Yuji Ikari

Background: Pentaspline pulsed‐field ablation (PFA) has emerged as a novel modality for pulmonary vein isolation (PVI) and posterior wall isolation (PWI) in persistent atrial fibrillation (AF). However, the learning curve associated with combined PVI + PWI and the contribution of anatomical factors to acute lesion completeness remain incompletely defined. Methods: We analyzed 182 consecutive patients with persistent AF undergoing fluoroscopy‐ and intracardiac echocardiography–guided PFA. Acute incomplete lesions were defined as residual conduction requiring additional PFA applications after systematic remapping. Learning phases were identified using segmented logistic regression and sequential block analysis. Multivariable models evaluated procedural phase, left atrial volume (LAV), and posterior wall geometry. Results: A learning curve was observed, with the greatest reduction in incomplete lesion formation occurring during the first 20 procedures. Incomplete lesions occurred in 55% of cases 1–20 (early phase), 32% of cases 21–160 (intermediate phase), and 4.8% of cases ≥ 161 (late phase). Procedural phase was the dominant determinant of incomplete lesions, with approximately 24‐fold and 9‐fold higher odds during the early and intermediate phases, respectively. Larger LAV showed a modest independent association (OR 1.009 per mL), whereas posterior wall geometry was not independently associated. Mid‐term arrhythmia recurrence did not differ across phases, although follow‐up duration was shorter in the late phase. Conclusions: Lesion completeness during pentaspline PFA with PVI + PWI improves progressively with increasing procedural experience. Although larger LAV may contribute modestly to incomplete lesion formation, procedural experience appears to play the predominant role. Systematic electroanatomical remapping remains important for identifying and correcting incomplete lesions even after substantial procedural experience has been accumulated.

Background: Open‐window mapping (OWM) using the extended early‐meets‐late (EEML) algorithm is widely used for accessory pathway (AP) localization. However, the impact of catheter design on mapping performance and procedural outcomes remains unclear. Methods: We retrospectively analyzed 43 consecutive patients undergoing OWM‐guided AP ablation. Twenty‐four patients were mapped using an 8‐spline high‐density catheter and 19 using a conventional 5‐spline catheter. Mapping characteristics and ablation outcomes were compared. Results: The 8‐spline catheter was associated with higher mapping density (7974 [5618–11 017] vs. 2485 [734–12 577] points; p < 0.001) and shorter mapping time (11.0 [8.5–14.8] vs. 28.0 [15.0–40.0] minutes; p < 0.001). The EEML‐derived AP gap width was narrower in the 8‐spline group (6.1 [5.3–7.5] vs. 8.3 [2.9–15.6] mm; p = 0.023), with lower EEML thresholds required for visualization ( p = 0.003). First‐pass AP elimination was achieved in all patients in the 8‐spline group (100%) compared with 68% in the 5‐spline group ( p = 0.004), with fewer radiofrequency applications required ( p = 0.048). Fluoroscopy time and complication rates were similar between groups. Conclusion: Compared with a conventional 5‐spline catheter, an 8‐spline high‐density catheter was associated with improved mapping efficiency and a higher rate of first‐pass ablation success. These findings suggest that catheter design may influence procedural performance in OWM‐guided AP ablation.

Background: Artificial intelligence (AI)‐enhanced electrocardiography (ECG) has been developed to detect paroxysmal atrial fibrillation (AF) from sinus rhythm ECGs (SR‐ECGs). For broader applicability, model development across institutions and ECG systems is essential. Methods: We developed an AI‐enhanced ECG model using SR‐ECGs from Tokai University ( n = 172,613; Nihon Kohden [NK] system) and The Cardiovascular Institute ( n = 19,170; GE MUSE system). AF‐labeled SR‐ECGs were defined as recordings within 31 days of an AF episode, while SR‐labeled SR‐ECGs were those with ≥ 1095 days of AF‐free follow‐up. Three datasets were constructed: Dataset 1 (Tokai University, all departments, NK), Dataset 2 (Tokai University, Cardiology Department, NK), and Dataset 3 (The Cardiovascular Institute, Cardiology Department, MUSE). We developed five models: scratch models (S1–S3) trained on Datasets 1–3, and fine‐tuned models (F1, F2) trained on Datasets 1 and 2 after pretraining on Dataset 3. Models were evaluated using A1–A3 (same as Datasets 1–3) and B1–B3, which differed in ECG resolution and compression (B1: original MUSE, B2: MUSE‐NK intermediate, B3: NK‐converted). Results: Model F2, fine‐tuned on homogeneous datasets from cardiology departments, showed consistently high performance (AUC: A1 = 0.885, A2 = 0.829, A3 = 0.845). Model F1, fine‐tuned on heterogeneous datasets, demonstrated lower performance (AUC: A1 = 0.837, A2 = 0.726, A3 = 0.660). Model performance was consistent across different ECG format variants (B1–B3). Conclusion: Fine‐tuning on homogeneous data improved performance and generalizability, whereas heterogeneous data led to reduced performance. ECG system format differences had minimal impact on model accuracy.