Overview Statistic: PDF-Downloads (blue) and Frontdoor-Views (gray)

Beyond Benchmarks: Towards Robust Artificial Intelligence Bone Segmentation in Socio-Technical Systems

  • Despite the advances in automated medical image segmentation, AI models still underperform in various clinical settings, challenging real-world integration. In this multicenter evaluation, we analyzed 20 state-of-the-art mandibular segmentation models across 19,218 segmentations of 1,000 clinically resampled CT/CBCT scans. We show that segmentation accuracy varies by up to 25% depending on socio-technical factors such as voxel size, bone orientation, and patient conditions such as osteosynthesis or pathology. Higher sharpness, isotropic smaller voxels, and neutral orientation significantly improved results, while metallic osteosynthesis and anatomical complexity led to significant degradation. Our findings challenge the common view of AI models as “plug-and-play” tools and suggest evidence-based optimization recommendations for both clinicians and developers. This will in turn boost the integration of AI segmentation tools in routine healthcare.

Export metadata

Additional Services

Search Google Scholar Statistics - number of accesses to the document
Metadaten
Author:Kunpeng XieORCiD, Lennart Johannes GruberORCiD, Martin Crampen, Yao LiORCiD, André FerreiraORCiD, Elias TappeinerORCiD, Maxime GillotORCiD, Jan Schepers, Jiangchang XuORCiD, Tobias PankertORCiD, Michel BeyerORCiD, Negar ShahamiriORCiD, Reinier ten BrinkORCiD, Gauthier DotORCiD, Charlotte Weschke, Niels van NistelrooijORCiD, Pieter-Jan VerhelstORCiD, Yan Guo, Zhibin XuORCiD, Jonas Bienzeisler, Ashkan RashadORCiD, Tabea FlüggeORCiD, Ross Cotton, Shankeeth VinayahalingamORCiD, Robert IlesanORCiD, Stefan Raith, Dennis Madsen, Constantin Seibold, Tong Xi, Stefaan Bergé, Sven NebelungORCiD, Oldřich KodymORCiD, Osku SundqvistORCiD, Florian ThieringerORCiD, Hans LameckerORCiD, Antoine Coppens, Thomas PotrusilORCiD, Joep KraeimaORCiD, Max Witjes, Guomin Wu, Xiaojun ChenORCiD, Adriaan LambrechtsORCiD, Lucia H Soares CevidanesORCiD, Stefan ZachowORCiD, Alexander HermansORCiD, Daniel TruhnORCiD, Victor AlvesORCiD, Jan EggerORCiD, Rainer RöhrigORCiD, Frank Hölzle, Behrus PuladiORCiD
Document Type:Article
Parent Title (English):Expert Systems With Applications
Year of first publication:2025
URL:https://www.medrxiv.org/content/10.1101/2025.06.11.25329022v1
DOI:https://doi.org/10.1016/j.eswa.2025.130031
Accept ✔
Diese Webseite verwendet technisch erforderliche Session-Cookies. Durch die weitere Nutzung der Webseite stimmen Sie diesem zu. Unsere Datenschutzerklärung finden Sie hier.