|Comment to Draft for Consultation on
2005 Recommendations of the ICRP
M. S. Sasaki, Kyoto University
I congratulate the Commission for the successful and continued great efforts in evolving the ICRP recommendation. I am confident that the new recommendation serves as a platform of the setting of radiation protection policy in the 21st century. However, I still find an ambiguity in the setting of radiation weighting factors addressed in paragraph 3.4.1. The subsequent comments are focused on this point.
(I cannot attach the figures here. I would appreciate if you could let me know your mailing address so that I can send the complete text with figures and equations)
1. Choice of reference radiation
RBE-LET relationship constitutes the basis of radiation weighting factors. The RBE has been estimated as effectiveness of various endpoints relative to those for X- or ƒÁ-rays as reference radiations. Current derivation of RBE has assigned RBE=1 to all photon radiations. However, evidence has been accumulated strengthening that conventional X-rays are 2-3 times more effective than ƒÁ-rays (ICRP 92). Microdosimetric analysis is also in favor of the difference (Chen, Int J Rad Biol 80:577, 2004). Therefore, it may be timely that the new ICRP recommendation adopts a less complicated single reference radiation, i.e., ƒÁ-rays. Reasons are also due to that most animal experiments use ƒÁ-rays as reference radiation, cancer risk in A-bomb survivors is expressed in ƒÁ-ray equivalent dose, and sufficient accuracy and avoidance of underestimation of risk are the principles of the ICRP recommendation.
2. Choice of endpoint
The endpoints have been chromosome aberrations, mutations, transformation, cell killing or life shortening. However, the dose-response kinetics in the low-dose range is usually poorly defined for the last four endpoints. Cell killing is also affected by the apoptotic cell death, which shows a weak LET dependency. The inclusion of all those endpoints could be a major source of variability in estimating RBEM. Chromosome aberrations are sensitive parameters for stochastic effects, show strict dose-response relationship down to several cGy and provide the RBEM with sufficient accuracy. Currently, considerable number of experiments using various radiation quality has been accumulated for human cells.
3. Modeling and formalism of radiation weighting factors
Data in literature on dicentric chromosome aberrations in human lymphocytes were reviewed, and the dose-responses were re-evaluated by fitting to a unified linear-quadratic model in a function of dose (D), Y=k(ƒÌ+D)D, where ƒÌ is a cross-over dose, ƒÌ=ƒ¿/ƒÀ and k is constant, k=ƒÀ, which is set a priori to k=5.6∙10^2/Gy^2. The ƒ¿-terms were thus obtained for photons, protons and ƒ¿-particles, and used for the calculation of LET-dependent quality factor, Q(L) (see Appendix A). The Q(L) function was then applied to the derivation of quality factor, Qn, effective quality factor, qE, and radiation weighting factor, wR, of neutrons (see Appendix B).
4. General features
Q(L) can be expressed in a simple function of LET. Q(L) against 220 kVp X-rays is similar to those of ICRP 60. The use of ƒÁ-rays as reference radiation almost doubles the Q(L) value as expected by ICRP 92. A major difference is the long tailing in the high LET range in ICRP 60, which may be relevant to the inclusion of high Z elements. The inactivation cross-section is not affected by atomic number at LET below 100 keV/ƒÊm, but differs among ions with different atomic number and hence Q(L) (Kraft and Kraft-Weyrather, GSI Report 87-11, 1983). The exact mechanism is not established. However, HZE has a large penumbra and could deposit substantial amount of energy to the cell nucleus even by a glancing hit while occasional direct hit results in an efficient cell killing. This questions the validity of a unified Q(L) function over heavy particles.
Quality factor, Qn, effective quality factor, qE, and radiation weighting factor, wR, of neutrons against ƒÁ-rays are qualitatively similar to those of Schuhmacher and Siebert, Rad Prot Dosim 40:85, 1992, Leuthold et al., Rad Prot Dosim 40: 77, 1992 and those presented in the ICRP 60 and ICRP 92 except for their magnitude. Here, the qE was obtained after modification by relative contribution of dose from charged particles, DhL, as adopted in ICRP 92, and converted to wR for spectral neutrons. These two quantities are essential for the risk evaluation; Qn is a special case of qE, such as in small organisms or in the calculation at a critical site independently of the captured ƒÁ-rays, where DhL=1.
In the ICRP 60, wR is expressed in quotient of ambient dose equivalent, q*, in 10 mm depth of ICRU sphere. This causes some problems in the energy range of thermal and epithermal neutrons, which are strongly attenuated and hence cannot be reflected to the ambient dose. The thermal and moderate energy neutrons are produced during the slowing-down process, which eventually results in less change in the overall energy spectrum as compared to that of incident neutrons. The quantity of qE is more appropriate for practical estimation of wR.
Thermal neutrons show considerably high RBE; about 52 % of the dose is given by high LET components in the tissue equivalent material. In view of their significant contribution to the quality factors, all relevant quantities should be presented at energy range down to 10^-8 MeV. Here, Figs. B-1 and B-2 are the line drawing by semi-spline interpolation, but may be expressed in a simple function like Fermi-Lorenz type functions for practical application (e.g., Wagner et al., Rad.Prot Dosim. 12:231, 1985).
Appendix A: Formalism of LET dependence of Q(L) of light charged particles
1. Reaction kinetics of dicentric aberrations
Dicentric yield is expressed by the following linear-quadratic function of dose, D.
Y=ƒ¿D+ƒÀD^2 or Y=k(ƒÌ+D)D , (1)
where k=ƒÀ is constant and ƒÌ=ƒ¿/ƒÀ. At small doses such as those of interest in radiation protection, the dose-quadratic term is negligible and the yield is expressed as Y=ƒ¿D. Since cross-section for dicentric formation is proportional to the square of LET (L), ƒÐdic=ƒ¿∙ƒÃ=ƒÈ∙L^2, the linear coefficient ƒ¿=a∙L may hold (Fig. A-1). (ƒÃ is a mean energy or dose imparted by a single traversal of charged particle, which is proportional to LET. Furthermore, the observed frequency of aberrations is modified by the removal of hit cells from cell population either by apoptosis or reproductive death, which is also dependent on LET of a charged particle. A generalized expression may be Y=ƒ¿0∙L∙D∙exp(k-ƒÁ∙L-ƒÂ∙L^2). RBEM or QF of a given radiation is expressed by RBEM=Q(L)=ƒ¿(L,k,ƒÁ,ƒÂ)/ƒ¿0(L0,k0,ƒÁ,ƒÂ), where ƒ¿ and ƒ¿0 are the observed linear coefficients of the test and reference radiation, respectively.
Cross-section of dicentric formation in a function of track average LET, L∆=100,T.
The data collection was restricted only for the experiments made in a track segment mode to avoid rapid change of dose within the cell by Bragg peak.
Dark yellow: oxygen and carbon ions
2. Formalism of quality factor, Q(L) (Heavy ions are not included)
The RBEM was calculated from the published dose-response relationships of dicentrics in human peripheral blood lymphocytes irradiated with photons, protons and ƒ¿-particles (Figs. A-2 and A-3). In the case of photons, the LET is that for the Compton electrons and photoelectrons. The RBEM-LET relationship was fitted to the following semi-empirical formula, which works irrespective of the choice of reference radiation.
Q(L)=(0.32∙L∙exp(0.67/(1+L0)-ƒÁL-ƒÂL^2)/(L0^0.2∙exp(-ƒÁL0-ƒÂL0^2)), . . . .(2)
where L0 and L are track average LET with cut-off energy of 100 eV, L∆=100,T, for reference and test radiation, respectively. L0=0.22 keV/ƒÊm for 60Co ƒÁ-rays and L0=1.5 keV/ƒÊm for 220 kVp X-rays (ICRU 1970). ƒÁ and ƒÂ are inactivation constants, where ƒÁ=2.8∙10^-5 and ƒÂ=7.5∙10^-5. The rationale of the formula is that the dicentric yield increases in proportion to LET but its overall frequency is modified by cell inactivation by traversal of charged particle (modified from: Takatsuji and Sasaki, Int J Rad Biol 45: 237, 1984; Takatsuji et al. J Rad Res 40: 59, 1999).
Figs. A-2 (left) and A-3 (right)
Quality factor, Q(L), of charged particles.
RBEM of dicentric formation was fitted to equation (2) (red curve). Fig. A-2: reference radiation is 60Co ƒÁ-rays. Fig. A-3: reference radiation is 220 kVp althovoltage X-rays. Gray line is ICRP 60 (1991).
Appendix B: Quality factor, Qn, and effective quality factor, qE, of neutron
1. Quality factor of neutron (Qn)
Quality factor of neutrons has been calculated by considering high LET charged particles resulting from exposure of ICRU soft tissue by neutrons. (In capture reactions at thermal energy, 14C from 14N(n,p)14C and 14C and ƒ¿-particles from 17O(n,ƒ¿)14C are not included, which constitute about 3.5 % of the dose from neutron capture). The LET of charged particles was determined by assuming that they completely stop in the tissue. Then the neutron energy dependent Qn has been calculated using Q(L)-LET function given in equation (2). The results are shown in blue lines in Fig. B-1 (reference radiation: ƒÁ-rays) and Fig. B-2 (reference radiation: X-rays; numerical values taken from Schuhmacher and Siebert, Rad Prot Dosim 40:85, 1992). The value thus obtained represents Qn in the critical volume on which neutrons are entering (incident site).
2. Effective quality factor neutrons incident on the human body (qE)
The incident energy equivalent quality factor at the cite of critical organs or tissues in the depth is modified by capture ƒÁ associated with the neutron moderation in the body (ICRP 92 and therein Auxier et al. 1968, Dietze 1995). Qn was modified by a energy-dependent relative dose contribution of high LET radiation, DhL(E), in MIRD phantom reported by Dietze (1995). The effective quality factor, qE, thus obtained is shown in red line in Fig. B-1 (reference radiation: ƒÁ-rays) and Fig. B-2 (reference radiation: X-rays). The following relations may hold for neutrons incident on the human body, where k(E) is neutron kerma and ƒÓ(E) is neutron fluence rate.
qE=Qn∙DhL and wR=(çƒÓ(E)∙qE(E)∙k(E)∙dE)/(çƒÓ(E)∙k(E)∙d(E)) . . . (3)
Figs. B-1 (left) and B-2 (right)
Quality factor (A) and effective quality factor (B) of neutron calculated basing on different reference radiation: ƒÁ-rays (left panel) and X-rays (right panel). The Qn against X-rays was reproduced from Schuhmacher and Siebert 1992.