3D measurements for tumours malignancies early diagnosis

Several medical studies have shown that tumor malignancy can be evaluated by two different measurements concerning to the volumetric growth of tumors and the irregularities of their surfaces. Today Doubling Time (DT) is widely used for quantification of tumor growth rate, but may not be appropriate for some tumors, because it does not concern to the fact that benign tumors are expansive, while malignant ones are infiltrative. This work is then focused on: 3D rendering of tumor, reconstruction of a synthetic model that approximates its shape, and extraction of geometric features that can attest malignancy both in terms of volumes and of a proposal of new irregularity indexes. In particular, we refer to colorectal and lung cancers and then present the description and the implementation of a new software. This paper is organized as follows: description of tumor segmentation from a dataset of Computer Tomography (CT) axial slices, 3D tumor reconstruction, Genetic Algorithm synthetic model implementation and evaluation of regularity of tumor shape.