Automated detection of active and structural MRI lesions in the sacroiliac joints in axial spondyloarthritis: training and validation across 3 phase III clinical trial datasets.
Authors
Affiliations (9)
Affiliations (9)
- University of Oxford, Oxford, UK.
- Oxford Clinical Artificial Intelligence Research (OxCAIR) Kadoorie Centre, Oxford University Hospitals NHS Foundation Trust, Headley Way, Oxford, OX3 9DU, UK.
- Rheuma Praxis Berlin, Berlin, Germany; Ruhr-Universität Bochum, Bochum, Germany.
- Division of Rheumatology, University of California San Francisco, San Francisco, CA, USA.
- Copenhagen Center for Arthritis Research, Center for Rheumatology and Spine Diseases, Rigshospitalet, Glostrup, Denmark; Department of Clinical Medicine; University of Copenhagen, Copenhagen, Denmark.
- Division of Rheumatology, University of Toronto and University Health Network, Toronto, ON, Canada; Department of Gastroenterology, Infectiology and Rheumatology (including Nutrition Medicine), Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Berlin, Germany.
- Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA.
- Selka Health, Oxford, UK.
- Department of Neuromuscular Diseases, University College London, London, UK; National Institute for Health Research (NIHR) University College London Hospitals Biomedical Research Centre, London, UK; Department of Rheumatology, Northwick Park Hospital, London North West University Healthcare NHS Trust, London, UK. Electronic address: [email protected].
Abstract
This study aims to assess the performance of a fully automated deep learning (DL) system for detecting active and structural magnetic resonance imaging (MRI) lesions of the sacroiliac joints (SIJs) in axial spondyloarthritis (axSpA), and validate its generalisability across independent clinical trial datasets. A 2-stage automated pipeline was developed to delineate left and right SIJs and detect 5 MRI-defined lesion types: 1 active lesion: bone marrow oedema (BMO), 4 structural lesions: erosions, fat lesions, sclerosis, and ankylosis. Lesions were assessed at the quadrant or joint level using paired T1-weighted and Short Tau Inversion Recovery sequences. Models were trained on the MEASURE 1 trial (132 patients) using consensus-based labels to address multireader variability and evaluated on 2 independent datasets (PREVENT-555 patients; SURPASS-414 patients). Performance was assessed using area under the curve (AUC), balanced accuracy, sensitivity, specificity, and kappa, and interpreted against expert reader evaluations, done according to the Berlin SIJ scoring method. Across all datasets, automated SIJ lesion detection against expert evaluations achieved performance comparable with expert interreader agreement. Structural lesions showed the strongest performance, particularly ankylosis (MEASURE 1 AUC: 0.97, SURPASS AUC: 0.99; balanced accuracy: 0.95 and 0.97). Robust results were also observed for erosions and fat lesions. BMO detection showed consistently high AUCs (0.85-0.93) with lower balanced accuracy (0.72-0.74). Model performance generalised across datasets without additional training. A fully automated DL-based approach can reliably detect active and structural SIJ MRI lesions in axSpA with robust external validation, supporting its potential use to enhance the consistency and scalability of MRI assessment in clinical trials and observational studies.