We evaluated a commercial three-dimensional (3D) electron beam treatment pl
anning system (CADPLAN V.2.7.9) using both experimentally measured and Mont
e Carlo calculated dose distributions to compare with those predicted by CA
DPLAN calculations. Tests were carried out at various field sizes and elect
ron beam energies from 6 to 20 MeV. For a homogeneous water phantom the agr
eement between measured and CADPLAN calculated dose distributions is very g
ood except at the phantom surface. CADPLAN is able to predict hot and cold
spots caused by a simple 3D inhomogeneity but unable to predict dose distri
butions for a more complex geometry where CADPLAN underestimates dose chang
es caused by inhomogeneity. We discussed possible causes for the inaccuracy
in the CADPLAN dose calculations. In addition, we have tested CADPLAN trea
tment monitor unit and electron cut-out factor calculations and found that
CADPLAN predictions generally agree with manual calculations. (C) 1999 Amer
ican Association of Physicists in Medicine. [S0094-2405(99)00412-5].