Encrypted login | home

Program Information

Validation Study of the Accuracy of the Transform Method for Clinically Intuitive Quality Assurance

H Norris

H Norris*, A Thomas, M Oldham, Duke University Medical Center, Durham, NC

SU-C-213AB-6 Sunday 1:30:00 PM - 2:15:00 PM Room: 213AB

Purpose: Per-beam quality-assurance (QA) methods (e.g. portal dosimetry) have been shown to be inadequate for reporting the clinical significance of delivery errors or deviations. Recently we presented a novel 'transform' method which can solve this problem utilizing 3D dosimetry techniques. Here, we evaluate the accuracy of this method under various conditions.

Methods: A range of known mechanical and delivery errors were introduced on three stereotactic IMRT head-and-neck patient treatment plans (e.g. gantry, collimator, and couch +/-5°, Dose normalization +/-5%). Verification plans were then created where the error plans were recomputed onto an RPC head-phantom containing a 3D dosimeter. The phantom dose distribution was then transformed back to the patient CT data using the new transform method. The accuracy of the transform method was evaluated in 3D by comparison of this transformed distribution with a direct calculation from the original patient plan but incorporating the error. Comparison metrics included DVH, gEUD, and gamma.

Results: Gamma passing rates were calculated with 2% dose difference/ 2mm distance-to-agreement criteria. Comparing the transformed dose distribution estimate with the true error dose distribution, global gamma passing rates were above 97% for all plans with an average passing rate of 99.33%. The passing rates were above 99% for all organ and target structures. A gEUD was calculated for each structure, and the average percent difference was found to be less than 4%.

Conclusions: Even when presented with substantial errors, the new transform method robustly predicted the clinical significance of these errors through DVH, gamma, and other clinical metrics (e.g. gEUD). These promising verification results are an important step toward validating a new approach to QA which enables the clinical significance of QA data to be evaluated.

Contact Email