Atrial fibrillation is the most common cardiac arrhythmia, affecting 8.5% of the general population in Hong Kong. Systemic embolization, and particularly stroke, is the most frequent major complication of atrial fibrillation. Long-term oral anticoagulation is recommended for most atrial fibrillation patients for prevention of embolism. However, such therapy is associated with an increased risk of bleeding and not all individuals are candidates for this therapy.

The left atrial appendage (LAA) is the usual source for clot that embolizes. Occlusion of the LAA ostium (LAAO) with percutaneous device in patients with nonvalvular atrial fibrillation has emerged as an alternative to anticoagulation for prevention of embolism.

Procedural success requires careful planning and understanding of the anatomy of LAA, as well as the interaction between the appendage and the occlusion device. However, complexity and variability of LAA anatomy exists and these anatomical variations among individuals poses challenges to accurately sizing and positioning the device. Certain anatomic variations of the appendage, for instance the presence of a sharp bend in the proximal or middle portion of the dominant lobe, or prominent pectinate muscles, pose particular challenge to device implantation and demands pre-specified implantation technique. Assessment of LAA anatomy relies on imaging, usually combining peri-interventional 2D transesophageal echocardiography (TEE) with fluoroscopy guidance and, less frequently, with pre-interventional computed tomography (CT). Because lobes of the LAA exist in different planes, imaging must be done in multiple planes to visualize the entire LAA. Even with advanced imaging, complete understanding the 3D geometry of the appendage is challenging, and the mechanical interaction between the device and the anatomy is difficult to predict or quantify.

The limitation of imaging in the assessment of LAA anatomy may lead to inaccurate device sizing. Over- or under-sizing increase the chance of pericardial effusion, incomplete occlusion, and device embolization. Incomplete occlusion of the LAA ostium is common and may jeopardise the procedural efficacy in embolism prevention. Device re-sizing during procedure prolongs procedural time as well as radiation exposure; moreover, device re-positioning within the appendage may cause inadvertent tissue injury and increase the risk of cardiac perforation. Indeed, procedural complication rate of LAAO remains fairly high in real-world practice.

3D printing (3DP) is a novel technology able to create a patient-specific model of any given anatomical portion of the heart for preoperative device testing and procedural simulation. The simulation "rehearsal" experience can enhance the operator's confidence, allowing the operator to anticipate difficulties before the actual intervention; this potentially reduces the procedural time (hence cost and radiation hazard), device re-sizing, the number of deployment attempts, and promotes procedural success.

The aim of this project is to evaluate the effect of 3D-printed patient-specific LAA model compared with standard imaging planning on procedural efficacy and safety of LAAO.

The project will be divided into two parts:

Part I Technical validation of 3D-printed LAA models Validation of anatomical accuracy and material properties of the 3D-printed LAA model will be conducted in 30 patients referred for clinically indicated surgical excision of the LAA.

Part II Randomised clinical trial on additive benefit of 3D printing for LAAO. The investigators will conduct a randomized, controlled, open-label, trial in 200 patients undergoing LAAO for stroke prevention in our hospital.

Eligible subjects will be randomly assigned by a computer program in a 1:1 ratio to either preoperative planning using 3D-printed LAA model (3D printing arm) or standard imaging planning alone (no-3D printing arm).

The investigators hypothesize that preoperative 3DP planning could help in finding the correct position within LAA, sizing the device and guiding the choice of the closure device despite the measurements provided by imaging alone.

Inclusion Criteria:

• Age ≥18 years
• Non-valvular atrial fibrillation
• CHA2DS2-VASc ≥2
• Absolute or relative contraindications to long-term oral anticoagulation (e.g. history of intracranial or life-threatening bleeding, HAS-BLED ≥3, triple anticoagulant therapy)
• Planned for LAAO procedure

Exclusion Criteria:

• Intracardiac thrombus is visualized by echocardiographic imaging
• An atrial septal defect repair or closure device or a patent foramen ovale repair or closure device is present
• The LAA anatomy will not accommodate a device
• Any of the customary contraindications for other percutaneous catheterization procedures (e.g., patient size too small to accommodate TEE probe or required catheters) or conditions (e.g., active infection, bleeding disorder) are present.
• Known hypersensitivity to any portion of the device material or the individual components
• Comorbidities other than atrial fibrillation that required oral anticoagulation
• Suboptimal LAA image on TEE or CT that is inadequate for 3D modelling and printing
• Refusal to participate the study