First trimester cervical angles for the prediction of spontaneous preterm birth

Objective: First trimester cervical angles for the prediction of spontaneous preterm birth (sPTB) remains unclear. The objective is to explore the potential value of first trimester cervical angles for the prediction of sPTB. Study design: This was a secondary analysis of data derived from a prospective cohort study for sPTB screening in singleton pregnancies at 11 + 0-13 + 6 weeks in women attending routine Down's syndrome screening at Prince of Wales Hospital, Hong Kong SAR, between June 2018 and July 2020. Methods: Using archived images of the sagittal view of the cervix, eight types of cervical angles were measured: Angle-A (two lines crossing the greatest curvature of cervix); Angle-B (isthmus and endocervical canal); Angle-C1 (anterior wall and cervix including isthmus); Angle-C2 (anterior wall and endocervical canal); Angle-D1 (posterior wall and cervix including isthmus); Angle-D2 (posterior wall and endocervical canal); Angle-E (anterior wall and isthmus); Angle-F (posterior wall and isthmus). Likelihood ratios for cervical angle multiples of the median (MoMs) were computed and combined together with maternal variables to estimate the patient-specific risk of sPTB for each case. The screening performance of sPTB at <34 and <37 weeks was assessed by receiveroperating characteristic (ROC) curve analysis. Detection rates (DRs) for sPTB at <34 and <37 weeks were determined at a 10 % fixed false positive rate (FPR). The areas under the ROC curves (AUCs) were compared using DeLong test. Results: Among a total of 3658 included pregnancies, sPTB at <37 weeks occurred in 19 cases (0.52 %) and 154 cases (4.21 %) respectively. In the term birth group, cervical angles were affected by log10 cervical length and maternal factors (age, height, weight, method of conception, previous preterm birth, previous cervical surgery, use of progesterone). When compared to term birth group, median Angle-E was significantly increased in sPTB at <34 weeks (P = 0.017); while median Angle-D1 and D2 were significantly decreased in sPTB at <34 weeks (P = 0.049 and 0.025, respectively). The a priori risk for sPTB at <34 weeks was provided by body mass index, previous miscarriage, and previous PTB. Similarly, the a priori risk for sPTB at <37 weeks was provided by maternal height, and previous PTB. For the prediction of sPTB at <34 weeks, the best AUC was achieved by a combination of maternal factors and Angle-E MoM (AUC: 0.786, DR, 30.0 % at a FPR of 10 %). However, the difference between the AUCs of maternal factors only, Angle-E, and the combined model did not reach statistical significance. For the prediction of sPTB at <37 weeks, the best AUC was achieved by a combination of maternal factors and Angle-D2 MoM (AUC: 0.599, 95 % CI: 0.539-0.658, DR, 18.7 % at a FPR of 10 %). The difference between the AUCs of maternal factors only, Angle-D1, Angle-D2 and the combined models did not reach statistical significance. Conclusions: Whilst there are associations between cervical angles and sPTB, these indices have limited value for prediction of sPTB in the first trimester. Further prospective studies are needed to identify other effective markers for the prediction of sPTB in the first trimester. Further prospective studies are needed to identify other effective markers for the prediction of sPTB in the first trimester.