99mTc-PYP uptake. These data are important for helping clinicians learn about the echocardiographic function features related to 99mTc-PYP uptake and can help generate hypotheses for future studies.

National Cardiovascular Data Registry Model Predicts Long-Term Mortality in Patients Undergoing Percutaneous Coronary Interventions.

National Cardiovascular Data Registry (NCDR)-based logistic regression model is available for clinicians to predict in-hospital all-cause mortality after a percutaneous coronary intervention (PCI). However, this model has never been used to predict long-term all-cause mortality after PCI. Therefore, we sought to test the ability of the NCDR model to predict the short- and long-term risk of all-cause mortality in patients undergoing PCI. All patients undergoing PCI in the Mayo Clinic Health System were enrolled in the Mayo Clinic CathPCI registry. Patient-level demographic, clinical, and angiographic data from January 2006 to December 2017 were extracted from the registry. Patients who underwent coronary artery bypass graft surgery (CABG) were excluded. The area under the receiver operator characteristic curve (AUC) was calculated to assess the ability of the NCDR model to predict outcomes of interest (6-month, 1-year, 2-year, and 5-year all-cause mortality) after PCI. A total of 17,356 unique patients were included for the final analysis after excluding 165 patients who underwent CABG surgery. The mean age was 66.9 ± 12.5 years, and 71% were men. The 6-month, 1-year, 2-year, and 5-year all-cause mortality rates were 4.2% (n = 737), 5.8% (n = 1,005), 8.06% (n = 1,399), and 14.2% (n = 2,472), respectively. The AUCs of the NCDR model to predict 6-month, 1-year, 2-year, and 5-year all-cause mortality were 0.84 (95% CI: 0.82-0.86), 0.82 (95% CI: 0.80-0.84), 0.80 (95% CI: 0.79-0.81), and 0.78 (95% CI: 0.77-0.79), respectively. The NCDR model was able to accurately predict both short- and long-term all-cause mortality after PCI.

Effects of Pectus Excavatum Repair on Right and Left Ventricular Strain.

BACKGROUND: The cardiopulmonary benefits of pectus excavatum repair have been debated. Echocardiographic speckle-tracking strain and strain rate have been used to evaluate and detect subclinical myocardial dysfunction in patients receiving cardiotoxic chemotherapy, and patients with valvular heart disease. This technology was applied to evaluate the effects of pectus excavatum surgery on left ventricular (LV) and right ventricular (RV) function. METHODS: Speckle tracing strain evaluation was performed on intraoperative transesophageal echocardiographic images acquired immediately before and after Nuss repair in adult patients (aged 18 years or more) from 2011 to 2014. Standard severity and compression indices were measured on chest imaging performed before pectus excavatum repair. RESULTS: In total, 165 patients with transesophageal echocardiographic images during repair were reviewed (71.5% male; mean age 33.0 years; range, 18 to 71; Haller index 5.7; range, 2.3 to 24.3). Significant improvement after repair was seen in global RV longitudinal strain (-13.5% ± 4.1% to -16.7% ± 4.4%, p < 0.0001) and strain rate (-1.3 ± 0.4 s-1 to -1.4 ± 0.4 s-1, p = 0.0102); LV global circumferential strain (-18.7% ± 5.7% to -23.5% ± 5.8%, p < 0.0001) and strain rate (-1.5 ± 0.5 s-1 to -1.9 ± 0.8 s-1, p = 0.0003); and LV radial strain (24.1% ± 13.5% to 31.1% ± 16.4%, p = 0.0050). There was a strong correlation between preoperative right atrial compression on transesophageal echocardiogram and improvement in RV global longitudinal strain rate immediately after pectus repair. CONCLUSIONS: Mechanical compression and impaired RV and LV strain is improved by Nuss surgical repair of pectus deformity.

Success of Minimally Invasive Pectus Excavatum Procedures (Modified Nuss) in Adult Patients (≥30 Years).

BACKGROUND: Minimally invasive repair of pectus excavatum (MIRPE) has become standard for pediatric and young adult patients, but its use for older adults is controversial. METHODS: We retrospectively reviewed electronic medical records of adults (≥18 years of age) who underwent MIRPE from January 1, 2010, through April 30, 2015, and collected demographic data, operative details, and information about outcomes. Cardiac function was measured before and after repair by intraoperative transesophageal echocardiography. We divided patients by age: 18 to 29 years of age and 30 years of age and older. RESULTS: Of 361 patients, 207 were 30 or older (mean, 40 years; range, 30 to 72 years; 71.5% men). Of the older patients, 151 had primary repairs. MIRPE was successfully used in 88.7% of patients older than 30 years of age versus 96.5% of those 18 to 29 years of age. For patients 30 years of age and older, open-cartilage resection, sternal osteotomy, or both was more common with increasing age (mean, 47.8 years versus 39.5 years; p = 0.0003) and higher mean Haller index (7.7 versus 5.5; p = 0.0254). Mean operative time for MIRPE was significantly longer for older patients (≥30 years of age) compared with younger adults (121 [60 to 224] minutes versus 111 [62 to 178] minutes; p = 0.0154). Right ventricular output increased 65.2% after repair in older adults. Although greater, the frequency of bar rotation requiring reoperation was not significantly increased in the older patients (p = 0.74). CONCLUSIONS: The majority of adult patients with PE can have successful repair with modified MIRPE. The use of cartilage or sternal osteotomy, or both, increased with patient age and defect severity.

Surgical repair of pectus excavatum relieves right heart chamber compression and improves cardiac output in adult patients--an intraoperative transesophageal echocardiographic study.

BACKGROUND: Cardiac compression in pectus excavatum (PE) deformity and effect of PE surgery on cardiac function in adults have been debated. We examined the effect of PE correction on right heart size and cardiac output. METHODS: A retrospective evaluation was performed of 168 adult patients who underwent a modified Nuss PE repair with intraoperative transesophageal echocardiography from 2011 to 2014. Seventeen patients with prior PE repair undergoing bar removal acted as controls. RESULTS: Mean age was 33.0 years (range, 18 to 71 years). There was an increase in right atrium (15.1%), tricuspid annulus (10.9%), and right ventricular outflow tract (6.1%) size after surgery (all P < .0001). Right ventricular cardiac output measured in a subset of 42 patients improved by 38%. No change in chamber size or cardiac output occurred before and after bar removal surgery in the control group. CONCLUSIONS: Surgical correction of PE deformity caused a significant improvement in right heart chamber size and cardiac output.