关键词: 卵巢癌高风险人群, 糖类抗原125, 人附睾蛋白4, 列线图模型, 受试者工作特征曲线, 决策曲线分析

Abstract: Objective To analyze the risk factors associated with the incidence of ovarian cancer in a high-risk population and to construct a nomogram model to optimize existing screening pathways. Methods This study retrospectively analyzed data from 550 individuals at high risk for ovarian cancer who were seen at the Shanghai First Maternity and Infant Hospital between January 2020 and October 2021. The cohort was randomly divided into a training set (n=385) and a validation set (n=165) at a 7∶3 ratio. Patients in the training set were subsequently categorized into an occurrence group and a non-occurrence group based on confirmed ovarian cancer diagnosis following a 36-month follow-up period. Univariate and multivariate logistic regression analyses were conducted to identify independent predictors for the development of ovarian cancer in the training set. A nomogram was constructed using R software. The model's performance was evaluated through internal and external validation using Receiver Operating Characteristic (ROC) curves, calibration curves, the Hosmer-Lemeshow goodness-of-fit test, and Decision Curve Analysis (DCA). Results There were no statistically significant differences in the distribution of baseline characteristics, such as age, BMI, and family history of ovarian cancer, between the training and validation sets (P>0.05). Both univariate and multivariate logistic regression analyses identified elevated levels of carbohydrate antigen 125 (CA125) (>35 U/mL; OR=4.705), carbohydrate antigen 199 (CA199) (>37 U/mL; OR=3.322), and human epididymis protein 4 (HE4) (pmol/L; OR=12.272), a high-risk classification on the risk of ovarian malignancy algorithm (ROMA) index (OR=7.615), and low levels of the Pulsatility Index (PI) (OR=3.646) and resistance index (RI) (OR=2.475) as independent risk factors for ovarian cancer incidence in the high-risk population (all P<0.05). The evaluation and validation of the predictive nomogram demonstrated excellent discriminatory power, with an area under the curve (AUC) of 0.930 (95% CI: 0.887-0.973) in the training set and 0.921 (95% CI: 0.861-0.981) in the validation set. Furthermore, the calibration curves and the Hosmer-Lemeshow test signified a high degree of model calibration (Training set: R²=0.576, χ²=11.347, P=0.124; Validation set: R²=0.539, χ²=6.603, P=0.471). Decision Curve Analysis indicated that the model offered a substantial net benefit across a wide range of threshold probabilities (0.00-0.91 in the training set and 0.00-0.74 in the validation set). Conclusions CA125, CA199, HE4, the ROMA index, PI, and RI are significant predictors for the development of ovarian cancer in high-risk individuals. The nomogram constructed based on these factors can accurately predict the risk of ovarian cancer. Therefore, its application for risk stratification may serve as a valuable tool to facilitate precision clinical screening for ovarian cancer.

Key words: High-risk ovarian cancer population, Carbohydrate antigen 125, Human epididymis protein 4, Nomogram model, Receiver operating characteristic curve, Decision curve analysis