A New Deep Learning Algorithm for Detecting Spinal Metastases on Computed Tomography Images

This study introduces a new computer program based on deep learning to help doctors find bone cancer that has spread to the spine. The program looks at CT scans to spot damage to the bones. The AI model was tested and showed it could find bone metastases with a sensitivity of 0.78, precision of 0.68, and F1-score of 0.72 per slice, and 0.75, 0.36, and 0.48 per lesion. The AI model's comparable sensitivity to experts and improved the diagnostic accuracy of residents highlights its potential as a valuable tool in clinical practice.

• 1
  The AI model achieved a sensitivity of 0.78, precision of 0.68, and F1-score of 0.72 per slice in detecting osteolytic bone metastases.
• 2
  The observer studies indicated that the AI model had comparable sensitivity to orthopedic or radiology experts.
• 3
  The AI model improved the sensitivity and F1-score of residents in detecting osteolytic bone metastases.