Three-Dimensional Breast Tumor Visualization on MRI Based on α-Shapes

Nowadays cancer is the leading cause of death worldwide, to determine the specific treatment for cancer tumors, it is common to get their 2D representation. To obtain a more accurate representation of the tumor a 3D model is used. In this paper a 3D reconstruction of all the models in the Breast Tumor Patient Models Repository, was made by using the a-shape technique. To generate the respective shape from every tumor model, the images were transform into point clouds considering the vertices and the central point of the voxels generated using different slices. The volume of each tumor obtained from this method was compared to the ones measured by five other techniques: stl model, the Trace method, the Frustum model, the convex hull method, and by counting the voxels quantity in every model. In 68% of the dataset the a-shape method had the best volume approximation, in 23% of it the best estimation was from the Trace method, and the Frustrum model was better for only 9% of the cases. The 3-D representations and the volume estimations are intended to be implemented in the planning of breast cancer treatment by microwave ablation, so it is necessary improve this process to provide trustworthy information.