RESUMEN
PURPOSE: Measurement of the modulation transfer function (MTF) is performed by evaluating the response of an imaging system to a predefined input. To obtain accurate results when using an edge phantom, the detector input signal must resemble an ideal step function. The MTF of megavoltage (MV) imagers used in radiotherapy has been measured with highly absorbing edge phantoms fabricated from thick metal blocks. This study investigates the influence of the edge phantom design on the accuracy of the resulting MTF. METHODS: The MTF of an electronic portal imaging device (EPID) was measured at 6 MV beam quality with four edge phantoms made of lead with 1.3, 3.3, 5.0, and 10.0 cm thickness. Monte Carlo simulations were carried out for these and a selection of tungsten phantoms to determine the photon fluence at the imaging plane and quantify the systematic error in the MTF introduced by the edge phantom design. RESULTS: The measured MTF depends on the design of the edge phantom. The detector input signal of a thin phantom is affected by secondary radiation from the phantom itself, causing an overestimation of the MTF. The amount of secondary radiation can be reduced by increasing the phantom thickness or introducing an air gap between the phantom and the detector. Both methods introduce geometric unsharpness, which can result in an underestimation of the true MTF. Edge phantoms made from 4.0 cm thick tungsten or 5.0 cm thick lead induce comparatively small systematic errors of below 3% or 5%, respectively. CONCLUSIONS: When MTF measurements are conducted at MV energies, even a highly absorbing edge phantom will introduce a systematic error of several percent. Direct comparison of MTFs obtained with different edge phantoms should therefore be treated with caution.