Clinical Application of Deep Learning for Spine MRI Interpretation: A Multicenter Evaluation of Artificial-Intelligence-Assisted versus Manual Reading on Diagnostic Agreement with the Reference Standa
4 hours ago
- #Spine diagnostics
- #AI in healthcare
- #MRI interpretation
- Lumbar spine diseases significantly affect quality of life, requiring accurate diagnostic tools.
- Lumbar VNet Pro (LVP) is the first real-time AI-assisted system integrated with MRI hardware for lumbar spine analysis.
- LVP was trained on 2,453 MRI datasets and validated internally and externally across multiple centers.
- LVP showed high performance in localization (Dice = 0.93), segmentation (Dice = 0.92), labeling (identification rate = 0.90), and timeliness (average inference time = 1.1 s).
- Internal testing with 100 patients showed LVP recognition accuracy at 100% and 97% consistency with manual assessment.
- External testing with 1,522 patients compared LVP to manual and human-machine-assisted methods, showing better performance for AI-assisted approaches.
- AI-assisted methods performed better for pathologies like lumbar disc herniation, spinal canal stenosis, and lateral recess stenosis (AUC >0.95 vs. >0.90 for manual).
- Real-time integration of LVP with MRI improved positioning accuracy and reduced interobserver variability.
- Further studies are needed to assess LVP's generalizability across diverse clinical settings.