Real-time beam monitoring for error detection in IMRT plans and impact on dose-volume histograms

This study aimed to test the sensitivity of a transmission detector for online dose monitoring of intensity-modulated radiation therapy (IMRT) for detecting small delivery errors. Furthermore, the correlation of changes in detector output induced by small delivery errors with other metrics commonly employed to quantify the deviations between calculated and delivered dose distributions was investigated. Transmission detector measurements were performed at three institutions. Seven types of errors were induced in nine clinical step-and-shoot (S&S) IMRT plans by modifying the number of monitor units (MU) and introducing small deviations in leaf positions. Signal reproducibility was investigated for short- and long-term stability. Calculated dose distributions were compared in terms of gamma passing rates and dose-volume histogram (DVH) metrics (e.g., D-mean, D-x%, V-x%). The correlation between detector signal variations, gamma passing rates, and DVH parameters was investigated. Both short- and long-term reproducibility was within 1%. Dose variations down to 1 MU (a dagger signal 1.1 +/- 0.4%) as well as changes in field size and positions down to 1 mm (a dagger signal 2.6 +/- 1.0%) were detected, thus indicating high error-detection sensitivity. A moderate correlation of detector signal was observed with gamma passing rates (R-2 = 0.57-0.70), while a good correlation was observed with DVH metrics (R-2 = 0.75-0.98). The detector is capable of detecting small delivery errors in MU and leaf positions, and is thus a highly sensitive dose monitoring device for S&S IMRT for clinical practice. The results of this study indicate a good correlation of detector signal with DVH metrics; therefore, clinical action levels can be defined based on the presented data.