@misc{oai:repo.qst.go.jp:00071760,
 author = {Miki, Kentaro and M, Zenklusen Silvan and Mori, Shinichiro and 三木 健太朗 and 森 慎一郎},
 month = {Jun},
 note = {Purpose:
Using the carbon ion beam for range measurements instead of photons would avoid the uncertainty involved in the conversion of electron density to stopping powers, one of the largest uncertainties present in charged particle therapy. While the Bragg peak of a mono-energetic carbon beam that is stopping in water is steep, and can be easily analyzed, the depth dose distribution of a beam passing through inhomogeneous tissue is generally degraded as reported by Urie et al 1986. For this study, simulations were used to calculate the dose distribution of a thin carbon ion pencil beam passing through inhomogeneous tissue in order to see if it is possible to obtain an accurate range measurement by using carbon ions.
Material & Methods
To simulate the degraded depth dose profiles we calculated the dose distribution of a thin pencil beam passing through a patient liver CT and a virtual CT thorax phantom created with XCAT. We calculated the initial average range of the degraded depth dose profile by folding the 2D Gaussian spot profile with the initial pencil beam depth dose curve and the position dependent range shifts along the beam path. 
The depth dose profile was then unfolded by a simulated annealing algorithm in order to check if the initial range spectrum could be reproduced. The average range <Ru> was calculated from the unfolded spectra and compared to the average range of the initial spectra to test the validity of the simulating annealing algorithm.
Results
The initial range <Ri> was in average (over the whole thorax area) about 3 mm shorter than the actual range RCT calculated directly from the CT image. For larger beam sizes the standard deviation of this difference grew from ɛ = 1mm (pencil beam size, σB = 1 mm) to ɛ = 4mm (σB = 5 mm), see Figure 1a. The largest differences, up to a few cm in range are observed in regions of high gradient in the tissue densities, for example at the intersection of lung and chest wall or at the edge of bones, as shown on Figure 1b for a beam size of σB = 3 mm.
The difference between <Ri> and <Ru> was observed to be in average 0.5 mm ± 0.15 mm, Figure 1a. This value did almost not show any dependency on the initial beam size and was almost always between ± 2 mm, reflecting the very good fitting of the simulated annealing algorithm.
Discussion & Conclusions
The calculation showed us that within a region of relative homogeneity (well within the lung or the liver) the described method of averaged range works well and it could be used to measure the range at a single point within the target directly by using carbon ions. This would give us valuable information at very low cost about the range within the tumor., 53rd Annual Conference of the Particle Therapy Co-Operative Group (PTCOG53)},
 title = {Feasibility study of using a carbon ion pencil beam for pinpoint range measurements in charged particle therapy},
 year = {2014}
}