Correlation between 3D morphometric characteristics of kidney tumors and their histological structure

Abstract

Aim. To assess the correlation between clinical characteristics (sex, age, and maximum tumor size), 3D morphometric characteristics of renal lesions obtained through processing of multispiral computed tomography data, and renal tumor histology.Materials and methods. Evaluation of kidney tumor malignancy on the basis of comparative analysis of primarily morphometric characteristics of the lesion obtained through processing of multispiral computed tomography data and histological tumor structure is presented. Data of 308 patients (175 (56.8 %) males and 133 (43.2 %) females) with unilateral renal tumors were analyzed.Results. Multivariable analysis showed that malignancy of kidney tumor is associated with sex (male), location in the middle segment, tumor size, tumor shape (spherical with conical base) (24.8 %), while mushroom-like lesion shape was more common in benign tumors (35.2 %). In univariate models, only two variables were statistically significant predictors: patient sex and tumor shape.Conclusion. The developed logistic model based on analysis of such predictors as sex and kidney tumor shape has a high percentage (87.6 %) of correct predictions of tumor histological structure.