Online dose delivery verification in small animal image-guided radiotherapy

To accomplish novel radiation treatment techniques in preclinical radiation research, small animal image-guided radiotherapy systems have been increasingly integrated into preclinical radiation research over the last decade. Although such systems have sophisticated tools (such as cone-beam computed tomography-based image guidance, robotic couch, treatment planning system (TPS), and electronic portal imaging devices [EPIDs]). To our knowledge, no established technique can perform independent and online verification of the delivered dose during radiotherapy. In this study, we implement an online EPID dosimetry for each administered SA-IGRT fraction in a rat orthotopic model of prostate cancer.

To verify the accuracy of delivered dose to rat, we compared the two-dimensional (2D) calculated dose distribution by TPS as the planned dose, with online dose distribution estimated using an EPID as the delivered dose. Since image acquisition software was not capable of acquiring integrated images over a long period of time, we used the EPID to estimate dose rate rather than dose. The central axis (CAX) dose rate values at the beam's exit surface were compared. In addition, 2D dose distributions were also compared under different gamma criteria. To verify the accuracy of our EPID dosimetry technique, we measured transit and exit doses with film during animal treatment. In this study, 20-mm cone was used to collimate beam. We previously observed that the EPID response was independent of collimator size for collimator size ≥15-mm, we did not apply for additional correction factor.

Comparison of exit CAX dose rate values of TPS-calculated and EPID-estimated showed that the average difference was 3.1%, with a maximum of 9.3%. Results of gamma analysis for 2D comparison indicated an average of 90% passing rate with global gamma criterion of 2 mm/5%. We observed that TPS could not calculate dose accurately in peripheral regions in which the penumbra effect was dominant. Dose rate values estimated by EPID were within 2.1% agreement with film at both the imager plane and the beam's exit surface for 4 randomly selected animals for which film measurement verification was performed.

The small animal radiation research platform (SARRP) system's built-in EPID was utilized to estimate dose delivered to rats at kilovoltage energy x-rays. The results of this study suggest that the EPID is an invaluable tool for verifying delivered dose to small animal to help validate conclusions made from preclinical radiation research.