Ultrasound-Trained Artificial Intelligence for Postmenopausal Adnexal Lesion Characterization: A Multicenter Diagnostic Accuracy Study

Announcing a new publication in BIO Integration journal. The purpose of this work was to establish and assess deep learning (DL) models based on ultrasound images for discriminating between benign and malignant adnexal lesions in postmenopausal women.

In this retrospective multicenter study, a total of 662 adnexal lesions from 662 postmenopausal women between January 2020 and December 2024 were included. Five DL models (model_Resnet50, model_{swin_transformer}, model_vit, model_{Convnext_tiny}, and model_{Regnet_y_8gf}) were trained and validated. Model performance was assessed with area under the curve (AUC), sensitivity, specificity, positive predictive value, and negative predictive value. The Assessment of Different NEoplasias in the adneXa (ADNEX) model without CA-125 was applied for comparison. The diagnostic performance of junior radiologists was assessed without or with DL model assistance. In addition, subgroup analysis was performed to assess the robustness of the DL models.

Model_{swin_transformer} yielded the highest AUC (0.964) among DL models in the external validation cohort (n = 93), with high sensitivity (0.952) and specificity (0.903). No statistical difference was observed between the AUCs of model_{swin_transformer} and the ADNEX model (AUC: 0.968; P = 0.819). Junior radiologists assisted by the DL model exhibited improved diagnostic performance, with higher AUCs (0.938 vs. 0.819; 0.944 vs. 0.838) and sensitivity (0.905 vs. 0.667; 1.000 vs. 0.857), while maintaining comparable specificity. However, model_{swin_transformer} did not significantly improve the diagnostic performance of attending radiologists and senior radiologists. Subgroup analyses revealed that model_{swin_transformer} presented superior diagnostic accuracy in purely cystic lesions, solid lesions, and lesions with a maximum diameter < 100 mm.

The proposed DL model has potential to assist radiologists in classifying adnexal lesions in postmenopausal women by effectively enhancing the diagnostic performance of junior radiologists in resource-limited healthcare settings.

Read More: https://www.scienceopen.com/hosted-document?doi=10.15212/bioi-2025-0175

BIO Integration is fully open access journal which will allow for the rapid dissemination of multidisciplinary views driving the progress of modern medicine.

As part of its mandate to help bring interesting work and knowledge from around the world to a wider audience, BIOI will actively support authors through open access publishing and through waiving author fees in its first years. Also, publication support for authors whose first language is not English will be offered in areas such as manuscript development, English language editing and artwork assistance.

Please visit www.bio-integration.org to learn more about the journal.

Editorial Board: https://bio-integration.org/editorial-board/

BIOI is available on the ScienceOpen platform (https://www.scienceopen.com/search#collection/55f45bac-fd64-431c-9fe7-50fd6a7f8781).

Submissions may be made using ScholarOne (https://mc04.manuscriptcentral.com/bioi).

There are no author submission or article processing fees.

Follow BIOI on Twitter @JournalBio; Facebook (https://www.facebook.com/BIO-Integration-Journal-108140854107716/) and LinkedIn (https://www.linkedin.com/company/bio-integration-journal/).

ISSN 2712-0074

eISSN 2712-0082

Wu M, Wu J, Wang Y et al. Ultrasound-Trained Artificial Intelligence for Postmenopausal Adnexal Lesion Characterization: A Multicenter Diagnostic Accuracy Study. BIO Integration 2025; 6: 1–12 DOI: 10.15212/bioi-2025-0175.