Aim: to improve efficacy of predicting preterm labor in the first trimester of pregnancy by combining diverse parameters of cervical ultrasound examination.Materials and methods. A prospective cohort study of 1517 women with uncomplicated pregnancy was performed. Inclusion criteria were: females underwent cervicometry at 11–14 weeks of gestation, singleton pregnancy, revealing no complaints at the onset of examination. All women were subdivided into four groups: Group 1 – 27 pregnant women with shortened cervix (less than 30 mm); Group 2 – 24 pregnant women without cervical gland area (СGA); Group 3 – 30 pregnant women with two risk factors (shortened cervix less than 30 mm and lacked СGA); Group 4 (control) consisted of 1436 pregnant women with cervix length exceeding 30 mm and presence of СGA.Results. Average delivery term in Group 1 was 35.7 weeks (95 % CI = 34.7–36.8), in Group 2 – 34.7 weeks (95 % CI = 33.59–35.0), in Group 3 – 33.23 weeks (95 % CI = 31.6–34.8), in Group 4 (control) – 38.11 weeks (95 % CI = 38.06–38.17). A significant moderate correlation (Rxy = 0.534) between shortened cervix, absence of СGA and delivery term was found (p < 0.001). A regression model consisting of cervical length and presence of СGA was simulated based upon 50.8 % factors underlying probability of preterm birth, revealing 42.6 % sensitivity, 99.1 % specificity, and 96.6 % overall diagnostic value. The area under the ROC curve was 0.902 ± 0.022 (95 % CI = 0.860–0.945). The data obtained reflect diverse biochemical changes such as collagen decomposition, altered glucosamine level and fluid amount within cervical tissues. Such processes result in shortened, softened and expanded (matured) cervix. Uncovering markers for preterm cervical maturation underlies a logical strategy to predict miscarriage.Conclusion. Untrasound cervical measurement in the first trimester of pregnancy allows for revealing valuable miscarriage predictors (shortened cervix and absence of GI). Risk assessment by combining diverse ultrasound cervix parameters would allow to improve overall predictive efficacy.