PURPOSE: To develop and evaluate a correction strategy for prostate rotation using gantry and collimator angle adjustments.
METHODS AND MATERIALS: Gantry and collimator angle adjustments were used to correct for prostate rotation without rotating the table. A formula to partially correct for left-right (LR) rotations was derived through geometric analysis of rotation-induced clinical target volume (CTV) beam's-eye-view shape changes. For 10 prostate patients, intensity-modulated radiotherapy (IMRT) plans with different margins were created. Simulating CTV LR rotation and correcting each beam by a collimator rotation, the corrected CTV dose was compared with the original and uncorrected dose. Effects of residual geometric uncertainties were assessed using a Monte Carlo technique. A large number of treatments representative for prostate patients were simulated. Dose probability histograms of the minimum CTV dose (D min) were derived, with and without online correction, resulting in a more realistic margin estimate.
RESULTS: Dosimetric analysis of all IMRT plans showed that, with rotational correction and a 2-mm margin, D min was constant to within 3% for LR rotations up to +/-15 degrees . The Monte Carlo dose probability histograms showed that, with correction, a margin of 4 mm ensured that 90% of patients received a D min >or=95% of the prescribed dose. Without correction a margin of 6 mm was required.
CONCLUSIONS: We developed and tested a practical method for (online) correction of prostate rotation, allowing safe and straightforward implementation of margin reduction and dose escalation.