Assessment of Random Error in Phantom Dosimetry with the Use of Error Simulation in Statistical Software

Area of Science:

• Medical Physics
• Radiation Dosimetry
• Computational Modeling

Background:

• Phantom dosimetry is crucial for radiation therapy quality assurance.
• Quantifying random error (RE) is essential for reliable dosimetry results.
• Existing methods for RE quantification can be complex and time-consuming.

Purpose of the Study:

• To assess the practicality of software simulation for quantifying random error in phantom dosimetry.
• To determine if software modeling can generate credible confidence intervals for effective dose (E).

Main Methods:

• Applied software error simulation to an existing dosimetry study.
• Incorporated study specifications, measurement values, and effective dose calculation algorithms into R software.
• Modeled four sources of RE: calibration factor, background radiation correction, dosimeter readout, and X-ray generator fluctuation.

Main Results:

• Calculated RE amounts using experimental values and error propagation rules.
• Performed 10,000 simulations of effective dose (E) calculations with applied RE.
• Generated a distribution of E values, yielding a confidence interval around the expected value.

Conclusions:

• Software modeling provides credible confidence intervals for effective dose in phantom dosimetry studies.
• These confidence intervals allow for substantiation or rejection of statistical significance between protocols.
• The modeling software is valuable for power analysis in planning future phantom dose experiments.