Radiation effects on cancer risks in atomic bomb

Radiation effects on cancer risks in atomic bomb survivors Chelyabinsk October 2, 2012 Dale L. Preston Hirosoft International Eureka, CA

Outline • Historical background of atomic bomb survivor studies background • Outline of major survivor cohorts • Dosimetry • Solid cancer risk estimates • Leukemia risk estimates • (Informal) comparison of survivor and Techa River cohort risks 2

A-bomb Survivor Studies National Cohorts Atomic T 65 D bombings Census Established RERF Joint Francis Tumor created Comm. Registries Comm 1945 1947 1950 1955 1958 1968 1975 DS 86 Blue Ribbon Comm. 1987 1995 DS 02 2004 ABCC Studies Mortality Study (LSS) Cancer Incidence Study (LSS) Clinical Study (AHS) F 1 Mortality Study F 1 Clinical Study (FOCS) 3

Early Studies • Genetic effects (malformations, premature birth, sex-ratio) – 72, 000 registered pregnancies 1948 -1953 – No apparent radiation effects • Leukemia – Initial reports from Japanese physicians in late 1940’s – First published report in 1952 – Initial crude risk estimates in 1956 (not from ABCC) • Solid Cancer – First studies in late 1950’s and early 1960’s – Some indication of effects • Hampered by crude dose estimates and limitations of statistical methods 4

Saving the Survivor Studies • Calls for end to ABCC studies in early 1950’s – Major genetic studies were completed with no compelling evidence of hereditary effects – Leukemia excess risk appeared to be declining – Studies being carried out in ad-hoc manner – Costs for program rising – Staff morale low 5

Saving the Survivor Studies • US National Academies of Science organized committee to assess what should be done about ABCC research • Recommendations – Reorganized program should continue – Unified study plan • Focus on fixed cohorts of survivors and their children with internal comparison groups • Mortality follow-up • Pathology (autopsy) program • Clinical studies • Highlighted need for dose estimates 6

ABCC/RERF Cohorts Life Span Study (LSS) Original LSS includes groups of non-military Japanese for whom followup data could readily be obtained: 1) All survivors' < 2 km with acute effects 2) Matched group of other survivors < 2 km 3) Matched group of people who were 2. 5 -10 km 4) Matched group of unexposed (not-in-city) individuals A-bomb Survivors 284, 000 1950 Census Master Sample 195, 000 Life Span Study 121, 320 Adult Health Study 22, 000 1958 - 7

ABCC/RERF - F 1 study cohorts F 1 Mortality 80, 000 Born between 1947 and 1953 Born between May 1946 and December 1984 FOCS 25, 000 selected, 12, 000 examined Biochemical Genetic studies 28, 000 Untoward pregnancy outcomes 77, 000 8

ABCC-RERF cohorts In-utero cohort Pooled IU cohort 3, 638 people • Pooled cohort combines overlapping clinical (1, 606 members) and mortality (2, 802 members) cohorts. • Mortality and cancer incidence data are available for all members of the cohort. 9

ABCC/RERF Follow-up Programs • Mortality – Based on mandatory nation-wide family registration – Updated on a three-year cycle • Cancer incidence – Hiroshima & Nagasaki tumor registries (1958 – present) – ABCC pathology program 1958 – 1972 – Hiroshima & Nagasaki tissue registries 1973 - present • Leukemia and related disorders – Leukemia registry 1950 – 1987 – Hiroshima & Nagasaki Tumor Registries 1958 – present • Clinical Examinations – Biennial exams – 70 -80% participation through 25 AHS exam cycles – Adapted for use in F 1 clinical study (FOCS) • Mail Surveys – 1965 (Ni-hon-san study men), 1968 (women), 1978, 1991, 201? 10

Dosimetry • Location – Specified as coordinates on fairly crude US army maps • Sought corroboration of location • Recorded to nearest 10 m in each coordinate if detailed shielding history obtained and nearest 100 m for others • External Shielding – Crude shielding category information available on virtually all people of interest – Detailed shielding histories for most survivors within 1. 6 km in Hiroshima and 2 km in Nagasaki • Self shielding (organ dose) – Available for survivors with detailed shielding histories 11

Evolving Dosimetry • Distance and occurrence of acute effects • “Air dose” curves with crude shielding adjustments (T 57 D) • Gamma and neutron air kerma equations with external shielding models (T 65 D) – Based on weapons tests with limited validation from measurements (TLD and Co 60 activation) • Monte-Carlo transport codes for transport and shielding including organ doses (DS 86) – More physical measurements • Updated transport and shielding models (DS 02) – Extensive validation efforts especially for neutrons 12

LSS Survivors within 3 Km Hiroshima 　Nagasaki Hypocenter Dose (m. Sv) ●　< 5 ●　500 – 1000 ●　5 – 100 ●　1000 + ●　100 – 200 ●　200 - 500 ▲　unknown * LSS: Life Span Study Cohort 13

Describing Excess Risks Excess relative risk (ERR) model Excess absolute rate (EAR) model Baseline (zero dose) risk function a age at risk; s gender; and b birth cohort Dose-response shape , e. g. linear, linear-quadratic, threshold, … Effect modification function e age at exposure 14

LSS Solid Cancer Incidence 1958 -94 • Information on gender and age-time patterns depends (only) on radiation-associated (“excess”) cases • Excess cases not explicitly identified • Number of relevant cases is relatively small, especially for specific sites * Attributable risk % for people with doses > 0. 005 Gy 15

LSS Solid Cancer Incidence Dose Response • • • Linear ERR/Gy 0 – 2 Gy 0. 49 No evidence of non-linearity (LQ model on 0 – 2 Gy) P > 0. 5 Curvature 0. 12 LSS-LDEF 1. 14 Using RERF public dataset lssinc 07. csv (www. rerf. or. jp) Proximal zero dose baseline (adjusted for distal and NIC) 16

LSS Solid Cancer Incidence Dose Response 0 – 0. 5 Gy • • Linear ERR/Gy 0 – 2 Gy 0. 49 0 – 0. 1 Gy 0. 51 0 – 0. 15 Gy 0. 51 Test for trend 0 – 0. 1 Gy P = 0. 08 0 – 0. 15 Gy P = 0. 01 Using RERF public dataset lssinc 07. csv (www. rerf. or. jp) Proximal zero dose baseline (adjusted for distal and NIC) 17

LSS Solid Cancer Excess Relative Risk Temporal Patterns Age at exposure -29% per decade (90% CI -39%; -18%) Attained age Age-0. 9 (90% CI -1. 5; -0. 2) Gender * M: 0. 29 F: 0. 58 F: M: 1. 9 * 18 (90% CI 0. 21; 0. 39) (90% CI 0. 42; 0. 68) (90% CI 1. 4; 2. 7) ERR per Sv at age 70 following exposure at age 30

LSS Solid Cancer Excess Rate Temporal Patterns Age at exposure -20% per decade (90% CI -30%; -10%) Attained age Age 3. 5 (90% CI 2. 9; 4. 1) Gender * M: F: 26 28 F: M: 1. 1 * 19 (90% CI 18; 34) (90% CI 23; 34) (90% CI 0. 8; 1. 6) Excess cases per 10000 PY at age 70 following exposure at age 30

Leukemia Incidence 1950 -2001 • • CLL is much less common in Japan than in Russia (or other countries) ATL cases mostly in Nagasaki with no indication of a dose response

LSS Leukemia Incidence • Almost half of cases among exposed associated with exposure 21

Leukemia Incidence • Significant non-linearity in dose response • Low dose risks about half of linear model estimates 22

Leukemia Incidence Temporal Patterns • ERR model with age and age at exposure dependence describes data as well as EAR model • EAR model with only attained age effect fits better than age at exposure and time since exposure model from 1994 (gray curves) 23

Informal Comparison of LSS and Techa River Risks • Solid cancer – ERR risk estimates similar around age 60 – Temporal patterns not significant in TRC but estimates exhibit somewhat different patterns than in LSS (increasing with attained age or age at exposure) – TRC has very limited power to detect effect modifications • Leukemia – ERR risk estimates for non-CLL similar with similar temporal patterns 24

Summary and Conclusions • Accumulating data and modern analytical methods make it possible to investigate radiation effect modification in some detail • Data are limited even in the largest cohort – Especially true when examining a site-specific risks and modeling interactions • Both ERR and EAR descriptions provide equally important and complementary information – Attained age is an important factor in both – Generalization of age at exposure and gender effects can be difficult • Pooled analyses may be useful in looking at effect modification – But LSS results may dominate such comparisons 25

Acknowledgments • We stand on the shoulders of giants Gil Beebe, Seymour Jablon, Jim Neel, Jack Schull • ABCC/RERF scientists and staff who made the ideas a reality George Darling, Howard Hamilton, Tetsuo Imada, Hiroo Kato, M. Kanemitsu, Bob Miller, Kenji Omae, Itsuzo Shigematsu and hundreds more • Collaborators Japan Akio Awa, Harry Cullings, Saeko Fujiwara, Shochiro Fujita, Sachiyo Funamoto, Kyoji Furukawa, Kazunori Kodama, Charles Land, Kiyo Mabuchi, Nori Nakamura, Don Pierce, Elaine Ron, Yukiko Shimizu, Michiko Yamada URCRM / NCI / JCCRER Lyudimila Krestinina, Marina Degteva, Alexander Akleyev, Sara Schonfeld, Elaine Ron, Faye Davis 26