Heart Rhythm Society

Background: The application of predictive analytics, particularly through artificial intelligence (AI) and machine learning (ML), has the potential to transform risk assessment in Acute Ischemic Stroke (AIS). This study aims to compare the performance of deep learning (DL) models against traditional machine learning (ML) algorithms in predicting AIS risk, using clinical data from patients with a history of cerebrovascular stroke events. Methods: In this retrospective analysis, we examined data from 5120 patients with documented AIS events, sourced from Rhythm Heart and Critical Care. Key predictive variables included hypertension, history of heart disease, occupation, average glucose levels, body mass index (BMI), and smoking status. To ensure unbiased model development, stroke history was blinded during training. We assessed the performance of various DL architectures. These were benchmarked against established traditional models, such as Logistic Regression (LR), AdaBoost, and Gaussian Naïve Bayes (GaussianNB). Model efficacy was evaluated using key metrics including accuracy, precision, prediction rate, and area under the receiver operating characteristic curve (AUC). Results: The RADAIS model, incorporating deep learning techniques, successfully generated stroke predictions for 5017 patients (97% of the cohort). The model demonstrated superior performance with an accuracy of 0.95, precision of 0.99, prediction rate of 0.97, and AUC of 0.97. In contrast, traditional machine learning models exhibited lower performance, with Logistic Regression (accuracy: 0.71, precision: 0.71, prediction: 0.69, AUC: 0.79), Gaussian Naïve Bayes (accuracy: 0.70, precision: 0.68, prediction: 0.73, AUC: 0.81), and AdaBoost (accuracy: 0.78, precision: 0.75, prediction: 0.82, AUC: 0.76). These results underscore the superior accuracy and precision of deep learning models in AIS prediction. Conclusion: The findings from the RADAIS study highlight the significant advantages of deep learning models over traditional machine learning methods for predicting the risk of Acute Ischemic Stroke. With a remarkable accuracy of 0.95 and an AUC of 0.97, deep learning-based models offer a more precise, reliable, and clinically relevant tool for stroke risk assessment, ultimately facilitating more informed decision-making in stroke management.

BackgroundConcomitant epicardial ablation of ventricular tachycardia (VT) remains a clinical challenge in cardiac surgery due to the need for intra-operative mapping. A preoperative patient-tailored epicardial guide could provide an intraoperative ‘blueprint’ of arrhythmogenic target substrate, thereby facilitating the workflow. Thus far, no 3D printing material or technique has fully met the requirements for this application. This study explores the feasibility of high-consistency rubber (HCR) silicone molding to produce low-cost, customized guides for VT scar ablation.MethodsAn inverted mold was created in Meshmixer using merged LGE–CMR and cardiac CT images of a 61-year-old VT patient, and printed using fused deposition modeling. HCR silicone was milled, sculpted, and trimmed to fit the negative mold. The guide was cured, subjected to autoclaving, and bench-tested on an ex vivo porcine heart model using radiofrequency and cryo-ablation. Various durometers and thicknesses were tested to determine the optimal fit for our application.ResultsFive surgical guides were made using NuSil™ MED-4072 and MED-4080 silicone (thickness range: 2.0–3.4 mm). Models with 2.1–3.0 mm thickness and 70 Shore A hardness achieved the best balance between flexibility and rigidity for application on a beating heart. The thinnest model (2.0 mm) was too pliable for stable placement. The guides withstood autoclaving and ablation procedures (radiofrequency; cryo-energy) without deformation or compromising structural integrity.ConclusionsThe HCR silicone molding technique allows for the production of flexible, cost-effective epicardial guides for VT ablation, minimizing the need for a full electrophysiology team throughout the entire procedure.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12938-025-01435-z.

Background: Conduction system pacing (CSP), including His bundle (HBP) and left bundle branch area pacing (LBBAP), is a growing alternative to traditional right ventricular or biventricular pacing. However, there is limited comparative data for stylet-driven (SDLs) and non-stylet-driven (NSDLs) leads implantation for CSP. Objective: To evaluate differences with SDLs versus NSDLs for HBP and LBBAP using data from the National Cardiovascular Data Registry (NCDR). Methods: We analyzed 11,412 CSP lead implantations, with 10,005 NSDLs and 1,407 SDLs, from April 2021 to December 2023 within the NCDR Electrophysiology Device Implant Registry. Baseline characteristics, procedural data, and in-hospital outcomes were compared between groups using multivariable logistic regression adjusted for demographics, comorbidities, implant indication, and hospital characteristics. Results: NSDLs remained the predominant lead type, though SDL use increased significantly during the study period. Compared with NSDL patients, those undergoing SDL implantation had higher rates of heart failure (61.6% vs. 48.6%, p<0.001), lower left ventricular ejection fraction (42.3% vs. 47.4%, p<0.001), and more frequent prior device implantation. Implant success was slightly higher with SDLs (98.4% vs. 96.2%, p<0.001), but overall complication rates were low and comparable (2.8% vs. 2.1%, p=0.089). No significant differences were observed in cardiac perforation, lead dislodgement, or in-hospital mortality. In multivariable analysis, lead type was not associated with increased odds of complications or prolonged hospitalization. Conclusion: In the largest analysis to date, both SDL and NSDL strategies achieved exceptionally high implant success rates for CSP, with slightly higher rates for the former, accompanied by low complication rates. These findings support the safety and effectiveness of both strategies and suggest that lead selection may be guided by operator preference.

Abstract Background Integration of intracardiac echocardiography (ICE) with electroanatomic mapping software can potentially improve procedural efficiency and safety, increase procedure confidence, and reduce radiation exposure. Here, we present the initial clinical experience of a novel ICE system integrating a sensor-enabled ICE catheter into a 3D electroanatomic mapping software. Purpose Evaluate the experience with a novel integrated ICE system during electrophysiology procedures. Methods The ViewFlex X study is a prospective, first-in-human, non-significant risk study evaluating integration of a sensor-enabled ICE catheter (ViewFlex X ICE Catheter, Sensor Enabled) with a 3D electroanatomic mapping system (EnSite X EP System) in patients undergoing an EP procedure. Per protocol, ICE was used to visualize the chamber of interest; additional visualization was at operator’s discretion. Following ICE, high-density mapping catheters were used to create voltage-based models that informed treatment. Subjects were followed through discharge for safety monitoring. This study is actively enrolling. Results At the time of this report, 85 patients have enrolled in the ViewFlex X study with data available for 73 patients. There were 73 ablation procedures, including 4 concomitant ablation procedures with LAAO device placement and 1 lead placement. The patient population was 68% (n=50) male (mean 67 years, BMI 29.7) and 32% (n=23) female (mean 65 years, BMI 31.2). The most common indication was paroxysmal AF (30/73, 41%), followed by persistent AF (22/73, 30%), atrial flutter (7/73, 10%), PVC (6/73, 8%), SVT (4/73, 5%), and VT (4/73, 5%). A flexible-tip RF ablation catheter was the most common ablation catheter (n=37/73, 51%), followed by PFA catheters (n=30/73, 41%). Acute success was achieved in 100% of procedures. Procedural uses of ICE are shown in Table 1. Physician feedback highlighted the ease of operability of the ICE catheter with one hand, ability to reach desired intracardiac locations, and catheter responsiveness. The average perceived accuracy of the ICE models compared to maps created using a high-density mapping catheter was 85±11%. Feedback from several cases noted the ICE catheter was able to visualize anatomical locations not accessible with the mapping catheter, likely resulting in more accurate chamber geometry in the ICE model. Case images of the real-time simultaneous ICE and 3D electroanatomic map are shown in Figure 1. Conclusions A novel integrated ICE system was used to support left atrial and ventricular procedures in a real-world setting with a high degree of efficacy for a variety of workflows, which may speak to its utility in minimizing fluoroscopy usage. Operator feedback emphasized the high degree of accuracy of the ICE models and operability of the ICE catheter. These initial results highlight the versatility of a novel integrated ICE system in a variety of real-world workflows and indications.