Unencrypted login | home

Program Information

Modulation Severity as a Predictor for Patient-Specific IMRT QA Delivery Failure

R Price

R Price*, I Veltchev, C Ma, Fox Chase Cancer Center, Philadelphia, PA

WE-G-108-9 Wednesday 4:30PM - 6:00PM Room: 108

Purpose: he goal of this work was to determine if highly modulated fields contribute to decreased passing rates for patient-specific IMRT delivery QA.

Methods: Data for 10 Varian Eclipse-planned patients in each of 3 categories were collected: IMRT, carriage shift, and RapidArc. Total plan modulation was calculated as the ratio of total MU to dose per fraction. Applied modulation, in cGy, was determined for each field as the ratio of total MU per field to total plan modulation. The difference between applied modulation and dose from the planning system on a per beam basis was calculated and the mean values found for each patient and each category. Limits were applied based on the category mean plus 3 standard deviations for 2 beams as well as maximum limits for a single beam. Data were evaluated with respect to these limits for 889 patients. The individual results were compared with the IMRT delivery QA data collected with the IBA Matrixx using absolute 3% dose and 3mm DTA acceptance criteria.

Results: Modulation limits were exceeded for IMRT plans in 64%, 44% and 22% of cases and for RapidArc plans in 43%, 22% and 2% of cases for delivery QA passing rates of 92%, 95% and 98%, respectively. The mean passing rate for delivery QA was approximately 97% (range 90.4-100%).

Conclusion: There is a definite trend indicating lower delivery QA passing rates with severe modulation for both modalities. However, given an arbitrary acceptable passing rate of 95% with the modulation test failing 44% and 22% of the time for IMRT and RapidArc, respectively, modulation may not be a strong predictor of delivery QA acceptability. The cause(s) of severe modulation should be evaluated on a case-by-case basis with appropriate time and attention allotted for all phases of the QA process.

Contact Email: