SU-E-T-526: Evaluation of Dose Mapping Errors via Use of a Volume-Based Dose Mapping Method

To quantify dose mapping errors (DMEs) of a point-based dose mapping method for 4D lung treatment plans.

Point-based dose mapping methods utilize deformation vector fields (DVFS) to interpolate dose from a deformed image. Volume-based dose mapping methods consider the volume overlap between deformed and reference voxels; defining dose as the integral energy divided by the integral mass of the voxel, and conserving integral dose . DME is defined as the dose differences between volume-based and point-based mapped dose (DME=(DpointBased-DvolumeBased)/DRx). The DME for a 4D lung case is compared with a bitmap DME method, both using a Pinnacle research version 8.1y DVF. DME is computed for ten 4D lung cases (five 10 phases, five 3 phase) with Pinnacle research version 9.100 DVFs. Multi-phase accumulated 4D DMEs are also evaluated.

For all cases, the largest DMEs are located in the dose/density gradient regions. With Pinnacle 8.1y DVF, mapping dose from phase 9 to phase 0, results in a DME=-0.2%±6.1% (range of -76%∼112%). The same case with Pinnacle 9.100 DVFs, DME=0.3%±4.8%(-41%∼32%). Locations of large DME are consistent with those from the bitmap method. For the ten 4D lung cases, accumulated mean DME are within ±0.07% (std. deviations: 1∼5%, range -102%∼64%). Maximum tumor DMEs are less than 30cGy (DRx=7200cGy) for all patients.

Due to its inherent integral dose conservation, volume-based dose mapping methods can quantify errors in point-based dose mapping methods. While mean DME values are small for the cases tested, standard deviations near 5% indicate that a substantial number of voxels have ∼5% dose mapping errors, however these dose errors do not occur in the target structures. Work supported by NIH P01CA116602.