The Effect of Low-Dose CT Protocols on Shoulder Model-Based Tracking accuracy Using Biplane Videoradiography

Area of Science:

• Orthopedics
• Medical Imaging
• Biomechanics

Background:

• Model-based tracking is crucial for quantifying shoulder kinematics.
• Computed tomography (CT) is commonly used for 3D bone reconstruction, but increases radiation exposure.
• Lower-dose CT protocols may be feasible due to bone-soft tissue contrast.

Purpose of the Study:

• To evaluate the dose-accuracy trade-off between low-dose CT scans and kinematic tracking accuracy.
• To assess the impact of reduced CT dose on the tracking of humerus, scapula, and clavicle.
• To determine the feasibility of using intensity-based registration algorithms with low-dose CT for shoulder motion analysis.

Main Methods:

• Investigated full-dose and four low-dose CT protocols on three fresh-frozen cadavers.
• Modulated X-ray tube current and peak voltage for low-dose protocols.
• Compared model-based tracking results against radiostereometric analysis (reference standard).
• Quantified accuracy using root-mean-square error (RMSE), bias, precision, and worst-case errors for segmental and joint kinematics.

Main Results:

• Achieved average CT dose reductions of 70.6-92.8%.
• Observed slight increases in RMSE for glenohumeral and acromioclavicular joints with decreased dose.
• Found minimal differences (≤0.3°, ≤0.1 mm) in joint kinematic errors between highest and lowest dose protocols.

Conclusions:

• Substantial CT dose reduction is achievable for shoulder motion analysis.
• Low-dose CT protocols do not significantly compromise data fidelity for kinematic tracking.
• Model-based tracking with reduced CT dose offers a viable alternative for shoulder motion analysis.