EUVIMEDEUVIMED entdecken
Crossref

Artificial Intelligence-based Detection and Reporting of Degenerative Lumbar Spine Pathology on MRI

M. P. Talabard, I. Bouaddi, M. Durteste, R. Naidji, N. E. Regnard, D. Vittecoq

Seminars in Musculoskeletal Radiology · 2026-06 · 30 (S 01) · S1-S22

DOI: 10.1055/s-0046-1824367 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Seminars in Musculoskeletal Radiology, online erschienen am 02.06.2026. Verfasst von M. P. Talabard und 5 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose or Learning Objective: To assess the performance of an artificial intelligence (AI) system (LumbarMR, Gleamer, Paris, France) for the automated detection and reporting of degenerative pathologies on lumbar spine magnetic resonance imaging (MRI), and to compare the efficiency of a radiologist when unaided versus aided by AI. Methods or Background: This multicenter retrospective study included consecutive adult patients referred for lumbar MRI due to axial low back pain, lumbar radiculopathy, or compressive neuropathy. The MRI protocols used were comprised sagittal T1- and T2-weighted sequences and axial T2-weighted sequences. Ground truth was established by two expert radiologists who evaluated degeneration at each intervertebral level (T8–T9 to L5–S1), including Modic changes, disk degeneration, disk herniation, disk bulging, facet arthropathy, and central and foraminal stenosis. An independent radiologist annotated the same examinations while recording interpretation time. The primary unit of analysis was the observation, defined as the intersection of patient, intervertebral level, and pathology. All observations were recoded again as binary outcomes for the presence or absence of pathology. Accuracy, sensitivity, and specificity were computed for the AI system and for the radiologist. Interpretation times and performance metrics were compared between AI-unaided and aided conditions using Wilcoxon signed- rank and McNemars tests, respectively. Results or Findings: A total of 199 patients (mean age: 47.2 ±– 14.5 years; 116 women) and 9,918 analyzable observations were included. The AI achieved an overall accuracy of 92.9%. It showed strong diagnostic performance for disk degeneration (sensitivity, 0.82; specificity, 0.94; accuracy, 0.91), Modic changes (0.87, 0.96, 0.96), facet arthropathy (0.85, 0.92, 0.90), and central canal stenosis (0.94, 0.97, 0.97). Lower sensitivities were observed for disk herniation (0.71), disk bulging (0.70), left foraminal stenosis (0.66), and right foraminal stenosis (0.64), although specificity remained consistently > above 0.93 for all conditions. The radiologists interpretation time was reduced by 17% per examination ( P < 0.001); while sensitivity remained unchanged between AI-unaided and aided conditions ( P = 0.87). Conclusion: The AI system demonstrated high accuracy and excellent specificity for detecting lumbar spine pathologies. Moreover, AI assistance reduced interpretation time without affecting radiologist sensitivity. These findings support the use of AI as an adjunct in spine MRI interpretation, possibly which may contributing to faster and more standardized reporting of degenerative disease. Publication History Article published online: 02 June 2026 © 2026. Thieme. All rights reserved. Thieme Medical Publishers, Inc. 333 Seventh Avenue, 18th Floor, New York, NY 10001, USA

Zitieren

APA

Talabard, M. P., Bouaddi, I., Durteste, M., Naidji, R., Regnard, N. E., & Vittecoq, D. (2026). Artificial Intelligence-based Detection and Reporting of Degenerative Lumbar Spine Pathology on MRI. Seminars in Musculoskeletal Radiology, 30(S 01), S1-S22. https://doi.org/10.1055/s-0046-1824367

Vancouver

Talabard MP, Bouaddi I, Durteste M, Naidji R, Regnard NE, Vittecoq D. Artificial Intelligence-based Detection and Reporting of Degenerative Lumbar Spine Pathology on MRI. Seminars in Musculoskeletal Radiology. 2026;30(S 01):S1-S22. doi:10.1055/s-0046-1824367

BibTeX
@article{Talabard2026,
  author = {Talabard, M. P. and Bouaddi, I. and Durteste, M. and Naidji, R. and Regnard, N. E. and Vittecoq, D.},
  title = {Artificial Intelligence-based Detection and Reporting of Degenerative Lumbar Spine Pathology on MRI},
  journal = {Seminars in Musculoskeletal Radiology},
  year = {2026},
  volume = {30},
  number = {S 01},
  pages = {S1--S22},
  doi = {10.1055/s-0046-1824367}
}
Kostenlos starten →