Impact of region of interest size and location in Gafchromic film dosimetry

Introduction: Accurate film dosimetry requires careful consideration of sources of uncertainty. Some of the sources of uncertainty are dependent on the size and location of region of interest (ROI), especially in small fields. Avoiding the penumbra is often a reason for using a small ROI. In contrast, choosing very small ROIs may increase uncertainty due to the reduction of the sampling data. Using a dose profile as flat as possible for the ROI becomes more significant in small fields in which the flatness of profile is greatly affected by the penumbra. This study aims to evaluate dependency of measured dose on size and location of region of interest (ROI) in Gafchromic EBT3 film for small field dosimetry. Materials and Methods: Gafchromic EBT3 films were irradiated using the 6MV beam from a linear accelerator at 10 cm depth (100 cm SSD) of a solid water phantom for a range of field sizes of 6×6 to 100×100 mm2. For scanning, 48 bits RGB mode with a spatial resolution of 72 dpi was used. The red channel of ImageJ software was used for evaluation. The appropriate location of ROIs were found by two methods. First, the ROI was visually placed at the center of image. Second, the profile of pixel value versus distance was plotted and the center of profile was used. Each scanned film was read using both methods and for three ROI sizes (1, 2, and 4 mm). In every scan, the film was re-positioned in the center of scanner. This uncertainty was compared with scan-to-scan variability when using piece of glass on top of the film. A plastic scintillator, Exradin W1, was used as a reference dosimeter. To obtain doses from W1 scintillator, a calibrated ion chamber with the same setup as W1 was used. Results: Comparing the three ROI sizes using both methods showed that there is less than 2% difference from reference in output factor measurements for field sizes larger or equal to 10×10 mm2. For field sizes of 15×15 and 10×10 mm2, the smaller mean percentage differences were observed in profiled-ROI (4×4 mm2) and centered-ROI (4×4 mm2). For the field sizes of 8×8 and 6×6 mm2, the profiled-ROI (2×2 mm2) had smallest mean percentage difference, which was 0.88%. The scan to scan variability was less than 0.31%. Conclusion: The ROI size of 4×4 mm2 is appropriate for dose measurements in field sizes of 100×100 mm2 to 10×10 mm2, regardless of the method of finding location of ROI. In field sizes smaller than 10×10 mm2, finding location of the ROI by profile of pixel values increases the accuracy of measurement, and ROI size of 2×2 mm2 has the smallest difference from the reference dose measurements.