[Evaluation of risk of progression of endometrial hyperplasia in patients with metabolic syndrome].

A mathematical model based on principles of multifactor analysis was developed to predict clinical outcome of endometrial hyperplasia (EH) in patients with metabolic syndrome (80). Seventy-seven factors--anthropometric, clinical, anamnestic, hormono-metabolic, immunohistochemical, etc.--were included. Evaluation of the most informative indices integrated with the discriminative model showed that anthropometric (waist and hip circumference, sagittal diameter, etc.) and clinico-anamnestic (age, age of secondary sexual characters appearance, body weight at birth, suckling pattern, etc.) ones are of similar significance. A profile of hormono-metabolic parameters (cholesterol-low density lipoprotein, leptin, testosterone, progesterone and fasting glucose levels) helped identify a wide range of EH-related disorders in patients with metabolic syndrome. Consistently with the literature data, level of PTEN expression pointed to the presence of this tumor's suppressor in most EH cases which was matched by absence of its expression in endometrial carcinoma. Our model provided high sensitivity (89%) and specificity (82%) in predicting risk of progression in patients with endometrial hyperplasia and metabolic syndrome.