Monte Carlo treatment planning for stereotactic radiosurgery

Objectives: In radiosurgery treatment planning there is general acceptance that the target volume can be assumed to be homogeneous and that corrections for variations in contour are unnecessary. Thus, dose algorithms employed in radiosurgery treatment planning are quite unsophisticated; in almost every case the algorithms are the simple product of tissue-maximum and off-axis ratios and an output factor. In small photon beams, however, the lack of side scatter equilibrium compromises these assumptions. Methods: In this work we have employed Monte Carlo techniques in an attempt to obtain a more accurate representation of radiosurgical dose distributions. Specifically, the Monte Carlo system which we have devised traces the paths of primary and secondary radiation through a patient-specific anatomical representation defined by computed tomography data. In this manner the perturbation effects from external contour changes and internal tissue heterogeneities are accounted for completely. The ability to precisely mimic multi-beam multi-are stereotactic delivery has been incorporated into our Monte Carlo treatment planning interface. Results: Subsequent calculations show that substantial differences can exist when homogeneity is not assumed. Tissue heterogeneities produce a lateral broadening of the beam, resulting in a smaller volume contained within the higher isodose levels (80-90%) with a corresponding increase in the volume treated at the lower isodose levels (<50%). Conclusions: These results suggest that further investigation and refinement of radiosurgery dose algorithms is in order. (C) 1998 Elsevier Science Ireland Ltd. All rights reserved.