A new model for "accept or reject" strategies in off-line and on-line megavoltage treatment evaluation

Purpose: With megavoltage on-line imaging, a temporal record of the treatment setup variation is made available during the course of radiotherapy. This work describes the theoretical framework where the imaging data is incorporated as feedback for deriving off-line and on-line "accept or reject" decisions to improve the quality of radiotherapy on an individual patient basis. Methods and Materials: A dynamic model of treatment setup variation is developed to describe: (a) the variations of multifraction treatment position and (b) the variation of measurement. Based on this model, a confidence region of subsequent treatment position is estimated from measurements of previous treatment position. The confidence region is compared with an allowance region of the treatment position, representing the acceptable margin of setup variation. The difference of the two regions are then evaluated to derive off-line and on-line "accept or reject" treatment decisions. Results: The model was simulated retrospectively using sequential daily images from different treatment sites. "Accept or reject" decisions were produced for on-line and off-line applications. The results also demonstrate the ability of the model to include the effects of setup variation that drifted during the course of treatment. Conclusion: The preliminary result demonstrates that "accept or reject" treatment decisions can be derived quantitatively from on-line imaging data. The model shows potential to guide the physician and therapist in implementing efficient and accurate treatment adjustments. Further work is required to test the model in the clinical setting.