Objectives: Many types of congenital heart disease are amenable to surgical repair or palliation. The procedures are often challenging and require specific surgical training, with limited real-life exposure and often costly simulation options. Our objective was to create realistic and affordable 3D simulation models of the heart and vessels to improve training.

Methods: We created molded vessel models using several materials, in order to identify the material that best replicated human vascular tissue. This material was then used to make more vessels to train residents in cannulation procedures. Magnetic resonance imaging views of a 23 months-old patient with double-outlet right ventricle were segmented using free open-source software. Re-usable molds produced by 3D printing served to create a silicone model of the heart, with the same material as the vessels, which was used by a heart surgeon to simulate a Rastelli procedure.

Results: The best material was a soft elastic silicone (Shore A hardness 8). Training on the vessel models decreased the residents' procedural time and improved their grades on a performance rating scale. The surgeon evaluated the molded heart model as realistic and was able to perform the Rastelli procedure on it. Even if the valves were poorly represented, it was found to be useful for preintervention training.

Conclusions: By using free segmentation software, a relatively low-cost silicone, and a technique based on re-usable molds, the cost of obtaining heart models suitable for training in congenital-heart-defect surgery can be substantially decreased.