@misc{oai:repo.qst.go.jp:00065474,
 author = {安藤, 興一 and et.al and 安藤 興一},
 month = {Jun},
 note = {Carbon-ion radiotherapy started Year 1994 in Chiba, Japan and since then more than 8000 patients were　treated by 2014. Bragg peaks of mono-peak carbon-ion beams are spread out and provide various LET　along with a beam path. Physical dose within the Spread-Out Bragg peak (SOBP) are designed to produce　10% cell kill at any position within a SOBP so that biological effectiveness would be homogeneous. The　design is based on in vitro cell kill of a human tumor cell line HSG. There are two questions remain to be experimentally examined; one question is whether the SOBP with inclined and therefore heterogeneous physical dose distribution produce homogeneous cell kill also for other cells and tissues. Another question
is whether homogeneous cell kill could be obtained after doses producing larger than 10 % cell kill, as the RBE of high LET beams depends on cell kill level. We here employed early skin reaction in mice as an endpoint to answer the above stated questions. For each dose group, 5 mice were immobilized on a PMMA plate to place right hind feet in a rectangular field of dimensions 28 x 100 mm, and received single and fractionated radiation doses. The foot reaction was initiated on the 8th post-irradiation day and continued daily until the 35th post-irradiation day. The scores in this period were averaged in each animal and the mean was calculated for each radiation dose group. Skin reaction caused by gamma-ray irradiation is detectable after single doses of 25 Gy or larger in C3H mice legs, and shows dose responses good enough to
discriminate between low and high LET (linear energy transfer) radiation. Isoeffect doses after various LET of mono-peak carbon-ion beams were obtained and used to calculate alpha and beta terms. Fe-plot, i.e., 
regression between dose per fraction and reciprocal total dose, showed fairly good linearity through all doses while the regression ran off linearity when LET increased up to 80 keV/　m. As the run off was primarily
caused by single doses, we used doses of all fractionated irradiation but not single doses to obtain alphaand
beta terms from these Fe-plots. Surprisingly, the depth dose distribution of the new ridge filter thus
designed and produced was almost identical to that of the ridge filter previously designed based on the in
vitro HSG cell kill data. Using the new ridge filter, we conduced fractionated irradiations with various LETs of carbon ions. Similar to mono-peak beams, fractionation-dependent increase of isoeffect doses was observed more clearly for lower LETs than higher LET carbon ions. The Fe-plots again ran off linearity for high LET carbon beams. Alpha and beta values for each Fe-plot were calculated by fitting against saturation-corrected lineal energy (y*) values of all fractionated doses, excluding the single doses. RBE for each LET of the SOBP was calculated using thus obtained alpha and beta. Biological dose distribution of the new ridge filter was calculated by multiplying physical doses and RBE, and was almost homogeneous within the SOBP.
Why biological dose distribution for skin reaction is homogeneous within the SOBP even though design of which is based on in vitro cell kill? Analyzing relation between survival parameters (i.e., alpha and beta) and
LET up to 100 keV/　m for either HSG cell kill or skin reaction, we found that the regression lines of alpha
values linearly increased and had slopes similar to each other, i.e., 0.023 or 0.026 　m/ keV. However, beta values showed least dependency on LET so that the slope of regression was -0.027 or 0.01 　m/ keV for HSG or skin. The difference of LET dependency between alpha and beta values was also observed for V79 cells and hypoxic HSG cells in vitro. RBEs for HSG and skin increased linearly with LET with slopes smaller than alpha values. The similar increase of alpha thus observed indicates that LET affects single track events, irrespective of cell or tissue. Taken together that beta is rather independent of LET, our observation supports microdosimetric kinetics model (MKM) being used in the ongoing carbon-ion radiotherapy at NIRS.
\nLineal energy in the formula is a coefficient of D but not that of D2. Alpha expected by MKM increases with an increase of lineal energy. Suppose that target size and beta were independent of cell types, alpha would linearly increase with an increase of lineal energy. The coefficient of D at y*=0 becomes 　0, which corresponds to intrinsic radiosensitivity of individual cells and tissues.
Single doses in the Fe-plots deviated from linearity for high LET carbon ions but not for gamma rays. As Fe-plot simply employs a formula
\nthe observed deviation is puzzling. One possible explanation for the deviation is any change of radiosensitivity induced by the first dose during fractionation. We previously reported that isoeffect doses of 20 keV/　m carbon ions to cause a given skin reaction in mice abruptly increased after 2 fractions of 5.2
Gy each were given to legs before the following test doses. The abrupt increase was not observed for gamma rays. In clinical trials of non-small cell lung cancer conducted during past 20 years, 1-day treatment required larger total doses than those expected from protocols that employed 4 through 18
fractions by using Linear-Quadratic model. These observations suggest that radiosensitivity of tissues may change after high LET radiation, possibly due to adaptive response or stimulated repair of sub-lethal damage.
In addition to our own data stated above, we shortly introduce 2 reports recently presented from NIRS. Dr.
Yoshitaka Matsumoto et al. have investigated effects of carbon ions metastasis-related biological phenomena in a murine melanoma cell line. Using Boyden chamber assays, they show that RBE values of carbon ions for tumor cell migration and invasion as well are larger than that for cell kill. Invasion decreases with an increase of doses above 2 Gy, but rather increases after doses below 1 Gy of photon.
Dr. Mayumi Fujita et al. have compared 32 tumor cell lines for their invasiveness, and have found that nitric oxide closely relates to tumor invasion after 2 Gy of carbon-ion irradiation.
We conclude that it is not necessary to design different SOBPs for individual tissues in carbon-ion radiotherapy as far as homogeneous biological effectiveness within a SOBP is concerned, and that biology will further advance high LET radiotherapy in future., PTCOG 53},
 title = {Recent Progress in Biology of Heavy ions},
 year = {2014}
}