INDIVIDUALIZATION OF NEOADJUVANT THERAPY IN THE TRATMENT OF PATIENTS WITH LOCALLY ADVANCED RECTAL CANCER

Abstract

Summary. Aim: to increase the effectiveness of combined treatment of patients with local advanced rectal cancer by individualizing neoadjuvant therapy and prevention of surgical complications and relapses. Objects and methods: the analysis of the results of treatment of 71 patients of rectal cancer, who received treatment in the KNP “Precarpathian Clinical Oncological Center of the Ivano-Frankivsk Regional Council” in the period 2016-2021. The study includes patients with uncomplicated local advanced rectum cancer. An analysis of the immediate and distant results of treatment of patients was carried out using neural network technology (PyTorch software). Results: it is determined that the complication of the surgical stage of treatment was in 16 patients under study. The total 3-year survival of patients with stage II-III cancer without complications of the surgical stage of treatment was 92.5%, and the total 5-year – 84%. For patients, with the existing complications, 3-year total survival was 79.5%, 5-year – 68.0%. Only with the help of mathematical analysis methods, taking into account the factors of the prognosis, you can choose the most optimal method of treatment of each particular patient. The proposed methodology for the use of neural networks was to serve the following steps: selection of many informative parameters, guided by target parameters, correlation and regression analysis. The points of view of accuracy of determining them and the target parameters and the possibility of their measurement, formulation of the method. Conclusion: neural networks were applied and informative criteria were selected for choosing the optimal variant of the neoadjuvant stage of treatment of patients with stage II-III rectal cancer, which was confirmed by their high accuracy of 80–95%.