Purpose: It is well established that using image guidance for prostate motion allows reduction of margin, dose escalation, decreased toxicity and recently improved outcomes. However, current methods only account for translational motion, not rotational variations. The purpose of this study is to assess whether rotations in anatomy lead to significant changes in the delivered dose for prostate patients. Methods: Under an IRB approved protocol, 11 consecutive patients underwent prostate IMRT using IGRT with implanted metal‐oxide semiconductor field‐effect transistors (MOSFETs); the Dose Verification System (DVS) manufactured by Sicel Technologies. Two dosimeters were implanted per patient. From conebeam CT (CBCT) registration, corrections were applied to all translational errors. For rotations larger than 3 degrees, patient were repositioned and realigned to attempt to correct the rotation. Both translational and rotational errors based on the CBCT were documented. The daily DVS readings were compared to CBCT rotations about each axis (pitch, roll and yaw) and the root‐mean square (RMS) rotation. Results: 372 CBCT images were acquired. The correlation between rotation and DVS measurement was analyzed using linear regression. The R2 value for pitch was 0.059 and 0.144 for each dosimeter, respectively. For roll, the R2 values were 0.049 and 0.001. For yaw, the values were <0.001. For the RMS rotation, R2 was 0.034 and 0.038. As it could confound results, the angular dependence of the dosimeters was measured during commissioning and found that it was approximately 0.5% for 5 degree rotations. Conclusions: We did not find any significant correlation between prostate rotation around any axis and discrepancy in DVS reading. These results show that rotations seen clinically do not have a substantial effect on the dose delivered to the prostate. Further studies will attempt to determine at what angle rotations begin to affect the dose distribution, if at all.
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging