Accurate triage of ovarian masses is facilitated in many centers by the guideline-endorsed, extensively validated IOTA-ADNEX risk model, yet the model still relies on manual measurements of key tumor features. We developed ADNEX-AI, a multi-task deep-learning system that automatically segments four ADNEX ultrasound predictors - lesion, locules, solid tissue, papillary projections - and outputs their quantitative values. The network was trained on 816 annotated images from 369 consecutive women recruited at 11 centers (43% malignancies) and prospectively evaluated on a temporally separate cohort of 1088 patients scanned at 10 of those centers (8008 images; 35% malignancies). ADNEX-AI discriminated benign from malignant tumors with an AUC of 0.930 (95% CI 0.913-0.943), less than but close to examiner-derived ADNEX (0.945; 0.930-0.957; P = 0.004) while delivering better calibration and markedly lower inter-center variability. By removing manual caliper work yet preserving full interpretability, ADNEX-AI could extend high-quality ovarian-cancer risk stratification to clinics that lack specialized ultrasound expertise.

Department of Biomedical Data Sciences Leiden University Medical Centre Leiden The Netherlands

Department of Clinical Science and Education Södersjukhuset Karolinska Institutet Stockholm Sweden

Department of Clinical Sciences Malmö Lund University Lund Sweden

Department of Development and Regeneration KU Leuven Leuven Belgium

Department of Gynaecological Oncology and Gynaecology Medical University of Lublin Lublin Poland

Department of Gynaecology and Obstetrics Hôpital Erasme Hôpital Universitaire de Bruxelles Brussels Belgium

Department of Life Science and Public Health Universita' Cattolica del Sacro Cuore Rome Italy

Department of Obstetrics and Gynaecology Imperial College London London UK

Department of Obstetrics and Gynaecology Skåne University Hospital Malmö Sweden

Department of Obstetrics and Gynaecology Södersjukhuset Stockholm Sweden

Department of Obstetrics and Gynaecology University Hospitals Leuven Leuven Belgium

Department of Obstetrics and Gynaecology Ziekenhuis Oost Limburg Genk Belgium

Department of Obstetrics Gynecology and Reproductive Biology Brigham and Women's Hospital Harvard Medical School Boston MA USA

Department of Perinatology and Oncological Gynaecology Faculty of Medical Sciences Medical University of Silesia Katowice Poland

Department of Woman Child and Public Health Fondazione Policlinico Universitario Agostino Gemelli IRCCS Rome Italy

Gynaecological Oncology Centre Department of Obstetrics and Gynaecology 1st Faculty of Medicine Charles University and General University Hospital Prague Czechia

Institute for Maternal and Child Health IRCCS Burlo Garofolo Trieste Italy

Laboratory of Tumour Immunology and Immunotherapy Department of Oncology KU Leuven Leuven Belgium

Leuven Unit for Health Technology Assessment Research KU Leuven Leuven Belgium

Processing Speech and Images KU Leuven Leuven Belgium

Saint Camillus International University of Health and Medical Sciences Rome Italy

STADIUS Center for Dynamical Systems Signal Processing and Data Analytics Department of Electrical Engineering KU Leuven Leuven Belgium

Zobrazit více v PubMed

Bray, F. et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. PubMed

Siegel, R. L., Giaquinto, A. N. & Jemal, A. Cancer statistics, 2024. PubMed

Arnold, M. et al. Progress in cancer survival, mortality, and incidence in seven high-income countries 1995–2014 (ICBP SURVMARK-2): a population-based study. PubMed DOI PMC

Timmerman, D. et al. ESGO/ISUOG/IOTA/ESGE Consensus Statement on preoperative diagnosis of ovarian tumors. PubMed DOI

White, K. M., Seale, H. & Harrison, R. Enhancing ovarian cancer care: a systematic review of guideline adherence and clinical variation. PubMed DOI PMC

Van Calster, B. 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. PubMed DOI PMC

Van Calster, B. et al. Validation of models to diagnose ovarian cancer in patients managed surgically or conservatively: multicentre cohort study. PubMed DOI PMC

Sundar, S. et al. Risk-prediction models in postmenopausal patients with symptoms of suspected ovarian cancer in the UK (ROCkeTS): a multicentre, prospective diagnostic accuracy study. PubMed DOI

Andreotti, R. F. et al. O-RADS US risk stratification and management system: a consensus guideline from the ACR ovarian-Adnexal Reporting and Data System committee. PubMed DOI

Timmerman, S. et al. External Validation of the Ovarian-Adnexal Reporting and Data System (O-RADS) Lexicon and the International Ovarian Tumor Analysis 2-Step Strategy to Stratify Ovarian Tumors Into O-RADS Risk Groups. PubMed DOI PMC

Barreñada, L. et al. ADNEX risk prediction model for diagnosis of ovarian cancer: systematic review and meta-analysis of external validation studies. PubMed DOI PMC

Pajouheshnia, R., van Smeden, M., Peelen, L. M. & Groenwold, R. H. H. How variation in predictor measurement affects the discriminative ability and transportability of a prediction model. PubMed DOI

Van Calster, B., Timmerman, S., Geysels, A., Verbakel, J. Y. & Froyman, W. A deep-learning-enabled diagnosis of ovarian cancer. PubMed DOI

Spisak, T. Statistical quantification of confounding bias in machine learning models. PubMed DOI PMC

Gao, Y. et al. Deep learning-enabled pelvic ultrasound images for accurate diagnosis of ovarian cancer in China: a retrospective, multicentre, diagnostic study. PubMed DOI

Christiansen, F. et al. International multicenter validation of AI-driven ultrasound detection of ovarian cancer. PubMed DOI PMC

Li, J. et al. A deep learning model system for diagnosis and management of adnexal masses. PubMed DOI PMC

Whiting, P. F. et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. PubMed DOI

Barcroft, J. F. et al. Machine learning and radiomics for segmentation and classification of adnexal masses on ultrasound. PubMed DOI PMC

Xie, W. et al. Developing a deep learning model for predicting ovarian cancer in Ovarian-Adnexal Reporting and Data System Ultrasound (O-RADS US) Category 4 lesions: a multicenter study. PubMed DOI PMC

Sladkevicius, P. & Valentin, L. Interobserver agreement in describing the ultrasound appearance of adnexal masses and in calculating the risk of malignancy using logistic regression models. PubMed DOI

Collins, G. S. et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. PubMed DOI PMC

Timmerman, D. et al. Terms, definitions and measurements to describe the sonographic features of adnexal tumors: a consensus opinion from the International Ovarian Tumor Analysis (IOTA) group. PubMed DOI

Education and Practical Standards Committee. Minimum training recommendations for the practice of medical ultrasound. PubMed

Install‚, A. J. F., Van Den Bosch, T., Moor, B. De. & Timmerman, D. Clinical data miner: an electronic case report form system with integrated data preprocessing and machine-learning libraries supporting clinical diagnostic model research. PubMed DOI PMC

Berek, J. S., Renz, M., Kehoe, S., Kumar, L. & Friedlander, M. Cancer of the ovary, fallopian tube, and peritoneum: 2021 update. PubMed DOI PMC

Timmerman, D. et al. Logistic regression model to distinguish between the benign and malignant adnexal mass before surgery: a multicenter study by the International Ovarian Tumor Analysis Group. PubMed DOI

Fedorov, A. et al. 3D Slicer as an image computing platform for the Quantitative Imaging Network. PubMed DOI PMC

Isensee, F., Jaeger, P. F., Kohl, S. A. A., Petersen, J. & Maier-Hein, K. H. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. PubMed DOI

Kang, H. et al. A comprehensive benchmarking of a U-Net based model for midbrain auto-segmentation on transcranial sonography. PubMed

Cao, W., Mirjalili, V. & Raschka, S. Rank consistent ordinal regression for neural networks with application to age estimation. DOI

He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In

Verbakel, J. Y. et al. ROC curves for clinical prediction models part 1. ROC plots showed no added value above the AUC when evaluating the performance of clinical prediction models. PubMed DOI

Van Calster, B. et al. Calibration: the Achilles heel of predictive analytics. PubMed DOI PMC

Wynants, L. et al. Clinical utility of risk models to refer patients with adnexal masses to specialized oncology care: multicenter external validation using decision curve analysis. PubMed DOI

Lundberg, S. M. & Lee, S.-I. A unified approach to interpreting model predictions. In

White, I. R., Royston, P. & Wood, A. M. Multiple imputation using chained equations: Issues and guidance for practice. PubMed DOI

Pedregosa, F. et al. Scikit-learn: machine Learning in Python.

Seabold, S. & Perktold, J. statsmodels: Econometric and statisticalmodeling with Python. In

Debray, T. P. A. et al. A framework for meta-analysis of prediction model studies with binary and time-to-event outcomes. PubMed DOI PMC

Campo, B. D. C. Towards reliable predictive analytics: a generalized calibration framework. Preprint at http://arxiv.org/abs/2309.08559 (2023).