Register for Updates | Search | Contacts | Site Map | Member Login

ICRP: Free the Annals!

View Comment

Submitted by Albrecht Kellerer, ICRP C1
   Commenting as an individual
Document 2005 ICRP Recommendation
1 - Diagrams supplied are available to ICRP but cannot be displayed here
2 - The length of the comment exceeds the available size, therefore the comment is split into 3 parts

PART 1 OF 3:
Summary Remarks
The ¡§Draft For Consultation, 2005 Recommendations of the ICRP¡¨ (Draft for Consultation) will serve as an important platform for the necessary discussions before finalisation of the 2005 Rec-ommendations of ICRP. Specifically it offers a chance to correct the major problem, the relation between wR and Q(L), that has caused concern since ICRP 60. The subsequent comments are focus-sed on this problem.
Ralph H. Thomas has addressed the issue in his Commentary on ICRP 92 (draft 10 Aug. 2004 to the Chairman of ICRP). He goes further than the recent report ICRP 92 and favours ¡V with good argu-ments ¡V a more radical correction than the one suggested in ICRP 92. The ICRP Draft for Consul-tation, on the other hand, follows the suggestion in ICRP 92 only halfway, it incorporates the nu-merical changes of the radiation weighting factor wR that ICRP 92 has recommended, but disregards the more essential request in ICRP 92 to re-establish a link between wR and an LET-dependent weighting factor.
Having been co-responsible for ICRP 92, I feel obliged to point out that ¡V in spite of the arguments for a radical correction ¡V the proposition in ICRP 92 is still a viable solution that avoids striking changes in consecutive ICRP recommendations. It is merely necessary to insert into the Draft for Consultation the linkage between wR and an LET-dependent weighting factor. Whether this LET-dependent weighting factor is chosen to be Q(L) (option 1) or ¡V as suggested in ICRP 92 ¡V is taken to be a modification of this function (option 2) that requires less change of the current wR for neu-trons can be left to the decision of the Main Commission.
The Critical Point: wR vs. Q(L)
ICRP 60 introduced in the definition of effective dose, E, a simplified radiation weighting factor, wR. It was initially meant to be a standard value that results in terms of the quality factor, Q(L), for a standard human phantom and a standard directional distribution of a specified incident radiation. No problem would have arisen, if the use of the standard values wR had been encouraged wherever it is convenient, while still admitting the explicit procedure in terms of LET.
However, in the important case of neutrons the numerical derivation of wR from Q(L) had to be crude at the time, and ¡V perhaps for this reason ¡V wR was subsequently dissociated from Q(L). This removed what used to be a common element of the primary quantity and the secondary, operational quantities. The primary quantity, effective dose, E, contains now a radiation weighting factor, wR, with no defined connection to the weighting factor, Q(L) . Only wR is now supposed to be used to determine E or the radiation weighted organ doses. The alternative of using an LET-dependent ¡§lo-cal¡¨ weighting factor, such as Q(L), is currently ruled out, although in complex radiation fields of very high energy it would be expedient. The LET-dependent weighting factor Q(L) is used only in the special operational quantities ambient dose equivalent and personal dose equivalent.
There have been repeated complaints about the resultant dual system of ¡§computed¡¨ and ¡§meas-ured¡¨ dosimetric quantities. A recent article by Thomas [Standards for Standard-Makers? A Testing Time. Rad. Prot. Dosimetry (2004), Vol.109, No.4, 277-289] contains a pointed discussion of the problem, and the issue is specifically addressed in his Commentary on ICRP 92.
The points in the Commentary on ICRP 92 are well taken, but not all relate to decisive issues. Thus there are objections against the notation qE that is used in ICRP 92 for the ¡§effective qual-ity factor¡¨. This is a matter of minor consequence, because the notation qE is chosen ad hoc and not necessarily for general use. Secondly the Commentary on ICRP 92 notes that when the cur-rent convention Q(L) is ¡§compared¡¨ to the values wR for neutrons in ICRP 92, this may be mis-understood as a comparison of individual values Q(L) with values wR, while it refers, of course, to the comparison of the applicable integrals over Q(L) with values wR. But apart from these formal issues, there is emphasis in the Commentary on ICRP 92 on the critical problems that have arisen due to the separation of wR from Q(L).
While the commentary of Thomas speaks for itself, it ought to be pointed out at the outset that both the relevant sections in ICRP 92 and in the Commentary on ICRP 92 (also the comments by other members of the radiation dosimetry community) relate to the same flaw of the current dual system of radiation protection quantities. ICRP 92 and the Commentary on ICRP 92 agree in the diagnosis of the problem, but differ in the proposed solution. The Commentary on ICRP 92 advocates a radi-cal correction. In fact ¡V as far as I can judge ¡V Thomas favours abandoning the concept of wR in favour of the LET-dependent weighting factor Q(L) alone. In effect, this would be a return to ICRP 26, not necessarily in terms of the numbers, but certainly in terms of the concepts. He makes a good case for this proposed radical reversion.
ICRP 92 suggests a compromise, it does not propose to abandon the radiation weighting factor wR altogether. It merely proposes to re-establish coherence between wR and an LET-dependent weight-ing factor, so that both concepts can be used where they are best applicable. The LET-dependent weighting factor is here termed w(L) rather than Q(L), because it need not necessarily be equal to the current convention Q(L). Instead it is suggested that the weighting factor be scaled up from Q(L) in order to minimise the required change in the present values wR for neutrons. The choice of w(L) is partly a matter of judgement. The current radiobiological evidence admits equally the present function Q(L) (option 1), the scaled up function w(L)=1.6 Q(L) ¡V 0.6 (option 2), or some similar relation. It is not precise enough to reject any of these options.
The compromise offered in ICRP 92 differs from the radical proposal for change only insofar as it favours the retention of wR as a convenient standard approximation in all those cases where the ac-tual integration over the LET spectra in the organs of interest is not required. Both approaches agree in offering the use of an LET-dependent weighting factor in refined computations and in certain measurement procedures.
The subsequent Specific Comments to the Draft for Consultation contain on pages 15 and 16 pro-posed textual changes under option 2. If the need for linking wR to an LET-dependent weighting factor is understood, and option 2 is chosen, it will be sufficient to consult the Specific Comments. Otherwise the subsequent considerations may be helpful.

Detailed Discussion of Some Topics Related to the Current Problems
ICRP 92 has emphasised the need to identify the LET-dependent weighting factor ¡V here termed w(L) ¡V that is compatible with wR. This can not be done without certain modifications of the current values wR, because these values are ¡V as seen for neutrons ¡V not compatible with w(L)=Q(L) nor with any other function w(L). A considerable reduction of the values wR for neutrons would be required to make them compatible with Q(L) (see final section, option 1). To avoid this substantial reduction, the suggested LET dependent weight-ing factor w(L) was patterned in ICRP 92 after Q(L), but was scaled up essentially by the factor 1.6. (see final section, option 2).
The ICRP draft document Draft for Consultation accepts, in essence, the numerical changes of wR that are required with option 2, but the function w(L) is not identified. To insert this missing element requires no major change in the Draft for Consultation; the attached specific comments contain a possible formulation. In view of the continued discussions and the more radical solutions that are being proposed, some of the intricacies of the issue deserve to be considered.

The notion of operational quantities:
The effective dose, E, depends not only on the incident radiation, but also on the individual characteristics of the person and on the exposure geometry. This makes it too complex for routine monitoring which is, in-stead, performed in terms of suitably simplified, operational quantities. They provide a (usually conserva-tive) approximation to the primary quantity. The secondary quantities need to be clearly defined, and, to be practical, they ought to be fairly stable. On the other hand, they must be sufficiently flexible to admit adjust-ments as dosimetry techniques improve. As Thomas suggests in his Commentary on ICRP 92:
¡§An even better option would be to abandon the dual concept of protection and operational quantities al-together and define only protection quantities and , ..., leave it to the ingenuity of dosimetrists to deduce the means of measurement¡¨
The particular operational quantities ambient dose equivalent and personal dose equivalent relate to a single reference depth, usually 10mm in tissue. This is crude, but adequate for familiar, comparatively simple radia-tion fields. The relative values of E and of the operational quantities are then known and, in spite of the fact that the quantities are unrelated, one can roughly infer for these fields the effective dose from ambient or personal dose equivalent.
But the increasing importance of complex high energy radiation fields ¡V in the vicinity of accelerators, in flight altitudes, or in space ¡V demands a more flexible approach.

Computation of effective dose:
The earlier quantity effective dose equivalent was defined as:
HE = ƒÃT wT HT (1)
with the organ equivalent dose:
HT = „Å Q(L) DT(L) dL (2)
where DT(L) is the distribution, of absorbed dose in LET in organ T.
With this definition the effective quality factor, i.e. the ratio qE=HE/DE between effective dose equivalent and the tissue weighted absorbed dose, was dependent not just on the composition of the incident radiation but also ¡V although to a rather minor degree ¡V on its directional distribution and on the characteristics of the exposed individual.
ICRP 60 then introduced ¡V as simplification ¡V the radiation weighting factor wR which was initially meant to be a standard effective quality factor, qE, that depends only on the composition of the field. It also introduced the simplified name effective dose and the symbol E:
E = ƒÃT wT HT (3)
with the organ weighted dose now being defined as:
HT = ƒÃR wR DT, R = ƒÃR wR „Å DT,R(L) dL (4)
where DT, R = „Å DT,R(L) dL is the average absorbed dose due to incident radiation of type R, and wR is the radiation weighting factor for radiation type R. DT,R(L) is the distribution, of absorbed dose in LET in organ T from the incident radiation of type R.
Whether Eq(2) or Eq(4) is simpler depends on circumstances. Eq(4) requires the derivation of the values
DT, R , the mean absorbed doses to the various organs T from the different components, R, of the incident radiation. For an incident radiation field with few components R and energies sufficiently low to disregard the charged particle ranges (kerma approximation) the integral in Eq(4) can be replaced by an integral over the initial energies of the liberated charged particles. The equation is then somewhat simpler than Eq(2). On the other hand, with high energy radiations ¡V where the kerma approximation is not applicable ¡V the integra-tion over DT(L) is required anyhow to determine the organ absorbed doses DT, R. The inclusion of the quality factor into the integral is then trivial, and Eq(2) is simpler in so far as it requires no subdivision into radiation types.
Conclusion: The current definition of effective dose offers a certain computational simplification for simple radiation fields of moderate energies. For mixed radiation fields of high energy the computational effort is less with the former definition.