On-line optimization of intraoperative electron beam radiotherapy of the breast.

Abstract PURPOSE: To optimize the dose delivery to the breast lumpectomy target treated with intraoperative electron beam radiotherapy (IOERT). MATERIALS AND METHODS: Two tools have been developed in our MU calculation software NEMO X to improve the dose homogeneity and the in-vivo dosimetry effectiveness for IOERT treatments. Given the target (tumor bed) thickness measured by the surgeon, NEMO X can provide auto dose normalization to cover 95% of the target volume with 95% of the prescription dose (PD) and a "best guess" of the expected dosimeter dose (EDD) for a deep seated in-vivo dosimeter. The tools have been validated with the data of 91 patients treated with IOERT on a LIAC mobile accelerator. In-vivo dosimetry has been performed with microMOSFETs positioned on the shielding disk inserted between the tumor bed and the chest wall. RESULTS: On average the auto normalization showed to provide better results if compared to conventional normalization rules in terms of mean target dose (|MTD-PD|/PD⩽5% in 95% vs. 53% of pts) and V107 percentage (〈V107〉=19% vs. 32%). In-vivo dosimetry MOSFET dose (MD) showed a better correlation with the EDD guessed by our tool than just by assuming that EDD=PD (|MD-EDD|/EDD⩽5% in 57 vs. 26% of pts). CONCLUSIONS: NEMO X provides two useful tools for the on-line optimization of the dose delivery in IOERT. This optimization can help to reduce unnecessary large over-dosage regions and allows introducing reliable action levels for in-vivo dosimetry.