Room: Exhibit Hall | Forum 7
Purpose: Patient-specific quality assurance (QA) measurements for VMAT SRS are challenging and labor intensive, particularly as target size decreases. An accurate independent 3-dimensional dose calculation has the potential to replace patient-specific measurements. We evaluated a commercial system based on the superposition-convolution algorithm.
Methods: 75 measurements using radiochromic film (EBT-XD, Ashland Chemical, Covington, KY) in an acrylic phantom were selected from our VMAT SRS patient-specific QA database. The plans used a 10 MV flattening filter free beam and a multileaf collimator having 2.5 mm leaf width. The measurements were for 51 plans, of which 24 had multiple targets. There were two measurements for each multiple target plan, corresponding to the largest and smallest target. The plans were recalculated using Mobius3D v2.1 (Mobius Medical Systems, Houston, TX) and four sets of parameters: 1 mm and 2 mm grid size, and 1 mm grid size with dosimetric leaf gap offset Â±0.5 mm from the default value (DLGÂ±0.5). The calculation-to-film dose ratio and gamma with criteria 1%/1mm were calculated.
Results: The median target size was 15.2 mm (range 2.3 to 39.2). For targets larger (smaller) than the median size, the mean calculation-to-film ratio was 1.01 (0.91), 0.99 (0.83), 1.04 (0.99), and 0.98 (0.82) for 1mm, 2mm, DLG+0.5, and DLG -0.5, respectively, compared to 0.98 (0.98) for the treatment planning system. The mean fraction of pixels having gamma>1 was 13.3% (30.1%), 16.0% (46.4%), 15.6% (45.9%), and 15.2% (24.7%) for 1mm, 2mm, DLG+0.5, and DLG-0.5, respectively, compared to 3.6% (3.0%) for the treatment planning system.
Conclusion: Better agreement was obtained for 1 mm grid size; however, the calculation was not sufficiently accurate to replace patient-specific measurements, particularly for smaller targets. Adjustment of the dosimetric leaf gap did not improve the agreement. Improvement in small field modeling is needed to improve the accuracy for small targets.