Lichter MD, Karagas MR, Mott LA, Spencer SK, Stukel TA, Greenberg ER, for the New Hampshire Skin Cancer Study Group. Therapeutic Ionizing Radiation and the Incidence of Basal Cell Carcinoma and Squamous Cell Carcinoma. Arch Dermatol. 2000;136(8):1007-1011. doi:10.1001/archderm.136.8.1007
To estimate the relative risk of developing basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) after receiving therapeutic ionizing radiation.
Population-based case-control study.
A total of 592 cases of BCC and 289 cases of SCC identified through a statewide surveillance system and 536 age- and sex-matched controls selected from population lists.
Main Outcome Measures
Histologically confirmed BCC and invasive SCC diagnosed between July 1, 1993, through June 30, 1995, among New Hampshire residents.
Information regarding radiotherapy and other factors was obtained through personal interviews. An attempt was made to review the radiation treatment records of subjects who reported a history of radiotherapy. Overall, an increased risk of both BCC and SCC was found in relation to therapeutic ionizing radiation. Elevated risks were confined to the site of radiation exposure (BCC odds ratio, 3.30; 95% confidence interval, 1.60-6.81; SCC odds ratio, 2.94; 95% confidence interval, 1.30-6.67) and were most pronounced for those irradiated for acne exposure. For SCC, an association with radiotherapy was observed only among those whose skin was likely to sunburn with sun exposure.
These results largely agree with those of previous studies on the risk of BCC in relation to ionizing radiation exposure. In addition, they suggest that the risk of SCC may be increased by radiotherapy, especially in individuals prone to sunburn with sun exposure.
TREATMENT WITH prolonged administration of low-dose ionizing radiation is associated with subsequent nonmelanoma skin cancer (NMSC), particularly basal cell carcinoma (BCC).1 Implicated treatments include radiotherapy of inflammatory dermatoses such as eczema, psoriasis, acne vulgaris, and tinea capitis.2,3 Additionally, an increased risk of BCC has been related to ionizing radiation treatment of goiters,4 ankylosing spondylitis,5- 8 acute lymphocytic leukemia, and astrocytoma.9 Whether radiotherapy enhances risk of squamous cell carcinoma (SCC) is less clear. A population-based study among men in Alberta, Canada, reported a 5- to 6-fold increase in incidence of BCC and SCC associated with nondiagnostic x-ray exposure.10 Using data collected as part of a large case-control study in New Hampshire, we also had the opportunity to evaluate the risk of BCC and SCC in relation to previous therapeutic ionizing radiation. We examined risks according to age at first exposure, time since exposure, condition for which irradiation was administered, and site of exposure. We further assessed the frequency and number of treatments, and, for those who were irradiated for cancer, the amount of exposure by dose per week and by dose per treatment course.
The case-control study group consisted of patients with BCC and SCC, aged 25 to 74 years at diagnosis, who were identified through a population-based incidence survey of skin cancer conducted in New Hampshire covering diagnoses from July 1, 1993, through June 30, 1995.11 Age- and sex-matched controls were selected from state drivers' license lists for those younger than 65 years and from Medicare enrollment lists for those 65 years or older.
An in-person interview that took 90 minutes to 2 hours to complete was usually conducted in the subject's home. Interviewers were masked to the study hypotheses and were not told the case-control status of the participant. Interviewers asked about sun exposure history, outdoor recreation, skin reaction to sunlight, personal use of tobacco and alcohol, educational background, and medical history. To document previous exposure to therapeutic ionizing radiation, we asked subjects if they were ever treated with x-rays along with the condition and anatomical location treated, age at and duration of treatment, and the frequency of treatments. We requested permission to obtain the medical records of those who reported a history of radiotherapy. Medical records review was done without knowledge of the subject's case-control status or other data collected at the interview.
We estimated the odds ratio (OR) and 95% confidence intervals (CIs) of BCC and SCC associated with radiotherapy using unconditional logistic regression.12 We restricted our analysis to the 592 cases of BCC, 289 cases of SCC, and 536 control subjects who did not have a history of radiotherapy for skin cancer. Patients treated with ionizing radiation for skin cancer are more likely to be in the case group than in the control group. Therefore, the inclusion of these cases could bias analyses toward an association with radiotherapy. We adjusted for age (as a continuous variable) and sex in all models. In addition, we examined the potentially confounding effects of outdoor exposure (hours per week spent outdoors in the summer—recreationally and occupationally); cigarette smoking history (never, former, current); level of education (less than college, college, graduate or professional school); and sun-sensitive skin type defined according to skin reaction to strong sunlight for the first time in summer for 1 hour (severe sunburn with blistering, painful sunburn, mild sunburn and some tanning, and tanning with no sunburn). We also explored the possibility that the effects of ionizing radiation exposure may be modified by skin type.
We grouped anatomical location of both skin cancer and radiation field to the head and neck, trunk, or limbs. For controls, we randomly assigned a site based on the anatomical distribution of the overall case group. Data were too sparse to study specific anatomical sites of involvement. Therefore, we computed separate ORs for skin cancers in the radiation field and at unirradiated sites. In this analysis, we excluded individuals whose tumor or irradiation occurred at more than 1 anatomical location.
Lastly, we attempted to analyze the dose per week and the dose per treatment course among those who reported treatment for cancer and for whom irradiation records were available. Treatment for nonmalignant conditions generally had occurred before 1970; thus, the medical records were no longer available.
By design, cases and controls were comparable for age and sex. The overall mean age of cases with BCC and SCC was 58.5 years and 64.7 years, respectively, and 60.5 years for controls. About 40% of the subjects were women and 98% were white. Seven percent of the controls and 12% of the cases reported previous ionizing radiation therapy for reasons other than skin cancer.
For BCC, we found a significantly increased risk associated with radiotherapy (age and sex-adjusted OR, 1.88; 95% CI, 1.24-2.87) and for SCC a borderline significant increased risk (age- and sex-adjusted OR, 1.56; 95% CI, 0.95, 2.55) (Table 1). Adjustment for other potentially confounding factors had no appreciable effect on the risk estimates.
The magnitude of the ORs differed according to the reported reason for treatment. Radiotherapy for acne was associated with a pronounced and statistically significant increased risk for BCC and SCC (Table 1). A 2-fold increase in BCC risk was found in relation to radiotherapy for cancer, but no increased risk of SCC was found (Table 1). No relation was observed among those who had radiotherapy for other benign skin conditions.
Odds ratios for BCC and SCC according to time since the initial exposure and age first treated with radiation are given in Table 1. Persons whose first radiation treatment occurred before they were 20 years old had a greater risk of BCC and SCC than those treated at older ages. An association with BCC and SCC was largely confined to radiotherapy 40 years or longer prior to diagnosis. However, risk of BCC was modestly elevated among those who were treated in the past 20 years or who were 40 years or older when first treated. The ORs for BCC and SCC seemed to increase with the frequency of treatments and the reported number of treatments, but the trends were not statistically significant.
Risk of BCC and SCC was related specifically to tumors arising at the site of prior radiotherapy (Table 2). The difference between the ORs for the exposed and unexposed sites was statistically significant for BCC (P = .01) and borderline significant for SCC (P = .07). When we examined whether the risks associated with radiotherapy were modified by skin type, we found that for BCC, the ORs for radiation treatment were comparable among those who tend to burn and those who tend to tan with first exposure to sunlight in the summer (among burners, OR,1.87; 95% CI, 0.86-4.03 and among tanners, OR, 1.96; 95% CI, 1.18-3.26) (P value for interaction, .97). Whereas for SCC, risk was increased only among those with a tendency to burn (OR, 3.02; 95% CI, 1.37-6.68) and was not elevated among those who tend to tan (OR, 0.73; 95% CI, 0.35-1.52) (P value for interaction, .01).
Medical records were reviewed for 44 (79%) of the 56 subjects who reported receiving therapeutic ionizing radiation for cancer. Of the 39 patients receiving radiation beam therapy alone, the type of devices used were cobalt 60 (n = 5), linear accelerator with photons (n = 27), linear accelerator with electrons (n = 2), and linear accelerator with photons and electrons (n = 5). Two patients had radiation implants; 3 persons were missing device information. Restricting to patients with a single course of beam therapy, the ORs for BCC were elevated for exposures less than or equal to 10 Gy/wk (OR, 3.09; 95% CI, 1.10-8.70), and when the dose per treatment was less than or equal to 2 Gy (OR, 3.53; 95% CI, 1.28-9.79). We were unable to compute ORs for higher doses of treatment because only a few individuals had them. There were too few cases of SCC for which we obtained radiotherapy records to examine risk by dose of treatment for this cell type.
Our data indicate that previous therapeutic ionizing radiation leads to an increased risk of BCC and SCC. For both histological types, risks were confined to the region of the body to which radiotherapy was delivered. Risk also was highest among those who had radiotherapy for treatment of acne. For SCC, risk was principally elevated among those with a sun-sensitive phenotype.
Although risk of BCC has been consistently related to radiation exposure in prior studies, results regarding SCC are less clear. Studies of atomic bomb survivors in Japan have found a relation for BCC, but not SCC.13,14 Prior cohort studies of individuals previously treated with ionizing radiation for tinea capitis found 5- and 6-fold increases in the relative risk for head and neck BCC, but no increase in SCC.13 The lack of increased risk for SCC possibly could be because of the younger age of those studied especially in the follow-up studies of childhood exposures. A Canadian case-control study of men identified through the Alberta Cancer Registry reported ORs of 5.7 and 4.8 for BCC and SCC, respectively, following nondiagnostic radiation exposure.10 In a recent cohort study of patients with NMSC, a relative risk of 1.7 was found for new BCC occurrences, and a relative risk of only 1.0 for new SCC occurrences15; a limitation of this study is that all subjects had at least 1 prior NMSC.
In our current case-control study, an association between radiotherapy and risk of SCC was found specifically among those with a sun-sensitive phenotype. To our knowledge, this finding has not been observed previously, and in part, may explain the absence of an association with SCC among atomic bomb survivors in Japan. In a cohort study of children irradiated for tinea capitis,14 the relation with NMSC was lower among Israelis of Asian or African descent than other Israelis. Further studies may help to clarify the possibility of individual susceptibility to radiation-induced NMSC.
Younger age at first ionizing radiation exposure has been associated with greater risk of subsequent NMSC. Japanese atomic bomb survivors from Hiroshima or Nagasaki who were aged 40 years or older at detonation had no apparent increased risk of NMSC.15 Among children irradiated for tinea capitis, there was an inverse relation with age at treatment, although the oldest treated age group was 10 to 14 years old.13 In our data, earlier age at first treatment also was more strongly related to BCC and SCC risk.
The latency period between first exposure to therapeutic ionizing radiation and the appearance of NMSC is presumed to be at least 20 years.2,14,16 However, it is difficult to separate the effects of latency from those of age at treatment and type of therapy received. In our data, risks of BCC and SCC were highest among those treated 40 years or longer before diagnosis. Basal cell carcinoma risk also was increased within 20 years of treatment; this finding may be explained in part by the association we observed between radiotherapy for cancer, a more recent exposure. To fully tease out the effects of latency from other effects would require an even larger study than ours.
In epidemiological studies, cancer risks associated with radiotherapy are specific to the sites or organs in the radiation field. Indeed, most skin cancer studies were restricted to the site of exposure (eg, head and neck skin cancers for tinea capitis treatment). Thus, our data confirm that the risk of BCC and SCC is confined to the region of the body to which the therapeutic ionizing radiation was delivered.
Potential limitations of our study are that our results are based mainly on self-report of prior therapy and that New Hampshire state driver's licenses were used to obtain age- and sex-matched controls. Regarding the first limitation, we did not find evidence of differential misclassification based on our medical record review. In the subset of subjects for whom we obtained medical records, the agreement with self-report was virtually 100% for previous history of radiotherapy (data not shown). For the second limitation, all except 5 study participants younger than 65 years had valid driver's licenses. Therefore, exclusion of these individuals did not affect our results (data not shown).
We were able to obtain total radiation doses from the medical records of a high percentage of study participants who received linear accelerator photon radiotherapy for cancer. All but 2 of these patients had total radiation doses higher than 30 Gy, which is consistent with previous estimates for the total radiation dose necessary to induce NMSC. In these patients, the risks of BCC and SCC were increased specifically among those treated with 10 Gy/wk or less and less than or equal to 2 Gy per treatment. Study participants previously irradiated for acne had especially high ORs for BCC and SCC, and these individuals probably received highly fractionated doses over longer periods.17 More highly fractionated doses of ionizing radiation, involving moderate doses for each individual fraction, is also more carcinogenic in animal models.1
Data from this population-based case-control study indicate that persons previously exposed to therapeutic ionizing radiation have an increased risk for BCC and SCC. Our data also suggest that SCC risk following irradiation may be highest in persons with a sun-sensitive phenotype.
Accepted for publication March 2, 2000.
This study was supported in part by grants CA58290, CA57494, and CA23108 from the National Cancer Institute, National Institutes of Health, Bethesda, Md, and grant ACS SIG-17 from the American Cancer Society, Atlanta, Ga.
We are grateful to the pathology laboratories and physicians who took part in the study. We also thank Mark Brauning, MD, for his assistance reviewing the medical records, Virginia Stannard, MEd, for coordination of the study, and Jennifer Miglionico, MS, for assistance in obtaining the medical records.
Duane R. Anderson, MD; Robert W. Averill, MD; Anthony J. Aversa, MD; Bruce A. Bairstow, MD; Richard D. Baughman, MD; Lawrence G. Blasik, MD; Carolyn Carroll, MD; William E. Clendenning, MD; Daniel W. Collison, MD; George L. Crespo, MD; Stephen M. Del Guidice, MD; Robert L. Dimond, MD; Wilmot S. Draper, MD; Jeremy P. Finkle, MD; William E. Frank, MD; John L. Fromer, MD; Norman C. Goldberg, MD; David Goldminz, MD; Robert Gordon, MD; David S. Greenstein, MD; Thomas P. Habif, MD; Charles Hammer, MD; Tom Hokanson, PA; Steve A. Joselow, MD; Michael D. Lichter, MD; Maritza O. Liranzo, MD; Michael A. Mittleman, MD; Jose Peraza, MD; Robert B. Posnick, MD; Warren M. Pringle, MD; Mark Quitadamo, MD; Pauline B. Reohr, MD; N. Chester Reynolds, MD; Peter Sands, MD; Mitchell E. Schwartz, MD; Steven K. Spencer, MD; James C. Starke, MD; Susan Sullivan, MD; N. Hakan Thyresson, MD; Andrew P. Truhan, MD; Mauray J. Tye, MD; K. William Waterson, MD; Kathryn Zug, MD.
Corresponding author: Margaret R. Karagas, PhD, Dartmouth Medical School, Section of Biostatistics and Epidemiology, 7927 Rubin 462M-3, One Medical Center Drive, Lebanon, NH 03756-0001 (e-mail: firstname.lastname@example.org).