Quantitative analysis of the cervical texture by ultrasound in the mid-pregnancy is associated with spontaneous preterm birth

Objective New tools are required to identify women at increased risk of spontaneous preterm birth (sPTB). Quantitative texture analysis by ultrasound has been used to extract robust features from the ultrasound image of a given tissue to detect subtle changes in its microstructure. This may also be applied to the cervix. We aimed to determine if there is an association between quantitative analysis of the cervical texture (CTx) by ultrasound in the mid-trimester of pregnancy and sPTB <37+0 weeks. Methods This was a single centre, nested case–control study based on a prospective cohort of 677 consecutive singleton pregnancies assessed between 19+0 and 24+6 weeks. Women at increased risk for sPTB were included unless they received any treatment to prevent sPTB. Women who delivered <37+0 weeks (sPTB) were considered as cases (n = 32) and were matched with 320 contemporary controls randomly selected who delivered at term (10:1 for cases). Cervical images were assessed for quality, cervical length (CL) was measured off-line and a region of interest was delineated in the mid portion of the anterior cervical lip to be used for quantitative analysis of the CTx. A learning algorithm was developed to obtain the best combination of cervical textures features associated with sPTB based on feature transformation and discriminant analysis-regression. The ability of the learning algorithm to predict sPTB was evaluated using a leave-one-out cross-validation technique, obtaining a CTx-based score for each individual. Receiver–operating characteristic (ROC) curves were drawn and sensitivity, specificity, positive and negative likelihood ratios were calculated for the optimal cut-off based on the ROC curve. The results were compared to those obtained with the CL. At all points the investigators studying the images were blinded to pregnancy outcome. Results Images of 310 women (27 cases and 283 controls) were of sufficient quality and included in the study. Median CTx-based score (−1.01 vs −0.07, p = <0.0001) was significantly lower in cases compared to controls. CTx-based score maintained its significant association with sPTB after adjusting for possible confounders (history of sPTB, conization or Mullerian malformation and CL<25 mm). CTx-based score AUC’s ability to identify women who had a sPTB (0.77; 95% CI 0.66-0.87) was significantly better than CL AUC’s (0.60; 95% CI 0.47-0.72), p = 0.03. Median CL (mm) was similar (37.7 vs 38.6, p = 0.26) for cases and controls yet cases were more likely to have a CL<25 mm than controls (18.5% vs 0.4%, p < 0.0001). Conclusion Quantitative analysis of the cervical texture is able to extract information from the ultrasound images related to sPTB and generate a CTx-based score which is independently associated to sPTB.