Diagnostic Value of Spinal Cord Lesion Metabolism on [ 18 F ] FDG PET / CT for the Classification of Myelopathy
Abstract
Introduction [ 18 F]FDG PET/CT scan is sometimes performed as part of the diagnostic evaluation of non‐compressive myelopathies, but the spinal cord lesion metabolism is not systematically evaluated. We aimed to investigate the diagnostic value of spinal cord lesion metabolism on [ 18 F]FDG PET/CT for differentiating inflammatory myelitis from neurosarcoidosis and tumoral lesions. Methods We conducted a retrospective cohort study of patients presenting with spinal cord lesions who underwent [ 18 F]FDG PET/CT. Patients were classified as primary inflammatory myelitis, neurosarcoidosis, tumoral lesions, or other etiologies. PET/CT examinations were reviewed by a nuclear medicine physician, masked to final diagnosis, who assessed the presence of hypermetabolism (qualitative analysis) and measured lesion SUV max , lesion‐to‐liver ratio, and level‐standardized Z ‐score to account for physiological uptake variability (quantitative analysis). Results A total of 106 patients were included: 60 (56.6%) with inflammatory myelitis, 20 (18.9%) with neurosarcoidosis, 10 (9.4%) with tumoral lesions, and 16 (15.1%) with other etiologies. Hypermetabolism was found in 80.0% of tumoral lesions, 50.0% of neurosarcoidosis, and 23.3% of inflammatory myelitis. PET/CT showed good performance in differentiating inflammatory myelitis or neurosarcoidosis from tumoral lesions using qualitative analysis (hypermetabolism: sensitivity 0.77, specificity 0.80, and 0.50, respectively) and quantitative analysis (SUV max : AUC 0.89 and 0.84, respectively). The performance of SUV max for distinguishing inflammatory myelitis from neurosarcoidosis was low (AUC 0.55), but improved with the lesion‐to‐liver ratio (AUC 0.64) and level‐standardized Z ‐score (AUC 0.65). Conclusion Spinal cord hypermetabolism on [ 18 F]FDG PET/CT demonstrated good performance for distinguishing tumoral lesions from inflammatory or neurosarcoidosis, but lower performance for differentiating inflammatory myelitis from neurosarcoidosis.
Related
- MaxCAT-Net: Hybrid Transformer-Based Skin Lesion Classification and Lightweight Segmentation with Calibration and Cross-Dataset Validation
- Case Report: Paralytic rabies masquerading as traumatic spinal cord injury: a diagnostic pitfall in trauma surgery
- Inflammatory biomarkers in pyogenic spinal infection: evidence, limitations, and integrated diagnostic perspectives
- Diagnostic, Therapeutic, and Prognostic Implications of Carpal Tunnel Syndrome and Spinal Stenosis in Wild-type ATTR-Cardiomyopathy
- Identification of Mitochondrial-Lipid Metabolism-Related Diagnostic Targets in Osteoarthritis: Integrative Multi-omics Data and Experimental Validation
BibTeX
@article{gavoille2026diagnostic,
title = {{Diagnostic Value of Spinal Cord Lesion Metabolism on [ 18 F ] FDG PET / CT for the Classification of Myelopathy}},
author = {Antoine Gavoille and Anthime Flaus and Matthieu Dietz and Julie Pique and Antoine Seyve and Pierre Lardeux and Romain Marignier and Marc Janier and Pascal Sève},
journal = {European Journal of Neurology},
year = {2026},
month = jul,
doi = {10.1111/ene.70667},
url = {https://hal.science/hal-05751753v1},
note = {CC BY-NC 4.0},
}
RIS
TY - JOUR
TI - Diagnostic Value of Spinal Cord Lesion Metabolism on [ 18 F ] FDG PET / CT for the Classification of Myelopathy
AU - Antoine Gavoille
AU - Anthime Flaus
AU - Matthieu Dietz
AU - Julie Pique
AU - Antoine Seyve
AU - Pierre Lardeux
AU - Romain Marignier
AU - Marc Janier
AU - Pascal Sève
PY - 2026
DA - 2026/07/14
JO - European Journal of Neurology
DO - 10.1111/ene.70667
UR - https://hal.science/hal-05751753v1
AB - Introduction [ 18 F]FDG PET/CT scan is sometimes performed as part of the diagnostic evaluation of non‐compressive myelopathies, but the spinal cord lesion metabolism is not systematically evaluated. We aimed to investigate the diagnostic value of spinal cord lesion metabolism on [ 18 F]FDG PET/CT for differentiating inflammatory myelitis from neurosarcoidosis and tumoral lesions. Methods We conducted a retrospective cohort study of patients presenting with spinal cord lesions who underwent [ 18 F]FDG PET/CT. Patients were classified as primary inflammatory myelitis, neurosarcoidosis, tumoral lesions, or other etiologies. PET/CT examinations were reviewed by a nuclear medicine physician, masked to final diagnosis, who assessed the presence of hypermetabolism (qualitative analysis) and measured lesion SUV max , lesion‐to‐liver ratio, and level‐standardized Z ‐score to account for physiological uptake variability (quantitative analysis). Results A total of 106 patients were included: 60 (56.6%) with inflammatory myelitis, 20 (18.9%) with neurosarcoidosis, 10 (9.4%) with tumoral lesions, and 16 (15.1%) with other etiologies. Hypermetabolism was found in 80.0% of tumoral lesions, 50.0% of neurosarcoidosis, and 23.3% of inflammatory myelitis. PET/CT showed good performance in differentiating inflammatory myelitis or neurosarcoidosis from tumoral lesions using qualitative analysis (hypermetabolism: sensitivity 0.77, specificity 0.80, and 0.50, respectively) and quantitative analysis (SUV max : AUC 0.89 and 0.84, respectively). The performance of SUV max for distinguishing inflammatory myelitis from neurosarcoidosis was low (AUC 0.55), but improved with the lesion‐to‐liver ratio (AUC 0.64) and level‐standardized Z ‐score (AUC 0.65). Conclusion Spinal cord hypermetabolism on [ 18 F]FDG PET/CT demonstrated good performance for distinguishing tumoral lesions from inflammatory or neurosarcoidosis, but lower performance for differentiating inflammatory myelitis from neurosarcoidosis.
SN - 1351-5101
SN - 1468-1331
ER -