Issued September 30, 1992
Ionizing radiation is not generally known to leave a characteristic marker in those cells that are malignantly transformed and ultimately destined to become an overt malignancy. Thus, the most comprehensive medical examination, and accompanying laboratory tests done on a patient with a malignancy, however valuable they may be in determining the type and extent of the malignancy as well as its optimal treatment and prognosis, rarely provide definite information as to its causation. As a result, it is not possible, on the basis of medical evaluation, to unequivocally prove or disprove a claim that a specific malignancy was caused by a specified radiation exposure.3 Therefore, another basis for judgment as to causation of the malignancy must be sought, despite the fact that new developments in molecular biology may ultimately link specific cancers with specific radiation exposures.
The primary basis must be human data on the excess frequency of malignancies following exposure to ionizing radiation, i.e., risk coefficients derived from epidemiological studies of populations exposed to ionizing radiation. The human data suitable for use in developing risk coefficients have been summarized in several reports, e.g., UNSCEAR (1988), NAS/NRC (1990) and NCRP (1992).4 Such reports review information on the excess incidence and mortality of specific malignancies as a function of radiation dose, sex, age at exposure, and time since exposure.
The probability of causation (PC) approach is an epidemiology-based method of developing an estimate of the probability, rather than proof, that, but for the specific radiation exposure, the specified malignancy would not have appeared.5 The probability that the individual’s malignancy was caused by radiation exposure, the PC, may be estimated from the ratio of the increased incidence of the malignancy attributable to the specified radiation exposure to the sum of the spontaneous malignancy rate and the increase attributable to the specific radiation exposure, i.e., the PC6 in percent is:
PC = AD / (B + AD) × 100
where A is the lifetime radiation risk coefficient for the individual (taking account of age, sex, time since exposure, etc.), B is the appropriate baseline lifetime malignancy rate for the individual, and D is the radiation dose received by the individual.7
The advantages of the PC method include the following: (1) it provides a logical procedure for addressing the issue of radiation induced malignancy that otherwise has no apparent scientific or medical solution; (2) if the radiation dose to the individual concerned can be reliably established, then the PC can be determined by means of a formula; and (3) with group exposures, it provides a quantitative means of determining the merit of a presumption of causation or a ranking of those in a group as to the relative likelihood that their malignancy was radiation induced.
The main limitation of the PC method is that the approach applies epidemiological and demographic data to the individual which involves the assumption that the individual is of average susceptibility, and, in so doing, provides a probability rather than a proof of causation. However, as noted above, the probability can be tailored to the particular individual, in that the age, sex, time since exposure, type of malignancy, and dose (and possibly demographic data related to ethnic groups) can be used to obtain the radiation risk coefficient and the baseline risk that are appropriate for the individual in question.8 In using these data the underlying assumptions must be recognized, i.e., risk estimates are transported from one population to another, observed data are projected to obtain lifetime risks and the risks of high radiation doses are extrapolated to lower radiation doses and dose rates.
Until such time as the knowledge of the etiology of malignancies is improved, for example, through the identification of characteristic molecular markers that indicate the precise cause of a malignancy, the PC approach is a reasonable way to address the problem of evaluating the likelihood that a prior radiation exposure was to blame for the occurrence of a malignancy in an individual. The Council recommends that the PC approach be used as an aid in deciding the question of cause and effect between a malignancy and a specified previous exposure to ionizing radiation.9
1 This statement results from the earlier work of NCRP Scientific Committee 71 on The Probability that an Individual’s Cancer is Causally Related to Previous Radiation Exposure. Serving on Scientific Committee 71 were: V.P. Bond, Chairman; B.D. Breitenstein, R. Catlin, J.C. Hickman, S. Jablon, E.W. Madeira, Jr., B.L. Rich, L.A. Sagan, W.G. Schaffer, G.L. Voelz, Members; and M.A. Bender, P. Groer, M. Miles, and B.E. Titus, Consultants. This statement was drafted by V.P. Bond, B.D. Breitenstein, and W.M. Beckner. Although a draft report prepared by Scientific Committee 71 was circulated to the Council members for review and approval in August of 1986, many of the concepts developed by the Committee had by then been captured in NIH Publication No. 85-2748, Report of the National Institutes of Health Ad Hoc Working Group to Develop Radioepidemiology Tables (NIH, 1985) and the National Academy of Sciences/National Research Council’s Final Report, Assigned Share for Radiation as a Cause of Cancer (NAS/NRC, 1984). As a result, the Council now believes that the remaining need is simply a Statement outlining the rationale for an approach to using the probability of causation (PC) for establishing the relation between an individual’s previous radiation exposure and their later development of a cancer.
2 The many benefits to individuals, and society in general, from the use of ionizing radiation are not the subject of this Statement.
3 There are a few possible exceptions such as an hemangioma of the liver following known thorotrast injection and mastoid carcinoma following ingestion of radon precursors such as that found in the radium dial painters.
4 The studies on which risk estimates have been derived have necessarily involved radiation doses higher than those normally encountered as a result of occupational or diagnostic medical/dental exposure and thus, extrapolation from high doses and dose rates to low doses and dose rates is required.
5 The accuracy of the PC calculation depends on the quality of the input data, the uncertainties of which must be evaluated in each case.
6 The PC calculation does not result in a “true” probability in a statistical sense, i.e., the ratio of the number of outcomes of a particular result to the total number of possible outcomes. Rather, the PC calculation is closer to a conditional probability where the result is conditional on a cancer occurring and there having been a previous radiation exposure.
7 If multiple radiation exposures are involved, the total PC is the sum of the PC for the individual exposures. If carcinogenic agents in addition to radiation exposure are involved, the PC for the radiation exposure may be calculated by adding the risk of the other agents, to the denominator of the PC equation.
8 The question has arisen, are risk coefficients for the U.S. population, when tailored to the individual, applicable to those individuals who have had a previous cancer? Available information is not sufficient to provide an answer except in those cases of known genetic predisposition for a particular cancer; in which case, the use of the tailored, U.S. population risk coefficients would be inappropriate.
9 Medical uses of ionizing radiation include: (1) high doses for the treatment of cancer, (2) moderate doses for other therapeutic and certain diagnostic procedures and (3) minimal doses for most radiologic procedures. The decision to use ionizing radiation is based on the clinical judgment of the patient’s physician as is any other diagnostic or therapeutic procedure. As a result, the anticipated benefit to the patient can be expected to exceed any known potential radiation risk to the patient. However, situations may arise where a patient later develops a condition alleged to be associated with an earlier medical radiation exposure. Provided that an appropriate baseline risk and radiation risk for the patient can be estimated, considering age at exposure, sex, time since exposure, possible genetic predisposition or resistance and medical history, and that the radiation dose(s) and portion(s) of the body exposed are known, the PC method can provide a reasonable estimate of the probability that the earlier medical exposure caused the alleged condition.
NAS/NRC (1984). National Academy of Sciences/National Research Council. Assigned Share for Radiation as a Cause of Cancer (National Academy Press, Washington).
NAS/NRC (1990). National Academy of Sciences/National Research Council. Health Effects of Exposure to Low Levels of Ionizing Radiation, Report of the Committee on the Biological Effects of Ionizing Radiations, BEIR V (National Academy Press, Washington).
NCRP (1992). National Council on Radiation Protection and Measurements. Evaluation of Risks Estimates for Radiation Protection Purposes, Report of Scientific Committee 1-2 (National Council on Radiation Protection and Measurements, Bethesda, Maryland) (in press).
NIH (1985). National Institutes of Health. Report of the National Institutes of Health Ad Hoc Working Group to Develop Radioepidemiology Tables, NIH Publication No. 85-2748 (U.S. Government Printing Office, Washington).
UNSCEAR (1988). United Nations Scientific Committee on the Effects of Atomic Radiation. Sources. Effects and Risks of Ionizing Radiation, 1988 Report to the General Assembly, with Annexes, E.88.IX.7 (United Nations Publications, New York).