Program Information
Independent and Direct Absolute Dosimetric Validation for Portal Dose Prediction Algorithm Used in Portal Dosimetry
T Li*, W Byant , L Fountain , E Duffy , Palmetto Health Tuomey, Sumter, SC
Presentations
PO-BPC-Exhibit Hall-12 (Saturday, March 18, 2017) Room: Exhibit Hall
Purpose: To propose a method as to independently validate the portal dose prediction algorithm by direct absolute dosimetric measurement.
Methods: An aSi1000 electronic portal imaging devices (EPID) on TrueBeam linear accelerator (Varian Medical Systems, Palo Alto, CA) and portal dose prediction (PDP) algorithm in Eclipse treatment planning system (TPS, v11.0) is commissioned for portal dosimetry. Five plans are created using analytic anisotropic algorithm (AAA) using sliding window technique with 6 MV photons. The beams are painted with uniform fluence across field, with field size ranging from 3x3, 5x5 10x10, and 15x15 to 20x20. Each beam delivers 100.0 cGy central axis dose to Dmax (1.5 cm) with 150 cm source to axis distance (SAD) setup. All five plans are delivered on PTW729 (PTW Inc., Freiburg, Germany) with 150 SAD using TrueBeam. These five plans are then used by PDP algorithm to predict dose for EPID at 150 SAD. The predicted dose are compared with PTW729 measurements. The same setup 150 SAD is used throughout this study so that the direct comparison is possible.
Results: The comparison of PTW729 measurement with PDP algorithm shows good agreement (about 3.0%) for field sizes larger than 5x5, for filed sizes smaller than 5x5, the PDP algorithm underestimates the central axis dose by about 5.0%. The cause of PDP algorithm dose underestimation for small field is currently under investigation.
Conclusion: Our method uses identical setup (150 cm SAD) for AAA calculation, PDP algorithm and PTW729 measurement. This makes direct comparison possible. The PDP algorithm predicted dose and PTW729 measured dose shows good agreement for field size larger than 5x5. As PTW729 measurement is not dependent on TPS or EPID, this method can be used to independently validate PDP algorithm within predetermined clinical limits.
Contact Email: