Construction of a nomogram model for predicting benignity and malignancy in adnexal masses

This study aimed to construct a simple nomogram model to predict the benign or malignant nature of ovarian masses.

A total of 794 patients with ovarian masses from two independent medical centers between 2017 and 2024 were included. Among them, 700 patients from center 1 were randomly divided into training and internal validation datasets, and 94 patients from center 2 comprised the external validation dataset. Clinical information, ultrasound images, and sonographic characteristics were collected for all patients. Independent risk factors were screened and incorporated into the nomogram. The diagnostic performance of the nomogram was compared with that of the Risk of Malignancy Index 4 (RMI4) and Ovarian-Adnexal Reporting and Data System (O-RADS) models.

Ultrasound score, menopausal status, maximum tumor diameter, elevated carbohydrate antigen 125, and elevated human epididymis protein 4 were identified as independent risk factors for distinguishing benign from malignant ovarian masses. In the training, internal validation, and external validation datasets, the areas under the receiver operating characteristic curve (AUCs) were 0.912, 0.906, and 0.949, respectively. The corresponding sensitivity values were 79.9%, 71.2%, and 93.3%, and the specificity values were 87.6%, 90.4%, and 86.7%, respectively. Compared with RMI4 and O-RADS, the nomogram demonstrated the highest AUC and greater clinical net benefit.

The nomogram is simple to apply and demonstrates significantly higher diagnostic performance than RMI4, while achieving comparable performance to O-RADS. It may serve as a practical and widely applicable tool for predicting the benign or malignant nature of ovarian tumors.