Toward a clinical lumbar CAD: herniation diagnosis

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Spinal Diagnostics

Background:

• Lumbar disc degeneration and herniation are common conditions requiring accurate diagnosis.
• Clinical Magnetic Resonance Imaging (MRI) is crucial for visualizing spinal structures.
• Computer-aided diagnosis (CAD) systems can enhance diagnostic efficiency and accuracy.

Purpose of the Study:

• To develop and evaluate a robust, efficient, and accurate CAD system for diagnosing lumbar disc herniation from clinical MRI.
• To aid clinicians in diagnostic decision-making for lumbar spine conditions.

Main Methods:

• A Bayesian-based classifier utilizing a Gibbs distribution was implemented.
• Discs were segmented using a gradient vector flow active contour model (GVF-snake) to extract shape features.
• The classifier processed T2-SPIR weighted sagittal MRI slices, identifying herniated discs slice-by-slice.

Main Results:

• The system achieved an average diagnostic accuracy of 92.5% in cross-validation experiments.
• Experiments involved 65 clinical cases, with a random leave-out strategy for training and testing.
• The method demonstrated effectiveness in detecting herniated discs regardless of their location (central or lateral).

Conclusions:

• An automatic and robust method for diagnosing lumbar disc herniation in clinical MRI was successfully developed and validated.
• The developed system is intended for clinical practice to support reliable diagnostic decision-making.