An improved method for evaluating LINAC isocenter

Area of Science:

• Medical Physics
• Radiation Oncology
• Quality Assurance

Background:

• Traditional medical linear accelerator (LINAC) isocenter quality assurance relies on the Winston-Lutz (WL) test, a scalar metric with limited insight into beam-to-target accuracy during gantry and couch motion.
• Existing methods do not fully account for geometric errors, potentially overlooking critical inaccuracies in treatment delivery.

Purpose of the Study:

• To develop a more sensitive method for verifying LINAC isocenter accuracy.
• To account for geometric errors typically overlooked by conventional WL techniques.
• To improve prediction of beam-to-target accuracy across various gantry and couch positions.

Main Methods:

• Constructed a 3D coordinate system from WL images to determine radiation beam axis and marker shift locations.
• Developed the M-matrix, a novel performance metric predicting beam-to-target errors based on gantry and couch positions.
• Introduced 'clinical isocenter' as an optimized target for tumor positioning, derived from the M-matrix cost function.
• Demonstrated the technique on a clinical LINAC, comparing pre- and post-repositioning marker shifts, radiation isocenter, and M-matrix.
• Validated the M-matrix against traditional WL techniques using Monte Carlo simulations with variations in LINAC geometry, marker position, and measurement noise.

Main Results:

• Successfully demonstrated the M-matrix technique on a Varian LINAC, with repositioning to the clinical isocenter resulting in an error matrix magnitude below 0.81 mm.
• Marker position had minimal impact on radiation and clinical isocenter locations and radius.
• Verification confirmed the M-matrix's accuracy in predicting geometric inaccuracies between the radiation beam and tumor.
• Monte Carlo simulations showed the M-matrix is more sensitive and specific than traditional WL techniques for detecting potential treatment errors.

Conclusions:

• Developed and validated a framework for LINAC isocenter verification using a 3D model derived from WL images.
• The M-matrix replaces the scalar isocenter radius, offering insights into couch and beam contributions to isocenter quality and revealing errors missed by traditional methods.
• The clinical isocenter offers an alternative to physical isocenter for tumor positioning, particularly in cases of suboptimal LINAC geometry.