Background: TAVR is a safe alternative to surgical valve replacement. 3D printing is an emerging technology transforming 2D medical images to tangible anatomically accurate models. Hypothesis: To determine the accuracy of 3D printed models of the aortic valve (AV) complex for TAVR planning compared to CTA. Methods: 10 pts with pre TAVR contrast enhanced CTA were randomly selected. The blood pool of the AV complex from proximal ascending aorta above the coronary arteries to the AV annulus was automatically segmented. A 3D printable STL model was exported to Computer Aided Design software and further processed. Models were printed with a material extrusion 3D printer. One reader blinded to CTA measured the max and min diameter of the AV annulus and coronary ostia takeoff height of the printed model with a caliper. Two independent CV imagers performed these measurements in double-oblique CTA reformats. Model measurements were compared to CTA findings by Pearson’s correlation and via the mean and 95% confidence interval (CI) of the absolute difference to CTA compared to CTA interobserver differences. Results: 3D printed model measurements had excellent correlation to CTA (r=.86-.97). Mean absolute difference in max diameter between 3D-printed model and CTA was 0.85 mm (CI: [.43, 1.28]), compared to 1.17mm (CI: [.59, 1.75]) between CTA readers; for min diameter it was 0.49mm (CI: [.2, .78]) compared to 0.45mm (CI: [.17, .73]) between CTA readers. Mean absolute difference of height to left main ostium was 1.62mm (CI: [.93, 2.3]) compared to 1.47 mm (CI: [.78, 2.16]) for CTA readers, and to right coronary ostium it was 1.0 mm (CI: [0.53-1.46]) compared to 1.98 mm (CI: [1.38, 2.58]) for CTA readers. Conclusion: CTA-derived 3D printed models of the AV complex provided excellent correlation to CTA measurements of the annulus and ostia height. 3D printing can assist towards planning of TAVR sizing and minimizing procedural complications such as paravalvular leak and coronary occlusion.