The Effect of Treatment Plan Complexity On Optically-Stimulated Luminescent Dosimetry
A Kuruvilla MS*, A Rea MS, A Riegel PhD, A Kapur PhD, A Jamshidi PhD, L Potters MD, Long Island Jewish Medical Center, New Hyde Park, NYSU-E-T-215 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
The purpose of this study was to characterize uncertainties in radiation dose to optically-stimulated luminescent dosimeters (OSLDs) for varying levels of field modulation.
We considered 7 conformal treatment plans (25 fields), 9 enhanced dynamic wedge plans (21 fields), and 4 IMRT plans (28 fields). Five OSLDs arranged in a plus shape on 9 cm of solid water buildup were covered with 5 mm bolus and scanned using a large bore CT scanner. The images were imported into the treatment planning system. Each field of each plan was projected onto the phantom surface at gantry angle 0°, the field isocenter was placed at 10 cm depth, and dose was calculated using collapsed cone convolution and a 4 mm dose grid. The expected point dose to each OSLD was measured in the phantom plan. The OSLD geometry was replicated on a linear accelerator, and all 3D conformal, wedged, and IMRT fields were shot using 6 MV photons. All exposed OSLDs for each field were read approximately 18 hours later and measured doses were compared with planned doses.
For open, wedged, and IMRT fields, measured doses were 2.3% ± 4.1% higher, 0.9% ± 6.1% higher, and 2.6% ± 14.0% higher than planned doses respectively. The standard deviation was greatest for IMRT plans and lowest for 3D conformal plans. A few IMRT fields demonstrated substantial differences between planned and measured doses for individual OSLDs (over 60%). After further analysis, these large discrepancies occurred when leaf edges fell on or near the dosimeter position.
Uncertainties in dose were much higher with IMRT fields than 3D and wedged fields. Increasing modulation increases dose uncertainty. Further research will investigate the effect of dose grid resolution and calculation model, specifically the advantages of Monte Carlo calculations near the skin surface.