311 / 1224
Type B uncertainty … scoring
➢
CPE can be assumed and only photons need be simulated
Photon-only simulation introduces errors >2% @ distances at or below 1.6, 3, and
7mm for Ir-192, Cs-137, and Co-60 sources, respectively (Perez-Calatayud et al.
2012) but this is source specific and clinically irrelevant in most cases.
➢
Scoring cell dimensions are important! See Taylor et al. (2007) for details and
Rivard et al. (2004) and Perez-Calatayud et al. (2012) for recommendations
➢
For analogue MC increasing voxel size will reduce type A but increase volume
averaging
➢
For track length estimators reducing voxel thickness while preserving surface
will reduce volume averaging without affecting type A