Introduction: Syncope was reported to be associated with sudden cardiac death in heart failure patients regardless of the etiology of syncope. We attempted to use unsupervised machine learning to identify the prognostically distinct phenotypes in a large population of heart failure patients presenting with syncope. Hypothesis: Unsupervised machine learning can identify prognostically distinct phenotypes in a large population of heart failure patients presenting with syncope. Methods: Patients with a diagnosis of systolic or diastolic heart failure and syncope at Mayo Clinic sites were identified for baseline and follow-up data. Variables with ≥ 20% missing data were excluded; a total of 57 variables were used for k-means clustering analysis. Patients were grouped according to 4 distinct k-means determined clusters for statistical analysis. Results: A total of 3,557 patients with proper follow-up and echocardiography data were included in the final analysis. The median follow-up duration was 36.3 months, and 2908 (81.7%) patients died. Cluster 2 (n= 680) was identified as the cluster with the highest mortality (92%), which contained patients with high proportions of cardiovascular comorbidities, especially chronic kidney disease. The Kaplan-Meier survival curves are shown in Figure 1. In multivariate Cox regression analysis, the presence of CKD independently predicts mortality (HR 1.62, 95%CI 1.50-1.76, p <0.0001) after adjusting for age and sex. Conclusions: Unsupervised machine learning identified patients with CKD, CAD and DM are high-risk phenotypes in heart failure patients with syncope, and can be used to guide future studies in this highly heterogeneous, unexplored population.
Congenital right coronary artery-superior vena cava (RCA-SVC) fistula is rare and typically does not manifest any symptoms until the fifth decade of life. The present case demonstrates a 48-year-old woman who developed Sinus node dysfunction of unknown cause after Percutaneous coil embolization of the RCA-SVC fistula requiring permanent pacemaker. (Level of Difficulty: Intermediate.).
Carotid intima-media thickness (CIMT) is regarded as a controversial risk marker for cardiovascular disease (CVD). We aimed to evaluate the role of CIMT and carotid plaque progression as predictors for the progression of coronary plaque and compositions.In the Garlic 4 study, asymptomatic patients with intermediate CVD risk (Framingham risk score 6-20%) were recruited for a serial carotid ultrasound, and coronary artery calcium score (CAC)/coronary computed tomography angiography (CCTA) studies for subclinical atherosclerosis at a baseline and 1 year. The association between progression of quantitatively measured coronary plaque compositions and the progression of CIMT/carotid plaque was analyzed. A P value <0.05 is considered as statistically significant.Forty-seven consecutive patients were included. The mean age was 58.5 ± 6.6 years, and 69.1 % were male. New carotid plaque appeared in 34.0 % (n = 16) of participants, and 55.3 % (n = 26) of subjects had coronary plaque progression. In multilinear regression analysis, adjusted by age, gender, and statin use, the development of new carotid plaque was significantly associated with an increase in noncalcified coronary plaque [β (SE) 2.0 (0.9); P = 0.025] and necrotic core plaque (1.7 (0.6); P = 0.009). In contrast, CIMT progression was not associated with the progression of coronary plaque, or coronary artery calcium (CAC) (P = NS).Compared to CIMT, carotid plaque is a better indicator of coronary plaque progression. The appearance of a new carotid plaque is associated with significant progression of necrotic core and noncalcified plaque, which are high-risk coronary plaque components.
Abstract Objective To accurately predict post-TAVR mortality, we proposed a family of new echo-parameters (augmented blood pressure) derived from blood pressure and aortic valve gradient measurements and examined them in this study. Patients and Methods Patients in the Mayo Clinic National Cardiovascular Diseases Registry-TAVR database who underwent TAVR between January 1, 2012, and June 30, 2017, were identified to retrieve baseline demographics, echocardiographic and mortality data. Augmented blood pressure parameters and valvulo-arterial impedance were evaluated by univariate and multivariate Cox regression. Receiver operating characteristic curve analysis was used to assess the model performance against the Society of Thoracic Surgeons (STS) risk score. Results The final cohort contained 974 patients with a mean age of 81.4±8.3 years old, and 56.6% were male. The mean STS risk score was 8.2±5.2. The median follow-up duration was 354 days, and the one-year all-cause mortality rate was 14.2%. Both univariate and multivariate Cox regression showed that augmented systolic blood pressure and augmented mean arterial pressure (AugMAP) parameters were independent predictors for 1-year post-TAVR mortality (all p<0.0001). A univariate model of AugMAP1 supersedes the STS score model in predicting 1-year post-TAVR mortality (area under curve: 0.700 vs. 0.587, p=0.0051). Conclusion Augmented mean arterial pressure provides a simple but effective approach for clinicians to quickly estimate the clinical outcome of TAVR patients. It can be incorporated in the assessment of TAVR candidacy.
Abstract Background Automated data extraction from echocardiography reports could facilitate large-scale registry creation and clinical surveillance of valvular heart diseases (VHD). We evaluated the performance of open-source Large Language Models (LLMs) guided by prompt instructions and chain of thought (CoT) for this task. Methods From consecutive transthoracic echocardiographies performed in our center, we utilized 200 random reports from 2019 for prompt optimization and 1000 from 2023 for evaluation. Five instruction-tuned LLMs (Qwen2.0-72B, Llama3.0-70B, Mixtral8-46.7B, Llama3.0-8B, and Phi3.0-3.8B) were guided by prompt instructions with and without CoT to classify prosthetic valve presence and VHD severity. Performance was evaluated using classification metrics against expert-labeled ground truth. Mean Squared Error (MSE) was also calculated for predicted severity’s deviation from actual severity. Results With CoT prompting, Llama3.0-70B and Qwen2.0 achieved the highest performance (accuracy: 99.1% and 98.9% for VHD severity; 100% and 99.9% for prosthetic valve; MSE: 0.02 and 0.05, respectively). Smaller models showed lower accuracy for VHD severity (54.1-85.9%) but maintained high accuracy for prosthetic valve detection (>96%). CoT reasoning yielded higher accuracy for larger models while increasing processing time from 2-25 to 67-154 seconds per report. Based of CoT reasonings, the wrong predictions were mainly due to model outputs being influenced by irrelevant information in the text or failure to follow the prompt instructions. Conclusions Our study demonstrates the near-perfect performance of open-source LLMs for automated echocardiography report interpretation with purpose of registry formation and disease surveillance. While larger models achieved exceptional accuracy through prompt optimization, practical implementation requires balancing performance with computational efficiency.
Quantitation of regurgitation severity using the proximal isovelocity acceleration (PISA) method to calculate effective regurgitant orifice (ERO) area has limitations. Measurement of three-dimensional (3D) vena contracta area (VCA) accurately grades mitral regurgitation (MR) severity on transthoracic echocardiography (TTE).We evaluated 3D VCA quantitation of regurgitant jet severity using 3D transesophageal echocardiography (TEE) in 110 native mitral, aortic, and tricuspid valves and six prosthetic valves in patients with at least mild valvular regurgitation. The ASE-recommended integrative method comprising semiquantitative and quantitative assessment of valvular regurgitation was used as a reference method, including ERO area by 2D PISA for assigning severity of regurgitation grade.Mean age was 62.2±14.4 years; 3D VCA quantitation was feasible in 91% regurgitant valves compared to 78% by the PISA method. When both methods were feasible and in the presence of a single regurgitant jet, 3D VCA and 2D PISA were similar in differentiating assigned severity (ANOVAP<.001). In valves with multiple jets, however, 3D VCA had a better correlation to assigned severity (ANOVAP<.0001). The agreement of 2D PISA and 3D VCA with the integrative method was 47% and 58% for moderate and 65% and 88% for severe regurgitation, respectively.Measurement of 3D VCA by TEE is superior to the 2D PISA method in determination of regurgitation severity in multiple native and prosthetic valves.
'/%3e%3cuse xlink:href='%23a' transform='rotate(60)'/%3e%3ccircle r='17' fill='%23000095'/%3e%3ccircle r='15' fill='%23fff'/%3e%3c/svg%3e)