ADNEX-AI: automated extraction of ultrasound predictors for interpretable ovarian cancer risk stratification
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.
Authors
- D. Timmerman (ORCID: https://orcid.org/0000-0002-3707-6645)
- Axel Geysels (ORCID: https://orcid.org/0000-0002-5922-1030)
- A. C. Testa (ORCID: https://orcid.org/0000-0003-2217-8726)
- E. Epstein (ORCID: https://orcid.org/0000-0003-2298-7785)
- Tom Bourne (ORCID: https://orcid.org/0000-0003-1421-6059)
- F. Moro (ORCID: https://orcid.org/0000-0002-5070-7245)
- Yvette S. Groszmann (ORCID: https://orcid.org/0000-0002-5484-3404)
- Marek Kudła (ORCID: https://orcid.org/0000-0003-0928-3433)
- L. Valentin (ORCID: https://orcid.org/0000-0002-3830-6414)
- G. Garofalo (ORCID: https://orcid.org/0000-0003-0728-2066)
- P. Sladkevicius (ORCID: https://orcid.org/0000-0001-9563-8135)
- Stefan Timmerman (ORCID: https://orcid.org/0000-0003-2253-7522)
- L. Jokubkiene (ORCID: https://orcid.org/0000-0002-0758-3507)
- Artur Czekierdowski (ORCID: https://orcid.org/0000-0001-6481-4271)
- Matthew B. Blaschko (ORCID: https://orcid.org/0000-0002-2640-181X)
- F. Buonomo (ORCID: https://orcid.org/0000-0002-6587-2622)
- Jolien Ceusters (ORCID: https://orcid.org/0000-0001-5422-5668)
- Wouter Froyman (ORCID: https://orcid.org/0000-0002-1398-9124)
- Ben Van Calster (ORCID: https://orcid.org/0000-0003-1613-7450)
- Bart De Moor (ORCID: https://orcid.org/0000-0001-5836-2037)
- Caroline Van Holsbeke
- Daniela Fischerová
Institutions
- Università Cattolica del Sacro Cuore (IT)
- Medical University of Lublin (PL)
- Brigham and Women's Hospital (US)
- Université Libre de Bruxelles (BE)
- Harvard University (US)
- Medical University of Silesia (PL)
- Lund University (SE)
- Charles University (CZ)
- IRCCS Materno Infantile Burlo Garofolo (IT)
- Leiden University Medical Center (NL)
- Agostino Gemelli University Polyclinic (IT)
- Karolinska Institutet (SE)
- Stockholm South General Hospital (SE)
- Saint Camillus International University of Health and Medical Sciences (IT)
- Skåne University Hospital (SE)
- Ziekenhuis Oost-Limburg (BE)
- Imperial College London (GB)
- KU Leuven (BE)
Publication Details
- Journal
- npj Precision Oncology
- Published
- 2025-12-11
- DOI
- https://doi.org/10.1038/s41698-025-01215-x
- Primary Topic
- Ovarian cancer diagnosis and treatment
- Type
- article
- Field-Weighted Citation Impact
- 0.00