Experimental verification of a CT-based Monte Carlo dose-calculation method in heterogeneous phantoms

To further validate the Monte Carlo dose-calculation method [Med. Phys. 25, 867-878 (1998)] developed at the Memorial Sloan-Kettering Cancer Center, w e have performed experimental verification in various inhomogeneous phantom s. The phantom geometries included simple layered slabs, a simulated bone c olumn, a simulated missing-tissue hemisphere, and an anthropomorphic head g eometry (Alderson Rando Phantom). The densities of the inhomogeneity range from 0.14 to 1.86 g/cm(3) simulating both clinically relevant lunglike and bonelike materials. The data an reported as central axis depth doses, dose profiles, dose values at points of interest, such as points at the interfac e of two different media and in the "nasopharynx" region of the Rando head. The dosimeters used in the measurement included dosimetry film, TLD chips, and rods. The measured data were compared to that of Monte Carlo calculati ons for the same geometrical configurations. In the case of the Rando head phantom, a CT scan of the phantom was used to define the calculation geomet ry and to locate the points of interest. The agreement between the calculat ion and measurement is generally within 2.5%. This work validates the accur acy of the Monte Carlo method. While Monte Carlo, at present, is still too slow for routine treatment planning, it can be used as a benchmark against which other dose calculation methods can be compared. (C) 1999 American Ass ociation of Physicists in Medicine. [S0094-2405(99)01912-4].