A protocol for the commissioning and quality assurance of new planning computers

Any new radiotherapy planning system needs to be thoroughly tested. Besides checking the accuracy of the algorithm by comparing plans done on the system with measurements done in a phantom, it is desirable for the user to compare the new equipment with a tried and tested system before it is used clinically. To test our recently purchased planning systems, a protocol was developed for running a comparison between these and our existing planning computer, an IGE RTPLAN. A summary of the test protocol that was developed is as follows: (1) A series of plans is created on the old system, to include at least one plan of each common type. The series includes at least one plan with a bone inhomogeneity, and one with an air or lung inhomogeneity, and these plans are computed both with and without inhomogeneity correction. Point dose calculations are made for a number of positions on each plan, including the dose at the centre of the treatment volume. (2) Each of these plans is reproduced as accurately as possible on the new system using the original CT data and patient outlines. (3) The old and new plans, including those with and without inhomogeneity correction are overlaid and compared using the following criteria: (a) how well the volumes of interest coincide, (b) how accurately the positions of the points of interest are reproduced, (c) the doses at the points of interest, (d) the distances between the isodoses defining the dose plateau, (e) the maximum displacement between the corresponding pairs of isodoses in the dose gradient around the tumour. The protocol has been used to test two systems: the (newly developed) Siemens Axiom and the Helax TMS (running on a DEC Alpha). A summary of the results obtained will be presented. These were sufficient to show up several minor problems, particularly in the Axiom system