Next | ToC | Prev

Early-Onset Breast-Cancer:   Evidence on Radiation-Induction

          An enormously important issue is breast-cancer which becomes clinically manifest in very young women --- before age 35, for instance. Recent evidence, discussed below, underscores the importance of acting upon the finding of this book, if prevention is everyone's goal. So that all readers (including the "easy readers") can contemplate the meaning of this new evidence, we introduce the units in which radiation doses are measured.
Part 1.   Dose-Units, Especially "Medical Rads"

          The term "medical rad" is the one which we use in reaching the finding of this book. To show what it means, we have to cover some other units.

          Rads.   The "rad" is a unit in which amounts (doses) of ionizing radiation are measured, the way "dozen" measures an amount of eggs. Rad is the abbreviation for "radiation absorbed dose." A rad is really just a ratio of energy delivered by ionizing radiation, per gram of irradiated cells or tissue. A rad is 100 ergs of energy per gram --- a definition which no reader needs to remember. One thousandth of a rad, or 0.001 rad, is a milli-rad. An effort is underway to re-name the rad as a centi-gray (cGy) and to call 100 rads a gray (Gy). We and many others prefer to stick with rads.

          Roentgens.   The Roentgen is a unit for measuring ionization in air;   a few reports also use it for doses inside the body. The abbreviation for Roentgen is R (but a small "r" was customary in the older literature). Under common medical circumstances, an entrance dose of 1 Roentgen at the skin gives a breast-tissue dose of about 0.69 rad, when the beam is traveling from front-to-back of a patient. The Roentgen is the dose-unit often used to describe x-ray dosage from fluoroscopy. In the 1930-1935 era, many fluoroscopy machines could generate beams with dose-rates like 100 Roentgens per minute (Braestrup 1969).

          Rems.   Another unit is called the rem, an abbreviation for "roentgen equivalent, man." The rem can indicate that adjustments have been made for the non-standard "quality" of some radiations. Usually (but not always) the standard is an x-ray of 100 to 400 KeV. An effort is underway to re-name the rem as a centi-sievert (cSv) and to call 100 rems a sievert (Sv).

          Medical Rads.   Per rad, gamma rays from an atomic bomb are about half as harmful as x-rays from medical irradiation, so 1.0 rad of A-bomb gamma radiation could be called 0.5 rem. We avoid rems in this book by converting all doses into "medical rads." Hence, 1.0 rad of A-bomb gamma radiation is called 0.5 medical rad.

Some Very Common Radiation Dose-Levels

          o - Natural Background Radiation. The typical annual dose from natural background radiation, excluding doses from inhaled radon, is about 0.1 rem or 100 milli-rems (BEIR 1972, p.50;   BEIR 1990, p.18). The dose per year rises with altitude. Because the natural background dose is mostly from gamma and cosmic radiation, its medical equivalent per year is about 0.05 medical rad (50 medical milli-rads). In a 70-year lifespan, the cumulative dose is about 3.5 whole-body medical rads. A whole-body dose is received by all parts of the body, in contrast to a partial-body dose.

          The natural dose-rate per minute is of interest, for comparison with dose-rates from fluoroscopy. Nature's dose-rate per minute is (0.05 medical rad / year) times (1 year / 525,600 minutes), or 0.000000095 medical rad per minute from natural background radiation. If we round this off, it is about one ten-millionth of one medical rad per minute. At this dose-rate, only a tiny fraction of cells is irradiated per minute.

          o - Airline Travel.   The extra radiation dose, from flying between the east and west coasts of the USA, is about 0.3 milli-rem (0.0003 rem) per hour of commercial flying. For a ten-hour roundtrip, the extra dose would be about 3 milli-rems (0.003 rem). It would require about 3,300 flying-hours to receive 1.0 extra rem of whole-body irradiation --- equivalent to about 0.5 extra medical rad. Dose-rates from flying vary with altitude and latitude.

Part 2.   The Dose Which Doubles the Rate of Early-Onset Breast-Cancer

          Ordinarily in epidemiological studies of cancer-development from radiation, one faces the "small numbers problem":   An insufficient series of cases in the relatively early follow-up period to permit a reliable conclusion. So time is allowed to run, to accumulate additional cases ("statistical power"). And then all the cases --- short latency and long-latency cases, combined --- are examined together for such issues as cancer-increase per rad of dose. But what is lost, within the accumulated total, is any evaluation of whether the "early-onset" cases are different from the other cases. From here on, the cases diagnosed before age 35 will be called "early-onset" cases.

          With follow-up of the A-bomb survivors now complete for 1950 through 1985, there may be enough total cases of breast-cancer to separate the early-onset cases from the others, and to compare them for induction-rate per rad of irradiation. In 1993, Charles Land, Masayoshi Tokunaga, and additional RERF analysts, made a brief report in the Lancet, entitled "Early-Onset Breast Cancer in A-Bomb Survivors" (Land 1993). By extracting the pertinent data from the figure in their 1993 paper, we can tabulate their findings in the nearby box. They are remarkable results, to say the least!

          In their 1993 paper, Land et al are reporting exclusively on female A-bomb survivors who received the bomb-exposure before the age of 20 years. Although Land et al do not say so, the number of females who were exposed by the bombs below age 20 was approximately 12,000, and their average age at the time of bombing was about 10 years old (calculated from Gofman 1990, Table 26-F). There are 205 incident cases of breast-cancer, reported between 1950-1985 for this group. The cases are segregated by Land et al into two main groups:   Women whose breast-cancers occurred before age 35 years constitute one group, and women whose breast-cancers appeared after age 35 years, the other group. We repeat:   All the women were less than age 20 at the time of exposure.

          In the boxed tabulation, the entries for "fractional increase ... per rad" indicate a spectacular difference between early-onset cases versus cases occurring at age 35 and beyond. The difference is treated as real (not spurious) in both Land 1993 and Tokunaga 1994. We shall propose an explanation of the difference in Part 4. But here, we will focus on the breast-dose which, if delivered sometime before age 20, can double the rate of early-onset breast-cancer. This value can not change with any future observation of the A-bomb survivors, because the story for early-onset cases was over when the youngest survivors (newborn in 1945) passed the age of 35 --- in 1980.

* - Tabulation Based Upon Land et al, 1993.

                Average Age     Number of   Fractional Increase
                When Breast-    Breast-     in Breast-Cancer Rate
                Cancer Occurs   Cancer      over Spontaneous Rate
                                Cases         per rad
Breast Cancers
Occur Before
Age 35 years    32 years         27           0.136  13.6 %
Breast Cancers  41 years         80           0.013   1.3 %
Occur After     49 years         85           0.024   2.4 %
Age 35 years    57 years         13           0.017   1.7 %
Sum of cases, bomb-exposed women 205 cases.

The "Doubling Dose" for Early-Onset Breast-Cancer

          The dose which doubles the spontaneous frequency of early-onset breast-cancer is the dose which causes a 100 % increase in its spontaneous rate. Therefore, to estimate how many rads would cause a 100 % increase, we divide 100 % by 13.6 % per rad (from the boxed tabulation). Thus, 7.35 rads, received before age 20, is the approximate doubling dose for early-onset breast-cancer.

          The doubling dose for early-onset breast-cancer is even lower when we consider medical x-ray radiation, or beta particles of energy comparable to such x-rays. The medical rad has approximately two times the effectiveness of A-bomb radiation from which Land, Tokunaga, and colleagues developed their findings. Therefore, we must warn that the dose of medical rads required to double the spontaneous rate of early-onset breast-cancer must be in the neighborhood of only 3.68 medical rads, received sometime before age 20. As readers will see in Sections 2 and 3 of this book, such doses and far higher ones have been commonly received during childhood from certain medical procedures --- and we do not mean doses from radiation therapy after a cancer has already occurred.

Part 3.   A Test for Agreement about the Magnitude of Risk

          Below, we will show the good agreement between RERF's quantification of risk (by Land, Tokunaga, et al) and our own independent analysis of 1990. The "easy readers" of this book may wish to skip to Part 4 or Part 5, but others will find Part 3 very interesting.

          Our analysis used the data on cancer deaths (all types) from 1950-1982 in the A-Bomb Survivors, and the RERF analysis used exclusively the breast-cancer incidence from 1950-1985 among the A-Bomb Survivors. Both analyses (Gofman and RERF) are finding out the "fractional increase or percent increase above the spontaneous rate, per rad" --- a concept which we named "the K-value" simply for the sake of brevity. Tokunaga et al call the same thing "Excess Relative Risk," or ERR.

          Based on the linear dose-response in 15-G and 15-H of Gofman 1990, our K-value estimate for females exposed at ages 0-9 years is 0.01922 (the same as 1.922 %). Age 0 means from birth to the first birthday. For exposure at ages 10-19, the K-value is 0.01097 (or 1.1 %). When the two values are weighted by the number of cases, the average is 0.01265, or 1.265 % per rad for exposure at ages 0-19. This value is for all types of cancer, combined. For breast-cancer alone, the K-value is 2.524 times the value for combined types (Thompson et al 1994, pages S26, S49, S61). Multiplying the 1.265 % per rad for all cancers, by 2.524 to adjust for breast-cancer alone, we obtain a K-value of 3.2 % per rad for breast-cancer, when bomb-exposure occurred at ages 0-19 years.

          The comparable K-value from RERF's own analysis is 2.41 % per rad --- from Tokunaga 1994, p.215, Table VI. RERF's value is derived from observations of cancer incidence, whereas our value is derived from observations of cancer mortality. When our 3.2 % is divided by their 2.41 %, we see that one estimate differs from the other by only 33 % . The two separate analyses are in remarkably close agreement.

          The similarity in estimates is supportive of the concept that, when extra radiation induces extra cancers, it induces cancers which are fatal and non-fatal in the same proportion as occurs without the extra exposure to radiation.

Part 4.   Early-Onset Cases vs. Later Cases:   Why Such a Difference?

          If readers look back in Part 2 at the boxed tabulation, they will see the column for "average age when breast-cancer occurs" for 205 of the A-bomb survivors irradiated between birth and age 20. For cases diagnosed at an average age of 32, the percent increase per rad above the spontaneous rate is 13.6 %, whereas the percent increase per rad is dramatically lower if diagnosis occurs at the average ages of 41 years, or 49 years, or 57 years.

          What accounts for the striking difference? Is there some big biological difference between the radiation-induction of breast-cancer for cases which appear before age 35, compared with the radiation-induction of cases which appear after age 35? Is there some altered response (to irradiation during childhood) in the irradiated hosts after they pass age 35? Various possibilities are discussed in Tokunaga 1994 (pp.221-222).

          We propose a relatively simple explanation. Our figure, "Three Curves with a Story," demonstrates how the observed difference could arise.

Three Curves with a Story

Three Curves with a Story

Age at Breast-Cancer Diagnosis, in Years.

 Age      A       B       C      |   Age      A       B       C
 29       0       0              |   50      42.5    102     2.4
 30       0.1     1      10.0    |   51      44.0    104     2.4
 31       0.2     2.5    12.5    |   52      45.0    107     2.4
 32       0.2     4.0    20.0    |   53      46.5    108     2.4
 33       0.6     7.5    12.5    |   54      47.0    110     2.4
 34       1.0    11.0    11.0    |   55      47.5    112     2.4
 35       2.0    14.0     7.0    |   56      48.0    113     2.4
 36       3.5    20.0     5.7    |   57      48.5    114     2.4
 37       5.05   25.5     5.0    |   58      49.0    114     2.3
 38       7.5    32.0     4.3    |   59      49.3    114     2.3
 39       9.5    39.5     4.2    |   60      49.6    114     2.3
 40      12.0    45.5     3.8    |   61      49.7    114     2.3
 41      15.0    54.0     3.6    |   62      50.0    114     2.3
 42      18.0    61.0     3.4    |   63      50.0    114     2.3
 43      21.5    68.0     3.2    |   64      50.0    114     2.3
 44      25.0    75.0     3.0    |   65      50.0    114     2.3
 45      28.0    80.0     2.9    |   66      50.0    114     2.3
 46      32.5    86.0     2.6    |   67      50.0    114     2.3
 47      35.5    90.0     2.5    |   68      50.0    114     2.3
 48      38.5    94.0     2.4    |   69      50.0    114     2.3
 49      41.5    98.0     2.4    |   70      50.0    114     2.3

Who Are These Curves?

          o - Curve A:   In the nearby figure, Curve A depicts a spontaneous breast-cancer rate beginning to rise when women reach the age of about 30. We wish to emphasize that Curve A is a "generic" or illustrative curve, with imaginary rates of new cases (per 10,000 women) on the vertical axis. The reality-based aspect of Curve A is its depiction of a rise which is gradual, then steep, and then gradual again.

          For ages 20 through 29, the spontaneous frequency of breast-cancer is very low (shown as "zero" rate). Then, for ages 30, 31, and 32, we have elevated each data-point (really, data-symbol) slightly above the baseline. At age 33, we show the rate of new cases starting to increase in each successive year, with the amount of annual increase greater (steeper) between ages 35 and 50 than between ages 50 and 60, for example. The rates plotted as Curve A are tabulated in the box underneath the figure.

          o - Curve B:   Curve B depicts the corresponding breast-cancer rate in a comparable group of women who received breast-irradiation between birth and age 20. Again, we show imaginary rates of new cases (per 10,000 women) on the vertical axis, and we have not specified any particular radiation dose. What we are showing, however, is reality-based:   Starting at age 30, the irradiated group shows a rate of breast-cancer which is higher than the spontaneous rate. The real-world observations of a higher rate were discussed in Part 2. Among the irradiated A-bomb survivors who were ages 0-19 during the bombings, there are 27 cases of early-onset breast-cancer --- and all but 2 or 3 of those cases were diagnosed when the women were ages 30 through 34 (Tokunaga 1994, p.214). There are 178 additional cases diagnosed after age 35 in that group.

          What we want to explain is why the percent increase per rad is so much higher for the cases diagnosed at the average age of 32 (the early-onset cases) than for cases diagnosed later. Curve C depicts the question.

          o - Curve C:   Curve C shows the result when each cancer-rate of Curve B, for irradiated women, is divided by the corresponding (and lower) cancer-rate of Curve A, for non-irradiated women. The values are also tabulated beneath the figure. These ratios are the relative rates, or Relative Risks (RR), which quantify how many times higher the exposed rate is than the spontaneous rate. We plot the RR values as Curve C, which uses the scale of the vertical axis as an "all-purpose" scale (just don't say "per 10,000 women").

          At age 32, the tabulation shows that the irradiated rate is 20 times higher than the spontaneous rate. To transform this Relative Risk of 20 into percent increase per rad (K-value), we just subtract 1.0 from the Relative Risk and then divide 19 by the dose in rads. Example:   If the average dose which produced Curve B were 110 rads, then the fractional increase per rad for cases diagnosed at age 32 would be (19 / 110 rads), or 0.173 per rad --- which is the same as 17.3 % per rad.

What Is the "Story" of A, B, and C?

          Curves A and B are nice, smooth curves suggesting nothing biologically exotic. Nonetheless, their relationship generates a stunning peak in Relative Risk, and therefore in percent increase per rad, for early-onset cases diagnosed at an average age of 32. After its peak value of 20, Relative Risk declines dramatically --- down to 4.3 for diagnosis at age 38, down to 3.0 for diagnosis at age 44, and then remaining above 2.0 for diagnosis beyond age 44.

          What is the meaning of this peak for early-onset cases?

          When two curves (for irradiated and non-irradiated groups) have been near the zero-rate during a follow-up study, the addition of just a few cancer-cases to one curve earlier than to the other curve has to cause very high Relative Risks --- even when the higher rates are not very high at all. Such events are illustrated for ages 30, 31, and 32 at diagnosis. At age 29 and younger, the breast-cancer rates are shown as equal in the irradiated and non-irradiated groups. Curves A and B are right on top of each other, at the zero-rate. At age 30, both rates increase a little (see the tabulated values):   Non-irradiated moves to 0.1 case per 10,000 women, and irradiated moves to 1 case per 10,000 women --- a very low rate compared with what is coming later. Nonetheless, this slight change means that the risk goes "overnight" from being equal in both groups, to being 10-fold higher in the irradiated group --- all because of 1 case per 10,000 women. At ages 31 and 32, additional small changes drive the Relative Risk to its peak at 20.

          In computing Relative Risks, the spontaneous rates are the denominators of the fractions. As long as the the spontaneous rates remain near zero, even modest growth in the exposed rates will generate enormous Relative Risks. The phenomenon does not happen when analysts look at the women exposed at older ages, because the spontaneous rate is already well above zero when analysts begin comparing the irradiated and non-irradiated groups.

The Key:   A Baseline Rate Near Zero

          For A-bomb survivors irradiated below age 20 and diagnosed with early-onset breast-cancer before age 35, the exceedingly high Relative Risks (and percents increase per rad) are real --- but their initial magnitudes turn out to be temporary. Relative risks compute at a less spectacular level as soon as the spontaneous cancer-rates have their own rapid climb (away from zero) in the denominator of such ratios. We have been able to mirror these observations, in a generic way, with Curves A and B, which generate Curve C. The three curves also mirror the observation that the less spectacular Relative Risk (for cases diagnosed after age 40) persists at an approximately constant level through the 1985 follow-up (see Land's data in Part 2;   confirmation in Tokunaga 1994, p.221).

          In our opinion, "Three Curves with a Story" means this:   Analysts of the A-Bomb Study need not invoke special concepts about the cancers or the hosts, in order to explain the much higher Relative Risk observed for early-onset cancer-cases than for the cases diagnosed at older ages. The observed difference in Relative Risk, due to radiation, can be explained by the fact that both curves rise from a baseline rate which is very close to zero.

True Meaning of "Less Spectacular"

          The "less spectacular" Relative Risk really reflects a much greater number of radiation-induced breast-cancers than the peak Relative Risk. In figures like ours, it is the area under a curve which reflects the aggregate number of cases observed. The area under Curve B minus the area under Curve A reflects the excess number of cases induced by radiation. Clearly, the difference in the two areas for ages 30 through 35 is very small compared with the difference in the two areas beyond age 35. The "less spectacular" Relative Risk not only endures much longer, but it reflects the multiplication (by radiation exposure) of a much higher spontaneous cancer-rate.

Part 5.   The Need for Action

          How many women who have developed early-onset breast-cancer over the past half-century know whether or not they were irradiated in infancy?

          We wonder. Readers of this book are going to learn about a very large number of female children who suffered the fate of such undesirable irradiation of the breasts --- even before they left the nursery of hospitals where they were born. Others suffered nearly the same fate in their first few years of life.

          No one can alter the past. But from the past, we can learn the key to preventing many, many cases of early-onset and later breast-cancer. Prevention.

Time Would Tell ... And Time Has Told

          Once upon a time (1977), the United Nations Scientific Committee on the Effects of Atomic Radiation, known as UNSCEAR, speculated that breast-irradiation during infancy and childhood might have "minimal" cancer-consequences because "only a few breast cells" exist to be irradiated before puberty (UNSCEAR 1977, p.389). Since these few cells must be the progenitors of all future breast-tissue cells, we rejected that line of reasoning (Gofman 1981, p.249-250).

          In 1994, the issue seems settled. Tokunaga et al (1994, p.215, Table VI) report statistically significant excess breast-cancer rates in the bomb-irradiated women who were ages 0-9 years old in 1945 --- as well as in the group which was 10-19 years old in 1945. Both groups, each analyzed separately for the period 1950-1985, show the excess. And both groups have much higher risks per rad than risks for women who were 20 years of age and older at the time of the bombings.

The Irradiation of Children Today

          The era of irradiating children is far from past. The diagnostic use of x-rays can be extremely helpful in pediatric medicine. For example, we are aware that a high number of x-rays may be taken of premature infants and others in neonatal intensive care units (NCRP 1989). The use of x-rays is also high in association with birth defects (especially heart problems). The use of x-rays for children involved in automobile and other accidents can also be high, especially if there are insurance battles and lawsuits.

          The task mandated by the evidence in this chapter is not to stop such examinations, but rather, it is to make sure that the frequency and doses are kept to the minimum really needed. In Section 4 of this book, we have some suggestions about what concerned parents, physicians, and medical schools can do.

# # # # #

Next | ToC | Prev
back to PBC | CNR | radiation | rat haus | Index | Search