AI automated grid placement in the OMERACT knee inflammation MRI scoring system (KIMRISS) for bone marrow lesion assessment: A multi-reader exercise.
Authors
Affiliations (12)
Affiliations (12)
- Department of Radiology & Diagnostic Imaging, University of Alberta, Edmonton, AB, Canada. Electronic address: [email protected].
- Department of Radiology & Diagnostic Imaging, University of Alberta, Edmonton, AB, Canada.
- Canadian Institutes of Health Research, Institute of Musculoskeletal Health and Arthritis, Toronto, ON, Canada.
- Department of Family Medicine, University of Alberta, Edmonton, AB, Canada.
- CARE Arthritis Ltd., Edmonton, AB, Canada.
- Department of Radiology, University of Calgary, Calgary, AB, Canada.
- Department of Internal Medicine, Rheumatology Institut de Recherche Expérimentale et Clinique, Cliniques Universitaires Saint-Luc, Université Catholique de Louvain, Rheumatology, Brussels, Belgium.
- Medical Centre Zenit, Department of Rheumatology, Bleicheplatz 3, Schaffhausen, Switzerland.
- Qscan Radiology Clinics, Gold Coast, QLD, Australia.
- Department of Radiology & Diagnostic Imaging, University of Alberta, Edmonton, AB, Canada; Medical Imaging Consultants, Edmonton, AB, Canada.
- Department of Radiology & Diagnostic Imaging, University of Alberta, Edmonton, AB, Canada; CARE Arthritis Ltd., Edmonton, AB, Canada; Medical Imaging Consultants, Edmonton, AB, Canada.
- CARE Arthritis Ltd., Edmonton, AB, Canada; Department of Medicine, University of Alberta, Edmonton, AB, Canada.
Abstract
To validate the reliability and feasibility of AI-automated grid placement for bone marrow lesion (BML) scoring in the Knee Inflammation MRI Scoring System (KIMRISS) using the OMERACT Filter. Eleven experts evaluated 40 MRI cases using manual and automated grid placement. Grids were compared both directly using spatial similarity metrics and indirectly using agreement metrics calculated on resulting KIMRISS BML scores. Feasibility was assessed using the System Usability Scale (SUS). In most regions, automatically- and manually-placed grids demonstrated strong spatial similarity (e.g., mean femur Dice Coefficient = 0.78) and KIMRISS BML score agreement (mean intraclass correlation coefficients of 0.86 and 0.89 for baseline and change scores, respectively). SUS scores for automated grid placement were moderate (mean = 66.1). Automated grid placement is a reliable and feasible improvement to KIMRISS that could improve the ease and reproducibility of quantifying osteoarthritis in clinical trials.