SUMMARY
Hyperplastic processes of endometrium represent an extremely important, complex and multi-faceted problem of practical gynecology, because it takes from 10 to 50% in the gynecological pathology structure, and its incidence is increasing steadily. At the same time they have tendency to a prolonged, recurrent course, and the absence of specific, pathognomonic symptoms, the complexity of differential diagnosis. Therefore, the effective diagnosis of these diseases and prevention of complications is particularly important. The use of the information technology in medicine, and especially in order to improve the diagnosis of gynecological diseases, has become increasingly important. The purpose of our study is to improve efficacy of the diagnostic of endometrial hyperplastic processes in perimenopausal women, based on in-depth analysis of clinical and epidemiological data in different types of hyperplasia using neural network clustering.Methods. There have been performed a retrospective analysis of 52 medical cards and biopsies of perimenopausal women: 1st group -28 women with simple endometrial hyperplasia (SEH); 2nd group – 24 women with complex endometrial hyperplasia (CEH), the control group consisted of 12 healthy women aged (45,5 ± 0,7)years. We analyzed patients' age, socio-economic factors, obstetrical and gynecological history, clinical and laboratory features of the course of climacteric syndrome and comorbid conditions for women with different types of hyperplasia. Histopathological diagnosis was verified histologically and clinically. ResultsRisk factors for hyperplastic processes of the endometrium for women in perimenopausal period include unfavorable socio-economic factors: living in rural areas, difficult working conditions, and bad habits. Endometrial hyperplasia was significantly more likely to develop for women with chronic diseases of the uterus, three or more pregnancies, a large number (4 or more) obstetrical and gynecological interventions and using intrauterine contraceptive. For women with different types of endometrial hyperplastic processes climacteric syndrome had moderate or severe course with anemia, vegetative, psycho- and endocrine-metabolic disorders. Among the comorbid conditions of women in perimenopausal period with endometrial hyperplasia more often observed the dishormonal pathology and the breast disorders of hepato-biliary tract. The results, based on analysis of combined changes with using multiparameter neural clustering have shown that the relatively older age and menopausal syndrome revealed a relatively higher proportion of carried surgical interventions of patients and based on these results, we can predict the progression of endometrial hyperplastic processes. The results have identified patterns based on clinical and anamnestic analysis using multiparameter neural clustering could be used to develop diagnostic criteria for predicting disease development in the region. Conclusion The results have identified patterns based on clinical and anamnestic analysis using multiparameter neural clustering could be used to develop diagnostic criteria for predicting disease development in the region.