SU-D-141-01: Quantification of Automatic Couch Correction Accuracy with An Optical Tracking System

Abstract

Purpose: To quantify the accuracy of automatic couch corrections using an in-house optical tracking system (OTS). Methods: First, system calibration accuracy was assessed with a ball-bearing phantom (0.01 mm resolution). An infra-red marker (IRM) was mounted on the phantom and shifted by known amounts (up to ±1cm) along all 3 directions. OTS observed shifts were compared to known shifts to evaluate system calibration accuracy. System stability was assessed by observing fluctuations of a stationary IRM for a 10-minute period. Second, accuracy of automatic couch correction was retrospectively analyzed on 4 head and neck (HN) patients (51 fractions) and 4 pelvis patients (94 fractions). To prevent patient setup error, an IRM was permanently mounted on the immobilization devices (aquaplastic face mask or body mold) of each patient. The OTS continuously tracked IRM throughout the treatment to validate setup accuracy. Residual errors of automatic couch correction were calculated as the difference between CBCT dictated shifts and OTS observed actual shifts after automatic couch adjustment. Results: The OTS calibration accuracy was within 0.18 mm along each direction. System stability was within 0.12 mm. For couch adjustments in patient treatment, the mean and maximum residual errors in any direction were found to be 0.20 mm and 2.3 mm for HN and 0.03 mm and 1.5 mm for pelvis cases, respectively. 21 cases (14%) had a residual couch correction error over 1.5 mm, and 3 cases (2%) had a residual error over 2 mm. Conclusion: Overall, automatic couch corrections are reliable and accurate. However, errors of over 2 mm have been found in our study. This suggests the importance of daily QA of automatic couch correction. An accurate OTS with real-time tracking ability should be considered in hypo-fractionated treatments to prevent accidental couch correction errors.