Projected fete cancer nsks @C E Lano er av 36% By population group, the projected number of radiation-related cancers 1s 41, or about 120% of pro- Jected baseline, for the heavily exposed RongelapIsland community (those exposed in 1954 on Rongelap Island and on Ailmgmae), 11% of baselme for the Utnk 2 quantified for any common environmental carcmogen, and we feel reasonably confident about our msk proyections, with a few caveats First, there 1s some evidence that Micronesians, mcludimg Marshallese, may share smmilar cancer patterns, mcludimg igh thyroid cancer community, 22% for the exposed mid-latitude group, and about 0 8% for the southern latitude group, which 1s estimated to have received the lowest radiation doses Except for thyroid cancer, which tends to be diagnosed at younger ages than the majority of cancers, most rates, with native Hawanans (Henderson et al projected to occur after 2008 An exception to this finding 1s for the Rongelap Island and Ailmgmae exposed community which does not include anyone born after 1954 For that cohort, the projected number of tahty m Pacific Islanders, with wide variations m the of the baselme (1e, non-radiation related) cancers are lifetime, radiation-related thyroid cancers (Table 4) 1s 12, or 20 trmes the 06 baseline cases projected in the absence of exposure to radioactive fallout The projected lifetime number of excess leukemia cases 1s | 6, nearly 4 timesthe projected baselme of 045 About 80% of both the excess and baselme thyroid cancers and leukemas are projected to have been diagnosed by the end of 2008 For stomach, colon, and remammg solid cancers, respectively, the excess cancers are estimated to equal 95%, 180%, and 75% of projected baseline values In the results for Utrik, the projected numbers of excess cancers for the relatively small percentage of community members who werenot present on theatoll at the tme of the Bravo test have been mcluded 1mthetotal In contrast to the Rongelap Island and Ailmgmae exposed community, and to a lesser extent the Utnk exposed community, we estrmate that among the members of the mid-lattude and southern latitude populations alive at some time durmg 1948-1970, about 20% were bor after 1954 This difference m age distribution 1s reflected m the fact that proportionally fewer baselme cancers and, except for leukemia, proportionally fewer radiation-related cancers among the mid-latitude and southern latitude populations are projected to have been diagnosed m 2008 or earlier (Table 4) In Table 5, the values m Table 4 have been converted to estimatesof attributable risk, 1 e , the projected proportion of cancers attributable to fallout-related radiation dose, calculated as excess risk divided by the sum of baseline and excess risk, and expressed m percent The values for attributable risk are considered to be the mam result of our analysis DISCUSSION The dose-response relationship between iomzing radiation and subsequent cancer risk 1s among the best 1985) However, an extensive review of published reports of cancersurveillance studies and epidemiological and clinical cancer studies m the native Hawanan and Pacific Islander populations (Hughes et al 2000) found a lack of systematic data collection on cancer meidence and morstatus of cancer research among ethnic groups Thus, baselme cancer rates used im our analysis, which were constructed to be representative of the native population of Hawan, are not necessarily perfectly representative of the MI population The second caveat1s that any static or tme-specific life table, like the US Decenmal Life Tables for 1989-1991 used here (NCHS 1997), corre- sponds to a snapshot m tme andreflects current mortality rates when the hfe table was constructed, which may differ from those 30 y before or 30 y later However, uncertamties in baselme cancer rates and age-specific, all-cause mortahty apply similarly to estimates of excess and baseline risk Therefore, the estimated proportion of cancers attributable to fallout-related radiation dose as presented m Table 5 should be relatively unaffected These considerations aside, our calculations project a substantial burden ofradiation-related cancer m the more heavily-exposed Marshallese population groups, and a correspondmgly lighter burden in the more populous but less exposed atolls m the mid-latitude and southern latitude regions of the MI We project that over half (35%, with 90% uncertamty limits 28% to 69%) of the cancers (since 1948) that have already been diagnosed or may be diagnosed m the future among members of the Rongelap exposed cohortare attributable to their fallout exposure, whereas radiation exposure accounts for less than 2% (1 6% with hmits 04% to 34%) of past and future cancer diagnoses among the exposed MI population as a whole In the exposed MI population, and 1mall population subsets represented m Table 4, the residual category, “other sohd cancers,” which makes up about 80% of basehne risk, 1s projected to account for the largest number of hfetrme radiaton-related cancers However, im terms of “attributable nsk,” or the fraction attributable to radiation exposure, the thyroid gland 1s the smgle organ projected to develop the largest attributable fraction of cancers In the exposed population as a whole, 21% (6% to 39%) ofthyroid cancers are projected to be radiation-related compared to 95% (87% to 97%) among membersof the Rongelap and Ailiginae exposed cohort,