Room: AAPM ePoster Library
Purpose: To implement a simple approach for accurate density modeling of the treatment couch in the treatment planning system.
Methods: TThe SunNuclear Stereophan and the A26 chamber was used to measure couch attenuation with 6MV and 15MV at three indexing positions (~H4, ~H1, ~F1) along the VARIAN PerfectPitch 6DoF IGRT couch. The Stereophan was setup isocentrically, and the gantry was set to various angles where radiation passes through the couch. Data collected with couch attenuation was subtracted from that without (the Stereophan can extend off the couch-end). This measured data was then compared to Eclipse predicted values. A single Hounsfield unit, for the Eclipse-modeled medium couch, was retroactively determined to minimize the differences between measurements and Eclipse predictions.
Results: 1) the percent difference in attenuation, with and without couch, is energy dependent. 2) Variation in couch thickness exists with respect to lateral position from the center line; this is largely ignored for the modeled couch in Eclipse. 3) If only a single HU and a single preset thickness in Eclipse is used to produce a best-fit density for the entire couch, Eclipse prediction will be slightly under-attenuated at all couch positions except the thickest part (~F1). Using the medium couch and the Varian default -300HU (physical density ~0.723g/cc), the percent attenuation difference is at most 52% from predictions, and 77% of data are within 20%. At -200HU (physical density ~0.831g/cc), the percent attenuation difference is at most 32% from predictions, and 92% of data are within 20%. 4) The Varian default -300HU is a suboptimal value for the Eclipse-modeled thick couch at the F1 couch position, typical for pelvis treatments.
Conclusion: The Stereophan creates a measurement setup, completely independent of machine and measurement device parameters (i.e. machine output, phantom modeling in Eclipse), which can reliably determine treatment couch modeling parameters.