Department of Radiation Oncology, Kimmel Cancer Center, Jefferson Medical College, Thomas Jefferson University Hospital, Philadelphia, PA 19107

Materials and Methods:
Dose inhomogeneity for treatment plans generated by inverse techniques often exceeds the 5 to 10% encountered in traditional treatment planning. This report investigates the hypothesis that forward planning can produce similar dose distributions when normal constraints on dose homogeneity are lifted. A second hypothesis is that the forward approach results in intensity maps that are simplified relative to inverse planning. The forward planning technique described here (implemented on the CMS/FOCUSä planning system) starts with conformal fields at nine equally spaced, co-planar gantry angles. For these same gantry angles, additional fields are added that only partially irradiate the target so that invaginations are created within the dose distributions produced by the conformal fields. The width and depth of the desired invagination determine the size and relative weight of the added fields. Simple rules for determining these parameters are discussed. In order to improve the overall dose homogeneity, missing tissue compensators are used for all fields. Instead of using any physical beam modifiers, compensator thicknesses generated with the FOCUS system are converted to intensities and combined with the added field segments to produce a final intensity map for each gantry angle. These final intensity distributions are compared to those produced for the same patient with the NOMOS/CORVUSä inverse-planning system with the optimization restricted to use the nine gantry angles. This comparison is accomplished by segmenting the final intensity maps and counting the total number of sub-fields as a measure of the complexity of the distributions. Additionally, the amount of the dose that is delivered as an open conformal field is determined. Dose-Volume Histograms (DVHs) for the PTV and all critical structures are also compared for the two planning techniques. The comparison was carried out for a total of 5 cases for patients with head and neck lesions that surround the spinal cord. In each case, the optimization included a boost volume.