Geometric distortion in computed tomography scans for radiotherapy: Quantification and quality assurance implications

Area of Science:

• Medical Imaging Physics
• Radiotherapy Physics
• Image Quality Assessment

Background:

• Computed tomography (CT) is often assumed to be geometrically accurate and used as a reference for other modalities like MRI.
• However, the geometric accuracy of CT itself has not been thoroughly investigated.

Purpose of the Study:

• To characterize CT geometric distortion using a phantom.
• To establish a reference scan for quantifying geometric distortion in CT for radiotherapy planning.

Main Methods:

• A CIRS Model 603A Distortion Phantom was used to quantify geometric distortion.
• Evaluated system reproducibility, phantom setup, scan parameters, and temporal variations.
• Compared CT geometric distortion to linac-based cone-beam CT (CBCT) and a Computer Aided Design (CAD) model.

Main Results:

• All eight CT scanners analyzed showed measurable geometric distortion.
• Mean distortion ranged from 0.13-0.52 mm (vs. CBCT) and 0.25-0.82 mm (vs. CAD).
• Longitudinal couch inaccuracy or sag likely contribute to CT geometric distortion.

Conclusions:

• CT scans for radiotherapy planning exhibit measurable geometric distortion.
• Current CT quality assurance may miss submillimeter distortions crucial for stereotactic radiotherapy.
• Quantifying CT geometric distortion is vital for risk assessment and margin design in high-precision radiotherapy.