Tumor growth is a complex process due to the uncontrollable proliferation of cells that invade neighboring tissues. This type of process is very relevant for the diagnosis and definition of adequate therapeutic strategies. This implies the assessment of its complexity according to the descriptors, as well as the scaling analysis and the fractal geometry, which in essence define the geometric growth of the tumor. In this work we can calculate the factors of local rugosity (αloc) and fractal dimension (dF) characterizing in vivo and in 3D the tumor growth of non-small cells, adenocarcinomas in lung and healthy lung, by means of tomographic axial images (CT). For the calculation of the fractal dimension, the three-dimensional point matrix of the tumor was initially determined, which was obtained by the method of segmentation of average K-images. Once the matrix was obtained, the fractal dimension of Hausdorff (dF) was calculated with it, applying the box counting algorithm. At the host tumor interface, the local roughness exponent (αloc) was calculated by using an algorithm that uses the width of the interface and the small areas subtended at the solid angle, which is generated between the center of mass of the injury and its periphery. The results obtained in terms of the fractal dimension and local roughness coefficient, showed that they are parameters that can be used for the characterization of this type of lesions, since the results showed similarity in each histological group studied (non-small cell tumors, adenocarcinomas of lung and healthy lung)