Development and validation of a model for individualized prediction of cervical insufficiency risks in patients undergoing IVF/ICSI treatment

Abstract

Background Women who conceived with in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) are more likely to experience adverse pregnancy outcomes than women who conceived naturally. Cervical insufficiency (CI) is one of the important causes of miscarriage and premature birth, however there is no published data available focusing on the potential risk factors predicting CI occurrence in women who received IVF/ICSI treatment. This study aimed to identify the risk factors that could be integrated into a predictive model for CI, which could provide further personalized and clinically specific information related to the incidence of CI after IVF/ICSI treatment. Patients and methods This retrospective study included 4710 patients who conceived after IVF/ICSI treatment from Jan 2011 to Dec 2018 at a public university hospital. The patients were randomly divided into development (n = 3108) and validation (n = 1602) samples for the building and testing of the nomogram, respectively. Multivariate logistic regression was developed on the basis of pre-pregnancy clinical covariates assessed for their association with CI occurrence. Results A total of 109 patients (2.31%) experienced CI among all the enrolled patients. Body mass index (BMI), basal serum testosterone (T), gravidity and uterine length were associated with CI occurrence. The statistical nomogram was built based on BMI, serum T, gravidity and uterine length, with an area under the curve (AUC) of 0.84 (95% confidence interval: 0.76–0.90) for the developing cohort. The AUC for the validation cohort was 0.71 (95% confidence interval: 0.69–0.83), showing a satisfactory goodness-of-fit and discrimination ability in this nomogram. Conclusion The user-friendly nomogram which graphically represents the risk factors and a pre-pregnancy predicted tool for the incidence of CI in patients undergoing IVF/ICSI treatment, provides a useful guide for medical staff on individualized decisions making, where preventive measures could be carried out during the IVF/ICSI procedure and subsequent pregnancy.

A Validated Model for Individualized Prediction of Live Birth in Patients With Adenomyosis Undergoing Frozen–Thawed Embryo Transfer

“…Statistical analyses were performed using the STATA 14.0 MP software and Regression Modeling Strategies (R version 3.6.3). To build up the nomogram and measure the area under the curve (AUC), we used the “regplot”, “pROC”, and “rms” in R software ( 19 ). Differences between groups were compared using Student’s t -test or chi‐square test as appropriate.…”

A Validated Model for Individualized Prediction of Live Birth in Patients With Adenomyosis Undergoing Frozen-Thawed Embryo Transfer

“…Statistical analyses were performed using the STATA 14.0 MP software and Regression Modeling Strategies (RMS, R version 3.6.3). For the nomogram establishment and the AUC measurements, we used the "regplot", "pROC" and "rms" in R software [19]. Differences between groups were compared using Student's t-test or Chi-squared test as appropriate.…”

“…The predictive accuracy of the models was measured using the average optimism of the area under the curve (AUC). A precise prediction model would result in a plot where the observed and predicted probabilities fall along the diagonal [19].…”

“…Variables entered into the nomogram model were age, uterine volume, stage of transferred embryo, twin pregnancy, and protocol of FET in the study. Values for each of the model covariates were mapped to points on a scale ranging from 0 to 100 and the total points obtained for each model corresponded to the probability of a live birth [19].…”

“…The predictive accuracy of the models was measured using the average optimism of the area under the curve (AUC). A precise prediction model would result in a plot where the observed and predicted probabilities fall along the diagonal [19]. Logistic regression analysis revealed blastocyst transfer, small uterine size and GnRH-a pretreated prior to FET improved live birth, but twin pregnancy negatively impacted live birth.…”