Improving the diagnosis of endometrial cancer in postmenopausal women in primary care settings using an artificial intelligence-based ultrasound detecting model

Objectives We aimed to develop a deep learning (DL) model based on ultrasound examination to assist in ultrasound-based assessment of confirmed endometrial cancer (EC) in postmenopausal women, with the goal of improving diagnostic efficiency for EC in primary care settings. Methods A novel DL system was developed to analyze comprehensive gynecological ultrasound images, specifically targeting the identification of EC based on ultrasound features, using the diagnosis made by ultrasound specialists as the reference standard. Ultrasound measurements were performed to assess endometrial thickness and tumor homogeneity in all patients using gray-scale sonography. Intertumoral blood flow characteristics were analyzed through the blood flow area (BFA), resistance index (RI), end-diastolic velocity (EDV), and peak systolic velocity (PSV). The system’s performance was assessed using both internal and external test sets, with its effectiveness evaluated based on agreement with the ultrasound specialist and the area under the receiver operating characteristic (ROC) curve for binary classification. Results A total of 877 patients with EC diagnosed by endometrial biopsy at Hospital of Traditional Chinese Medicine of Qiqihar between January 1, 2020, and December 31, 2024, were enrolled in this study. 877 ultrasound images were divided into three groups: 614 for training, 175 for validation, and 88 for testing. The AUC for the training set was 0.844 (95% CI: 0.784–0.893). In the validation set, the AUC for predicting EC was 0.811 (95% CI: 0.748-0.864), while in the testing set, the AUC reached 0.858 (95% CI: 0.800-0.905). Conclusions The DL model demonstrated high accuracy and robustness, significantly enhancing the ability to diagnostic assistance for EC through ultrasound in postmenopausal women. This provides substantial clinical value, especially by enabling less experienced physicians in primary care settings to effectively detect EC lesions, ensuring that patients receive timely diagnosis and treatment.