Here, we provide a workflow that allows the identification of healthy and pathological cells based on their 3-dimensional shape. We describe the process of using 2D projection outlines based on the 3D surfaces to train a Self-Organizing Map that will provide objective clustering of the investigated cell populations.