Encrypted login | home

Program Information

Use of a Commercial TPS for Deriving Exit Dose Distribution for Patient Specific QA with An EPID

no image available
P Haering

P Haering*, C Lang , M Splinter , DKFZ, Heidelberg, Germany

Presentations

SU-E-T-773 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall


Purpose:
A method to derive exit dose distributions with a commercial treatment planning system for comparison to Epid measured doses.

Methods:
In contrast to published methods, we present an approach that uses a treatment planning system (Raystation, Raysearch) to calculate exit dose patterns based on a modified patient CT dataset. The EPID is assumed as water equivalent and therefore is represented as water filled ring in the images. This is done by: -Export a renamed copy of the patient CT and plan data set. –Manipulate the CT-data to a field of view of 1000 mm. – Insert the water filled cylinder ring, centered to the target point. – Manipulate plan file with the new target and dose prescription point. – Import and recalculate dose on the manipulated data set. – Export and extract exit field dose matrix from the cylinder (optional entrance dose). – Calibrate matrix to size and rebin data from the cylinder to EPID equivalent data. – Correct Epid measurement for scatter, position and field size. -Evaluate and compare data in Verisoft (PTW). Data manipulation and extraction is done by a simple tool (IDL, Exelis).

Results:
This method was tested on a Siemens Artiste 6MV for field sizes up to 27x27cm2 limited by used geometry and EPID size. First phantom measurements show good results for fields up to 20x20cm (pass rate > 95% for 3%, 3mm Gamma index) while larger fields have higher discrepancies towards the field edges. This might result from off axis softening of the beam and the higher sensitivity of the detector to beam scatter.

Conclusion:
This method might simplify the use of exit dosimetry with the EPID for patient specific QA as it uses the dose calculation of a commercial treatment planning system. The concept was proven by phantom data sets, giving acceptable results.


Contact Email: