Analysis of various beamlet sizes for IMRT with 6 MV photons.

Application of intensity modulated radiation therapy (IMRT) using multileaf collimation often requires the use of small beamlets to optimize the delivered radiation distribution. Small-beam dose distribution measurements were compared to dose distributions calculated using a commercial treatment planning system that models its data acquired using measurements from relatively large fields. We wanted to evaluate only the penumbra, percent depth-dose (PDD) and output model, so we avoided dose distribution features caused by rounded leaf ends and interleaf leakage by making measurements using the secondary collimators. We used a validated radiochromic film dosimetry system to measure high-resolution dose distributions of 6 MV photon beams. A commercial treatment planning system using the finite size pencil beam (FSPB) dose calculation algorithm was commissioned using measured central axis outputs from 4.0x4.0 to 40.0x40.0 cm2 beams and radiographic-film profile measurements of a 4.0x4.0 cm2 beam at twice the depth of maximum dose (dmax). Calculated dose distributions for square fields of 0.5x0.5 cm2, and 1.0x1.0 cm2, to 6.0x6.0 cm2, in 1.0x1.0 cm2, increments were compared against radiochromic film measurements taken with the film oriented parallel to the beam central axis in a water equivalent phantom. The PDD of the smaller field sizes exhibited behavior typical of small fields, namely a decrease in dmax with decreasing field size. The FSPB accurately modeled the depth-dose and central axis output for depths deeper than the nominal dmax of 1.5 cm plus 0.5 cm. The dose distribution in the build-up and penumbra regions was not accurately modeled for depths less than 2 cm, especially for the fields of 2.0x2.0 cm2 and smaller. Using the gamma function with 2 mm and 2% criteria, the dose model was shown to accurately predict the penumbra. While for single small beams the compared dose distributions passed the gamma function criteria, the clinical appropriateness of these criteria is not clear for a composite IMRT plan. Further investigation of the cumulative impact of the observed dose discrepancies is warranted. We speculate that the observed differences in the penumbra regions arise from some energy dependent artifact in the radiographic-film profiles used for commissioning. In the future, radiochromic film based commissioning might provide a more accurate data set for dose modeling.

Determination of N(pp)(gas) and N(pp)(D) for a plane-parallel chamber.

The purpose of this work is to investigate the consistency of determining N(pp)(gas) and N(pp)(D) by using three independent calibration methods from the AAPM TG 39 and IAEA TRS 381 protocols: 1) calibration with a high-energy electron beam in a phantom; 2) in-phantom calibration in a (60)Co beam; and 3) in-air calibration in a (60)Co beam. The plane-parallel chamber considered was the PTW-Markus and the comparisons were made against a calibrated PTW cylindrical Farmer-type chamber 30001. The phantom material used for the electron beam and (60)Co in-phantom methods was a solid water phantom (RW3). For the electron beam method, the nominal energies were 18 and 21 MeV. An acrylic buildup of 0.5 g/cm(2) thickness was used for the (60)Co in-air method. For each method, N(pp)(gas) and N(pp)(D) were obtained for the plane-parallel chamber as proposed by the AAPM TG 39 and IAEA TRS 381 protocols. The absorbed doses were measured along the central axis at a distance of 100 cm (SSD=100 cm) with 10 x 10 cm(2) field size at the depth of the maximum for each electron beam. The values of N(pp)(gas) by the three independent calibration methods agreed to within +/-0.6%. This meant that any of the methods would give a fairly good value. Similar results were obtained for N(pp)(D). In comparing the results for the electron beam method at energies of 18 and 21 MeV, the latter gave better agreement. The ratios of N(pp)(gas) and N(pp)(D) for the three methods were in agreement within 0.7%. The results for the absorbed dose intercomparison in the AAPM TG 39 and IAEA TRS 381 protocols showed that they agreed to within +/-0.7% which meant that any of the calibration methods and two different protocols would give an accurate result.