A New Approach to Predict Cephalopelvic Disproportion, Based on Three-Dimension Magnetic Resonance Imaging Pelvic Reconstruction and Computer Simulated Labor: A Multicenter Prospective Study

Abstract

Background: Cephalopelvic disproportion (CPD) often causes emergency cesarean section (CS), which increases the risk of maternal morbidity and surgical complications than elective cesarean delivery. How to predict CPD before labor has been the focus of researchers for years. Based on the three-dimension (3D) reconstruction of magnetic resonance imaging (MRI) of maternal pelvis and fetal head, we developed a software of computer simulated labor named Natural birth prediction system, which could assess the proportionality of pelvis and fetus size. We conducted this research to assess the potential value of NBPS on predicting CPD and compare the value with a previous pelvimetry model and the fetal-pelvic index.

Methods: A multicenter prospective study was conducted between November 2023 and September 2024 in three hospitals in Zhejiang province,China. Singleton primiparas who were suspected with CPD and who were willing to attempt vaginal delivery were included. Exclusion criteria were nonvertex fetal presentation and contraindications for MRI. Women chose scheduled CS due to changes in their preferences or conditions such as preeclampsia, as well as those who underwent emergency CS for reasons like fetal distress, were excluded from the final analysis. Patients were categorized according to NBPS results (absence CPD, borderline CPD, and presence CPD). Demographic, clinical, and radiological characteristics were collected. 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 another MRI model involving maternal body mass index, pelvimetry, and fetal measurements. For each model, specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV) were calculated.

Findings: The final analysis included 303 patients. Among the 14 women identified as presence CPD by the NBPS, 12 underwent emergency CS due to CPD, while the rest two were deliverd vaginally with forceps assistance. 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 due to CPD. The area under the ROC curve (AUC) for NBPS model 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 fetal-pelvic index, the sensitivity and specificity were 72.7% and 42.6%, and the PPV and NPV were 13.4% and 92.7%, respectively.

Interpretation: This study demonstrates that the novel software based on three-dimension reconstruction of the female pelvis and fetal head shows promise as a reliable tool for predicting CPD in clinical practice.

Funding: This work was supported by Scientific Research Foundation of the National Health Commission (WKJ-ZJ-2126), supported by the National Key Research and Development Program of China (No. 2021YFC2700700, 2022YFC2704600, 2022YFC2704601), supported by Medical and Health Program of Zhejiang Province (2022KY1229) and supported by Science Program of Quzhou (2023K186).

Declaration of Interest: All authors declare no competing interests.

Ethical Approval: Approval was obtained from the ethics committee at the Women’s Hospital, School of Medicine Zhejiang University (IRB-20230253-R). In addition, written informed consent was obtained from all women who agreed to participate.

Keywords: cephalopelvic disproportion, magnetic resonance imaging, three-dimension reconstruction, prediction