High risk of breast carcinoma after irradiation of young women with Hodgkin's disease




Treatment-associated second neoplasms have emerged as a major threat to the continued survival of patients cured of Hodgkin's disease. In this study, the authors investigated the risk of breast carcinoma in an irradiated Hodgkin's disease population.


One hundred and eleven women younger than 60 years presenting between 1964 and 1984 with Stage I and II Hodgkin's disease who received mantle irradiation were retrospectively analyzed and compared with an age specific population. Median follow-up was 18 years (range, 10-30 years), and the median age at initiation of therapy was 24 years. Kaplan-Meier actuarial risks, relative risks (RRs) (the ratio of the observed to the expected cases) with 95% confidence intervals (CIs), and the log rank test for trends were calculated.


Fourteen women developed breast carcinoma: 8 of 33 patients younger than 20 years at the time of irradiation, 5 of 48 patients age 20 to 29 years, and 1 of 30 patients age 30 years or older. Actuarial calculation predicted a 34.0% (CI, 14.2-53.8) risk of breast carcinoma at 25 years after therapy for the youngest group, 22.3% (CI, 4.1-40.5) for the group of intermediate age, and 3.5% (CI, 0-10.1) for the oldest group. The RR of breast carcinoma was 56 (CI, 23.3-107) for those 19 years or younger at the time of treatment, 7.0 (CI, 2.3-16.4) for those age 20-29 years, and 0.9 (CI, 0-5.3) for those 30 years and older. Excluding 1 patient who was age 38 years at the time of irradiation, the remaining 13 breast carcinomas were tightly clustered in women irradiated between the ages of 14 through 25, and were detected in years 11 through 25 after treatment, with 7 occurring in years 15 through 18.


Women younger than 30 years, particularly those younger than 20 years, who have received mantle irradiation for Hodgkin's disease require meticulous follow-up for breast carcinoma. The high incidence of breast carcinoma in this patient population should be considered when making treatment decisions in young women with early stage Hodgkin's disease. Cancer 1997; 79:1203-10. © 1997 American Cancer Society.

Treatment-associated second neoplasms have emerged as a major threat to the continued survival of patients cured of Hodgkin's disease. Acute nonlymphocytic leukemia, a highly lethal disorder closely linked to alkylating agent use, was delineated first.1, 2 Because of their longer latency periods, the incidence of second solid neoplasms is only now being fully appreciated.3, 4 Carcinoma of the breast is a distressing radiation-induced neoplasm to which young women are particularly susceptible.5-19 The authors describe the emergence of this malignancy in an important fraction of young women irradiated for Hodgkin's disease at the Massachusetts General Hospital between 1964 and 1984. This study is noteworthy in its length of follow-up.


One hundred and eleven consecutive women younger than 60 years who received mantle irradiation for clinical or pathologic Stage I or II Hodgkin's disease at the Massachusetts General Hospital between July 1, 1964 and June 30, 1984 were retrospectively accessed and followed, with the exception of 6 patients, 5 of whom died of lymphoma and 1 who was lost to follow-up, within 5 years of treatment. Attempts to contact patients continued until September 30, 1994; recent reliable follow-up information was obtained for 106 of the 111 eligible women. The 5 women whose recent status could not be determined had been followed for 10 to 14 years, and are included in the analysis.

The Ann Arbor staging classification for Hodgkin's disease was employed.20 Diagnostic studies included bilateral lower extremity lymphangiography and bone marrow biopsy for those with systemic symptoms. Laparotomy and splenectomy were added to the diagnostic evaluation of most clinical Stage I and II patients in late 1969, and utilized throughout the remainder of the study period. Prior to 1973, the radiation technique was comprised of a shaped anterior mediastinal field treated with anterior dose three times that of the posterior field (37 patients). The maximum daily anterior dose was ≥300 centigray (cGy) or per fraction. After 1973, a conventional fractionation and field arrangement using equally weighted opposed anterior and posterior shaped mantle fields was employed (74 patients). The daily dose to the midplane of the mediastinum was between 150 and 180 cGy and the maximum daily dose was usually not more than 220 cGy. The calculated maximum dose to the mediastinum is the maximum dose received by a rim of breast tissue lying outside the radiation shielding. Less than 10% of that dose would be delivered to the major portion of breast tissue because of shielding of the breasts and lungs. All patients were treated with megavoltage X-rays, 2-megavolt (Mv) X-rays from a Van der Graaff generator before 1977 and 10-Mv X-rays from a linear accelerator for the remainder of the study. After 1975, the dose to the involved sites varied from 40 to 45 gray (Gy) in 4 to 6 weeks with uninvolved sites receiving 36 to 40 Gy.21 Twenty-eight patients received nitrogen mustard, vincristine, procarbazine, and prednisone (MOPP) chemotherapy, 10 as part of combined treatment and 18 for recurrence.

Radiation records of all patients were reviewed for this study, and the dose to the anterior mediastinum calculated. Clinical and pathologic records of those treated for breast carcinoma at the study (four patients) or outside (ten patients) institutions were also reviewed, with the exception of one patient for whom information was obtained from an outside oncologist.

The actuarial risks of the development of breast carcinoma were calculated from the initial treatment date with the technique of Kaplan and Meier.22 Data from patients without breast carcinoma were censored at the last follow-up interval or death (six patients: two each from heart disease and lung carcinoma, and one each from Hodgkin's disease and fulminant sepsis) in these analyses. To estimate the breast carcinoma risk of a control population, person-years of observation were compiled from the date of initial Hodgkin's disease therapy to the date of last follow-up, death, or diagnosis of or death from breast carcinoma.23 Annualized breast carcinoma incidence rates specific for age and gender, obtained from the SEER program,24 were multiplied by the accumulated person-years to estimate the expected number of breast carcinomas. Yearly incidence figures were calculated by extrapolation from the 5-year SEER rates for 1986-1990.

Statistical methods for risk estimation were based on the assumption that the observed number of breast carcinomas followed a Poisson distribution. Tests of significance and confidence intervals (CIs) for the risk ratios (RRs) (ratios of the observed to the expected number of cases) were calculated with the use of exact Poisson probabilities. Age group comparisons of breast carcinoma incidence were based on the log rank test.


The median follow-up of the 111 eligible women was 18 years, and the median age at irradiation was 24 years (range 9-59 years). Fourteen women developed breast carcinoma: 8 of 33 patients younger than 20 years at the time of treatment, 5 of 48 patients aged 20 to 29 years, and 1 of 30 patients 30 years or older (Table 1). Actuarial calculation predicted a 34.0% (CI, 14.2-53.8) breast carcinoma incidence 25 years after therapy for the youngest group, 22.3% (CI, 4.1-40.5) for the intermediate group, and 3.5% (CI, 0-10.1) for the oldest group (Fig. 1). The RR of breast carcinoma was 56 (CI, 23.3-107) for those 19 years or younger at the time of treatment, 7.0 (CI, 2.3-16.4) for those age 20-29 years, and 0.9 (CI, 0-5.3) for those 30 years and older (Table 1). Excluding a 38-year-old woman whose neoplasm occurred only 8.5 years after therapy and may have been unrelated to irradiation, the remaining 13 breast carcinoma cases were tightly clustered in women irradiated at ages 14 through 25 years (median 18 years; mean, 19.3 years) (Fig. 2; Table 2). The P value for the comparison between the actuarial curves of breast carcinoma incidence for the younger than 20 years and the 20 years or older groups was 0.05, and that for the comparison between the younger than 20 years and the older than 30 years groups was 0.09.

Figure 1.

Risk of breast carcinoma by age at time of irradiation (Kaplan-Meier analysis): ≤19 years (33 patients, 8 breast carcinomas); 20-29 years (48 patients, 5 breast carcinomas); ≥30 years (30 patients, 1 breast carcinoma).

Figure 2.

Breast carcinoma occurrence (dark-shaded square) versus age at time of Hodgkin's disease irradiation plotted at 2-year intervals.

Table 1. Risk of Breast Carcinoma by Age at Time of Irradiation
Age group (yr)No. at riskAge median (range) (yr)Median (range) (yr)Person-yearsObserved/expected casesRR (CI)Kaplan-Meier risk at 25 yrs (CI) (%)Absolute riska
  • RR: relative risk; CI: 95% confidence interval.

  • a

    Absolute risk was calculated by subtracting the expected number of cases from the observed number, dividing the difference by the person-years of observation, and multiplying the quotient by 104 to estimate the excess number of cases per 10,000 person-years.

< 203316 (9–19)20 (12–28)6638/0.1556 (23.3–107)34.0 (14.2–53.8)118
20–294824 (20–29)20 (10–30)9175/0.717.0 (2.3–16.4)22.3 (4.1–40.5)46.7
≥ 303038 (30–59)16 (10–29)4921/1.060.9 (0–5.3)3.5 (0–10.1)--
Table 2. Breast Carcinoma after Irradiation for Hodgkin's Disease
 Hodgkin's diseaseBreast carcinoma
Case no.Age (yr)YearIrradiation (Gy)MOPPInterval (yr)DescriptionFollow-up (yr)
  1. MOPP: nitrogen mustard, vincristine, procarbazine, and prednisone; Gy: gray; IDC: invasive ductal carcinoma; DCIS: ductal carcinoma in situ; UOQ: upper outer quadrant; 0/00+; number of positive axillary lymph nodes. Rt: right; Lt; left.

121196338--17Rt: UOQ, IDC, 3 cm, 2/16+10
219196538--16Rt: UOQ, IDC, 1.1 cm, 0/16+12
325196742+20No details 
417197044+11Lt: IDC, metastaticDeath 1.5
     12Rt: T1N0M0 
518197048--21Rt: Lateral, IDC, 0/8+4
622197045--24Rt: UOQ, IDC, 1.2 cm0.1
722197045--25Lt: IDC, 3.6 cm, 0/7+1
814197342--18Rt: Multiple foci, IDC, 0/7+3
     20Lt: IDC, 2 cm, 0/13+1
914197442--19Rt: Multiple foci, IDC, 0/11+ (Lt: prophylactic)2
1038197443--8.5Lt: IDC, 1.5 cm, 0/4+8
1118197553+19Rt: UOQ, IDC, 0.3 cm, 0/9+0.3
1217197743+17Rt: Multiple foci, IDC, 0/16+0.5
1325197942--16Rt: UOQ, IDC 3 cm, 0/5+0.2
1417198047--12Rt: UOQ, IDC, 3.6 cm, 4/12+1
      LT: DCIS

The maximum dose to the anterior mediastinum ± standard deviation for the 14 patients who developed breast carcinoma was 4401 cGy ± 386 (range, 3612-5284 cGy) and 4528 cGy ± 419 (range, 3788-6408 cGy) for the remaining 97 women. Seven breast carcinomas were observed in the 37 women (19%) treated prior to 1973 with 3 to 1 weighting of the anterior to posterior mediastinal fields and fractions of 300 cGy or greater (median follow-up, 25 years), and 7 in the 74 women (9%) treated after that date with equal anterior to posterior weighting and smaller fractions (median follow-up, 16 years). The higher risk in patients irradiated before 1973 is proportional to the number of cases expected with the longer period of observation (RR = 1.28) (Table 3), and the slightly higher risk of breast carcinoma in later years of observation (8 of 14 breast carcinomas in years 17-29 after treatment and 6 breast carcinomas in the first 16 years) (Table 4). Four of the 14 women (29%) with breast carcinoma received MOPP chemotherapy, as did 24 of the 97 (25%) breast carcinoma free individuals (RR = 1.24) (Table 3).

Table 3. Risk of Breast Carcinoma in Women Irradiated Prior to 1973a and in 1973-1984, and in Those Treated with and without Added MOPP Chemotherapy
   Breast carcinoma
Patient groupNo.Person-yearsNo.Number per 10,000 person-yearsRisk ratio
  • MOPP: nitrogen mustard, vincristine, procarbazine, and prednisone; RT: radiation therapy.

  • a

    Prior to 1973, 3:1 weighting of anterior to posterior mediastinal fields was employed.

Prior to 1973a37925775.71.28
RT alone8315521064.41.0
RT + MOPP28500480.01.24
Table 4. Breast Carcinoma Risk in Women 29 Years or Younger by Years of Observation
Year of observationNo. at riskNo. with breast carcinomaAnnual rate of breast carcinoma
1081 0
1364 0
1460 0
1937 0
2228 0
2326 0
2610 0
275 0
283 0
291 0

Table 4 records the occurrence of breast carcinoma in women 29 years or younger as a function of the length of follow-up. The first of the 13 breast malignancies was observed in the 11th year after radiation therapy, and the last in the 25th; 7 were detected in the 15th through 18th years. These findings suggest a somewhat longer latency period for breast carcinoma in women aged 20 to 29 years who underwent irradiation than in younger women (Fig. 1).

Clinical and pathologic details of the encountered breast carcinomas are outlined in Table 2. All cases presented as breast masses detected either by the patient or her physician. All tumors were invasive ductal carcinoma of the conventional type. They arose in the right breast in 9 of 11 women with unilateral involvement. One of the 2 left-sided tumors was the single short latency period case in the 30 years and older age group. (Significant differences between irradiation exposure of the left and right breasts were not detected). Two patients developed bilateral (asynchronous) invasive carcinomas; in one patient, ductal carcinoma in situ arose in the breast contralateral to the invasive carcinoma. In some cases staging information was incomplete. Of those that could be analyzed, five tumors were T1, three were T2, and three breasts contained multiple foci of invasive carcinoma. The sites of individual breast carcinomas could not be reliably correlated with the distribution of radiation dose to the breast. Axillary lymph nodes were involved in 4 of 11 cases. Only 1 patient died of breast carcinoma at last follow-up, but follow-up exceeded 4 years in only 3 individuals. A family history of breast carcinoma and age of menarche were not available for the entire cohort of women, but none of the 13 young breast carcinoma patients carried a pregnancy to term prior to her treatment for Hodgkin's disease.

Acute nonlymphocytic leukemia was not detected in this cohort at the time of last follow-up. The other solid neoplasms encountered are listed in Table 5. Except for the parotid tumors, which occurred early in the second decade, these occurred late in the second or in the third decade after treatment.

Table 5. Other Second Neoplasms
Case no.Age at Hodgkin's disease diagnosis (yr)Second neoplasmInterval (yr)
  • MALT: mucosa-associated lymphoid tissue.

  • a

    Died of second neoplasm.

117Mixed tumor of the parotid11
  Transitional cell carcinoma of the kidney25
225Carcinoma of the thyroid27
329Non-Hodgkin's lymphoma of the stomach (MALT lineage)26
426Carcinoma of the lung24a
526Mixed tumor of the parotid14
655Melanoma of the toe4
  Mixed tumor of the parotid10
  Osteogenic sarcoma of the tibia18
  Sarcomatoid carcinoma of the esophagus22a
725Sarcoma of the esophagus21a
843Carcinoma of the lung18a


An increased risk of breast carcinoma has been documented in women surviving the atomic bomb,25 after repeated chest fluoroscopies,26-28 after irradiation of the breast for postpartum mastitis29, 30 or an enlarged thymus,31 and in the contralateral breast after therapeutic irradiation for breast carcinoma.32 In these clinical contexts the dose-response curve appears to be linear over the 0- to 10-Gy range accessible to study,25, 27-29, 33 and is influenced little by dose fractionation or protraction.34 However, there is little published data regarding the carcinogenic risk of doses above 10 Gy, particularly for the doses in the 35-45-Gy range employed to treat Hodgkin's disease.35 Cell killing has been postulated to reduce the carcinogenic risk along an exponential curve at doses above 10 Gy.36

In addition to dose, age at time of exposure has received careful study in these earlier examples of radiation-induced breast carcinogenesis.25, 27-29, 33 Adolescent girls and young women are most sensitive, a carcinogenic effect of radiation to the breast is not evident in women exposed after age 40.24, 26, 32 In young women, a latency period of at least 10 years is usually required, with no discernible upper time limit.25, 27-31, 33 A shorter latency period (5-10 years) has occasionally been reported in the infrequent breast neoplasms after exposure in the fourth decade of life.27, 33

Sporadic cases of breast carcinoma after irradiation for Hodgkin's disease were reported beginning in 1978,5-7 but the frequency of this complication has been recognized only in the past several years.4, 11, 13, 14, 16-19 Only in the middle 1960s did a large cohort of women cured of Hodgkin's disease through modern radiation fields and dosage begin to accumulate and start along the pathway that would lead to breast carcinoma 10 or more years later.

The findings reported in this study confirm and extend earlier observations5-19 concerning breast carcinoma in women radiated for Hodgkin's disease. The authors' experience includes a cohort of Hodgkin's disease patients at very high risk for breast carcinoma; 81 young women (younger than 30 years) with long follow-up (10-30 years; mean, 19.5 years; median, 20 years). The authors have actually observed breast carcinoma supervene in approximately 25% (8 of 33) of those women treated before age 20 years, by actuarial calculation, one-third will be affected within 25 years of therapy. The authors' experience with women age 20-29 years at treatment is only slightly less disturbing; the occurrence of breast carcinoma was observed in 1 in 10 (5 of 48), with an actuarial estimate of 1 in 5 at 25 years. Although a minor role of confounding variables cannot be ruled out, the magnitude of the RRs in the current study far exceed those of other known breast carcinoma risk factors.36

Similar actuarial estimates (one-third affected after 25 years) have been projected37 for a much larger cohort of adult women14 followed for a shorter mean time period (10 years), and, most recently, for large cohorts who underwent irradiation in childhood and adolescence.17-19 The authors' observation of 14 breast carcinomas in 111 women treated for Hodgkin's disease can be compared with 25 of 885 women,14 17 of 483,18 and 16 of 67019 in these other recent studies.

In the current series radiation dose to the anterior mediastinum was the same in those women who developed breast carcinoma and those who did not, and the preponderant weighting of the anterior field and larger fractions employed before 1973 did not appear to play a significant role in carcinogenesis. The slight additional risk of breast carcinoma observed in the current study when MOPP chemotherapy was added to irradiation was not statistically significant. Earlier investigators suggested that added MOPP might shorten the breast carcinoma latency period.14

From the earlier investigations and this one, a coherent picture emerges of the risk of breast carcinoma in women irradiated for Hodgkin's disease. Mantle irradiation in the second and third decades of life carries the gravest risk of breast carcinoma.14 Indeed, the risk appears to abate in the second half of the third decade. In the current study, if the 38-year-old woman whose breast neoplasm possibly was unrelated to irradiation is excluded, all other breast carcinomas occurred in women treated at age 25 years or younger (Fig. 2; Table 2). The other investigation bearing on the risk in women in the third decade reveals a major decrease in RR when the 25-29 year group is compared with those younger than 15 years, 15-19 years, and 20-24 years at treatment, but does not list specific ages of the breast carcinoma patients.14

Two studies indicate that breast tissue is less sensitive to the carcinogenic action of Hodgkin's disease irradiation in the first decade of life than in the second and third,17, 18 a finding consistent with the experience in atomic bomb survivors.25 In agreement with other investigators,14 no significant increase in the breast carcinoma risk in women 30 years or older irradiated for Hodgkin's disease was found in the current study. In summary, the maximum risk of breast carcinoma is observed in women irradiated in the second and first half of the third decades of life, with shoulders of the risk peak extending into the first decade and second half of the third decade.

In unirradiated women, early age at the birth of their first child appears to markedly reduce the risk of breast carcinoma, as does, to a lesser degree, late menarche and early menopause.36 These observations are interpreted as revealing a window of risk (from menarche to first birth or menopause) in which breast epithelium, prior to its terminal differentiation, is susceptible to the carcinogenic effect of unbalanced estrogen stimulation.36, 38, 39 The observed tight clustering of breast carcinoma in women irradiated for Hodgkin's disease at ages 14 through 25 years in the current study is consistent with a comparable window of risk during which the incompletely differentiated breast epithelium is sensitive to the carcinogenic action of irradiation. The absence of full term pregnancies in the young women in the current study supports that hypothesis, but the absence of pregnancy in 13 women whose median age was only 18 has little statistical power. A detailed study of breast carcinoma risk factors in this irradiated population is a subject for future investigation.

The high risk of breast carcinoma in young women after Hodgkin's disease irradiation mandates meticulous and regular breast examination in these women beginning within 10 years of Hodgkin's disease treatment. Currently, the authors recommend yearly mammography despite the high false-negative rate of the technique in younger women.

It is also evident that radiation oncologists involved in the treatment of patients with Hodgkin's disease must continue their efforts to minimize dosage to the breast during mantle radiation.34 Present techniques, employing improved imaging modalities, three-dimensional planning, and sophisticated patient repositioning permit better shielding of normal tissue. For young women without axillary lymph node involvement, reduction of breast dose may be possible without jeopardizing the curative potential of mantle irradiation. Nonetheless, because significant carcinogenic risk may come from breast tissue at the field margins receiving radiation in the range of 5 to 10 Gy, the optimal reconfiguration of the mantle field is not intuitively obvious. The long latency period of radiation-induced breast carcinoma compounds the difficulty in addressing this problem.

The fact that so many women have survived so long after Hodgkin's disease radiation argues against the precipitous adoption of competing treatment modalities. However, until the problem is resolved, the high incidence of breast carcinoma reported in this study and observed by others4, 11, 13, 14, 16-19 should raise the consideration of chemotherapy or combined therapy with reduced radiation as alternate treatments for young women with early stage Hodgkin's disease. Pediatric oncologists have already adopted such combined regimens to minimize developmental abnormalities.40 Effective, nonleukemogenic, nonsterilizing drug combinations such as doxorubicin, bleomycin, vinblastine, and imidazole carboxamide,41 either alone or with low dose radiation, afford attractive alternatives to conventional radiation therapy for young women (age 15 to 29 years) who present with early stage Hodgkin's disease.