Do we need Monte Carlo treatment planning for linac based radiosurgery? A case study.

The accuracy of conventional empirical and semi-empirical dose calculation algorithms for radiation therapy treatment planning is limited. The main problem is that these algorithms fail to adequately consider the lateral transport of radiation. Most conventional algorithms use measured dose distribution data as input. These data induce an added inaccuracy to stereotactic radiosurgery dose calculations due to the difficulty of acquiring accurate dosimetric data for very small beams; however, since multiple arcs of large solid angles are usually used in stereotactic radiosurgery, the errors introduced by conventional dose algorithms are quite likely to be diluted. The use of Monte Carlo treatment planning for stereotactic radiosurgery has been investigated and described in the present paper. The OMEGA Monte Carlo code system is used as the dose engine in an in-house developed radiosurgery treatment planning system. The Monte Carlo treatment plans are done for two typical clinical cases. In one case, the collimator of 20 mm diameter is used and the lesion is located in the peripheral part of the brain. In the other case, the collimator diameter is 30 mm and the lesion is in the central part of the brain. The resultant dose distributions are compared with those calculated with a conventional dose algorithm which is based on the standard Tissue Maximum Ratio (TMR)/Off Axis Ratio (OAR) formalism. Without the inhomogeneity correction, the conventional algorithm yields accurate relative dose distributions for both cases compared with the Monte Carlo calculations. The absolute dose at the isocenter may be overestimated by the conventional algorithm by 1.5% for the first case and 2.6% for the second case; however, using the method of ratio of TMRs for inhomogeneity correction, the overestimation can be greatly reduced for both cases. The inclusion of the inhomogeneity correction into the conventional dose algorithm does not alter the relative dose distributions. Based on the clinical cases studied, it may be concluded that the conventional dose algorithm is sufficient for radiosurgery treatment planning and the Monte Carlo based radiosurgery treatment planning is unwarranted.