Image Segmentation and Analysis of Flexion-Extension Radiographs of Cervical Spines

Area of Science:

• Biomedical Engineering
• Radiology
• Orthopedics

Background:

• Cervical flexion-extension radiographs are crucial for assessing spinal stability.
• Accurate analysis of subtle changes in cervical spine alignment is challenging.
• Existing methods may not fully capture dynamic changes in spinal structures and implants.

Purpose of the Study:

• To introduce a novel machine vision-based analysis tool for cervical flexion-extension radiographs.
• To quantify the local curvature changes of the spinolaminar line and anterior cervical plates.
• To investigate the dynamic deformation of cervical implants during flexion and extension.

Main Methods:

• Semiautomatic image segmentation of cervical radiographs.
• Machine vision and computerized image processing for landmark detection (spinolaminar line, cervical plates).
• Quantitative analysis of local curvature changes during dynamic flexion-extension maneuvers.

Main Results:

• The tool successfully visualizes local curvature changes in the spinolaminar line and cervical plates.
• Extension radiographs showed greater spinolaminar line curvature changes; flexion radiographs showed greater cervical plate curvature changes.
• Unexpectedly, cervical plate lordosis decreased during extension and increased during flexion.

Conclusions:

• The developed analysis tool offers a more precise interpretation of cervical flexion-extension radiographs.
• Findings suggest dynamic deformation of cervical plates is significant and warrants further investigation.
• This method may improve the diagnosis of spinal instability and pseudoarthrosis in fused spines.