Effect of CT Number to the Dose Calculation in Heterogeneity Correction Algorithm
M Su1*, (1) Mount Sinai School of Medicine, New York, NYSU-E-T-581 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Purpose: To study the sensitivity of dose calculation to CT number in Hounsfield units (HU) in three heterogeneity dose correction algorithms.
Methods: A set of 24x16x20 cm³ ellipse cylinder phantom was created with smooth HU distribution (HU range [-810, 1000]) in Eclipse treatment planning system. Three heterogeneity dose corrected algorithms, Modified Batho (MB), Batho Power Law (BPL) and Equivalent TAR (eTAR), were applied for dose calculation with 6MV and SAD setup. The dose difference was evaluated by comparing the dose calculated at HU=0 (water) phantom with other HU value phantoms which all had same setup and fixed 100 MU delivered to each one of them for 1) three algorithms, 2) different field size, and 3) different depth. A set of double straight lines, percentage_dose_correction = slope*HU, was calculated based on the least squares error method to best fit the percentage dose correction vs. HU in HU<0 and HU>0 ranges for every algorithm and parameter date set.
Results and Conclusion: 1) A straight line fitted very well with every percentage dose correction vs. HU data set (the straight line coefficient of determination: r²>0.995). An over all average slope was -0.021 (standard deviation σ=0.0032) in HU<0 and -0.011 (σ=0.0017) in HU>0 regions. The dose correction showed more sensitive to HU in HU<0 than in HU>0 region. This was caused by the slope difference in HU to electron density conversion curve. 2) As irradiated field size increased, the sensitivity of dose correction to HU decreased in MB. 3) The dose correction sensitivity to HU was proportional to the calculated depth in MB. 4) BPL showed a least dose correction sensitivity to HU comparing with other two algorithms which was nearly 25% lower than MB and eTAB in HU<0 range, and 25% lower than MB and 14% lower than eTAB in HU>0 range.