ADNEX-AI: automated extraction of ultrasound predictors for interpretable ovarian cancer risk stratification
Year: 2025
Type: article
Abstract: 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.
Source: npj Precision Oncology
Authors: Axel Geysels, Giulia Garofalo, Stefan Timmerman, Jolien Ceusters, Francesca Buonomo, Daniela Fischerová, Antonia Carla Testa, Francesca Moro, Povilas Sladkevicius, Ligita Jokubkiene, Caroline Van Holsbeke, Marek Kudla, Artur Czekierdowski, Elisabeth Epstein, Yvette Groszmann, Matthew Blaschko, Tom Bourne, Bart De Moor, Lil Valentin, Ben Van Calster, Dirk Timmerman, Wouter Froyman
Institutions: KU Leuven, Université Libre de Bruxelles, Universitair Ziekenhuis Leuven, Erasmus Hospital, IRCCS Materno Infantile Burlo Garofolo, Charles University, General University Hospital in Prague, Università Cattolica del Sacro Cuore, Agostino Gemelli University Polyclinic, Saint Camillus International University of Health and Medical Sciences, Lund University, Skåne University Hospital, Ziekenhuis Oost-Limburg, Medical University of Silesia, Medical University of Lublin, Karolinska Institutet, Stockholm South General Hospital, Brigham and Women's Hospital, Harvard University, Imperial College London, Leiden University, Leiden University Medical Center
Language: English
Topic: AI in cancer detection
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $4,090
Funders: Agentschap Innoveren en Ondernemen, National Institute for Health and Care Research, European Commission, Fonds Wetenschappelijk Onderzoek, Agentschap voor Innovatie door Wetenschap en Technologie, KU Leuven, Fondation contre le Cancer, Horizon 2020 Framework Programme, Universitaire Ziekenhuizen Leuven, KU Leuven, Departement Economie, Wetenschap en Innovatie
Awards: Optimising ovarian tumour management: a multicentre study of the short and long term effects of the surgical versus conservative management of ovarian masses., G1969776114, G2998977510, G3450463239, G4599083061, G4869584417, Innovative research and valorization of plainchant through digital technology (THE SOUND OF MUSIC)., G6912332768, G7544918479, The ADNEX risk model for ovarian cancer diagnosis: improving performance across different hospitals and examiners
OpenAlex ID: W4417244479