903 Benefits versus Risks from Mammography A Critical Reasessment Fred A. Mettler, M.D., M.P.H. Arthur C. Upton, M.D. Charles A. Kelsey, Ph.D. Robert N. Ashby, M.D. Robert D. Rosenberg, M.D. Michael N. Linver, M.D. Department of Radiology, University of New Mexico, Health Sciences Center, Albuquerque, New Mexico. BACKGROUND. The use of mammography has increased rapidly over the last decade. The justification for mammographic examinations is the potential benefit they provide in detecting breast cancer at an early stage and reducing mortality. However, this benefit must be balanced against the associated potential risk of radiation carcinogenesis, economic costs, and a number of other factors. Most publications to date have used radiation risk factors and data from studies that were published over a decade ago, which now have been superseded by the results of more recent epidemiological studies. METHODS. This report examines the current literature regarding the benefits of cancer detection and the risk of radiation carcinogenesis, and calculates the ratio of benefit and risk for women who begin annual mammography screening at different ages. We have used current data to calculate the expected individual benefits and radiation risks associated with annual mammographic screening. RESULTS. It now appears that there is little risk of breast cancer associated with radiation exposure from annual mammography in women over the age of 35, although there is some indication that exposure of younger women may pose a risk for those women in a genetically sensitive subgroup. CONCLUSIONS. New data document that for a woman beginning annual mammographic screening at age 50 and continuing until age 75, the benefit exceeds the radiation risk by a factor of almost 100. Even for a woman who begins annual screening at age 35 and continues until age 75, the benefit of reduced mortality is projected to exceed the radiation risk by a factor of more than 25. Cancer 1996; 77:903-9. 0 1996 American Cancer Society. KEYWORDS mammography, breast cancer, radiation risk, radiation dosage. A Address for reprints: Fred A. Mettler, M.D., M.P.H., University of New Mexico, Health Sciences Center, Department of Radiology, Albuquerque, NM 87131-5336. Received April 20, 1995; revision received July 27, 1995; accepted July 27, 1995. 0 1996 American Cancer Society n increased incidence of breast cancer following exposure to ionizing radiation has been recognized since the 1960s.’ Over the last decade, the literature on radiation carcinogenesis in the breast has come to include new findings, and a number of combined analyses have clarified the benefits of mammography screening. The most common question women ask about mammography screening is “What are the benefits and risks to me?” rather than “What are the results of large population studies?” Published data now allow each woman to evaluate her radiation risk and her benefit in order to decide whether or not to have a mammogram. Recent updates of epidemiologic studies have delineated the risks of radiation-induced breast cancer for women of different ages at exposure. National accreditation with dosimetric evaluation of mammographic facilities over the last three years has provided reliable data concerning the average mammographic dose in the United States. Meta-analysis of a number of mammographic screening studies has also provided new data on the magnitude of the reduced mortality from screening mammography at CANCER March 1,1996 / Volume 77 / Number 5 904 I 25 30 35 40 45 50 55 80 85 70 75 80 85 Attained Age (years) FIGURE 1. Estimated excess relative risk per Sv. by interval of attained age (25-34 years, 35-44 years, 55-64 years, 65-74 years, and 2 7 5 DI exp (/&A) where D is equivalent years) with fitted model ERR (D;A) = C dose in Sv (neutron RBE = 10) and A is attained age. Estimates and 90% confidence limits stratified on city, age at the time of the bombings, attained age, and period. Total number of cases appears above the upper confidence limit for each interval of attained age. Sv: Sievert; CI: confidence interval. (Reprinted with permission from Tokunaga M, Land C, Tukapka S, Nishimori I, Soda M, Akiba S. incidence of female breast cancer in atomic bomb survivors, 1950-1985. Radiat Res 1994;138:209-23.) different ages. With these studies, it has been possible to quantify the cancer mortality risk-benefit ratio. EFFECT OF RADIATION ON THE INCIDENCE OF BREAST CANCER Women in a number of study populations have shown dose-dependent increases in the frequency of breast cancer after irradiation. The largest such population is comprised of women exposed to atomic bomb radiation at Hiroshima and Nagasaki, in whom 295 cases of breast cancer occurred between 1950 and 1987, versus the 200 cases expected in the 510,000 person-years of follow-up study.' The excess relative risk appears to have increased in proportion with the dose, averaging about 1.7 at a dose of 1 sievert (100 rem). It also appears to have varied inversely with age at the time of irradiation (Fig. 11, prompting speculation that the level of hormonal stimulation in women who were exposed at older ages was too low to promote the full expression of carcinogenic effects on their b r e a ~ t sAnother .~ possibility is that proliferating cells in the breasts of young women retain damage or are more susceptible to carcinogens. The Japanese data also showed that in women exposed younger than age 20 who developed cancer before the age of 35, the excess was more than 6 times greater than in those women whose cancers occurred at later ages, implying that women with early disease may represent a genetically susceptible subpopulation.4 Other study populations in whom similar dose-dependent excesses of breast cancer have been reported include women treated with radiation for various medical conditions in adult life, i.e., acute postpartum mastitis,s other benign breast disease^,^" and ankylosing spondylitis'; women treated with radiation for various medical conditions in childhood, i.e., hemangioma of the skin,' tinea capitis," enlargement of the thymus gland at infancy,' ' Hodgkin's disease," and other childhood malignanciesI3;women who received multiple fluoroscopic examinations of the chest in the treatment of pulmonary tuberculosis'4~15; women who received repeated X-ray examinations in the treatment of scoliosis"'; and early radium dial painters.".'* In contrast to the aforementioned populations, women exposed to low levels of radiation, such as those exposed to elevated environmental concentrations of radioactive weapons fallout,'"20those residing in the vicinity of nuclear power plants," those employed as radiation workers (other than radium dial painter^),".^" and those injected with diagnostic doses of radioactive iodine,24 have shown no consistent or significant increases in breast cancer rates. Risk of Causing Breast Cancer by Mammography Dose From Mammography The American College of Radiology has been accrediting mammography practices for several years. As part of this practice, dosimetry has been carried out, and at a May 1994 conference of the College, the average dose at accredited mammography facilities in the United States was reported to be 1.38 mGy per view, or typically double this dose (2.8 mGy) per examination, because 2 views of each breast are typically obtained (Pamela Wilcock, ACR, personal communication). These values are within the range reported in a number of other countries, for which the mean glandular dose ranged from 0.6-4.8 mGy.25 Dose-Incidence Relationship The mathematical relationship between incidence and dose, and the question of whether there is a practical threshold at low doses for radiation-induced breast cancer, has been evaluated by a number of investigators, including Land et al., who inferred that the best statistical fit to the available data was a dose-response curve of approximate linearity.26 The epidemiologic studies of breast cancer incidence after low-LET irradiation that are sufficient to enable risk estimates to be derived are summarized in Table 1." Although there may be some element of curvilinearity in the dose-incidence relationship, it appears from the available data that the error in risk estimates based on the assumption of linearity is likely to be small; however, in fitting potentially linear curves to the available data, Benefits vs. Risks in Mammography/Mettler et al. TABLE 1 Risk Factors for Female Breast Cancer Incidence from Various Epidemiologic Studies" Study Average excess absolute risk (I04PSY,,l~' Atomic Bomb survivors Postpartum mastitis Fluoroscopy (Massachusetts) Ankylosing spondylitis Swedish breast irradiation Thymic irradiation Skin hemangioma Scoliosis patients Cervical cancer Hodgkin's disease 6.8 (4.9-8.7) 9.1 (6.0-13) 8.0 (3.6-13) 9.0 (2.0-18) 2.7 (2.2-3.3) 4.9 (2.4-8.1) 4.1 (1.8-6.9) 17.8 (2.4-431 -0.3 (<0-0.2) 0.04 (0.03-0.07) Numbers in parentheses indicate 90% confidence intervals. PYSv: person year sieveri one should bear in mind that the confidence intervals (CI) from the Japanese mortality study encompass zero for doses of less than 0.5 Gy, although the corresponding incidence data demonstrate a statistically significant increase in the frequency of breast cancer at dose levels down to approximately 0.3 Gy for women of all ages. Although a large number of studies, which include natural background exposure and occupational exposure, do not demonstrate a statistically significant increase in breast cancer, this may be due to the relatively low doses and therefore low expected yields of tumors relative to the spontaneous incidence in such studies. Therefore, for purposes of an estimate of the potential detriment from mammography, it would appear that utilization of a linear extrapolation is not unreasonable, although it may provide a conservative scenario. Influence of Dose Rate A question that arises in risk estimation is the possible effect of dose rate on the induction of breast cancers by irradiation, and the applicability of various epidemiologic observations to mammography. In women exposed repeatedly to small doses of X-rays for months or years through fluoroscopic examinations of the chest for pulmonary tuberculosis, the relative risk of breast cancer per unit dose, adjusted for age at the time of exposure and duration of follow-up, is not significantly different from the risk in women exposed instantaneously to atomic bomb radiation and the risk in women exposed in one or more daily radiotherapy treatments for acute postpartum mastitis or ankylosing ~ p o n d y l i t i s . ~ As ' - ~ a~ result of the generally good agreement among studies (Table 11, the age specific excess risk factors derived from the atomic bomb survivors can probably be used justifiably without 905 a dose or dose-rate correction factor in estimating" the risks attributable to mammography. Influence of Age at the Time of Exposure Age at exposure appears to be the most significant modifying factor affecting the risks of screening mammography. Early reports in the literature indicate that radiation exposure during infancy may result in breast cancer in young women. In most studies of childhood and adult exposure, breast cancer risk decreases markedly with increasing age at exposure. In fact, the Life Span Study of atomic bomb survivors and studies of multiple fluoroscopies in Canadian tuberculosis patients show little, if any, detectable risk in women exposed older than the age of 40. The Japanese data indicate a three- to tenfold higher excess relative risk at 1 Gy for women exposed before age 20 than in those exposed after age 40. The data also indicate that survivors exposed before age 20 have a 13-fold excess relative risk at 1 Sievert (100 rem) for cancers occurring before the age of 35 compared with a twofold excess relative risk for cancers occurring after age 35.3 Because of the marked dependence of risk on age at exposure (Fig. l),utilization of a generalized risk estimate for all women in the population would significantly overestimate the frequency of potential cancer deaths in women exposed only after age 40. Efficacy of Screening The potential benefits of screening mammography have been the topic of numerous scientific studies over the last three decades. A number of these studies have recently been examined in detail, leading to the conclusion that screening of women over the age of 50 results in a 25-30% decrease in mortality.'" The value of screening in the 40-49 year age group has been a matter of debate, with some authors stating that it affords a benefit but that previous studies had not examined this age group specifically.3' This question has been addressed in a meta-analysis of 13 studies by Kerlikowske et al. who concluded that breast cancer mortality was reduced by 26% as a result of screening mammography in women aged 50-74 years (95% CI, 17-34%), but by only 7% in women aged 40-49 years (95% CI, -1324%)."l The latter does not represent a statistically significant benefit, but there is no reason to assume that the benefit from mammography suddenly begins at age 40 or 50. Another meta-analysis directed specifically at women 40-49 years of age estimated a 14% benefit from screening, which was not statistically significant at the 95% C1, but the same analysis, excluding the Canadian National Breast Screening (CNBS) Study, found a 24% benefit, which was statistically significant at the 95% CI.33 The Canadian National Breast Screening Study has been 906 CANCER March 1,1996 I Volume 77 I Number 5 judged as flawed on the basis of several factors, including selection bias, poor image quality, inadequate reader training, and a significant randomization e r r ~ r . ~ ~ - ~ ' TABLE 2 Breast Cancer BenefitlRisk Ratio for a Woman Having Annual Mammography Beginning at Age 35 Annual baseline breast cancer incidence per METHODS Individual Risk and Benefit We have compiled tables to show the benefit of screening and the potential radiation risk of mammography for women who begin screening programs at different ages. We have chosen variables that can be quantified with some degree of confidence. Using these tables, a woman considering whether or not to have a mammogram can assess the expected benefit in comparison with the radiation risk. In constructing these tables, we have necessarily been limited by a number of assumptions and constraints. We have not included the cost of mammography as a risk, but it should be considered in the decision-making process. The cost of mammography varies widely from a low of about $45 (through American Cancer Society screening programs) to a high of about $250. There is no evidence that the quality of the mammography examination is related to price. We also have not included any benefit as a result of the psychologic value of a negative mammogram, nor have we included any calculation of the costs of false-positive or false-negative results. These latter issues tend to be societal or postmammography concerns that are really not included in the decisionmaking process of a woman when she is deciding whether or not to get a mammogram. Age specific radiation risk factors that we have used are related to incidence (rather than mortality) and have been based on excess relative risk factors for the Japanese atomic bomb survivors (Fig. I ) , adjusted to account for the differences in baseline incidence between the Japanese and the U.S. populations. An excess relative risk of 1 means that there is a 100% increase in the natural breast cancer incidence. Based upon 1973-1990 SEER data," we have assumed that about 40% of naturally occurring or radiationinduced breast cancer cases are fatal and that radiationinduced breast cancer has a minimum latent period of ten years.z7Based upon American College of Radiology mammography accreditation data, we assumed a dose for screening mammography of 2.8 mGy per two-view examination. We have conservatively assumed a 5% reduction in breast cancer mortality as a result of annual mammographic screening in women who are screened from ages 35-39 and 25% in those who are 40 or older when screened."""0 RESULTS The results for women who begin annual screening at ages 35, 40, and 50 and continue until age 75 are shown Age 100,000 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 ti2 63 64 65 66 ti7 68 69 70 71 72 73 74 75 66 66 66 66 66 129 129 129 129 129 187 187 187 187 187 220 220 220 220 220 268 268 268 268 268 339 339 339 339 339 391 391 391 391 391 421 421 42 I 421 421 42 1 Fatal radiationinduced cases 0 0 0 0 0 0 0 0 0 0 <0.1 10.1 0.2 0.2 0.3 0.3 0.4 0.4 0.4 0.4 0.6 0.6 0.6 0.7 0.7 0.9 0.9 0.9 1 1 1.2 1.2 1.2 1.3 1.3 1.4 1.4 1.5 1.5 1.5 1.5 Fatal cases prevented by mammography 1.3 1.3 1.3 1.3 1.3 12.9 12.9 12.9 12.9 12.9 18.7 18.7 18.7 18.7 18.7 22 22 22 22 22 26.08 26.8 26.8 26.8 26.8 33.9 33.9 33.9 33.9 33.9 39.1 39.1 39.1 39.1 39.1 42.1 42.1 42.1 42.1 42.1 42.1 Benefit/ risk ratio >400 352 181 120 91 73 66 61 57 53 50 47 44 43 41 39 38 37 36 35 34 33 33 32 31 30 30 29 29 28 28 27 Tables assume reduction in mortality from breast cancer as a resulr of screening as follows: age 3539, 5 % 40-49. 15%: and 50-75. 25%. in Tables 2 , 3, and 4, respectively. The age 75 cutoff was chosen because this is close to the average female life span. The value of 40% mortality chosen for breast cancer has no effect on the calculated benefit/risk ratios because it applies to both naturally occurring and radiation-induced breast cancers. The value chosen for reduction of mortality for those women screened from ages 40-49 907 Benefits vs. Risks in Mammography/Mettler et al. TABLE 4 Breast Cancer BenefitlRisk Ratio for a Woman Having Annual Mammography Beginning at Age 50 TABLE 3 Breast Cancer BenefitlRisk Ratio for a Woman Having Annual Mammography Beginning at Age 40 ~ Annual baseline incidence per Age 100,oOO 40 41 42 '13 14 I 5 16 129 12!t Fatal radiation induced cases per 1 ~ , 0 0 0 0 18.i 56 1H7 0 0 <O.l 2iO 210 220 2:!0 .rn.i ?fiH 2liH 0.2 0.2 0.3 0.3 0.3 0.4 0.5 0.5 0.5 0.5 0.6 18.7 18.7 22 22 22 22 22 26.8 26.8 26.8 26.8 26.8 57 187 220 0 47 18 49 50 51 52 53 54 55 56 70 71 ?iH 258 26R 339 339 339 :!3Y ?39 39 1 29 1 391 391 19 I ,121 12 I 72 121 73 74 42 I .12 I 75 421 60 fil 62 ti3 64 65 66 67 6a 69 Age A 0 12rj I?9 0 58 59 Benefit/ risk ratio Annual incidence per 12.9 12.3 12.9 12.9 12.9 18.7 18.7 0 0 0 I23 I87 I87 1 87 r.I I Fatal cases prevented by mammography per 100,OOO 0 10.1 0.1 0 .I 0.i 0.i 0.7 0.7 0.8 0.9 0.9 0.9 0.9 1 50 51 52 >850 53 54 -- JJ 100,ooO B 220 220 220 220 220 268 268 268 268 268 339 339 333 339 Fatal radiation induced cases per 100,000 C 0 0 0 0 0 0 0 0 0 Fatal rases prevented by mammography per 100,000 D 22 22 22 22 22 26.8 26.8 26.8 26.8 26.8 33.9 33.9 33.9 33.9 33.9 33.9 39.1 Benefit/ risk ratio E , ' 1500 104 58 59 60 (i1 62 63 64 65 66 92 67 3YI 33.9 3Y.1 61 61 68 69 70 71 72 73 331 39 I 33.9 83 78 74 71 67 33.1 39.1 39.1 59 57 75 54 52 51 Explanation 01 Table 4. If a G:i-year-old ivomaii !who has liad ail amiual screening niaminographic. examinationsince age ,501 wanted to know hhar the ribk and henelir of her aiinual manimograni i\ou!d he lor this year, she ivorrld refer to lahle 1. Enrering the age colunin LA: and p i n g to age 63. she would first note [hat the baseline ["spuntancou<') risk :incidence)of breast cancer in this year is 33% 100.000 (Column R], ,411 her mammograms up to age 62 would haw given her an ;idditional fatal breast cancer risk of 0.07 cases\lUO,OW :Column Cl. One additional mmmographic examination rhis year 33.9 33.9 33.3 33.1 42. I 42.1 42.1 42.1 42.1 42.1 826 413 275 206 165 I38 118 49 17 46 41 43 only affects the benefitIrisk ratio for those specific years, regardless of whether a 40% or 15% value is used, the benefit/risk ratio remains in excess of SO. The results show that for any rational annual mammography screening program, the benefits substantially outweigh the risks. Even for the extreme example of a woman beginning annual screening at age 35 and continuing until age 75, we calculate a benefit that is at least 25 times greater than the potential radiation risk. IJsing a more common protocol of annual screening from age SO to age 75, the benefit in terms of reduced mortality as a result of screening should exceed the radiation risk by a factor of about 100. 74 339 39 1 39 1 421 42 1 421 121 421 421 0 0.1 0.1 0. I 0.1 0.1 0.2 0.2 0.2 0.2 0.3 0.3 0.4 0.4 0.4 0.4 0.5 39.1 39.1 39.1 12.1 12.1 12.1 42.1 42.1 12.1 1.140 ^, (20 480 360 288 240 206 1RO 160 144 131 1'0 111 103 9G 90 woiild rai% her total radiation risk to 0.09i100.000. These borh caii be rounded to 0.1 i100,ooO!Co!umn (:I. 'I he chances [hat the inammogram would prevenr a iatal canwr ivould he 33.'3~100.000;Column D;.'The herielitirisk ratio ~ o u l dbe 360 :Column El. In other w d c . her niamniogram exaniination rhis year w u l d provide 360 limes more benefit than the polenrialharm from 311 of lier nianimography screening examinations combined. DISCUSSION Screening for High Risk and Sensitive Subgroups In general, radiation-induced breast tumors do not appear in women until they have reached the age when spontaneous breast cancers usually occur. Thus, most radiation-induced cancers are indistinguishable in terms of age of appearance from spontaneously occurring breast cancers. It is possible that the reason for this is agedependent variations in hormonal stimulation or other factors. In atomic bomb survivors, there is apparently a markedly higher excess relative risk for those with early 908 CANCER March 1,1996 I Volume 77 I Number 5 onset breast cancers than for those with cancers that occur later, as noted above. This difference has been postulated to be due to the existence of a genetic subgroup with a high sensitivity to radiation-induced breast ~ a n c e rIn .~ the same population, a lower age at first pregnancy has been inferred to be protective not only against spontaneous breast cancer but also against radiation-induced breast ~ancer.‘“~~* The implication of this for mammography screening is not clear at the present time. Other genetic issues have been raised, particularly because an increased risk of breast cancer has been reported in families affected by ataxia-telangiectasia; Swift et al. have reported that heterozygous women who had been exposed to diagnostic fluoroscopic examination exhibited a fivefold increase in risk.43Although this observation would seem to be important, the diagnostic radiology involved was almost always to the abdomen, which would give only scattered radiation to the breast at very low doses (less than the annual dose from natural background); also, the exposed and control groups were of very different ages, and adjustment for this difference was not p o ~ s i b l e . Given ~~,~~ the numerous and major problems with the study, it is doubtful that it should affect risk estimates for screening mammography. Recently, several other breast cancer susceptibility genes, including BRCAl and BRCA2 in germ cell lines, have been identified.46.47 These two genes are thought to be responsible for most hereditary breast cancers (which account for 5-10% of all breast cancers). Women with the BRCAl mutations are estimated to have an 85% lifetime risk of breast cancer, and more than half of them will develop breast cancer before the age of 50:* The estimated incidence of each of these genes is 1 in 200 women.49The issue arises as to the potential impact of these discoveries on clinical practice and, in particular, on the indications for screening mammography. At present, there is little clinical impact as a result of genetic screening, in part because these genes are very large and many mutations have already been reported for the BRCAl gene alone. In addition, accurate and accessible tests far these mutations are not yet available. However, even without tests for these mutations being available, there are clearly persons with a strong family history of breast cancer, and whether such persons should have screening at an early age is of interest. The answer to this question depends on whether such “high risk” women, including those with BRCA mutations, are also at higher than normal risk for radiation-induced breast cancers. There is no reason to assume, however, that the excess risk in such women would be higher for radiation-induced cancers than for breast cancer in general and that the benefit/risk ratio would necessarily be different. If in the “high risk” groups the risk of radiationinduced cancer was not elevated disproportionately, the benefitlrisk ratio would shift in favor of the benefit. This might be offset by the increased difficulty of locating small cancers in the dense breast tissues of young women. CONCLUSION The possible harm from ionizing radiation should not be a significant factor in deciding whether annual mammographic screening should be performed. This is true even when considering screening in women as young as 35 years of age. REFERENCES 1. Mettler FA, Hempelmann L, Dutton A, Pifer JW, Toyooka ET, Ames WR. Breast neoplasms in women treated with xrays for acute post-partum mastitis. J Natl Cancer Inst 2. Thompson D, Mabuchi K, Ron E, Tokumaga M, Ochikubo S, Sugimoto S, et al. Cancer incidence in atomic bomb survivors, Part 11, solid tumors 1958-1987. Radiat Res 1964; 3. Tokunaga M, Land C, Tukapka S, Nishimori I, Soda M, Akiba S. Incidence of female breast cancer in atomic bomb survi- 1969;43:803- 11. 137:517-67. 4. 5. 6. 7. vors, 1950-1985. Radiut Res 1994;138:209-23. Tokunaga M, Land C, Aoki Y, Yamamoto T, Asano M, Sat0 E, et al. Proliferative and non-proliferative breast disease in atomic bomb survivors. Cancer 1993;72:1657-65. Shore RE, Hildreth N, Woodard E, Dvoretsky P, Hempelmann L, Pasternack B. Breast cancer among women given x-ray therapy for portpartum mastitis. J Natl Cancer Inst 1986;77:689-96. Beral E, Larsson LE, Mattson B. Breast cancer following irradiation of the breast. Cancer 1977;40:2905- 10. Mattson A, Ruden B, Hall P, Wilking N, Rutqvist LE. Radia- tion-inducedbreast cancer: long-termfollow-up of radiation therapy for breast disease. J Natl Cancer Inst 1993;85:167985. Darby SC, Doll R, Gill S, Smith PG. Long-term mortality after a single treatment course with x-rays in patients treated for ankylosing spondylitis. Br J Cancer 1987;55: 179-90. 9. Furst C, Lundell M, Holm LE. Tumors after radiotherapy for skin hemangioma in childhood: a case-control study. Actu Oncol 1990;29:557-62. 10. Ron E, Modan B, Boice JD Jr. Mortality after radiotherapy for ringworm of the scalp. AmJEpidernioll988;127:713-25. 11. Hildreth N, Shore RE, Dvoretsky P. The risk of breast cancer after irradiation of the thymus in infancy. N Eng2 J Med 1989;321:1281-4. 12. Hancock S, Tucker M, Hoppe R. Breast cancer after treatment of Hodgkin’s disease. J Natl Cancer Inst 1993;85:258. 31. 13. Li FP, Corkery J, Vawter G, Fine W, Sallan SE. Breast carcinoma after cancer therapy in childhood. Cancer 1983; 51:521-3. 14. Hrubec Z, Boice JD Jr, Monson R, Rosenstein M. Breast can- cer after multiple chest fluoroscopies: second follow-up of Massachusetts women with tuberculosis. Rudiat Res 1989;49:229-34. 15. Miller AB, Howe GR, Sherman GJ, Lindsay BA, Martin JY, Dinner PJ, et al. Mortality from breast cancer after irradia- tion during fluoroscopic examinations in patients being treated for tuberculosis. N Engl Jhled 1989;321:1285-9. Benefits vs. Risks in Mammography/Mettler et at. 16. Hoffman I), 1,onstein J , Morin M, Visscher W, Ilarris US11 I l l , 13oice JI) Jr. I3reast cancer in women with scoliosis exposed to multiple diagnostic x-rays. J N d Cancer Insf 1981);8 1: 1307- 12. 17. Adains E, Bntes A. Breast cancer in female radiunm dial workers first employed before 1930. 1 Okcup Med 1980;22:583-7. 18. Stehbings J , I.utxs H, Stehiey A. Mortality from cancers of major sites in female radium dial. Arti / I d Med 1984;5:43.5-59. 19. Madlado S, Land C,McKay F. Cancer niorlality rates and radioactive fallout in southwestern Utah. Am I Epiderniol 198;'; 125:44-61. 20. Wiklund K, Holm LE, Eklund G. Cancer risk in Swedish I.apps who breed reindeer. Ani 1 E+iderniol 1990: 132:107882. 21. Jablon S, tlruhec %, Uoice JI) Ir, Stone BJ. Cancer in populations living near nuclear facilities. Bethesda (MI)): National Cancer Institute; 1990. NIH Pub. No. 90-874. 22. Ueral V, Fraser P, Carpenter I., Booth M, Brown A, Rose G. Mortality of employees of the atomic weapons establishment 1951- 1982. BM1 1988;297:757-70. 23. Gilbert E, Cragle I), Wiggs 1.. Updated analysis of combined mortality data at the IIanford Site, Oak Ridge National Iaboratory, and Rocky Flats Weapons Plant. Radiaf Res 1993: 336:408-21. 24. Holm I.E, Wiklund Kf:, Lundel GE,Bergnian NA,Bjelkengren (;, Hriccson LJ(;, et al. Cancer risk in population examined with diagnostic doses of iodine-131. I Narl Carzcer Insr 1989;81:302-6. 25. United Nations Scientific Committee on the Effects of 26. 27. 28. 29. 30. 31. 32. Atomic Radiation. Sources and Effects of Ionizing Radiation. Annex C, Medical Radiation Exposures. 1993 Report to the General Assembly. 1993 Vienna, Austria. I.and CE, Boice J D Jr, Shore H, Norman JE, 'lokunaga M. Rreast risk from low-dose exposures to ionizing radiation: results of an analysis in parallel of data from three different exposed populations. J Ncitl Criricer Irisr 1980; 65353-76. IJnited Nations Scientific Committee on the Effects of Atomic Radiation (UNSCFAR). Epidemiological studies of radiation carcinogenesis. Report to the General Assembly. 1994 Vienna. .Austria. I3oice J D Jr, Land CE, Shore 11, Norman JE, 'I'okunaga M. Risk of breast cancer following low-dose radiation exposure. R ~ d i o l o g1979; 131~589-97. National Acatlemy of Sciences/National Research Council. I Iealth effects of exposure to low levels of ionizing radiation. I3EIfl V. Washington, DC: National Academy Press, 1990. Shapiro S. Screening: assessment of current studies. Cancer 1994;74(S~ppl):231-8. Kopans D. Questions and answers. AIR Ariz J Roenfgenol 1994; 163:1524. Kerlikowske K. Grady 11, Rubin S, Sandrock C, Ernster V1,. 909 Efficacy of screening mammography: a meta-analysis. JAMA 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. k fl, Ilutledge J, Smith R. Benefit of mainmography screening in women aged 40-49: current evidence from randomized screening trials. Carzcer. In press. Kopans I)H, Feig SA. The Canadian National Breast Screering Study: a critical review. AIR Am 1Roentgeriol 1993;IG1:755--60. Burhenne LJW, Burhenne HJ. l h e Canadian National Breast Screening Study: a Canadian critique. AJR Arri / I?oentgeriol 1993; 161:761-3. Boyd NF, long RA,Yaffe MI,'lritchler D. Lockwood G,Zylak CJ. A critical appraisal of the Canadian National Breast Cancer Screening Study. l?adioloa 1993; 189:661-3. 'Iarone RE. The excess of patients with advanced breast cancer in young women screened with mammography in the Canadian National Breast Screening Study. Grzcer 1995;75(4):997- 1003. Il. Cancer statistics review, 1973-1990. Uethesda (MMD): National Cancer Institute; 1993 N111 Pub. No. 93-2789. Tdbar I., 1:agerberg G, Chen 11, Duffy SW, Smart CR. Gad A, et al. Efficacy of breast cancer screening by age. Cancer 1995;75(10):25O7- 17. Feig SA. Estimation of currently attainable benefit from maminographic screening of women aged 40-49 years. Cart cer 1995;75(10):2412- 19. Land C. Breast cancer in the RERI: Life Span Study. REHF IJpdate, Spring Newsletter, 1993: 3-4. Hiroshima, Japan: Radiation Effects Research I h n d a t i o n , 1993. Land C, Hayakawa N,Machado S,Yaniada Y. A case-control interview study of breast cancer among Japanese A-bomb survivors. 11. Interactions with radiation dose. O'mcer Causes Control 1994;5(2):167-76. Swift M, Morrell D, Massey 1113, Chase CL. Incidence of cancer in 161 families affected by ataxia-telangiectasia. N Engl / Med 1991:325:1831-6. Boicc J Jr, Miller Il. Risk of breast cancer in ataxia-telangiectasia (letter to the editor]. N Big1 / Med 1992;326(20):1357-8. Bridges BA, Arlett CF. Risk of breast cancer in ataxia-telangiectasia [letter to the editor]. N Engl I Med 1992;326(20):1357-8. Miki Y, Swensen J, Shattuck-Eidens L), Futreal PA, IIarshman K, Tavtigian S, et al. Isolation of HRACI, the 17q-linked breast and ovarian cancer susceptibility gene. Sciertce 3994;266:66 -71. Wooster R, Neuhausen S,Mangion Y, Quirk Y, Ford I), Collins N, et al. I.ocali/.ation of a breast cancer susceptibility gene, BRAC2, to chromosome 13q 12- 13. Science 1994: 265:2088-90. Easton DI:, Bishop I), Ford I), Crockford G. Genetic linkage analysis in familial breast and ovarian cancer: results from 214 families. Arri I Hiirii Genet 3993;52:6378-701. Nowak H. Breast cancer gene offers surprises. Science 1994;265:1796-99.