|Home | About | Journals | Submit | Contact Us | Français|
Tara Henderson, MD, MPH, University of Chicago, 5841 S. Maryland Ave. MC 4060, Chicago, IL 60637, ude.ogacihcu.dsb.sdep@nosredneht
Alison Amsterdam, MD, Mount Sinai School of Medicine, 1 Gustave L. Levy Place, New York, NY 10029, firstname.lastname@example.org
Smita Bhatia, MD, MPH, City of Hope Comprehensive Cancer Center, 1500 Duarte Ave, Duarte, CA 91010, gro.hoc@aitahBS
Melissa M. Hudson, MD, St. Jude Children’s Research Hospital, 262 Danny Thomas, Place, MS 735, Memphis, TN 38105, gro.edujts@nosduH.assileM
Anna T. Meadows, MD, Children’s Hospital of Philadelphia, Division of Oncology, 34th & Civic Center Blvd., Philadelphia, PA 19104, ude.pohc.liame@swodaem
Joseph P. Neglia, MD, MPH, University of Minnesota Cancer Center, Division of Pediatric Hematology/Oncology, Rm D557 Mayo Bldg/MMC #484, 420 Delaware St, SE, Minneapolis, MN 55455, ude.nmu@ailgenj
Lisa R. Diller, MD, Dana-Farber Cancer Institute, Pediatric Oncology, 44 Binney Street, Boston, MA 02115, ude.dravrah.icfd@rellid_asil
Louis S. Constine, MD, University of Rochester Medical Center, Radiation Oncology, 601 Elmwood Avenue, Box 647, Rochester, NY 14642, ude.retsehcoR.CMRU@enitsnoC_siuoL
Robert A. Smith, PhD, Director, Cancer Screening, American Cancer Society, 250 Williams St, 6th Floor, Atlanta, GA 30303, gro.recnac@htimS.treboR
Martin C. Mahoney, MD, PhD, Cancer Prevention and Population Sciences, Department of Educational Affairs, Roswell Park Cancer Institute, Buffalo, New York 14263, gro.kraPllewsoR@yenohaM.nitraM
Elizabeth A. Morris, MD, Department of Radiology, Memorial Sloan-Kettering Cancer Center, 1114 1st Avenue, New York, NY 10065, gro.ccksm@esirrom
Leslie L. Montgomery, MD, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 424 East 68th Street, New York, NY 10065, gro.ccksm@lmogtnom
Wendy Landier, MSN, CPNP, City of Hope National Medical Center, 1500 East Duarte Road DPS 173, Duarte, CA 91010, gro.hoc@reidnaLW
Stephanie M. Smith, MPH, Department of Pediatrics, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10065, email@example.com
Leslie L. Robison, PhD, Epidemiology and Cancer Control, St. Jude Children’s Research Hospital, 262 Danny Thomas Place, MS 735, Memphis, TN 38105, GRO.EDUJTS@nosiboR.seL
Women treated with therapeutic chest radiation may develop breast cancer.
Summarize breast cancer risk and breast cancer surveillance in women following chest radiation for a pediatric or young adult cancer.
Studies from MEDLINE, EMBASE, Cochrane Library, and CINAHL (1966 through December 2008).
Articles selected to answer any of 3 questions: 1) What is the incidence and excess risk of breast cancer in women following chest radiation for a pediatric or young adult cancer? 2) For these women, are the clinical characteristics of the breast cancer and the outcomes following therapy different than for women with sporadic breast cancer in the general population? 3) What are the potential benefits and harms associated with breast cancer surveillance among women exposed to chest radiation?
Three investigators independently extracted data and assessed study quality.
Standardized incidence ratios ranged from 13.3 to 55.5; cumulative incidence of breast cancer by 40–45 years of age ranged from 13–20%. Risk of breast cancer increased linearly with chest radiation dose. Available limited evidence suggests that the characteristics of the breast cancers in these women and the outcomes following diagnosis are similar to those in the general population; these breast cancers can be detected by mammography, though sensitivity is limited.
Limitations include study heterogeneity, design and small sample size.
Women treated with chest radiation have a substantially elevated risk of breast cancer at a young age, which does not appear to plateau. Among this high risk population, there appears to be a benefit associated with early detection. Further research is required to better define the harms and benefits of lifelong surveillance.
An estimated 50,000 – 55,000 women in the United States have been treated with moderate to high-dose chest radiation (≥20 Gy) for a pediatric or young adult cancer (2–5), and these women are at significantly increased risk of breast cancer and breast cancer mortality following cure of their primary malignancy (1). Breast cancer risk is greatest among women treated for Hodgkin's lymphoma with high-dose mantle radiation, but it is also elevated among women who received moderate-dose chest radiation (e.g., mediastinal, lung) for other pediatric and young adult cancers, such as non-Hodgkin’s lymphoma, Wilms tumor, leukemia, bone cancer, neuroblastoma and soft tissue sarcoma.
Recognizing the high incidence of second cancers and other health problems affecting survivors of pediatric and young adult cancer, the Institute of Medicine in January 2002 charged the Children’s Oncology Group (COG) with the development of comprehensive clinical practice guidelines for the long-term follow-up care of childhood cancer survivors. In September 2003, the COG released the Long-Term Follow-Up Guidelines for Survivors of Childhood, Adolescent, and Young Adult Cancers (6). The following is a report of our background systematic review of the literature that was conducted to inform and update the COG breast cancer surveillance recommendations.
We focused the review on studies of girls and women with a pediatric or young adult cancer (diagnosis ≤30 years of age) who were exposed to moderate to high doses of therapeutic radiation targeted to mantle and modified mantle fields, mediastinum, lung, and chest (thorax). Hereafter, the term ‘chest radiation’ refers to any of these exposures. Women treated for Hodgkin’s lymphoma with mantle or modified mantle radiation represent about two-thirds of this population.
We structured the review around three key questions:
We searched MEDLINE, EMBASE, the Cochrane Library, and CINAHL between1966 -December 2008 using the search strategy detailed in Appendix A. Literature searches were supplemented by examining bibliographies of included studies and selected literature reviews.
Inclusion and exclusion criteria were developed for each question (Table 1). Three investigators (TOH, AA, KCO) reviewed the full text of papers that appeared to meet eligibility criteria based on Abstract review. Discrepancies regarding which studies to include were resolved by consensus. Appendix B summarizes the literature search and review process. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting criteria (7, 8) for all studies included in the review is provided in Appendix C.
For question 1, we extracted measures of breast cancer risk [standardized incidence ratio (SIR), odds ratio (OR), absolute excess risk (AER), cumulative incidence]. For question 2, we extracted information about affected women’s clinical characteristics: median interval from primary cancer to breast cancer, stage and location at breast cancer diagnosis, pathological features, and percent with bilateral breast cancer (metachronous and synchronous cancers). For question 3, we extracted data on the number of breast cancer cases detected by surveillance per person-year of follow-up.
This systematic review was not funded through an independent agency but was conducted through the COG.
Eleven retrospective cohort (9–19) and three case-control (20–24) studies met eligibility criteria and assessed the risk of breast cancer in women following chest radiation for a pediatric or young adult cancer.
The cohort studies included over 14,000 women, most of whom were treated for a pediatric or young adult cancer over a four decade period (1960–2000). Among these, about 7000 women were treated with chest radiation and 422 women subsequently developed breast cancer. Study objectives and designs varied. Most focused on the risks of Hodgkin’s lymphoma survivors; only four estimated risk among women treated with chest radiation for other cancer types (11, 13, 20, 22). The precision or generalizability of risk estimates of some of the cohort studies were limited by a lack of information regarding chest radiation (11, 14), relatively small sample sizes or number of breast cancer cases (9, 15–18), or single institution cohorts (9, 12, 17). Most assessed risk of different types of second malignancies (10, 11, 14, 15, 17, 18, 20), so detail regarding breast cancer risk was often limited. Six were well-designed with representative samples, adequate length of follow-up, and risk estimates based upon detailed treatment exposures (10, 13, 15, 16, 18, 19).
Heterogeneity notwithstanding, each of the studies reported a significantly increased incidence and/or absolute excess risk of breast cancer associated with chest radiation (Table 2). Among the higher quality cohort studies, the SIR ranged from 13.3 to 55.5 and the absolute excess risk ranged from 18.6 to 79.0 per 10,000 person-years (10, 13, 15, 16, 18, 19). Risk of breast cancer was found to increase as early as 8 years following chest radiation and did not plateau with increasing length of follow-up (9, 10, 13, 14, 16). The cumulative incidence of breast cancer by 40–45 years of age ranged from 13–20% and by 25–30 years of follow-up ranged from 12–26% (10, 13, 16, 19). This incidence is similar to that in women with a BRCA gene mutation, where by age 40 the cumulative incidence ranges from 10–19% (25–29) and is substantially higher than in young women in the general population in whom the cumulative incidence of invasive breast cancer by age 45 is only 1% (30–32).
Two of the three case-control studies focused on breast cancer, included diverse and large populations of cases and appropriately matched controls within a cohort, and had detailed treatment information including radiation dosimetry estimates (21–24).
In the first, a well-designed case-control study (21–23), Travis et al estimated that among women diagnosed with Hodgkin’s at age 15 years and counseled for screening at the age of 25, 9.2% of those treated with 20–39 Gy and 11.1% of those treated with ≥ 40 Gy would develop breast cancer by age 45 (22).
In the other well-designed case-control study, Inskip and colleagues reported that the odds of being diagnosed with breast cancer increased linearly with chest radiation dose, reaching 10.8 (95% CI, 3.8–31) for 40 Gy relative to no radiation (24). Their data suggested that women exposed to 20–30 Gy have odds ratios of breast cancer between about 6.0 and 9.0.
Yet another study suggests that risk may be associated with radiation field volume, given increased risk (odds ratio 2.7 (95% CI, 1.1–6.9)) in women treated with mantle field irradiation compared to women with similarly dosed mediastinal irradiation (omitting the axillary nodes) (19).
Certain factors appear to modify the risk of breast cancer associated with chest radiation. A history of breast cancer in a close relative modestly increases risk in women treated with chest radiation (13, 23). Inskip et al reported that risk was sharply reduced among women with ovarian radiation strongly associated with acute ovarian failure (OR per Gy 0.06 for women who received ovarian doses > 5 Gy compared with OR per Gy 0.36 for those who received doses less than 5 Gy) (24), a finding confirmed in other studies (13, 19, 21). High dose alkylating agent chemotherapy, which can lead to premature menopause, may also lower the risk of breast cancer (13, 19, 21, 22). Notably, risk in women treated in prepubertal years is not lower than those treated during adolescence, as suggested by some early studies (33) which did not account for the natural age-associated increase of risk (34). Updated studies with extended years of follow-up that have accounted for attained ages at follow-up have not found a difference in breast cancer risk among women treated with chest radiation in their prepubertal years compared with those treated in their adolescent years (10, 13, 24).
Only 3 studies reported on the clinical characteristics of breast cancer among women treated for a pediatric or young adult cancer with chest radiation (12, 13, 16), so we included five studies (35–39) of women with breast cancer following Hodgkin’s lymphoma diagnosed at any age. These eight studies reported on a total of 400 women with 451 cases of breast cancer with median ages of diagnosis of their primary cancers between 13 and 27 years.
The cohorts were predominantly from single institutions (12, 35–39). The amount of detail regarding the characteristics of the breast cancer cases varied substantially across studies, including the location of the cancer and the stage at diagnosis. Recognizing these limitations, available retrospective evidence of fair quality suggests that once the breast cancer is diagnosed, women (and their cancers) in our target population are similar to women in the general population (12, 13, 16, 35–39).
Table 3 summarizes the breast cancer characteristics of the cases reported in the eight cohort studies and general population estimates from Surveillance, Epidemiology and End Results (SEER) (3). The median age of breast cancer in women treated with chest radiation prior to age 20 was 35 years (13, 16). For women whose Hodgkin’s was diagnosed at any age, the median age at breast cancer diagnosis was 40–45 years (12, 35–39). In contrast, the median age of diagnosis for breast cancer in the general population is 61 years of age, with 1.9% of cases occurring between ages 20–34 and 10.6% between ages 35–44 (3).
The majority (77–85%) of breast cancer pathology in pediatric and young adult cancer survivors is invasive ductal carcinoma (12, 13, 35, 37–39). This is similar to the 70–80% invasive ductal carcinomas among cases in the general population (40). In a small but well-designed case-control study of 26 women with breast cancer following Hodgkin’s lymphoma and 26 women with sporadic breast cancer who were matched for age and stage of breast cancer, there were no statistically significant differences between groups in nodal status, histology, estrogen receptor status, or lymphatic vessel invasion (41).
Survival rates in women previously exposed to radiation appear to be strongly associated with stage of disease at diagnosis (35, 38, 39), similar to the general population (3). Importantly for women with node-positive breast cancer, therapy for the initial cancer may limit options for adjuvant therapy for secondary breast cancer, which in turn may be associated with poorer outcomes. For example, in a small retrospective study of 53 women with breast cancer following a lymphoma (66% Hodgkin’s, 34% Non-Hodgkin's lymphoma) and 103 controls matched for age, stage and year of breast cancer diagnosis, and estrogen and progesterone status, Sanna et al reported a statistically significantly lower 5-year overall survival among the lymphoma cases compared to the controls (86.6% vs 98.6%, P=0.03) (42). This difference was attributed to less aggressive adjuvant therapy used for node-positive lymphoma survivors; because of their previous lymphoma therapy, only 36% of the lymphoma survivors were treated with doxorubicin, a standard adjuvant option for patients with nodal involvement, in comparison to 69% of breast cancer patients in the control group (P=0.03).
Bilateral breast cancer incidence appears increased in women treated with chest radiation for a pediatric or young adult cancer. From the large Childhood Cancer Survivor Study cohort, Kenney et al reported bilateral disease in 17% of breast cancer cases: 5% synchronous, 12% metachronous (13). In the three Hodgkin’s lymphoma studies, information regarding bilateral cancer was provided (35, 38, 39). Of 219 women with breast cancer, 12.8% had bilateral disease; 5.5% synchronous, 7.3% metachronous. In contrast, in three population-based studies of the general population, 3.3–5.3% of the women had bilateral disease; 1.5–3.1% synchronous, 1.2–3.8% metachronous (43–45). However, the population at risk for breast cancer following chest radiation is still relatively young, so the percent of cases with metachronous disease will likely increase over time.
Two retrospective and three prospective surveillance studies provide information about the benefits and harms of breast cancer surveillance.
The two retrospective studies report the method of breast cancer detection for 92 women from two institutions at a time when breast magnetic resonance imaging (MRI) was not available (all breast cancers diagnosed before 1997) (38, 46). Dershaw et al reported in a retrospective review that mammography demonstrated 90% of the cancers in 27 women (55% under the age of 45 years), with 38% being detected only by mammography (46). Wolden et al retrospectively examined 71 cases of breast cancer in 65 women and noted 27% of breast cancers were initially detected by screening mammograms (38).Notably, 75% of the women at time of breast cancer diagnosis were premenopausal (median age, 43 years). They reported that following a 1990 institutional recommendation for mammographic screening in women with a history of mantle radiation, 27 of 37 cases (73%) were stage 0 or 1 at diagnosis compared to 13 of 28 women (46%) with stage 0 or 1 cases prior to 1990 (P=0.05) (38). Interpretation of these findings is limited by the retrospective nature of the studies and the very small sample sizes. Nevertheless, these two studies suggest that mammography can detect breast cancers in this population of women, including those who are premenopausal.
Three prospective surveillance studies assessed surveillance in 320 women, ages 24–55 years (47–49) (Table 4). Kwong et al (48) reported only data from the baseline mammogram of participants, while Diller et al report data from continued breast cancer surveillance for a median of 3.1 years and Lee et al report data for a median follow-up of 5 years (47, 49). Almost all of these women were screened with mammography alone; fewer than 5% were screened with MRI or ultrasound. None of the studies evaluated the cost of surveillance. Strengths of these studies included the prospective design, details regarding breast density, reporting of prevalent and incident (47, 49) cases of breast cancer, and recall/false positive rates (47, 48). The women in these studies were from three different regions of North America and had socioeconomic indicators similar to that of women in the Childhood Cancer Survivor Study, thus suggesting that the studies, while small, appear representative of the target population.
Of the 239 baseline mammograms reviewed for breast density, 60.6% were described as moderately to very/extremely dense (47–49). Following the baseline mammogram, 27 (15.2%) of 178 women were recalled for further testing (47, 48). Fifteen women had unremarkable magnification views and/or ultrasounds and did not require additional testing. The remaining 12 women underwent biopsy with the following results: invasive breast cancer (N=2), ductal carcinoma-in-situ (N=1), lobular carcinoma-in-situ (N=1), follicular small-cleaved cell lymphoma (N=1), benign conditions including atypical hyperplasia (N=7).
Among 205 women followed prospectively, 20 developed 22 incident cases of breast cancer (7 ductal carcinoma-in-situ, 15 invasive breast cancer) over 1074.8 person-years of follow-up (47, 49), for about 2.05 cases per 100 person-years of follow-up. In the United States during this time period, for women younger than 50, the incidence of breast cancer (ductal carcinoma-in-situ and invasive) was 0.05 cases per 100 person-years of follow-up (3). Including incident and prevalent cases, 58% of the breast cancers were detected initially by mammography while 42% were detected by palpation. Of the invasive breast cancers detected initially by mammography, all were T1 (size < 2 cm). At the time of last contact, only one woman from these two studies died secondary to her breast cancer (T size 5 cm; detected initially on physical examination) (47, 49).
Among the 178 women in the Diller and Kwong studies, 12.3% had a false-positive mammogram: 8.4% who were recalled and needed only additional imaging studies and 3.9% who required a biopsy (47, 48). While Lee et al (49) did not provide data regarding false-positive mammograms, they reported that the false-negative rate of annual mammography was 5%. Lastly, among the women in the Kwong study, a telephone-based counseling intervention aimed at increasing screening rates was tested. Notably, breast cancer worries and depressive symptoms did not increase from pre-intervention to post-intervention among women who were informed about their risk (50).
In summary, there is consistent observational evidence showing that women treated for a pediatric or young adult cancer with moderate- to high-dose therapeutic chest radiation (≥20 Gy) have a substantially elevated risk of breast cancer at a young age and that this excess risk does not appear to plateau with aging. Available limited evidence suggests that the characteristics of the breast cancers in these women and the outcomes following diagnosis are similar to those in the general population; and that there appears to be a benefit associated with early detection given that women diagnosed with early stage breast cancer following chest radiation have a high likelihood for a favorable outcome. However, many of these women may have fewer treatment options for their breast cancer (eg, additional radiation, doxorubicin chemotherapy) because of treatment exposures used to cure their first cancer (35, 36, 38, 39, 42, 51, 52).
Mammography appears to detect the majority of cancers in these women. However, more than half of mammograms in women who had previous chest radiation have moderate to very dense breast tissue, thus limiting the sensitivity of mammography in detecting early cancers in this population. A systematic review of 11 prospective studies reported that screening with both MRI and mammography among women with a hereditary risk of breast cancer appears to rule out cancerous lesions better than mammography alone (53). While all 11 studies reported a higher sensitivity for MRI than mammography for invasive cancer, mammography was more sensitive than MRI for ductal carcinoma-in-situ. It is not known whether combining mammography with MRI is superior to either test alone for detecting early cancers among women who have been treated with chest radiation. However, given the similar incidence rates among young women between these two high risk populations, the increased likelihood of dense breast tissue, and the similar response to curative therapy for early diagnosed breast cancer, it is seems reasonable to speculate that the data from the hereditary risk-focused studies apply to women treated with chest radiation. Given the relatively limited size of both populations, an adequately powered randomized clinical trial to determine if surveillance with mammography and/or MRI (versus no surveillance) is associated with a reduction in mortality is unlikely to be feasible or ethical (53–56). However, ongoing prospective and high quality retrospective studies that further assess the screening metrics of different imaging approaches for women who have been treated with chest radiation might still provide useful information.
It should be recognized that while women in this risk group may benefit from breast cancer surveillance at a younger age, there is still too little understood regarding the potential harms. For example, women initiating surveillance at 25 years of age would have at least 15 additional mammograms (prior to initiating screening at the usual age of 40) and thus an increased likelihood of experiencing false positive tests with the associated economic and emotional costs of additional testing and/or biopsies. Adding an MRI may further add to the false positive rate among these women and would substantially increase the economic cost of surveillance. Of note, the combination of mammography and MRI appears to be more cost-effective in screening young women with a hereditable risk of breast cancer than mammography alone (57–60).
Another potential harm with mammography is the additional radiation exposure and the potential for radiation-induced breast cancer. Among women in the general population (61–67) and those with a hereditary risk of breast cancer (68–70), much effort has been devoted to estimating the number of breast cancers induced by mammography. Because there are no studies that directly measure risk of breast cancer caused by radiation exposure with mammography, estimates have been based upon low-dose radiation exposure from other sources, such as atomic bomb radiation, chest fluoroscopy, and thymic irradiation (71). Combining this information with the number of breast cancers detected provides an indicator of the benefit/risk balance (63, 64). The estimated mean breast dose with a contemporary standard 2-view screening mammogram is about 3.85 – 4.5 mGy (68, 72, 73). Thus, in a woman treated with 20 Gy chest radiation, fifteen additional surveillance mammograms from age 25 to 39 would increase the total radiation exposure to 20.05775 Gy or by about 0.3%. Faulkner and Law estimate that among women in the general population between the ages of 30–34, radiation exposure from a 2-view mammogram induces about 82 cancers per million women screened and detects about 630 cancers, giving a detected/induced ratio of almost 8 (74). For women ages 35–39, the detected/induced ratio is about 19. Based upon the lifetime risks of women treated with moderate to high-dose chest radiation at a young age, they estimate that detection/induction ratios will be at least 3–4 times higher than the general population (74). In addition, they emphasize that the induction rates from mammography used in the general population are likely not applicable to this population of women treated with high doses of therapeutic radiation (74).
In addition to better understanding the potential harms and benefits of specialized surveillance in this population, we need to better understand the multiple factors that may modify breast cancer risk of women who have received low or moderate dose chest radiation, especially given recent modifications in therapy including reduction in the radiation dose and reduction in the volume of developing breast tissue exposed (75). While it is anticipated that the incidence of breast cancer may decrease in women treated with contemporary therapy, it is not known what the long-term effect of these lower radiation doses will be on overall risk. Potentially, lower exposure may still be highly carcinogenic but associated with a longer latency period to induction of breast cancer.
Many women treated with chest radiation are unaware of their risks, are not followed at a cancer center, and are seen by clinicians who may be unfamiliar with their long-term health risks. In a recent survey of women in North America treated for a pediatric cancer with chest radiation, only 49% were aware that chest radiation increased the risk of breast cancer. Among women ages 25–39, nearly half had never had a mammogram (or other imaging study) and less than 20% were in a regular pattern of BC surveillance (76). While most women in this high risk group who were 40–50 years of age reported some level of breast cancer surveillance, only 53% had at least two mammograms in the previous four years. Thus, interventions aimed at informing women and their clinicians regarding these risks and options for breast cancer surveillance are needed.
Finally, a continuing theme of research that aims to reduce the very serious long-term morbidity and premature mortality faced by pediatric and young adult cancer survivors (1, 77–79) is that there is consistent evidence linking the exposure to the late effect (i.e., chest radiation and breast cancer), but limited evidence that specialized surveillance will benefit this high risk population. In balancing the potential life-saving benefits of a young and productive population of women with the potential harms of specialized breast cancer surveillance, various national and international groups, based upon consensus, recommend initiating surveillance at a young age (55, 80–82) (Table 5). Current recommendations are based upon rather arbitrary radiation levels to the chest, which will need to be continually re-evaluated as research provides better estimates of dose-specific risks (24). Presently, for women treated for a pediatric or young adult cancer with chest radiation ≥20 Gy, the COG recommends annual surveillance mammography and MRI, starting at age 25 or eight years after completion of radiation therapy, whichever occurs last. These recommendations can be viewed in the Long-Term Follow-Up Guidelines document (pages 128–129) posted at www.survivorshipguidelines.org.
In summary, we identified consistent evidence that women treated for a pediatric or young adult cancer with moderate- to high-dose therapeutic chest radiation (≥20 Gy) have a substantially elevated risk of breast cancer at a young age and that this excess risk does not appear to plateau with aging. Available limited evidence suggests that the characteristics of the breast cancers in these women and the outcomes following diagnosis are similar to those in the general population; and that there appears to be a benefit associated with early detection given that women diagnosed with early stage breast cancer following chest radiation have a high likelihood for a favorable outcome. Further research is required to better define the harms and benefits of lifelong surveillance, and how estimates of risk and outcome might change given lower radiation doses use in contemporary treatment.
Grant support: Children’s Oncology Group is supported by Grant U10 CA098543 from the National Cancer Institute.
Publisher's Disclaimer: “This is the pre-publication, author-produced version of a manuscript accepted for publication in Annals of Internal Medicine. This version does not include post-acceptance editing and formatting. The American College of Physicians, the publisher of Annals of Internal Medicine, is not responsible for the content or presentation of the author-produced accepted version of the manuscript or any version that a third party derives from it. Readers who wish to access the definitive published version of this manuscript and any ancillary material related to this manuscript (correspondence, corrections, editorials, linked articles, etc…) should go to www.annals.org or to the print issue in which the article appears. Those who cite this manuscript should cite the published version, as it is the official version of record.”