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

Abstract

Objective: 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.

Materials and methods: 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.

Results: 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.

Conclusion: 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.

Keywords

Adnexal lesion / deep learning / postmenopausal

References

[1] Armstrong DK, Alvarez RD, Backes FJ, Bakkum-Gamez JN, Barroilhet L, et al. NCCN Guidelines® Insights: ovarian cancer, Version 3.2022. J Natl Compr Canc Netw 2022; 20(9): 972-80. [PMID: 36075393 DOI: 10.6004/jnccn.2022.0047]

[2] Huang J, Chan WC, Ngai CH, Lok V, Zhang L, et al. Worldwide burden, risk factors, and temporal trends of ovarian cancer: a global study. Cancers (Basel) 2022; 14(9): 2230. [PMID: 35565359 DOI: 10.3390/cancers14092230]

[3] Caruso G, Weroha SJ, Cliby W. Ovarian cancer: a review. JAMA 2025. [PMID: 40690248 DOI: 10.1001/jama.2025.9495]

[4] Webb PM, Jordan SJ. Global epidemiology of epithelial ovarian cancer. Nat Rev Clin Oncol 2024; 21(5): 389-400. [PMID: 38548868 DOI: 10.1038/s41571-024-00881-3]

[5] Nash Z, Menon U. Ovarian cancer screening: current status and future directions. Best Pract Res Clin Obstet Gynaecol 2020; 65: 32-45. [PMID: 32273169 DOI: 10.1016/j.bpobgyn.2020.02.010]

[6] Xiang H, Xiao Y, Li F, Li C, Liu L, et al. Development and validation of an interpretable model integrating multimodal information for improving ovarian cancer diagnosis. Nat Commun 2024; 15(1): 2681. [PMID: 38538600 DOI: 10.1038/s41467-024-46700-2]

[7] Van Holsbeke C, Daemen A, Yazbek J, Holland TK, Bourne T, et al. Ultrasound experience substantially impacts on diagnostic performance and confidence when adnexal masses are classified using pattern recognition. Gynecol Obstet Invest 2010; 69(3): 160-8. [PMID: 20016188 DOI: 10.1159/000265012]

[8] Christiansen F, Konuk E, Ganeshan AR, Welch R, Palés Huix J, et al. International multicenter validation of AI-driven ultrasound detection of ovarian cancer. Nat Med 2025; 31(1): 189-96. [PMID: 39747679 DOI: 10.1038/s41591-024-03329-4]

[9] Jacobs I, Oram D, Fairbanks J, Turner J, Frost C, et al. A risk of malignancy index incorporating CA 125, ultrasound and menopausal status for the accurate preoperative diagnosis of ovarian cancer. Br J Obstet Gynaecol 1990; 97(10): 922-9. [PMID: 2223684 DOI: 10.1111/j.1471-0528.1990.tb02448.x]

[10] Timmerman D, Testa AC, Bourne T, Ameye L, Jurkovic D, et al. Simple ultrasound-based rules for the diagnosis of ovarian cancer. Ultrasound Obstet Gynecol 2008; 31(6): 681-90. [PMID: 18504770 DOI: 10.1002/uog.5365]

[11] Wu M, Huang L, Chen Y, Wang Y, Zhang M, et al. Diagnostic accuracy of Ovarian-Adnexal Reporting and Data System, IOTA Simple Rules and Pediatric Risk of Malignancy Index for pediatric adnexal lesions: comparative study. Ultrasound Obstet Gynecol 2025; 66(3): 361-7. [PMID: 40643593 DOI: 10.1002/uog.29291]

[12] Van Calster B, Van Hoorde K, Valentin L, Testa AC, Fischerova D, et al. Evaluating the risk of ovarian cancer before surgery using the ADNEX model to differentiate between benign, borderline, early and advanced stage invasive, and secondary metastatic tumours: prospective multicentre diagnostic study. BMJ 2014; 349: g5920. [PMID: 25320247 DOI: 10.1136/bmj.g5920]

[13] Andreotti RF, Timmerman D, Benacerraf BR, Bennett GL, Bourne T, et al. Ovarian-adnexal reporting lexicon for ultrasound: a white paper of the ACR ovarian-adnexal reporting and data system committee. J Am Coll Radiol 2018; 15(10): 1415-29. [PMID: 30149950 DOI: 10.1016/j.jacr.2018.07.004]

[14] Strachowski LM, Jha P, Phillips CH, Blanchette Porter MM, Froyman W, et al. O-RADS US v2022: an update from the American College of Radiology’s ovarian-adnexal reporting and data system US committee. Radiology 2023; 308(3): e230685. [PMID: 37698472 DOI: 10.1148/radiol.230685]

[15] Wu M, Zhang M, Cao J, Wu S, Chen Y, et al. Predictive accuracy and reproducibility of the O-RADS US scoring system among sonologists with different training levels. Arch Gynecol Obstet 2023; 308(2): 631-7. [PMID: 35994107 DOI: 10.1007/s00404-022-06752-5]

[16] Meys EMJ, Jeelof LS, Achten NMJ, Slangen BFM, Lambrechts S, et al. Estimating risk of malignancy in adnexal masses: external validation of the ADNEX model and comparison with other frequently used ultrasound methods. Ultrasound Obstet Gynecol 2017; 49(6): 784-92. [PMID: 27514486 DOI: 10.1002/uog.17225]

[17] Meys EMJ, Kaijser J, Kruitwagen RFPM, Slangen BFM, Van Calster B, et al. Subjective assessment versus ultrasound models to diagnose ovarian cancer: a systematic review and meta-analysis. Eur J Cancer 2016; 58: 17-29. [PMID: 26922169 DOI: 10.1016/j.ejca.2016.01.007]

[18] Sayasneh A, Kaijser J, Preisler J, Johnson S, Stalder C, et al. A multicenter prospective external validation of the diagnostic performance of IOTA simple descriptors and rules to characterize ovarian masses. Gynecol Oncol 2013; 130(1): 140-6. [PMID: 23578539 DOI: 10.1016/j.ygyno.2013.04.003]

[19] Bi WL, Hosny A, Schabath MB, Giger ML, Birkbak NJ, et al. Artificial intelligence in cancer imaging: clinical challenges and applications. CA Cancer J Clin 2019; 69(2): 127-57. [PMID: 30720861 DOI: 10.3322/caac.21552]

[20] Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. Artificial intelligence in radiology. Nat Rev Cancer 2018; 18(8): 500-10. [PMID: 29777175 DOI: 10.1038/s41568-018-0016-5]

[21] tefan P-A, Lupean R-A, Mihu CM, Lebovici A, Oancea MD, et al. Ultrasonography in the diagnosis of adnexal lesions: the role of texture analysis. Diagnostics (Basel) 2021; 11(5): 812. [PMID: 33947150 DOI: 10.3390/diagnostics11050812]

[22] Chen H, Yang B-W, Qian L, Meng Y-S, Bai X-H, et al. Deep learning prediction of ovarian malignancy at US compared with O-RADS and expert assessment. Radiology 2022; 304(1): 106-13. [PMID: 35412367 DOI: 10.1148/radiol.211367]

[23] Gao Y, Zeng S, Xu X, Li H, Yao S, et al. Deep learning-enabled pelvic ultrasound images for accurate diagnosis of ovarian cancer in China: a retrospective, multicentre, diagnostic study. Lancet Digit Health 2022; 4(3): e179-87. [PMID: 35216752 DOI: 10.1016/S2589-7500(21)00278-8]

[24] Barcroft JF, Linton-Reid K, Landolfo C, Al-Memar M, Parker N, et al. Machine learning and radiomics for segmentation and classification of adnexal masses on ultrasound. NPJ Precis Oncol 2024; 8(1): 41. [PMID: 38378773 DOI: 10.1038/s41698-024-00527-8]

[25] American College of Obstetricians and Gynecologists’ Committee on Practice Bulletins-Gynecology. Practice bulletin no. 174: evaluation and management of adnexal masses. Obstet Gynecol 2016; 128(5): e210-26. [PMID: 27776072 DOI: 10.1097/AOG.0000000000001768]

[26] Grigore M, Popovici RM, Gafitanu D, Himiniuc L, Murarasu M, et al. Logistic models and artificial intelligence in the sonographic assessment of adnexal masses - a systematic review of the literature. Med Ultrason 2020; 22(4): 469-75. [PMID: 32905566 DOI: 10.11152/mu-2538]

[27] Koch AH, Jeelof LS, Muntinga CLP, Gootzen TA, van de Kruis NMA, et al. Analysis of computer-aided diagnostics in the preoperative diagnosis of ovarian cancer: a systematic review. Insights Imaging 2023; 14(1): 34. [PMID: 36790570 DOI: 10.1186/s13244-022-01345-x]

[28] Landolfo C, Bourne T, Froyman W, Van Calster B, Ceusters J, et al. Benign descriptors and ADNEX in two-step strategy to estimate risk of malignancy in ovarian tumors: retrospective validation in IOTA5 multicenter cohort. Ultrasound Obstet Gynecol 2023; 61(2): 231-42. [PMID: 36178788 DOI: 10.1002/uog.26080]

[29] Moro F, Vagni M, Tran HE, Bernardini F, Mascilini F, et al. Radiomics analysis of ultrasound images to discriminate between benign and malignant adnexal masses with solid morphology on ultrasound. Ultrasound Obstet Gynecol 2025; 65(3): 353-63. [PMID: 38748935 DOI: 10.1002/uog.27680]