In this study, the investigators identify the optimal AI value for achieving good acute outcomes in PVI.

Through the Phase 1 study, the investigators will prospectively analyze data including ablation parameters, AI and acute outcomes of 20 patients performed PVI with conventional ablation strategy (AI-blinded).

After Phase 1 study results, optimal target AI value for each predefined PV segment will be determined. Then, by applying optimal AI values derived from the Phase 1 study, the investigators will evaluate the feasibility of the AI-guided procedure and verify the optimal AI value for acute PVI outcomes in the prospectively enrolled patients (n = 30) in the Phase 2 study.

1. Phase 1 A total of 30 240 patients with AF will be prospectively and consecutively enrolled and applied conventional ablation strategy with AI-blinded using contact force sensing catheter (Thermocool SmartTouch Catheter, Biosense Webster Inc.). Conventional ablation will be performed signal reduction-guided, point by point ablation using Visitag automated annotation criteria (2520-35W, target contact force 10-20g30g, target time: 30-40sec for anterior/roof segments, 2015-30sec for posterior/inferior segments).

   AI value will be derived from comprehensive off-line analysis of ablation parameters including contact force, time, power and AI of these 30 240 patients by predefined 14 segments.

   Through comparison of AI of each segment with or without acute outcomes including PV residual potential after first encirclement/early reconnection (ER)/dormant conduction (DC), minimal AI value would be drawndetermined, and optimal AI value will be suggested for future AI-guided ablation strategy Phase 2 study (AI-guided ablation strategy).

2. Phase 2 The Phase 2 study aims to evaluate the feasibility of applying optimal AI values derived from the Phase 1 study, and to verify optimal values through AI-guided ablation. Thirty patients with AF will be prospectively and consecutively enrolled. AI information will be opened to the operator during ablation, and Visitag annotation criteria will also be used for AI-guided ablation with pre-set minimum target values for each segment derived from the Phase 1 study. After the procedure, the proportion of ablation points within the target AI value will be calculated by evaluation of the feasibility of the AI-guided ablation strategy. Acute outcomes of PVI including PV residual potential after first encirclement, ER, and DC will be evaluated based on predefined segments.

3. Comparison of conventional vs. AI-guided ablation The investigators will compare the acute outcomes of PVI using two different strategies: conventional vs. AI-guided ablation. During 1 year of follow-up in both groups, AF recurrence will be evaluated at 3, 6, 9 and 12 months using rhythm surveillance (ECG or 24-hour Holter monitoring). The AF recurrence rate at 1 year after PVI will be compared between the two groups. In addition, total ablation time, fluoroscopic time, procedure time, and complication rates for the index procedure will be compared between the two groups.

Inclusion Criteria:

• 20 to 80 years old
• Patients with symptomatic paroxysmal or persistent AF who had failed with anti-arrhythmic agents

Exclusion Criteria:

• Patients who had previous ablation for AF
• Patients with left atrial (LA) diameter more than 50mm
• Inability or unwillingness to receive oral anticoagulation with a Vitamin K antagonist (VKA)or non-VKA (NOAC) agent
• Known severe left ventricular systolic function (ejection fraction<35%)
• Pregnancy

Korean