Evaluation of the natural birth prediction system for assessing cephalopelvic disproportion: a multicentre...

To assess the potential value of the Natural birth prediction system (NBPS) on predicting cephalopelvic disproportion (CPD) and compare the value with a pelvimetry model and the foetal-pelvic index.

A multicentre prospective study was conducted between November 2023 and September 2024 in three hospitals in China. Singleton nulliparas suspected with CPD and willing to attempt vaginal delivery were included. Patients were categorised according to NBPS results (absence CPD, borderline CPD, and presence CPD). The software results were compared with actual delivery modes. Receiver operating characteristic (ROC) curve analysis was employed to compare the predictive value of NBPS with a pelvimetry-based nomogram model and foetal-pelvic index.

The final analysis included 303 patients. Among the 14 women identified as presence CPD by NBPS, 12 underwent emergency CS, while the rest two delivered vaginally with forceps. Of the 213 patients deemed absence CPD by NBPS, 190 had normal vaginal deliveries, 22 delivered with forceps, and only one required an emergency CS. The area under the ROC curve (AUC) for NBPS was 0.908, with a sensitivity of 97.0% and specificity of 78.5%. The PPV was 35.6%, and the NPV was 99.5%. In comparison, the AUC for the pelvimetry model was 0.827, with a sensitivity of 81.8% and specificity of 70.2%. The PPV was 28.4%, and the NPV was 96.4%. For the foetal-pelvic index, the sensitivity and specificity were 72.7% and 42.6%, and the PPV and NPV were 13.4% and 92.7%, respectively.

This study demonstrates that NBPS shows promise for clinically predicting CPD among those suspected of CPD and willing to attempt vaginal delivery. Our findings support its utility as an objective assessment tool in this specific population to help manage labour and anticipate the risk of dystocia.