z-logo
Premium
SU‐E‐T‐82: Position‐Dependent Discrete Point Spread Functions for EPID IMRT QA
Author(s) -
Nichita E,
Ostapiak O,
Farrell T
Publication year - 2011
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.3612033
Subject(s) - optics , point spread function , monte carlo method , image guided radiation therapy , physics , position (finance) , superposition principle , medical imaging , mathematics , computer science , artificial intelligence , statistics , finance , quantum mechanics , economics
Purpose: To devise a position‐dependent point spread function (PSF) to account for differences in EPID response to X‐ray scattering, glare, and photon energy compared to a water slab modeled within a treatment planning system.Method: Since the EPID image corresponds to a discrete pixel grid, the method calculates a discrete, position‐dependent, PSF based only on measurements and treatment‐planning‐system (TPS) results, without resorting to separate Monte Carlo calculations. The PSF is sought in the form of a linear combination of azimuthally‐symmetric functions (depending only on the distance between the interaction and scoring point) with position‐dependent coefficients (to account for increased off‐axis response). The PSF is calculated for a suitable test‐pattern, by minimizing the difference, in a least‐squares sense, between the acquired EPID image and the image obtained from the superposition of the PSF on the TPS result. Results: The method was tested using numerically‐simulated TPS results and EPID measurements. To provide a numerically rigorous test for the method, a 20% increase in off‐axis EPID response due to beam softening (higher than the real value) was assumed. An 8 by 6 checkerboard field was used as the test pattern to calculate the PSF which was subsequently applied to a simulated IMRT field consisting of a 20 by 20 array of 5‐mm square beamlets with random fluence values. The calculated PSF was able to reproduce the EPID image for the test pattern to within 1% (compared to 15% for a position‐invariant PSF) and the EPID image for the IMRT field to within 0.5% (compared to 16% for a position‐invariant PSF). Conclusions: A method for calculating a discrete, position‐dependent EPID PSF was developed. Preliminary tests show the method to be successful in reproducing simulated X‐ray scattering, glare, and off‐axis response of the EPID to within 1 %, which makes it promising for IMRT QA.

This content is not available in your region!

Continue researching here.

Having issues? You can contact us here