On measuring depth-dose distribution of range-modulated proton therapy fields

Depth-dose profile measurements are frequently performed for the characterization of spread-out Bragg peak treatment fields in proton therapy. The measured distribution often contains a significant amount of noise with a persistent pattern. As a result, substantial smoothing has to be applied which can alter a measurement. We explored the origin of the observed noise by closely examining the sampling process in the scanning algorithm. We will show that the time characteristics of the signal in the range-modulated proton therapy beams differ significantly from those in the photon beams, and as a result the measurement error is very sensitive to the sampling duration. The observed noise results mainly from inappropriate choices for this value, rather than from the dose distribution itself. Increasing the value arbitrarily may actually increase the noise magnitude. We will demonstrate that with an optimal value for the sampling duration and its accurate control, the noise magnitude can be reduced significantly, without increasing the measurement time.