The use of a compensator library to reduce dose inhomogeneity in tangential radiotherapy of the breast.

Background And Purpose: The dose variation throughout the volume of the breast from tangential fields can exceed 20% for large breasts. This is postulated to result in poor cosmesis [Radiother Oncol 16 (1989) 253], particularly at the inframammary fold, where the dose is highest. Compensators may be used to reduce this variation, but at the cost of the time to manufacture each unique compensator for the individual patients. This paper outlines the implementation and routine use of a library of reusable compensators.

Materials And Methods: For the period of December 1999 to May 2001, 94 patients attending for breast radiotherapy received treatment using breast compensators calculated from multiple outlines measured using the Osiris system. The compensators manufactured for the early patients were added to a library for possible reuse by later patients. Of the 94 patients, 28 patients' compensators formed the library and 66 subsequent patients have been treated using compensators derived from the library. Selection of the most appropriate library compensator was determined from the analysis of the distribution of the calculated dose-volume histogram for the whole breast, excluding lung, penumbra and build-up regions. Once the library was complete, approximately 50% of all subsequent breast patients were treated with compensators (46% from the library and 4% with individual compensators). This represented a usage rate of 92% for the library compensators for those patients requiring compensation.

Results: In all cases the compensators reduced the variation in the dose distribution. For example, the group treated with a library compensator demonstrated a mean reduction from 29 to 9% for the volume of breast tissue receiving more than 5% greater than the reference dose. If the same patients had been treated using their own individual compensators, the corresponding value would have been 7%. There is a small systematic, but negligible, difference in the two populations of dose variation for individual versus library compensators, but this difference (P=0.20) did not reach the level of statistical significance of P=0.05).

Conclusion: The method of creation and selection of library compensators has proved to be simple and reliable in practice. Every patient receiving radiotherapy for breast cancer is currently investigated under full software control to ascertain whether the use of a library compensator would be advantageous.