A scoring system predicts the adverse outcomes for cesarean scar pregnancy treated with transvaginal hysterotomy.

Background: Cesarean scar pregnancy (CSP) may lead to uterine rupture and heavy bleeding as the pregnancy progresses. Removal of the lesion by transvaginal hysterotomy is considered an effective first-line treatment for CSP, but there are few comprehensive prognostic models to determine which patients are candidates for transvaginal hysterotomy. This study aims to formulate a risk scoring system using a large cohort of CSP patients to evaluate the risk of adverse events when treated with transvaginal hysterotomy. Methods: A total of 273 CSP patients who underwent transvaginal hysterotomy at the First Affiliated Hospital of Sun Yat-sen University in China from January 2009 to February 2023 were studied, retrospectively. Patients' clinical characteristics, surgical outcomes and detailed ultrasound characteristics were recorded according to a newly developed consensus. Multivariable logistic regression was employed to identify the risk factors associated with adverse events, and subsequently, a predictive model was constructed. A scoring system was developed by stratifying these risk factors and assigning scores to them using a mathematical model. Calibration was assessed with the Hosmer-Lemeshow test, and internal validation was performed using stratified 5-fold cross-validation. The diagnostic accuracy, sensitivity, specificity, and the area under the receiver operating characteristics curve were calculated to evaluate the performance of both the developed prediction model and the scoring system. Results: Of 273 patients, 244 (89.4%) were successfully treated without experiencing any adverse events, while 29 (10.6%) had encountered at least one adverse event. The study identified gestational age [odds ratio (OR) =1.244, 95% confidence interval (CI): 1.052-1.422], the exact extent of the protrusion (OR =2.568, 95% CI: 1.627-4.047), and the vascularity of the lesion (OR =3.739, 95% CI: 1.551-9.016) as independent predictors for adverse events. Based on these findings, a prediction model and an 8-point scoring system were created. Both the prediction model and the scoring system demonstrated excellent performance. The average area under the receiver operating characteristics curve (AUC) of the prediction model was 0.828 (95% CI: 0.769-0.887), and the average accuracy was 0.901 (95% CI: 0.864-0.938). For the scoring system, the average area under the receiver operating characteristics curve was 0.829 (95% CI: 0.775-0.883), and the average accuracy of 0.898 (95% CI: 0.835-0.961). The model and scoring system had an acceptable goodness of fit according to the Hosmer-Lemeshow test (P=0.223 and P=0.707, respectively). Conclusions: The prognostic model and a scoring system derived from a large cohort of CSP patients who underwent transvaginal hysterotomy can effectively predict adverse events. The scoring system is easy-to-use and can intuitively help clinicians identify candidates for transvaginal hysterotomy. It facilitates the implementation of more individualized and effective management strategies for CSP, and potentially enhances patient prognosis.