ABSTRACT
BACKGROUND: In radiosurgery, the convention has been to prescribe radiation dose to a "covering isodose volume". This is presumed to be the minimum dose received by the entire tumor. Our purpose was to review our practice, assess different means of specifying a prescription isodose, and test how reliable they were in the face of various calculation methodologies. METHODS: Different minimum doses were calculated using three calculation methods for 20 brain targets and compared to the original physician intent. Monte Carlo (MC) calculations were run down to 2% (1 sigma) statistics. Voxel size depended on the imaging field of view and the calculation engine. For pencil beam convolution ("finite size pencil beam" - FSPB) calculations with or without heterogeneity correction (Methods 1-2), the calculation grid matched the CT scan resolution. For MC calculations (Method 3), the highest available in-plane resolution was 256 x 256 pixels. The median voxel volume was thus 0.58 mm(3) (0.47 to 0.95 mm(3)) for FSPB and 2.3 mm(3) (1.87 to 3.80) for MC. RESULTS: The absolute minimum target dose varied substantially between the three calculation methods - up to 25% difference between Methods 1 and 3. The differences were reduced when comparing near-minimum doses with absolute minimal volumes ΔV, DPTV-ΔV. The median difference in the isodose covering the PTV-0.03 cm(3) was 0% for Methods 1 and 2 and 3.6% for Methods 1 and 3. The median difference in the isodose covering the PTV-0.01 cm(3) was 0% for Methods 1 and 2 and 2.2% for Methods 1 and 3. In our data, the smaller the volume in which the minimum dose is calculated, the more sensitive this calculation was to dose calculation parameters. The standard deviation of the difference between physician intent and the isodose covering the PTV-0.01 cm(3) was 2.9% (range from -3.3% to 9.3%). CONCLUSION: In radiosurgery, absolute minimum doses are sensitive to changes in dose calculation grids and dose calculation algorithms. Based on our experience, standardizing dose prescription to the isodose volume covering the PTV-0.01 cm(3) or the PTV-0.03 cm(3) would have little impact on clinical practice and would be relatively insensitive to dose calculation parameters.
