Percutaneous left atrial appendage (LAA) occlusion has emerged as an alternative approach to medical therapy in patients with a contraindications for long-term oral anticoagulation. The correct sizing and optimal device positioning remain challenging due to the broad variability in LAA morphology. 3D printing issue from cardiac CT is a promising novel technology able to create an accurate patient specific model.

To explore in a large cohort the usefulness of 3D printed LAA models to predict adequate sizing and risk of peri-device leaks or malposition post-LAA closure (LAAC).

MDCT was performed in 68 patients (Chadsvasc 3.5±1.5) before and after (90±30 d) LAAC (AMPLATZER Amulet, St. Jude Medical). Segmentation was performed to include all LA. Segmented data sets were then converted to a STL file (V8Philips) and further post-processing (Geomagic). A 1.5-mm thick wall LAA models were 3D printed Formlabs, Cambridge, MA).

Significant peri-device leaks>5mm occurred in 23/68 patients. If 3D printed models were applied overall: 13% of patients would have received a smaller prosthesis and 25% a larger one. A device exchange was required in 11 (16%). The mismatch with 3D-printed model predicted size was a major predictor of peri-device leaks (AUC=0.91) VPP=89% VPN=93%. In comparing CT sizing and TEE sizing, had a predictive value significantly lower (AUC=0.57 and 0.59, respectively). Oblique device (off-axis) was observed in 28%. During 17±11 months fu, there were one TIA and three devices related thrombus in patient with off-axis prosthesis (Fig. 1).

A 3D print patient-specific LA model allows a more accurate sizing method compared to TEE and CT measurements while considering not only the LAA dimensions but all the anatomical parameters. In using 3D printed models may have important implications to minimize procedure-related complications, time of procedure, peri-device leaks, or malposition.