Performance of patient specific margins derived using a Bayesian statistical method

Abstract

In this paper a published Bayesian method is used retrospectively to estimate patient specific errors and margins for 470 prostate cancer patients after 5 fractions. The obtained margins show a small spread, with significantly larger margins for a small number of patients. These patient specific margins were compared with generic margins. Specific margins improve dose coverage but the difference with generic margins is small. The amount of patients that receive more than 90% dose to the CTV increases with 1%, while 2% of patients show 5% or more improvement in dose. We conclude that the Bayesian method is capable of generating correct patient specific margins but that the gain is small. © 2009 Springer-Verlag.