ORAU 95/F-29

CIRRPC

Science Panel Report No. 10

RECEIVED OCT 2 4 1995 OSTI

NEUTRON QUALITY FACTOR

June 1995

COMMITTEE ON INTERAGENCY RADIATION RESEARCH AND POLICY COORDINATION

(CIRRPC)

Office of Science and Technology Policy Executive Office of the President

Washington, DC 20506

nsnauntw OF « oa**« * " ^

I

The Committee on Interagency Radiation Research and Policy Coordination (CIRRPC) is chartered under the Office of Science and Technology Policy, Executive Office of the President, Washington, DC 20506.

This report was prepared under contract DE-AC05-760R00033 between the U.S. Department of Energy and Oak Ridge Associated Universities.

NOTICES

The opinions expressed herein do not necessarily reflect the opinions of the sponsoring institutions of Oak Ridge Associated Universities.

This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor any of the agencies that are members of CIRRPC,* nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, mark, or manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the U.S. Government or any agency thereof. The views and opinions of the authors expressed herein do not necessarily state or reflect those of the U.S. Government or any agency thereof.

Printed in the United States of America. Copies are available by referring to publication number ORAU 95/F-29 and writing to:

National Technical Information Service U.S. Department of Commerce 5285 Port Royal Road Springfield, VA 22161

NTTS price codes: Printed copy, A:03; Microfiche copy, A:01.

*The Departments of Agriculture, Commerce, Defense, Energy, Health and Human Services, Housing and Urban Development, Interior, Justice, Labor, State, Transportation, and Veterans Affairs; Environmental Protection Agency; Federal Emergency Management Agency; National Aeronautics and Space Administration; National Science Foundation; Nuclear Regulatory Commission; and the Office of Management and Budget.

DISCLAIMER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

CIRRPC

Science Panel Report No. 10

NEUTRON QUALITY FACTOR

June 1995

COMMITTEE ON INTERAGENCY RADIATION RESEARCH AND POLICY COORDINATION

(CIRRPC)

Office of Science and Technology Policy Executive Office of the President

Washington, DC 20506

COMMITTEE ON INTERAGENCY RADIATION RESEARCH AND POLICY COORDINATION

1019 Nineteenth Street, NW, Suite 700 Washington, D.C. 20036

July 20, 1995 MEMORANDUM

TO: Dr. John H. Gibbons, Assistant to the President for Science and Technology Office of Science and Technology Policy

THRU: Dr. Philip R. Lee, Chairman Committee on Health, Safety, and Food R&D Department of Health and Human Services

FROM: Dr. Alvin L. Young, Chairman Committee on Interagency Radiation Research apd'"FpJiey''Coordination

SUBJECT: Transmittal of Neutron Quality Factor Report

I am pleased to forward Science Panel Report No. 10, entitled Neutron Quality Factor, that was recently completed by the Science Panel of the Committee on Interagency Radiation Research and Policy Coordination (CIRRPC). The report was developed by a Science Panel working group, chaired during its final efforts by Dr. Randall S. Caswell, former Chairman of the CIRRPC Science Panel and former Chief of the Ionizing Radiation Division at the National Institute of Standards and Technology.

The initiative for this report was an early decision by the CIRRPC Science Panel to examine, as a collective body of Federal scientists, the scientific bases for certain recommendations promulgated by the International Commission on Radiological Protection and subsequently adopted by the U.S. National Council on Radiation.Protection and Measurements. Among their recommendations was to increase the quality factor for neutrons (QJ by a factor of two, i.e., to increase the value from 10 to 20. This radiation comparison is a normalizing parameter that is related to biological effectiveness and, physically, to linear energy transfer. The reference radiation for comparison is most often considered to be 60Co 7 rays or 250 kVp x rays.

After careful consideration of the related science, the working group concluded that there exists "...no compelling basis for a change in Q for neutrons at this time" and that the recommended nominal Qn value of 10 used by Federal agencies should be maintained "until such time as a need for change is firmly established." The Federal agencies are following this recommendation.

The CIRRPC report has been extensively reviewed by member agencies and numerous non-Federal scientists, and no substantive issues needing resolution remain. I believe that the report is scientifically sound and that it is an important contribution to the nation's radiation protection programs.

iii

COMMITTEE ON INTERAGENCY RADIATION RESEARCH AND POLICY COORDINATION

1019 Nineteenth Street, NW, Suite 700 Washington, D.C. 20036

July 19, 1995

MEMORANDUM

TO: Dr. A l v i n X ^ b u n o ^ h a i r m a n . OIRRPC

FROM: Dr: B. John Alfysworth, Acting Chairman, CIRRPC Science Panel

SUBJECT: Transmittal of the Neutron Quality Factor Report

I am pleased to transmit the latest CIRRPC Science Panel report, entitled Neutron Quality Factor.

The report reflects a long-term effort by the Science Panel to review the International Commission on Radiological Protection (ICRP) and the National Council or Radiation Protection and Measurements (NCRP) recommendations to increase the neutron quality factor (Q n). During this period, a number of Federal scientists have been involved in this effort. However, in spite of the different Federal scientists who worked on the report, there was continual agreement that a change in Q n , as had been recommended by ICRP and NCRP, was not warranted. We concluded that the selection of an appropriate value for Q n is confusing because of the lack of a properly defined "reference radiation for radiation protection." Based on this observation and our extensive examination of fundamental concepts related to the meaning of Q n and the methods used in determining its values, we concluded that there is no compelling reason to increase Qn by a factor of 2, and we recommended that Federal agencies should maintain a nominal Q n value of 10 "...until such time as a need for change is firmly established." Federal agencies are currently controlling exposures to neutrons accordingly.

It is noteworthy that the Federal agencies have chosen not to change Q n .

v

ACKNOWLEDGMENTS

The Committee on Interagency Radiation Research and Policy Coordination (CIRRPC) authors gratefully acknowledge the significant contributions of the many past and present participants in the effort to address the scientific basis for the selection of a neutron quality factor for use in radiation protection. CIRRPC is especially indebted to Elmer Eisenhower (formerly with the National Institute of Standards and Technology) for his extensive review and analysis of the experimental data on the relative biological effectiveness of neutrons. Special thanks are also extended to Victor Bond (Brookhaven National Laboratory), John Dennis (National Radiological Protection Board, UK), Leo Faust (Battelle Pacific Northwest Laboratory), Douglas Grahn (Argonne National Laboratory), Harald Rossi (Columbia University), Tore Straume (Lawrence Livermore National Laboratory), and Robert Ullrich (Oak Ridge National Laboratory) for their participation in a workshop that briefed CIRRPC on the latest scientific data related to this topic. A list of the many Federal employees who contributed to the development of this report can be found in Appendix B.

vii

TABLE OF CONTENTS

Page

A. INTRODUCTION 1 B. FUNDAMENTALS 2

1. RELATIVE BIOLOGICAL EFFECTIVENESS 2 2. QUALITY FACTOR (Q) IN THE DQH SYSTEM FOR RADIATION PROTECTION . . 5 3. THE CONCEPT OF REFERENCE RADIATION 7 4. THE DOSE AND DOSE-RATE EFFECTIVENESS FACTOR 8

C. REVIEW OF METHODS USED FOR DETERMINING VALUES OF HUMAN RBE 11

1. RBE FROM HUMAN EXPOSURE DATA 11 2. RBE FROM EXPERIMENTAL OBSERVATIONS 12 3. RBE DERIVED FROM "RADIATION DAMAGE" MODELS 15 4. SUMMARY 16

D. CONSIDERATIONS IN SELECTING Q 17

E. CONCLUSIONS 21

F. RECOMMENDATIONS 23

REFERENCES 25

APPENDIX A: ICRP RECOMMENDED VALUES A-l

APPENDIX B: AGENCY PARTICIPANTS IN THE DEVELOPMENT OF THE REPORT ON THE NEUTRON QUALITY FACTOR B-l

APPENDIX C: CURRENT MEMBERSHIP OF THE CIRRPC SCIENCE PANEL C-l

ix

LIST OF TABLES Page

Table 1. Summary of estimated constant maximum relative biological

effectiveness values (RBEM) versus gamma rays 2

Table 2. Chromosome aberrations in human lymphocytes 4

Table 3. DDREF for various endpoints as calculated from different dose points . . 19

Table Al. Relationship between collision stopping power in water (LJ

and quality factor (Q) A-l

Table A2. Effective Q values A-l

Table A3. Specified Q-L relationships A-l Table A4. Radiation weighting factors A-2

xi

LIST OF FIGURES

Page

Figure 1. RBE as a function of neutron absorbed dose for chromosome aberrations

in human lymphocytes 4

Figure 2. Schematic curves of incidence versus absorbed dose 9

Figure 3. Dose-rate effect of ^Co gamma dose on mean life shortening in B6CFi mice 10

Figure 4. Ratios of measured to calculated neutron activation in Hiroshima at various distances from the epicenter 12

Figure 5. Range of neutron RBEM values for a variety of tumors obtained from whole-body irradiation of B6CFj mice 13

Figure 6- Ratio of the effective dose equivalent, HE, to the individual dose equivalent, HP(10), as a function of photon energy 20

Figure 7. Ratio of the effective dose equivalent, HE, to the individual dose equivalent, HP(10), as a function of neutron energy 21

xm

A. INTRODUCTION

Both the International Commission on Radiological Protection (ICRP) (1985, 1990) and the National Council on Radiation Protection and Measurements (NCRP) (1987, 1993) have recommended that the radiation quality weighting factor for neutrons (Qn, or the corresponding new modifying factor, WR) be increased by a value of two for most radiation protection practices. This means an increase in the recommended value for Qn from a nominal value of 10 to a nominal value of 20. This increase may be interpreted to mean that the biological effectiveness of neutrons is two times greater than previously thought.

A decision to increase the value of Qn will have a major impact on the regulations and radiation protection programs of Federal agencies responsible for the protection of radiation workers. [See, for example, The Impact of a Proposed Change in the Maximum Permissible Limits for Neutrons to Radiation Protection Programs at DOE Facilities. PNL-3973/UC 41. September 1981.] Therefore, the purposes of this report are:

(a) to examine the general concept of "quality factor" (Q) in radiation protection and the rationale for the selection of specific values of Qn; and

(b) to make such recommendations to the Federal agencies, as appropriate.

This report is not intended to be an exhaustive review of the scientific literature on the biological effects of neutrons, with the aim of defending a particular value for Qn. Rather, the working group examined the technical issues surrounding the current recommendations of scientific advisory bodies on this matter, with the aim of detennining if these recommendations should be adopted by the Federal agencies. Ultimately, the group concluded that there was no compelling basis for a change in Qn.

The report was prepared by Federal scientists working under the auspices of the Science Panel of the Committee on Interagency Radiation Research and Policy Coordination (CIRRPC).1

In presenting our findings here, we will first review the basic philosophy and traditional use of relative biological effectiveness (RBE) and Q as parameters in implementing radiation protection for neutron sources. After having laid that groundwork, we will consider data, experimental methods, and representative models that have been proposed to determine the RBE for neutrons and to define the numerical value of the nominal quality factor for neutrons. Finally, we will look at the methods used to relate radiation dose to human health risk. In that discussion, we will consider how the assumptions and judgements made about the uncertainties

1 CIRRPC is a subcommittee of the National Science and Technology Council's Committee on Health, Safety and Food (Office of Science and Technology Policy, Executive Office of the President). CIRRPC membership currently consists of 18 Federal departments and agencies. It was chartered on April 9, 1984 to identify radiation issues of concern to the Federal agencies, to develop initiatives for proposed solutions, and to coordinate radiation research programs.

of biological effects from radiation at very low occupational doses affect both the selection of the Qn value and the assessment of risk.

B. FUNDAMENTALS

1. RELATIVE BIOLOGICAL EFFECTIVENESS

We begin with a discussion of one of the factors considered in establishing a value for the neutron quality factor (Qn), relative biological effectiveness (RBE). The RBE of one type of radiation (i) compared to another type of radiation (r), called the reference radiation, is defined as the ratio of the doses, D ; and D r, respectively, of the two types of radiation required to produce the same specific biological effect or endpoint. By convention, the dose required for the reference radiation (Dr) is placed in the numerator; thus,

R B E = ^ . (1)

The relative effectiveness of different types of radiation depends greatly on: (1) the specific exposure conditions for the radiations used (i.e., energy spectrum, dose rate, fractionation, etc.), (2) the choice of the reference radiation, and (3) the specific biological endpoint chosen. The RBE can, therefore, only be considered a useful parameter for a particular set of circumstances, rather than a universal characteristic of a particular type of radiation. Table 1 shows a range of RBE values for several biological endpoints.

Table 1. Summary of estimated constant maximum relative biological effectiveness values (RBEM) versus gamma rays (NCRP 1990).

Endpoint Range of Values2

Cytogenetic studies, human lymphocytes in culture 34-53

Transformation 3-80 b

Genetic endpoints in mammalian systems 5-70 c

Genetic endpoints in plant systems 2- 100

Life shortening, mouse 10-46

Tumor induction 16-59 a Values taken from larger tables or data given in the NCRP report. b The value of 80 was derived from one set of experiments only. c The value of 70, which was derived from data on specific locus mutations in mice, is not

necessarily an RBEM.

2

If the dose-response relationships are known for the given biological effect (U), functional relations for RBE can be derived. For example, consider two linear dose-response relationships for effects Ur and Uj,

(2) Ur = ar Dr and

v> = «i A , (3)

where at and ax are the linear coefficients in the dose-response relationships for the reference radiation (r) and "other" radiation (i), respectively. RBE at any dose can then be determined by making U r equal to U; (the isoeffect constraint) such that

or, D, = ar D r and (4)

(5)

Often, the dose-effect relationships are more complex than the simple linear relationships assumed in Eqs. (2) and (3). For example, a linear-quadratic relationship (U = aD + /3D2) has commonly been used to describe responses to exposures to ^Co gamma rays or to 250 kVp x rays, which are types of low linear energy transfer (LET) radiations widely used as reference radiations in radiobiology studies. Conversely, a linear dose-response relationship is often observed in experimental studies involving moderate doses of neutrons. Repeating the analysis described above, the RBE for neutrons, under these nonlinear conditions for the reference radiation, can be derived as follows:

U = or r D r + / 3 r t f = f f n D n ,

and, upon solving for D r, RBE becomes

(6)

RBE = a. 2# r D„ 1 +

^ i + Tl&tt

ar

(7)

Note that the RBE in this case (Eq.7) depends on the neutron dose. Using the linear and quadratic coefficients for chromosome aberrations in human lymphocytes (see Table 2), from the work of Edwards et al. (1982) and as reported in ICRU Report 40 (1986), the relationship between RBE and neutron dose is illustrated by Figure 1 on both a linear and a logarithmic dose scale. In the upper curve of each graph, the reference radiation is ^Co gamma rays. For comparison, the lower curves show the neutron RBE referenced to 250 kVp x rays. The differences in the curves are striking. When the neutron dose decreases, the RBE rapidly

3

increases, as seen in Figure 1A. Approaching very low neutron doses, RBE reaches a constant value, as shown in Figure IB. While the RBE may approach a constant value at very low doses, that value is highly variable because it is dependent on a number of factors, such as the choice of the reference radiation.

Table 2. Chromosome aberrations in human lymphocytes (Effect [chromosome aberrations per tell] = ciD + /?D2, where D is the absorbed dose).3

Radiation aCGy1) 0(Gy2)

^Co 1.57 ±0.29xl0" 2 5.00 ± 0.20 x lO"2

250 kVp x rays 4.76 ± 0.54 x 10"2 6.19 + 0.31 x lO"2

Fission neutrons'5 (E = 0.9 MeV)

72.8 ± 2.4 x lO'2 Fission neutrons'5 (E = 0.9 MeV)

72.8 ± 2.4 x lO'2

Fission neutronsb (E = 0.7 MeV)

83.5 + 1.0 xlO-2 Fission neutronsb (E = 0.7 MeV)

83.5 + 1.0 xlO-2

a Edwards et al. (1982), as reported in ICRU Report 40 (1986). b Average o;n = 78.2 x 10"2 Gy"1 for the two neutron energies.

Reference Radiation • '"Co gamma rays

250 kVp x rays

Figure 1A Neutron Dose (Gy)

Figure 1B Neutron Dose (Gy)

Figure 1. RBE as a function of neutron absorbed dose for chromosome aberrations in human lymphocytes, using a linear-quadratic dose-response relationship for the reference radiation (^Co gamma rays or 250 kVp x rays) and a linear dose-response relationship for neutrons. Figure 1A has dose plotted on a linear scale, and Figure IB has dose plotted on a logarithmic scale. The values of a r and /3r are given in Table 2, and a n is the average of the initial slopes for the two neutron energies listed in Table 2.

4

At very low reference radiation doses, when the quadratic term (j8r) is very small compared to cer, the value of the RBE reaches a constant maximum value sometimes referred to as RBEM

RBEM= lim RBE = ^ 2 - ( 8 )

For example, using only the linear terms shown in Table 2, RBEM is 50 with ^Co gamma rays as the reference radiation, and RBEM is 16 with 250 kVp x rays as the reference radiation. The projected maximum values of RBE in this system will be attained when the neutron dose is < 0.001 Gy for 250 kVp x rays and < 0.0001 Gy for ^Co gamma rays.

At high doses (e.g., > 0.1 Gy in Figure IB), the dependence of the estimated RBE on the neutron dose becomes inversely proportional to the square root of the neutron dose. This is described mathematically by

lim RBE = D„->oo A

n

(9)

For biological endpoints in which the contribution to the effect from the linear term is small (a r -> 0), the dose-effect relationship for the reference radiation approaches a dose-squared function. In this case, the RBE does not reach a maximum value with decreasing neutron dose, and Eq. (7) reduces to an expression identical to Eq. (9)

lim RBE = a..->0 *T

an &D n (10)

Note that the RBE for such a biological system would go to infinity as the neutron dose goes to zero. This observation is important in understanding, and in keeping perspective of, the significance of some values of RBEM reported in the literature. That is, extremely high values of RBEM c a n D e inferred in systems for which the dose-effect relationship for the reference radiation has a small or negligible linear term.

2. QUALITY FACTOR (Q) IN THE DQH SYSTEM FOR RADIATION PROTECTION

The product of the absorbed dose (D) and the quality factor (Q), both as determined at a specified point in tissue, is the quantity dose equivalent (H)

H = Q D. (11)

5

For radiation protection purposes, a limit is placed on the total dose equivalent (in units of rem or sievert) that may be accumulated by an individual in a specified period of time from all types of radiation. This is called the dose equivalent or DQH system.

Since 1977, Q has been defined by ICRP, ICRU, and NCRP in several publications (ICRP 1977; ICRU 1980, 1986; NCRP 1987) as a dimensionless factor used to account for differences in the biological effectiveness of different types of radiation; more or less explicitly, for the stochastic endpoints and dose range of concern in radiation protection activities. The relationship recommended by ICRP (1977) between Q and collision stopping power in water2 OU,) is shown in Table Al in Appendix A. Based on these values for L„, the nominal assigned values of Q for various types of radiation are shown in Table A2 in Appendix A. The nominal assigned value of Qn is 10.

More recently, ICRP modified its recommendations for the numerical value of Q (see Table A3 in Appendix A). This modification appears to be based on the judgement of the significance of higher RBEM values reported in the literature for intermediate energy neutrons and on the intent to base Q on radiation conditions likely to be experienced by radiation workers (ICRP 1990). ICRP also recommended the use of a new parameter, closely related to the quality factor, designated as the radiation weighting factor, wR. This dimensionless factor serves the same purpose in radiation protection as Q; that is, to normalize protection to that of low-LET radiation. For neutrons, however, its numerical value is determined from the energy of the radiation incident on the body rather than at a point in the organ, as is the case for the Q-based system. The product of the radiation weighting factor and the average absorbed dose to a particular tissue is called the equivalent dose (H T R,). That is,

H T , R = W R - D T ) R (12)

where the subscript T refers to a specific tissue of the body (e.g., lung), and the subscript R refers to the type of radiation (e.g., neutrons). ICRP recommended the use of specific values for the radiation weighting factors, as shown in Table A4 in Appendix A.

Alternatively, ICRP recommended that the following neutron energy-dependent mathematical relationship could be used in assigning wR values for neutrons of energy E (MeV)

wD = 5 + 17 e - ( ' n ( 2 ^ ) 2 / 6 - (13)

Another option recommended by ICRP for radiation types not included in Table A4 is to calculate the average, or effective, Q from the secondary particle spectrum using the Q(L)

2 Collision stopping power in water is used as a surrogate for L,,, in tissue and is considered equivalent to unrestricted linear energy transfer (ICRU 1993).

6

relationship of Table A3, at a depth of 10 mm in the ICRU sphere (ICRU 1985), as an approximation of wR.

NCRP also changed the focus of its recommendations on evaluating equivalent dose from consideration of dose at a point in tissue to the concept of average dose in a specific tissue or organ. The Council noted that the concept of averaging the dose in a specific tissue results from one's acceptance of the linear hypothesis that health detriment within a tissue or organ is proportional only to dose, and that variations of dose within a tissue of uniform sensitivity to cancer induction are considered unimportant. For this reason, with one exception, the ICRP-recommended values of wR given in Table A4 and the relationship between Q-L shown in Table A3 were endorsed by NCRP and were included in the recommendations in NCRP Report No. 116 (NCRP 1993). The one exception is the value of wR recommended for high-energy protons, as indicated in Table A4.

For clarity and simplicity, the remainder of the discussion herein will refer only to Q, rather than to both Q and wR. The modifications introduced by changing the location at which one determines each of these modifying factors (i.e., at a specific point in tissue for Q, and either incident on the body for external sources or at the point of origin of an internal source for WR) do not affect the subsequent discussions.

3. THE CONCEPT OF REFERENCE RADIATION

In the DQH system, the different types of measured radiation-absorbed doses are multiplied by what is judged to be the appropriate quality factor to account for differences in the effectiveness per unit absorbed dose for a given biological endpoint. Establishing a consistent radiation protection system, therefore, requires that the reference radiation be invariably defined and carefully selected and serve as the basis of protection limits. However, the use of three concepts in selecting a reference radiation leads to confusion when these concepts are considered to be compatible and the data derived from them interchangeable. These concepts, reference radiation for radiobiology, reference radiation for radiation protection, and assumed common effectiveness of all low-LET radiations, are discussed below.

A majority of radiobiology experiments designed to determine RBE utilize some type of photons as the reference radiation. The reference radiation for radiobiology used in many of the experiments to determine RBE values has been high-dose-rate orthovoltage x rays (200-250 kVp) (Boag 1953; ICRP 1955, 1960, 1972; ICRU 1957, 1986; NCRP 1971; UNSCEAR 1986). For many other RBE experiments, high- and low-dose gamma rays (particularly 137Cs or ^Co) have been utilized as the reference radiation (Grahn et al. 1983, 1986, 1992; Hill et al. 1985; Lloyd and Edwards 1983; Storer and Mitchell 1984; Thomson and Grahn 1988; Thomson et al. 1985, 1986; Ullrich 1983; Ullrich et al. 1976; Ullrich and Preston 1987; Upton et al. 1970; Vogel and Dickson 1981). Our most robust human data are from the A-bomb survivors exposed primarily to high dose/dose-rate gamma rays.

7

The second and more recent concept is reference radiation for radiation protection (ICRU 1976). This concept refers to the radiation conditions upon which the quality factor, Q, is based and, it is appropriate only for the low-dose levels of radiation associated with radiation protection. It may therefore be thought of as "reference radiation for Q." The reference radiation for radiation protection logically applies to the actual conditions of routine irradiations of workers and the public. Often, however, the reference radiation for radiation protection is not distinguished from the reference radiation for radiobiology (historically considered to be high- dose-rate orthovoltage x rays). Although not stated explicitly, ICRP, beginning in 1977, appears to consider low dose/dose-rate gamma rays as the reference radiation suitable for radiation-protection purposes. However, low dose/dose-rate x or gamma rays are not an ideal "reference radiation," because a standard reference should have properties of invariance and reproducibility. The effectiveness of protracted x or gamma rays strongly depends on both dose and dose rate; this dependence limits their use as the "reference radiation" to specific dose conditions.

The third concept utilized is the assumed common effectiveness of all low-LET radiations. For simplicity in health physics practice, the value of Q has been considered to be 1 for all low-LET radiations below 3.5 keV//xm (including gamma rays, x rays, and high- and low-energy beta rays), and independent of dose rate (ICRP 1990; NCRP 1987). This judgement was made "to reflect our lack of precise information in man," but with ICRP and NCRP acknowledging that effectiveness was dependent on photon energy and that there is a difference of 2 or more, due to changes in photon dose rates (ICRU 1986; NCRP 1990). The acceptance of having Q = 1 for all photon radiations does not mean, however, that the concept of assumed common effectiveness of all low-LET radiation can be applied to a "standard" or reference radiation. A factor of 2 or more between the effectiveness of two types of radiation judged to be equivalent "standard radiations" would lead to a difference of 2 or more in the quality factor for neutrons or for other types of radiation, depending on which "standard" is used in the RBE data being evaluated.

4. THE DOSE AND DOSE-RATE EFFECTIVENESS FACTOR

The concept of a dose and dose-rate effectiveness factor (DDREF) is employed to correct for the differences observed in the biological effectiveness determined at high dose/dose-rate conditions to low dose/dose-rate conditions. For radiations and endpoints that exhibit a linear-quadratic dose response, the risk estimate determined from the linear fit of high dose/dose-rate exposures is larger than the risk estimate determined at low dose/dose-rate exposures. The DDREF is defined as the ratio of the slope of the linear, no threshold fit to high dose/dose-rate response data, to the slope of the linear, no threshold fit to low dose/dose-rate response data (NCRP 1980; ICRP 1990).

According to the linear-quadratic model, the effect or risk can be expressed as U = aD + jSD2 for high dose-rate conditions. The linear extrapolation between the high dose/dose-rate at which the effect is observed and the origin gives a risk coefficient (risk per unit "high" dose) of an = a + )3D. At low doses, where the D 2 term can be considered negligible, the low dose-

8

rate risk coefficient (aL) equals the slope of the linear portion (a) of the linear-quadratic model (see Figure 2). One can then define the DDREF at any given dose as (ICRP 1990)

DDREF= or, H « T

a+pD a

1+H D a (14)

CD O

c 0)

o CD

U= d>

T3 0) O

Linear, No Threshold / Slope aH /

Absorbed dose (D) -

Figure 2. Schematic curves of incidence versus absorbed dose, with a H as the slope of the linear extrapolation between high dose and origin, and ah as the slope of the linear portion of the linear-quadratic model.

If the biological effectiveness, or dose response, is determined at high dose/dose-rate exposure conditions for a radiation type exhibiting a linear-quadratic dose response, as is often the case for x and gamma rays, then the dose response for low-dose protracted exposures is estimated by dividing the high-dose response by the DDREF. DDREF is highly dependent on the coefficients defining the linear-quadratic dose-response function, thus making DDREF and RBE interrelated. For example, as the linear coefficient (a r) of a linear-quadratic reference radiation decreases in value, RBEM increases in value, as does the corresponding DDREF. Since the choice of DDREF values influences the determination of both risk coefficients and RBEM, dose and dose-rate factors must be applied consistently when evaluating Qn and assessing risk.

For some endpoints, such as life shortening in mice (Thomson et al. 1981; Thomson and Grahn 1988), there is no evidence of a quadratic term in the dose-response relationship (see Figure 3). In this case, however, when the animals are exposed to the same total doses, but,

9

in a single exposure or in 60 weekly fractions, the slopes of the life-shortening curves decrease from 41 Gy"1 to 19 Gy"1 with fractionation of dose. A dose-rate factor may be required for systems with a strictly linear dose-response relationship.

400

350

300

CD 2 250 W I-rr o X CO UJ 200 UL

< III

150

100

Single exposure: • males • females

60 weekly fractions: x males + females

DOSE (Gy)

Figure 3. Dose-rate effect of ^Co gamma dose on mean life shortening in B6CF! mice. Male and female data for single exposure (Thomson et al. 1981) and for 60 weekly fractions (Thomson and Grahn 1988) were fit to linear functions of the form, U = ccy D. The slopes, as calculated for this report, were 41 Gy"1 and 19 Gy"1, respectively.

10

C. REVIEW OF METHODS USED FOR DETERMINING VALUES OF HUMAN RBE

A variety of methods have been used to estimate values for the neutron RBE in humans. These methods are based on evaluations of epidemiological information, experimental measurements, and a critical examination of radiation damage models. In this section, each of these methods will be addressed, and published data will be cited only as necessary to illustrate the approach taken. The most recent reviews of relevant scientific literature are provided in NCRP Report No. 104, The Relative Biological Effectiveness of Radiations of Different Quality (NCRP 1990), and in a report submitted to CIRRPC, Biological Effectiveness of Neutrons: Research Needs (Casarett et al. 1992). In these reports, the available data are summarized for cytogenetic effects, in vitro transformation, mutation, hereditary effects, carcinogenesis, and life shortening.

1. RBE FROM HUMAN EXPOSURE DATA

The revision of dosimetry data from the Hiroshima and Nagasaki bombings (Roesch 1987; NAS/NRC 1990) suggests that the neutron contributions to the dose were too small to support choosing a definitive value of neutron RBE in humans (NCRP 1990). Shimizu et al. (1989) estimated the neutron RBE ( + 1 SD) for leukemia and for all cancers except leukemia to be 52.0 (+ 59.6) and 10.1 (+ 38.8), respectively. Forcing the excess relative risk for the two cities to be identical, Shimizu et al. calculated the neutron RBE as 20-30 for leukemia; 30 or more for all cancers except leukemia; and less than 1 for cancers of the stomach, lung, and female breast. The disparity of these estimates is attributed by Shimizu et al. to the difficulty in deriving values of RBE from these data. Abrahamson (1989a; 1989b) points out that because RBE depends on dose (see Eq. 7), attempts to assign dose-independent values for neutron RBE to account for intercity differences are flawed.

To estimate the neutron RBE from the limited available human data, Zaider (1991) assumed a linear relative-risk model, with a gamma dose-dependent cell-killing term, having the form

e (D ) = (l+«nDn4ttYDY) e " Y ^ ,

where e(D) is the effect of the total dose (D), an is the linear-effect coefficient for neutrons, o^ is the linear-effect coefficient for photons, and yy is the cell-killing coefficient for photons. Using the Hiroshima/Nagasaki relative risk data for solid tumors, Zaider determined a maximum likelihood fit for the linear coefficients of this model. From the ratio of these linear coefficients, anlay, he computed an RBE value for neutrons of 70 + 50 (+1SD) applicable to solid tumors in humans.

More recently, measurements by Straume et al. (1992) have suggested that there is a systematic discrepancy between the calculated and measured neutron doses inferred for A-bomb survivors (see Figure 4). It appears, however, that at distances for which the discrepancies are

11

largest, total doses are sufficiently low that it is unlikely that these systematic differences, if correct, will result in major changes in risk estimates or provide adequate information for estimating neutron RBE values for humans (Preston et al. 1993).

100 F

CO

u n 3 tn CO CD

2

400 600 800 1000 1200 1400 Slant range (M)

1600 1800 2000

Figure 4. Ratios of measured to calculated neutron activation in Hiroshima at various distances from the epicenter (slant range). The dashed line is the least-squares best fit to all the data points (Straume et al. 1992).

2. RBE FROM EXPERIMENTAL OBSERVATIONS

There are two fundamental methods for relating neutron RBE data generated in the laboratory to the low-LET, high dose/dose-rate exposure data that form the basis for current human risk estimates.

In the first method, which could be termed the "conventional" method, human risk estimates for low-LET radiation are derived from available high dose/dose-rate human data (primarily, A-bomb survivors). These risk coefficients are corrected by the dose and dose-rate effectiveness factor (DDREF) inferred from the shape of the dose-response function from either the A-bomb survivor data or other human databases. RBEM values observed in a wide range of laboratory data are then applied to the DDREF-corrected human risk data. This approach relies on the assumption that RBEM values observed in the laboratory apply to humans. However, uncertainties in the selection of an appropriate human RBEM value a i e l a r g e because of the wide range of derived RBEM values observed in laboratory experiments. These uncertainties stem in part from the variety of different biological endpoints considered and from the enormous

12

sensitivity of RBEM values to changes in the effectiveness of the reference radiation at low doses. An example of how RBEM values derived from laboratory data may vary is shown in Figure 5. Average neutron RBEM values are plotted along with their corresponding range of values based on tumorigenesis data published by Grahn et al. (1992) for single and protracted whole-body irradiation of B6CF! mice. The types of tumors included in these data are lymphoreticular tissue, vascular tissue, epithelial (except ovarian), lung, liver, glandular and reproductive systems (except ovarian), Harderian gland, and ovarian. The average values of RBEM, a s w e u " as the range of these values, are shown to increase with increasing protraction of dose.

120 -,

100-

80

LU

CC 6 0

40-

20

Absorbed dose range for exposure conditions:

Samma Neutrons

Single exposure 0.225-7.88 Gy 0.01 -2.4 Gy 24 exposures/once per week 2.06-19.18 Gy 0.2 -2.4 Gy 60 exposures/once per week 1.0 -19.18 Gy 0.02-1.6 Gy

Average value of BB^, is the average of summing values for males (except for ovarian) and females in two time intervals of 600-799 and 800-999 days to death from first exposure. Intervals (days) from first exposure were: 400-599, 600-799,800-999 and 1000-1199.

I I Single Exposure 24 Exposures

One/Week 60 Exposures One/Week

Figure 5. Range of neutron RBEM values for a variety of tumors obtained from whole-body irradiation of B6CFJ mice (Grahn et al. 1992). The types of tumors included in this data are lymphoreticular tissue, vascular tissue, epithelial (except ovarian), lung, liver, glandular and reproductive systems (except ovarian), Harderian gland, and ovarian. Data points show average neutron RBEM values, and lines indicate range of RBEM values.

In the second method, the RBEM f° r humans is determined by the product of an appropriate neutron RBE value determined under high dose/dose-rate conditions and the DDREF selected for use with human low-LET, acute-exposure risk coefficients. This method has the advantage of being based on high dose/dose-rate data, where the range of laboratory RBE values are much smaller and less dependent on the choice and exposure conditions for the reference

13

radiation. This approach relies on the assumption that high dose/dose-rate RBE values observed in the laboratory apply to humans. Thus, the high dose/dose-rate data (RBEH) is modified for use in the case of low dose/dose-rate exposure (RBEM) by the DDREF selected for use with human low-LET, acute-exposure risk coefficients, or

RBEM = RBEH • DDREF . (16)

If it is true that the RBEH and the RBEM values are invariant across species, then, in principle, both methods should give the same and correct value of RBEM for human beings. However, because the spectrum of values of RBEM measured in the laboratory is considerably larger than the range of RBEH values, it would appear that the second approach has the advantage of introducing less uncertainty.

Under relatively high dose conditions, an increase in the absolute effectiveness of neutrons and other high-LET sources has been observed in laboratory experiments as the dose rate is reduced. This phenomenon has been referred to as the inverse dose rate effect. This effect has been observed in life-shortening experiments (e.g., Ainsworth et al. 1974 and Thomson et al. 1981), in the oncogenic transformation of mouse embryo-derived cells (e.g., Hill et al. 1982), and in Harderian gland tumor studies (e.g., Fry et al. 1976). In almost all cases, the observed effects suggest enhancement factors of 2 or less. It remains unclear at this time whether the mechanisms responsible for these observations are applicable to the human risk assessments at occupational exposure levels. Although it is important to continue to study these effects, there is insufficient evidence to support the use of an additional factor in the RBEM for humans to account for this effect.

A joint ICRP/ICRU Task Group chose to produce its recommendation for Qn values based on the first of these methods (ICRU 1986), but without providing full discussion of the magnitude of uncertainty introduced by using the conventional method. This Task Group proposed using a model for Q based on microdosimetry concepts used by Zaider and Brenner (1985); specifically, a relationship between Q and lineal energy3 (y). The general shape of this relationship was taken from data by Barendsen (1967) and Bird et al. (1980) showing that cell inactivation increased exponentially up to a maximum (at about 100 keV/^m). In specifying the functional relationship, Q(y), the ICRU/ICRP Task Group gave special consideration to RBE data from observations on chromosome aberrations in human lymphocytes by Edwards et al. (1982). The selection of these data was explained by the following reasons (ICRU 1986):

(a) ...an association between chromosome aberrations and genetic effects is well known and may also exist for carcinogenesis.

3 Lineal energy (y) is defined as the energy deposited from a single event in a defined volume, divided by the mean path length through the volume.

14

(b) .. .this is a set of data covering the widest range of radiation qualities investigated by a single group.

(c) ...it appears to be the best data on the effect of high-LET radiation on a human system which are available at this time.

The effective quality factor (Q) was defined as 1.0 for photons with energies on the order of 100 keV (i.e., orthovoltage x rays) in the Q versus photon energy relationship. Since the ICRU 40 model is considered invariant with respect to dose, it_can represent a value of Q applicable at occupational exposure levels. The model predicts a Q value of about 23 for fission neutrons (0.7 MeV), with reference to orthovoltage x rays. The ICRU/ICRP (1986) Task Group proposed a representative Qn value of 25. This value is consistent with the RBE data originally used to establish their model (Lloyd and Edwards 1983). .

Life shortening has been identified as another endpoint of particular relevance for radiation protection purposes because it represents the summation of the individual effects of many tumor-induction processes. As with other endpoints, NCRP noted that RBE values for life shortening were highly dependent on the dose rate of the low-LET reference radiation (NCRP 1990). Sometimes the RBE has been fitted to an inverse square root relationship to dose

RBE = A DB , (17)

where the constant A is highly variable and dependent on the species studied, the experimental conditions, and, in particular, the dose rate of the reference radiation. Values for A cited by NCRP ranged from 10 to 80. In most cases, B is approximately -0.5. This relationship is identical to Eqs. (9) and (10). Note that the low-dose limit for functions of this form goes to infinity. A subsequent reformulation of the relationship between life shortening and RBE suggested by Dennis (1987) is

RBE = A (K + D)B. (18)

This relationship ensures that the low-dose limit is well behaved. NCRP concluded that the low-dose limiting value of the neutron RBE for life span shortening was between 10 and 15 for single exposures and as high as 40 to 50 for continuous or long-term fractionated exposures (NCRP 1990). However, the differences observed between these limiting values may be more directly attributable to changes in the linear coefficients for the reference radiation, rather than to an increased effect of the neutrons.

3. RBE DERIVED FROM "RADIATION DAMAGE" MODELS

Because of the lack of definitive human biological data at the doses and dose rates typical of occupational exposures, it is hoped that some insight may be provided through the use of radiation damage models that would permit making more reliable extrapolations from either available human data or laboratory data. The complexity of biological systems makes the

15

development of such models extremely difficult. Nevertheless, the observation of some commonality in dose-response patterns provides insight into some understanding of mechanisms responsible for the radiation damage.

In some cases, simple observations about the stochastic nature of energy-deposition events in an irradiated biological system can provide useful insights. For example, if it is assumed that only a single biological event in a sensitive target is sufficient to either initiate or promote the cascade of events necessary to produce a tumor, then it is plausible to expect a linear dose-response relationship at doses below the point where cell nuclei are traversed, on average, by only a single ionizing particle. This point corresponds to roughly 50 mGy for 2 MeV neutrons and 0.7 mGy for ^Co gamma rays (based on an analysis of pp. 55-56 of ICRU 1983). If multiple events are necessary to produce the critical transformation, then more complex dose-response relationships would be expected.

While a full exposition of contemporary models of radiation damage is beyond the scope of this report, a summary of the most relevant models is warranted. One model that has had an enormous influence is the Theory of Dual Radiation Action first described by Kellerer and Rossi (1971). The model, in its present form, is based on the assumption that a cell is damaged by the formation of numerous sub-lesions in critical sites. Provided that these sub-lesions are close enough in time and space, they can interact to produce observable damage. The classical linear-quadratic dose-response relationship can be derived using this model.

The Katz model (Katz 1971) takes a more phenomenological approach. The probability of damage from a high-LET particle is assumed to be equal to the product of the probability of the effect caused by direct interaction of the particle with the sensitive part of the target cell and the probability of the effect caused by the dose distribution characteristic of delta rays from other passing high-LET particles. To use the model, four parameters must be determined. Two of these are derived from low-LET response data for the specific endpoint being modelled. The remaining two parameters are derived from heavy ion measurements. The ability to use the model depends on the availability of sufficient data to generate the parameters necessary to make the calculations.

Other models, such as the Lethal-Potentially Lethal model of Curtis (1986) or the Repair-Misrepair model of cell survival proposed by Tobias et al. (1980), Tobias (1985), and others, rely on compartmental relationships and associated rate constants relating the specific subpopulations of a targeted cell pool.

Great progress is being made in radiation damage models. Despite this, no single model has yet produced widely accepted human risk coefficients for neutron radiation exposure.

4. SUMMARY

We have briefly reviewed the various strategies that could be used to infer a human RBE for neutrons at occupational exposure levels. These strategies include an epidemiological

16

approach, the application of radiation damage models, and extrapolation methods for in vitro and in vivo laboratory RBE data to humans. Of these strategies, we believe that the laboratory data are the most promising at this time in establishing defensible neutron RBE data or, ultimately, human risk coefficients for neutron exposures. The two distinct methods used in applying laboratory data for neutron RBE to available high dose/dose-rate low-LET human data are:

(a) applying laboratory RBEM values directly to human risk factors that have been corrected by a DDREF; and

(b) applying laboratory high dose/dose-rate RBE values corrected by the DDREF for humans to the low dose/dose-rate case.

Of these two methods, it is our opinion that the second is probably superior because this estimate for human neutron RBE is not influenced by the large uncertainties associated with the RBEM values for laboratory animal data.

D. CONSIDERATIONS IN SELECTING Q

The ultimate decision in selecting a suitable value for neutron Q must rely on a substantial amount of subjective analysis and judgement, making the rationale for the decision somewhat difficult to explain. When few data were available, little analysis was required, and the basis for recommendations was weak. Over time, new laboratory data, as well as fundamental changes in the A-bomb survivor databases, have prompted a reevaluation of the scientific basis for Qn.

The joint ICRP/ICRU Task Group that prepared ICRU Report 40 (1986) provided a thoughtful and detailed discussion of the questions surrounding the choice of a numerical factor for Q. In considering these questions, the task group identified "certain basic principles along with consequent constraints" that "must be maintained for an acceptable system of radiation protection." These principles and constraints are as follows:

(a) The system should result in approximately equal risk (probability of effect) per unit dose equivalent of low- and high-LET radiations at or below the dose limits. The effects of principal concern are the induction of cancer and hereditary effects. Nonstochastic effects present a different case. Furthermore, the manner in which risk estimates are derived from human data must be taken into account (e.g., allowance, if any, for dose rate).

(b) Optimal use should be made of available information in experimental radiobiology and biophysical theory.

(c) Ideally, the system should permit implementation by both calculations and measurements. Either approach may be suitable when protection against

17

identified radiations is involved, but only the latter is possible in radiation fields of unknown composition.

(d) Simplifying assumptions should be permitted in view of the various unavoidable inaccuracies in the assessment of doses in radiation protection practice. However, these should not result in appreciable underestimates of the risk (e.g., by no more than a factor of about 2).

(e) The basic rationale for recommendations should be stated as clearly as possible.

These principles provide a framework for consideration prior to making any changes in the present system. The requirement that the "system should result in approximately equal risk (probability of effect) per unit dose equivalent of low- and high-LET radiations at or below dose limits" has not been demonstrated by either ICRP or NCRP. Although both NCRP (1993) and ICRP (1990) have recommended a doubling of the Q for neutrons, neither has provided a scientific basis for asserting that the biological effectiveness of neutrons is twice as great as was previously thought. Although both provided a table of RBEM values as the basis for the proposed change, neither included analyses of the uncertainties and assumptions associated with the chosen entries, of the completeness of these data, or of the basis for extrapolating these results to man.

Although it is recognized that large values of RBEM result mainly from the behavior of the reference radiation, little or no analysis has been reported comparing the shape of the dose-response function for low-LET reference sources used to generate RBEM values and the dose-response function observed for the A-bomb survivor cohort (Straume 1988; NCRP 1990). The apparent lack of a quadratic component in the human data (except for leukemias, when analyzed separately from all other cancers) suggests that RBEM values for the whole human system will be significantly smaller than for other species or for the human cellular endpoints cited by NCRP and ICRP.

In developing its recommendations in Publication 60 (ICRP 1990), ICRP assumed a relatively low DDREF for humans (DDREF=2). Table 3 below compares DDREF values, as calculated using Eq. (14), for A-bomb survivor data (high dose/dose-rate conditions) and for human chromosome aberration data. Because the DDREF depends on the dose from which the linear extrapolation is made, it is calculated for a variety of dose points (0.3, 1.0, and 3.0 Gy) corresponding to the range of the most important human risk data. Note that if the dose-response curve is linear, by Eq. (14) the DDREF is calculated to be 1.0. However, as shown in Table 3, DDREF values, as calculated from the A-bomb survivors data, vary from 1.3 to 4.4 in the dose range 0.3 to 3.0 Gy for leukemia, while for solid tumors the DDREF appears to be independent of dose and equal to 1.0 (NAS/NRC 1990). As indicated by this table, DDREF values are much larger for human chromosome-aberration laboratory data than for the A-bomb survivors data, particularly if ^Co is used as the reference radiation.

18

Table 3. DDREF for various endpoints as calculated from different dose points.2

Endpoint

Dose

Endpoint 0.3 Gy 1.0 Gy 3.0 Gy

A-bomb survivors Solid tumorsb Leukemiasc

1.0 1.3

1.0 2.1

1.0 4.4

Human chromosome aberrations'1 250 kVp x rays ^Co

1.4 2.0

2.3 4.2

4.9 10.5

a DDREF values calculated using Eq. (14) for low-LET radiations, where D is the dose from which one extrapolates to low dose; that is, to the linear range of the dose-response curve.

b In BEIR V (NAS/NRC 1990), human solid-tumor data was fit to a linear dose-response curve. However, the BEIR committee judged that some account should be taken of dose-rate effects and suggested a range of dose-rate reduction factors that may be applicable and provided a table of values for doing so. Although their suggested single best estimate for tumorigenesis is 4, based on laboratory animal studies, they appear to have endorsed a value of 2 for the DDREF.

c Coefficients used for human leukemia calculations are a = 0.243 Gy"1 and j3 = 0.271 Gy"2 (NAS/NRC 1990). The use of this linear-quadratic dose-response relationship contains an implicit DDREF of 2 for low doses.

d Coefficients used for human chromosome aberrations are listed in Table 2.

It should be noted, however, that the dose-response relationship for gamma ray-induced chromosome aberrations appears to have a larger quadratic component, and thus a larger inferred DDREF, than the dose response observed for cancer induction in the A-bomb survivors. This suggests that, for low-LET radiation, multi-track events are relatively more important in inducing chromosome aberrations than cancer. On the other hand, in the case of neutron radiation, damage induced by single tracks is expected to predominate, regardless of biological endpoint. In view of the difference in DDREFs, the use of the RBEM derived from data on chromosome aberrations is likely to overestimate the appropriate value of Q n for cancer induction.

To select an appropriate RBE value, on which to base Q, one must ensure that the functional relationships that underlie the RBE are compatible with those from which human risk is determined.

ICRU defined radiation protection quantities that apply to operational radiation protection measurements resulting from the use of personal dosimeters, such as Hp(d), the personal dose equivalent, and area monitoring instrumentation, such as H*(d), the ambient dose equivalent

19

(ICRU 1988; 1993). These new quantities are related to the effective dose equivalent, the quantity selected by ICRP and promulgated by the Environmental Protection Agency (EPA 1987), the Nuclear Regulatory Commission (NRC 1991), and the Department of Energy (DOE 1993). As can be seen in Figure 6, the ratio of the effective dose equivalent to the individual dose equivalent is close to unity for all orientations of the photon radiation field, except when the dosimeter is worn on the side of the body opposite to the direction of the radiation. The same ratio, however, is less than 0.6 for neutrons with energies below 2 MeV (Figure 7). For neutron spectra in the 10 keV to 1 MeV range (i.e., the energy range for which the bulk of occupational exposure to neutrons occurs), this ratio is between 0.2 to 0.5. For planar isotropic fields (i.e., conditions that often apply because of neutron scattering characteristics and the likelihood that a worker is moving and facing different directions in normal occupational environments), the ratio is lower by approximately a factor of 2. The implication of these definitions is that there is an inherent conservatism by a factor of at least 2 to 5 for neutrons in the energy range of 10 keV to 1 MeV, due to the response of a properly calibrated personal dosimeter. Except for energies below 25 keV, there is no equivalent conservatism for photons.

10*

102

_ 10 o X

ai X 10°

10"

10"

: Effective Dose Equivalent [ICRP 26/3o] X

X

* * R X

: 6 •

1 1 0

• a

'. a

• AP - Front X AP - Back V PA - Front D PA - Back • LAT - Front Cs. IS. - Front O PLIS. - Front

' i r i i n i l I I i I I n i l i i i

10 -2 10' 10" Photon energy / MeV

10'

Figure 6. Ratio of the effective dose equivalent, HE, to the individual dose equivalent, Hp(10), as a function of photon energy. Two locations for the personal dosimeter are considered: front of the body (Front) and back of the body (Back). Hp(10) is approximated by the dose equivalent at a depth of 10 mm along the central axis in the ICRU sphere. Five geometries are considered in the calculations: AP, broad parallel beam from front to back (anterior-posterior); PA, broad parallel beam from back to front (posterior-anterior); LAT, broad parallel beam from the side (lateral); IS, isotropic field; and PL.IS., planar isotropic field, perpendicular to body axis (ICRU 1988).

20

10' F

- Effective Dose Equivalent

10° k

o X0" 10'1

111

I

10 k-2

10"

•rjB

D D D • • •

DO •

• AP • LAT • PLIS.

iiniiW nniiitl i i mini M I mill iiinnil imiiiil iiiiiml l imn III I I IBI iiinnil i 11mill i iinnil i iiimil

10° 10? itf 103 10* 106 108 107 108

Neutron energy / eV

Figure 7. Ratio of the effective dose equivalent, HE, to the individual dose equivalent, Hp(10), as a function of neutron energy. The personal dosimeter is considered to be worn on the front of the body. Hp(10) is approximated by the dose equivalent at a depth of 10 mm in the ICRU sphere. Three geometries are considered in the calculations: AP, broad parallel beam from front to back (anterior-posterior); LAT, broad parallel beam from the side (lateral); and PL.IS., planar isotropic field, perpendicular to body axis (ICRU 1988).

E. CONCLUSIONS

1. The quality factor, Q, (or its replacement, the radiation weighting factor, WR) is defined by national and international scientific advisory bodies and by regulatory authorities in

21

the United States as that factor which normalizes the risks of different types of ionizing radiation. The types of radiation, for which the chosen value of the quality factor is unity, have been employed as "reference radiation for radiation protection" purposes. Although there are considerable differences in the effectiveness of different types of low-LET radiation in producing demonstrable biological or health endpoints, particularly at low exposure levels, the "reference radiation for radiation protection" has not yet been uniquely or authoritatively defined. This ambiguity has contributed significantly to the confusion in the debate regarding the most appropriate value for Qn.

2. Current assessments of health risks associated with exposure to ionizing radiation (both low- and high-LET) rest primarily on human exposure databases involving single acute exposures to primarily the low-LET radiation experienced by the Hiroshima/Nagasaki A-bomb survivors. The reliability of the extrapolation of these risk estimates from relatively high exposure levels to the much lower and protracted occupational exposure levels depends on the reliability of the assumed shape for the dose-response relationship (e.g., linear, linear-quadratic, etc.), as well as on the reliability of the assumptions regarding the magnitude of the corrections necessary to account for differences in dose and dose rate (DDREF). Dose and dose-rate effects are also important for extrapolation to RBEM values as applied to low-dose exposure conditions. These extrapolations result in the largest uncertainty associated with the assessment of risks to those exposed in the workplace.

3. Estimates of Q based on values of RBE are highly model-dependent. One approach to the problem of determining the quality factor for neutrons (QJ, applicable to exposure levels of interest to radiation protection, relies on an analysis of the maximum RBE (RBEM) values measured in the laboratory. This approach involves using a wide variety of in vivo and in vitro experimental systems with an equally wide variety of biological endpoints. The principal advantage of this approach is that the focus for the analysis of biological data is at exposure levels that are more relevant to those normally encountered in the workplace. However, this approach has several disadvantages. First, conclusions drawn using this approach are model-dependent because the values for RBEM are often determined by extrapolation rather than by direct measurement. Second, the range of RBEM values for a given reference radiation is largely due to the variety of biological systems and endpoints; that is, the values do not appear to converge toward a single number. This wide variation in numbers introduces the difficulty of determining an acceptable method for identifying biological endpoints with the greatest relevance to human risks, so as to reduce the wide range of values to a single quantity meaningful for use in radiation protection applications. Third, the quality factor, and, therefore, the risks due to neutrons are overestimated if the DDREF observed in the laboratory is larger than the DDREF used for human risk assessment for low-LET radiation.

4. As an alternative to using RBEM, the value of the quality factor can be based on RBE values obtained from linear fits to high dose/dose-rate data. The advantage of this approach is that it is based on data having relatively smaller uncertainties. Furthermore, this approach is more compatible with the exposure conditions from which radiation protection standards are derived (i.e., human risk data from the epidemiologic studies of the A-bomb

22

survivors). However, in the case of occupational radiation protection, it is necessary to correct this high-dose, high-dose-rate RBE by the human DDREF determined for low-LET radiation.

5. The "assumed common effectiveness of all low-LET radiations," while undoubtedly convenient for health physics-monitoring purposes, is not useful as a basis for a reference radiation used in establishing Q for high-LET radiations because the assumed commonality contains too much internal uncertamty. If one accepts that a "reference radiation for radiation protection" based on protracted gamma rays is the basis of Q values, then Q values greater than 10 for neutrons would probably follow due to the lower effectiveness of the reference radiation. It should then also follow that if Q is selected to be 1 for protracted gamma rays, one could also argue from experimental data that the Q for x rays should be 2 or 3. As NCRP states:

The RBE of some radiations of different LET, but within the range of LET encompassed by the "reference" radiation (j8~, j8+, e"1, y rays and x rays), are significantly different from unity. (This may require a more precise definition of the reference radiation for use in radiation protection, and perhaps the use of fractional values of Q for some of these radiations.) (NCRP 1990)

Therefore, in relation to the neutron quality factor, there appears to be an inconsistency in recommending a change by a factor of 2, based on a method that defines different reference radiations as equivalent, when the reference radiations may differ from one another in their biological effectiveness by a factor of 2 or more.

6. The current quantities used in the calibration and reporting of individual exposures to neutrons have a significant degree of conservatism not present when monitoring for photon exposures. This conservatism, estimated to be in the range of at least a factor of 2 to 5, is a consequence of the definition of the relevant radiation protection quantities and is not related to the characteristics of the dosimeters.

F. RECOMMENDATIONS

1. Recognizing that there are insufficient low-dose data to support a definitive value of Qn, we recommend that Federal agencies continue to explore theoretical, experimental, and epidemiological avenues to fill this void, or explore alternative systems to the DQH system in ensuring radiation protection associated with neutrons.

2. In view of the Concluding Remarks above, we see no compelling basis for a change in Q for neutrons at this time. The present values of the absorbed dose or dose-equivalent relationships for neutrons, nominally a Qn value of 10, used by the Federal agencies should be maintained until such time as a need for change is firmly established.

23

24

REFERENCES

Abrahamson, S. A radiobiological rationale for applying RBE to neutron exposures in A-bomb data. RERF Update 1(1); Spring 1989a.

Abrahamson, S. Reply to Dr. Madhvanath. RERF Update 1(2); Summer 1989b.

Ainsworth, E.J.; Fry, R.J.M.; Grahn, D.; Williamson, F.S.; Brennan, P.C.; Steamer, S.P.; Carrano, A.V.; Rust, J.H. Late effects of neutron or gamma irradiation in mice. In: Biological Effects of Neutron Irradiation. Vienna: International Atomic Energy Agency; 1974:359-379.

Barendsen, G.W. Mechanisms of action of different ionizing radiations on the proliferative capacity of mammalian cells. In: Cole A., ed. Theoretical and experimental biophysics. Vol I. New York: Dekker; 1967:167.

Bird, R.P.; Rohrig, N.; Colvett, R.D.; Geard, C.R.; Marino, S. Inactivation of synchronized Chinese hamster V-79 cells with charged particle track segments. Radiation Research 82:277; 1980.

Boag, J.W. The relative biological efficiency of different ionizing radiations. National Bureau of Standards Report 2946. 1953:147.

Casarett, G.W.; Braby, L.A.; Broerse, J.J.; Elkind, M.M.; Goodhead, D.T.; Oleinick, N.L, Biological effectiveness of neutrons: research needs. Report submitted to the Committee on Interagency Radiation Research and Policy Coordination. Report No. ORAU 94/D-58. Washington, D.C.: CIRRPC; 1992.

Curtis, S. Lethal and potentially lethal lesions induced by radiation — a unified repair model. Radiation Research 106:252-270; 1986.

DOE—see Department of Energy

Dennis, J. A. The relative biological effectiveness of neutron radiation and its implications for quality factor and dose limitation. Progress in Nuclear Energy 20:133-149; 1987.

Department of Energy. 10 CFR Part 835. Occupational Radiation Protection. 1993.

EPA—see Environmental Protection Agency

25

Edwards, A.A.; Lloyd, D.C.; Purrott, R.J.; Prosser, J.S. The dependence of chromosome aberration yields on dose rate and radiation quality. In: National Radiological Protection Board. Research and development report, 1979-1981. Chilton, Oxon, U.K.: NRPB; R&D 4:83; 1982.

Environmental Protection Agency. Radiation protection guidance to federal agencies for occupational exposure—recommendations approved by the president. Federal Register 52(17): 2822-2834; 1987.

Fry, R.J.M.; Garcia, A.G.; Allen, K.H.; Sallese, A.; Staffeldt, E.; Tahmisian, T.N.; Devine, R.L.; Lombard, L.S.; Ainsworth, E.J. The effect of pituitary isografts on radiation carcinogenesis in the mammary and Harderian glands of mice. In: Biological and environmental effects of low-level radiation. Vol. I, 213. IAEA Publication STI\PUB\409. Vienna: International Atomic Energy Agency; 1976.

Grahn, D.; Lee, C.H.; Farrington, B.F. Interpretation of cytogenetic damage induced in the germ line of male mice exposed for over 1 year to 2 3 9Pu alpha particles, fission neutrons, or ^Co gamma rays. Radiation Research 95:566-583; 1983.

Grahn, D.; Lombard, L.S.; Carnes, B.A. The comparative tumorigenic effects of fission neutrons and cobalt-60 7 rays in the B6CFj mouse. Radiation Research 129:19-36; 1992.

Grahn, D.; Thomson, J.F.; Carnes, B.A.; Williamson, F.S.; Lombard, L.S. Comparative biological effects of low dose, low dose rate exposures to fission neutrons from the JANUS reactor or to ^Co gamma rays. In: Maruyama, Y.; Beach, J.L.; Feola, J.M., eds. Proceedings of the international neutron therapy workshop. Vol 2. New York: Harwood Academic Publications; 1986:385-396.

Hill, C.K.; Buonoguro, F.J.; Myers, C.P.; Han, A.; Elkind, M.M. Fission-spectrum neutrons at reduced dose rate enhance neoplastic transformation. Nature 298:67-68; 1982.

Hill, C.K.; Carnes, B.A.; Han, A.; Elkind, M.M. Neoplastic transformation is enhanced by multiple low doses of fission-spectrum neutrons. Radiation Research 102:404-410; 1985.

ICRP—see International Commission on Radiological Protection

ICRU—see International Commission on Radiation Units and Measurements

International Commission on Radiation Units and Measurements. Report of the International Commission on Radiation Units and Measurements. ICRU Report 8 (National Bureau of Standards Handbook 62). Washington, D.C.: ICRU; 1957.

International Commission on Radiation Units and Measurements. Conceptual basis for the determination of dose equivalent. ICRU Report 25. Washington, D.C.: ICRU; 1976.

26

International Commission on Radiation Units and Measurements. Radiation quantities and units. ICRU Report 33. Bethesda, MD: ICRU; 1980.

International Commission on Radiation Units and Measurements. Microdosimetry. ICRU Report 36. Bethesda, MD: ICRU; 1983.

International Commission on Radiation Units and Measurements. Determination of dose equivalents resulting from external radiation sources. ICRU Report 39. Bethesda, MD: ICRU; 1985.

International Commission on Radiation Units and Measurements. The quality factor in radiation protection. Report of a joint task group of the ICRP and ICRU to the ICRP and ICRU. ICRU Report 40. Bethesda, MD: ICRU; 1986.

International Commission on Radiation Units and Measurements. Determination of dose equivalents from external radiation sources. Part 2. ICRU Report 43. Bethesda, MD: ICRU; 1988.

International Commission on Radiation Units and Measurements. Quantities and units in radiation protection dosimetry. ICRU Report 51. Bethesda, MD: ICRU; 1993.

International Commission on Radiological Protection. Recommendations of the International Commission on Radiological Protection (revised December 1, 1954). British Journal of Radiology (Supplement 6); 1955.

International Commission on Radiological Protection. Report of ICRP Committee U on permissible dose for internal radiation (1959). Health Physics 3:1-233; 1960.

International Commission on Radiological Protection. The RBE for high-LET radiations with respect to mutagenesis. ICRP Publication 18. Oxford: Pergamon Press; 1972.

International Commission on Radiological Protection. Recommendations of the International Commission on Radiological Protection. ICRP Publication 26. Annals of ICRP 1, No. 3. Oxford: Pergamon Press; 1977.

International Commission on Radiological Protection. Statement from the 1985 Paris meeting of the ICRP. In: ICRP Publication 45. Annals of ICRP 15, No. 3. Quantitative basis for developing a unified index of harm. Oxford: Pergamon Press; 1985:i-ii.

International Commission on Radiological Protection. 1990 recommendations of the International Commission on Radiological Protection. ICRP Publication 60. Oxford: Pergamon Press; 1990.

27

Katz, R.; Ackerson, B.; Homayoonfar, M.; Sharma, S.C. Inactivation of cells by heavy ion bombardment. Radiation Research 47:402-425; 1971.

Kellerer, A.M.; Rossi, H.H. RBE and the primary mechanism of radiation action. Radiation Research 47:15-34; 1971.

Lloyd, D.C.; Edwards, A.A. Chromosome aberrations in human lymphocytes: effect of radiation quality, dose, and dose rate. In: Ishihara, T.; Sasaki, M.S., eds. Radiation-induced chromosome damage in man. New York: A.R. Liss; 1983:23-49.

NAS/NRC—see National Academy of Sciences/National Research Council

NCRP—see National Council on Radiation Protection and Measurements

NRC—see Nuclear Regulatory Commission

National Academy of Sciences/National Research Council. Health effects of exposure to low levels of ionizing radiation (BEIR V). Report of the advisory committee on the biological effects of ionizing radiation. Washington, D.C.: National Academy Press; 1990.

National Council on Radiation Protection and Measurements. Protection against neutron radiation. NCRP Report No. 38. Bethesda, MD: NCRP; 1971.

National Council on Radiation Protection and Measurements. Influence of dose and its distribution in time on dose-response relationships for low-LET radiations. NCRP Report No. 64. Bethesda, MD: NCRP; 1980.

National Council on Radiation Protection and Measurements. Recommendations on limits for exposure to ionizing radiation. NCRP Report No. 91. Bethesda, MD: NCRP; 1987.

National Council on Radiation Protection and Measurements. The relative biological effectiveness of radiations of different quality. NCRP Report No. 104. Bethesda, MD: NCRP; 1990.

National Council on Radiation Protection and Measurements. Limitation of exposure to ionizing radiation. NCRP Report No. 116. Bethesda, MD: NCRP; 1993.

Nuclear Regulatory Commission. 10 CFR Part 20 et al., Standards for protection against radiation: final rule. Federal Register 56(98): 23360-23475; 1991.

Pacific Northwest Laboratory. The impact of a proposed change in the maximum permissible limits for neutrons to radiation protection programs at DOE facilities. Document: PNL-3973/UC 41. September 1981.

28

Preston, D.L.; Pierce, D.; Vaeth, M. Neutrons and radiation risk: a commentary. RERF Update 4(4); Winter 1992-1993.

Roesch, W.C., ed. U.S.—Japan joint reassessment of atomic bomb radiation dosimetry in Hiroshima and Nagasaki. Vol I. Hiroshima, Japan: Radiation Effects Research Foundation; 1987.

Shimizu, Y.; Kato, H.; Schull, W.J.; Preston, D.L.; Fujita, S.; Pierce, D.A. Studies of the mortality of A-bomb survivors. 9. Mortality: 1950-1985. Part 1. Comparison of risk coefficients for site-specific cancer mortality based on the DS86 and T65DR shielded kerma and organ doses. Radiation Research 118:502-524; 1989.

Storer, J.B.; Mitchell, T.J. Limiting values for the RBE of fission neutrons at low doses for life shortening in mice. Radiation Research 97:396-406; 1984.

Straume, T. A critical analysis of neutron RBE, risk and Q. Radiation Protection Dosimetry 23: 57-61; 1988.

Straume, T.; Egbert, S.D.; Woolson, W.A.; Finkel, R.C.; Kubik, P.W.; Gove, H.E.; Sharma, P.; Hoshi, M. Neutron discrepancies in the DS86 Hiroshima dosimetry system. Health Physics 63(4); October 1992.

Thomson, J.F.; Grahn, D. Life shortening in mice exposed to fission neutrons and gamma rays. VII: Effects of 60 once-weekly exposures. Radiation Research 115:347-360; 1988.

Thomson, J.F.; Williamson, F.S.; Grahn, D. Life shortening in mice exposed to fission neutrons and gamma rays. V: Further studies with single low doses. Radiation Research 104:420-428; 1985.

Thomson, J.F.; Williamson, F.S.; Grahn, D. Life shortening in mice exposed to fission neutrons and gamma rays. VI: Studies with the white-footed mouse, Peromyscus leucopus. Radiation Research 108:176-188; 1986.

Thomson, J.F.; Williamson, F.S.; Grahn, D.; Ainsworth, E.J. Life shortening in mice exposed to fission neutrons and gamma rays. I: Single and short-term fractionated exposures. Radiation Research 86:559-572; 1981.

Tobias, C.A. The repair-misrepair model in radiobiology: comparison to other models. Radiation Research 104:S77-S95; 1985.

Tobias, C.A.; Blakely, E.A.; Ngo, F.Q.H.; Yang, T.C.H. The repair-misrepair model of cell survival. In: Meyn, R.E.; Withens, H.R., eds. Radiation biology and cancer research. New York: Raven Press; 1980:195-230.

29

UNSCEAR—see United Nations Scientific Committee on the Effects of Atomic Radiation

Ullrich, R.L. Tumor induction in BALB/c female mice after fission neutron or gamma irradiation. Radiation Research 93:506-515; 1983.

Ullrich, R.L.; Preston, R.J. Myeloid leukemia in male RFM mice following irradiation with fission spectrum neutrons or gamma rays. Radiation Research 109:165-170; 1987.

Ullrich, R.L.; Jernigan, M.C.; Cosgrove, G.E.; Satterfield, L.C.; Bowles, N.D.; Storer, J.B. The influence of dose and dose rate on the incidence of neoplastic disease in RFM mice after neutron irradiation. Radiation Research 68:115-131; 1976.

United Nations Scientific Committee on the Effects of Atomic Radiation. Genetic and somatic effects of ionizing radiation. 1986 Report to the General Assembly (with annexes). New York: United Nations; 1986.

Upton, A.C.; Randolph, M.L.; Conklin, J.W. Late effects of fast neutrons and gamma rays in mice as influenced by the dose rate of irradiation: induction of neoplasia. Radiation Research 41:467-491; 1970.

Vogel, H.H.; Dickson, H.W. Mammary neoplasia following acute and protracted irradiation with fission neutrons and ^Co gamma rays. Radiation Research 87:453-454; 1981.

Zaider, M. Evidence of a neutron RBE of 70 (± 50) for solid-tumor induction at Hiroshima and Nagasaki and its implications for assessing the effective neutron quality factor. Health Physics 61(5):631-636; 1991.

Zaider, M.; Brenner, D.J. On the microdosimetric definition of quality factors. Radiation Research 103:302; 1985.

30

APPENDICES

APPENDIX A

ICRP RECOMMENDED VALUES

Table Al. Relationship between collision stopping power in water (Lm) and quality factor (Q) (ICRP 1977).

Lm in water (keV/^tm) Q

3.5 (and less) 7

23 53

175 (and above)

1 2 5

10 20

Table A2. Effective Q values (ICRP 1977).

X rays, y rays, and electrons 1

Neutrons, protons, and singly-charged particles of rest mass greater than one atomic mass unit of unknown energy

10

Alpha particles and multiply-charged particles (and particles of unknown charge) of unknown energy

20

Table A3. Specified Q-L relationships (ICRP 1990).

Unrestricted linear energy transfer, L, in water (keV//zm)

Q(L)

< 10 10-100 > 100

1 0.32L-2.2 300/VL

A-l

Table A4. Radiation weighting factors3 (ICRP 1990).

Type and energy rangeb Radiation weighting factor, (WR)

Photons, all energies 1

Electrons and muons, all energies0 1

Neutrons, energy < 10 keV 10 keV to 100 keV

> 100 keV to 2 MeV • > 2 MeV to 20 MeV

> 20 MeV

5 10 20 10 5

Protons,d other than recoil protons, energy >2MeV

5

Alpha particles, fission fragments, heavy nuclei

20

aAll values relate to the radiation incident on the body or, for internal sources, emitted from the source.

'The choice of values for other radiations is discussed in paragraph A14 of ICRP 1990. 'Excluding Auger electrons emitted from nuclei bound to DNA (see paragraph 26 of ICRP

1990). dNCRP assigns wR a value of 2 for >2 MeV protons and a value of 1 for >100 MeV

protons (see NCRP 1993).

A-2

APPENDIX B

AGENCY PARTICIPANTS IN THE DEVELOPMENT OF THE REPORT ON THE NEUTRON QUALITY FACTOR

(1985 TO PRESENT)

Department of Commerce

Dr. Randall S. Caswell Mr. Elmer Eisenhower

Department of Defense

CPT David George LTC J. Christopher Johnson (U.S. Army)

CAPT Eric E. Kearsley (U.S. Navy) CDR Kenneth Groves

Department of Energy

Dr. Judith D. Foulke Dr. Marvin E. Frazier Dr. Robert G. Thomas

Environmental Protection Agency

Mr. David E. Janes Dr. Jerome S. Puskin

Department of Health and Human Services

Dr. John D. Boice, Jr.

Nuclear Regulatory Commission

Mr. Robert E. Alexander

ORAU/CIRRPC (Technical Program Liaison)

Dr. William A. Mills Dr. Susan M. Langhorst

B-l

APPENDIX C

CURRENT MEMBERSHIP OF THE CIRRPC SCIENCE PANEL

Department of Agriculture

Mr. John T. Jensen

Department of Commerce

Dr. Bert M. Coursey Dr. Lisa R. Karam (Alt.)

Department of Defense

Mr. D. Michael Schaeffer Dr. E. John Ainsworth, Acting Chairman

Department of Energy

Dr. Matesh N. Varma Dr. Marvin E. Frazier (Alt.)

Department of Health and Human Services

Dr. Gilbert W. Beebe Dr. Bruce W. Wachholz (Alt.)

Department of Housing and Urban Development

Mr. Richard J. Alexander Mr. Joel Segal (Alt.)

Department of the Interior

Dr. Edward R. Landa

Department of Labor

Ms. Margie E. Zalesak (Alt.)

C-l

Department of Transportation

Mr. Richard W. Boyle (Alt.)

Department of Veterans Affairs

Dr. Neil S. Otchin, Acting Executive Secretary

Environmental Protection Agency

Dr. Jerome S. Puskin, Acting Vice Chairman

Federal Emergency Management Agency

Mr. Carl R. Siebentritt Mr. Michael S. Pawlowski (Alt.)

National Aeronautics and Space Administration

Dr. Frank M. Sulzman Dr. Donald E. Robbins (Alt.)

Dr. Walter Schimmerling

Nuclear Regulatory Commission

Dr. Donald A. Cool

National Science Foundation

Dr. Kamal Shukla

C-2