Semiautomatic tool for segmentation of hearts with congenital pathologies

Abstract

Introduction: Congenital cardiac anatomical anomalies are diseases that have an important mortality rate if they are not treated correctly and in time. Early diagnosis together with a good surgical planning increases the chances of success of the intervention and the post-surgical life expectancy. The realization of printed models from medical images favors preoperative planning, although there are few free computational tools capable of segmentation of the cardiovascular structures and its conversion to the tridimensional meshes. Methods: This paper proposes a semi-automatic segmentation tool of medical images to obtain a simple and fast three-dimensional model of the main organs and vessels affected in congenital heart disease. The tool was evaluated in a pediatric patient with Tetralogy of Fallot. Results: In a time suitable for clinical use, a three-dimensional digital model in stereolithographic format (STL) was created representing the anatomy of the patient’s right heart including the superior and inferior cava veins and the pulmonary artery. The resulting mesh was conditioned and 3D-printed on polylactic acid (PLA). The model was made available for physicians to plan a valve replacement. Conclusion: The proposed tool allowed to obtain presurgical models of the anatomical structures of interest in a simple and fast manner. It may be applicable to other scenarios of surgical planning, although this tool is expected to be improved in a near future by decreasing the manual user interaction.