2023 年 31 巻 1 号 p. 7-12
Introduction: fractal geometry has been implemented to study natures’ irregular objects with diverse applications in medicine, from characterizations of human structures such as the division of bronchial or arterial trees to the diagnosis of abnormal cells, as it has been evidenced in the study of neoplasms. In this investigation, a characterization of the degree of irregularity of histologic samples of normal and thyroid gland neoplasms was developed in the context of fractal and Euclidean geometry.
Methods: photographs of normal and neoplastic thyroid histologic samples obtained as surgical pieces were processed by a software that allows to delimit the borders of cellular structures. Then, grids of 5 and 10 pixels were superimposed over the images in order to measure their surfaces and occupation spaces through the Box-Counting method.
Results: For nucleus measures values between 14 to 34 in the 5 pixels grid and values between 7 to 16 in the 10 pixels grid suggest normality.
Conclusions: The surface of the nucleus and the cytoplasm measured by grids of 5 and 10 pixels, allow to carry out a more objective characterizations of both normal and abnormal thyroid cells.
