Improving the calibration of radiochromic films by the use of uncertainties in optical density and dose.

Purpose: To compare how the inclusion of the uncertainties in dose and optical density affects the results of the calibration of Gafchromic(®) EBT3 radiochromic films.

Methods: Five methods of least square minimization were compared for calibration of radiochromic films. These differed in the way in which the uncertainties of dose and optical density are taken into account. The comparison was made by simulating the calibration fit at an increasing number of points measured in the calibration table and in the gamma analysis of ten real IMRT verifications.

Results: According to the results of the simulations, the methods that take into account all the variability in dose and optical density are superior to those that do not use them. The improvements are reflected in lower bias of the parameters, lower variability and faster convergence when the number of the points increases in the calibration table. When these methods are employed in the gamma analysis comparison, the statistical significance of the results depends on how restrictive the parameters are that define this comparison.

Conclusions: Methods that take into account all the variability both in dose and in optical density should be used in the calibration process.