INVESTIGATION OF COMPUTER AUTOMATED ANALYSIS CAPABILITIES FOR THE EFFECTIVE DIAGNOSIS OF DISSEMINATED PROSTATE CANCER

Abstract

The aim. The possibility of applying a unified methodological approach to the problems of pattern recognition and information modeling of the process of diagnosis of various forms of malignant tumors with the help of CAD-analysis of medical images is substantiated.Material and methods. A retrospective analysis of medical histories and results of scintigraphy from 168 patients with newly diagnosed disseminated prostate cancer treated on the basis of the Khabarovsk regional clinical Oncology center in the period from 2003 to 2016. Considered original methodological approaches to the study of metastatic lesion of the skeleton using computer-aided analysis of planar osteoscintigraphy.Results. Volume of lesions of the skeleton were calculated with the use of computer systems for automated diagnostics based on the principles of image recognition and has the features of expert analysis. and the degree of tumor differentiation are independent of each other predictors of high risk of death in patients with disseminated prostate cancer. The high efficiency of CAD-analysis in the evaluation of bone metastatic index for this form of malignant advanced tumors is shown.