Squamous cell cancer of the anus is a rare malignancy that constitutes 1–2% of all gastrointestinal tumors. The American Cancer Society estimates that 5,260 incident anal cancer cases and approximately 720 attributable deaths will have occurred in 2010.1 The incidence of anal cancer has increased significantly over the past 2–3 decades.2 In a population-based review of the Surveillance, Epidemiology and End Results (SEER) cancer registry, a 2.8-fold increase in the incidence of anal cancer was identified from 1973 through 1998.2 Much of this increased incidence is thought to be attributable to infection with human papillomavirus (HPV), similar to the pathophysiology of cervical cancer. In fact, the presence of high-risk types of human papillomavirus, particularly HPV-16 or HPV-18, has been identified in a large proportion of anal cancer tissue specimens.3–6
Prevention of anal cancer with screening has been proposed to reduce the burden of disease from this malignancy. However, general population-wide screening with anal Papanicolaou smear would not likely be cost-effective or acceptable to most patients. Recommendations for anal cancer screening that target high-risk groups are needed. Presently, men who have sex with men and immuno-compromised hosts are considered to be at high risk for anal cancer.7 In addition, several population-based studies report an increase in anal cancer incidence among women with invasive and in situ cervical cancer8,9 as well as invasive vulvar and vaginal cancer.10 At this time, few data exist regarding the risk associated with in situ gynecologic neoplasm, and whether it differs from the risk associated with invasive gynecologic neoplasm.11 We used the SEER registry, a large population-based cancer registry, to measure the incidence of secondary anal cancers in a cohort of women with either in situ or invasive cervical, vulvar, or vaginal human papillomavirus–related neoplasm. We also characterized the duration between the diagnoses of the human papillomavirus–related gynecologic neoplasm and anal cancer and investigated whether radiation therapy for gynecologic cancer modifies the risk of a subsequent anal cancer.
MATERIALS AND METHODS
We used data from the SEER program for this study and included 17 registries from Atlanta, Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco–Oakland, Seattle–Puget Sound, Utah, Greater California, Kentucky, Louisiana, New Jersey, Georgia, Alaska, Los Angeles, and San Jose-Monterey.12 Information collected by SEER includes patient characteristics, tumor site, grade, stage, first course of treatment, and follow-up for vital status.13 We hold a data use agreement with the National Cancer Institute and our study protocols were reviewed and considered exempt by the Lahey Clinic Institutional Review Board. The SEER public access user file makes every effort to protect the identities of cancer patients. The data use agreement specifically requests that all research results must be presented or published in a manner that ensures that no individual can be identified.13 As per our data use agreement, we have hidden all sample sizes of fewer than five patients.
We identified all women diagnosed with invasive cervical, vaginal, or vulvar cancer from 1973 to 2007 or in situ cervical, vaginal, or vulvar cancer from 1973 to 1995. The SEER program ceased tracking in situ tumors of the cervix after 1995. The inclusion criteria were all girls and women older than 15 years of age at the time of a primary diagnosis of invasive cervical, vulvar, or vaginal cancer that was histopathologically categorized as squamous cell carcinoma. A total of 56,876 girls and women had an invasive gynecologic neoplasm, and 7,886 were excluded because of a history of previous primary cancer. Also excluded were 22,625 women with primary gynecologic neoplasm of the following cell types: adenocarcinoma, adenosquamous carcinoma, endometrioid adenocarcinoma, serous adenocarcinoma, and undifferentiated carcinoma. Study size was determined by number of patients who met inclusion and exclusion criteria.
We identified all women who developed squamous cell cancers of the anus after a primary diagnosis of either in situ or invasive primary gynecologic neoplasm. Two methods were used for selection of secondary anal cancer cases: the restrictive and the inclusive methods. In the restrictive method, we limited our cases to women who developed anal cancer as the second primary cancer. If another cancer developed between the gynecologic neoplasm and anal cancer, then it was excluded in the restrictive method. In the inclusive method, anal cancer could develop as a second, third, or fourth primary cancer. Standardized incidence ratios were calculated by both methods. Thirteen women who developed a second primary anal cancer within 1 year of the primary gynecologic diagnosis were also excluded to avoid potential bias of synchronous tumors, which may have made determination of time sequence difficult.
The computation of standardized incidence ratios was conducted using indirect standardization methods applied to the person-time accumulated among individuals meeting inclusion criteria from the SEER population. Person-years at risk for the development of subsequent cancers for each woman began at 2 months after the date of diagnosis of the gynecologic cancer and ended at the date of last contact with the patient, death, or end of the study period on December 2006, whichever was earliest. The person-years and observed cases of subsequent anal cancers were stratified according to patient age at initial gynecologic neoplasm diagnosis (5-year groups), race or ethnicity, and 5-year calendar intervals. Anal cancer incidence rates among the female population were calculated and stratified by age, race, and calendar-year group, and were multiplied by the person-years accrued by the gynecologic cancer cases to estimate the expected numbers of subsequent cancers for each stratum. The observed and expected numbers of subsequent cancers for each stratum were then summed. The standardized incidence ratio represents the ratio of the observed number divided by the expected number of subsequent cancers. Briefly, the standardized incidence ratio reports the incidence of cancer in a population at risk compared with an expected incidence of cancer in a population determined to be at average risk or “normal.” For example, in our manuscript a standardized incidence ratio of 1.0 indicates that the observed number of anal cancer cases in those patients with a history of gynecologic neoplasm is equal to what would be the expected number of anal cancer cases from a comparison “average” population in the SEER program. A standardized incidence ratio greater than 1.0 would indicate that more cancer cases occurred than expected; thus, a standardized incidence ratio of 2.0 should be interpreted as observing twice as many anal cancer cases among a population with gynecologic neoplasia than the expected number among an “average” population of women. Confidence intervals (CIs) for the standardized incidence ratios were calculated using the Vandenbroucke method.14 Standardized incidence ratios were calculated separately for in situ and invasive cancers. SEER*STAT software and SAS 9.2 were used for data analysis.13
In an effort to assess the potential effects of radiotherapy on the subsequent development of primary anal cancers, gynecologic cancer cases were stratified by whether they received radiation therapy (yes or no) and standardized incidence ratios were calculated as described above within these strata. A total of 1,103 women with unknown radiotherapy status were excluded from this analysis.
We also described the time period between the primary gynecologic neoplasm and subsequent anal cancer diagnosis using means and standard deviations and tested for differences between invasive and in situ neoplasm with the Student t test. Finally, we used Kaplan-Meier curves to characterize the risk of anal cancer over the duration of follow-up.
We identified a total of 189,206 cases: 132,330 cases of in situ human papillomavirus–related gynecologic neoplasm (124,075 cervical, 6,792 vulvar, and 1,463 vaginal) and 56,876 cases of invasive gynecologic neoplasm (43,669 cervical, 9,950 vulvar, and 3,257 vaginal). Demographic characteristics of the index cohort are presented in Table 1.
Using the inclusive method, we identified 255 cases of anal cancer, 58 anal cancers among women with invasive primary gynecologic neoplasm and 197 anal cancers after an in situ gynecologic neoplasm, during a follow-up of 138,553,519 person-years (Table 2). The aggregate standardized incidence ratio for the incidence of anal cancer in all patients with gynecologic neoplasm was 13.6 (95% CI 11.9–15.3).
The standardized incidence ratio for the incidence of anal cancer was 6.2 (95% CI 4.1–8.7), 17.4 (95% CI 11.5–24.4), and 1.8 (95% CI 0.2–5.3) for women with invasive cervical, vulvar, and vaginal cancer, respectively (Table 3). The standardized incidence ratios for women with in situ gynecologic neoplasm were higher than those for women with invasive cancers (cervical cancer: standardized incidence ratio=16.4, 95% CI 13.7–19.2; vulvar cancer standardized incidence ratio=22.2 (95% CI 16.7–28.4), vaginal cancer standardized incidence ratio=7.6 (95% CI 2.4–15.6) (Table 3).
The analyses were repeated using the restrictive method to remove the potential of confounding from other cancer treatment. Using this approach, the standardized incidence ratio for anal cancer was 5.1 (95% CI 3.2–7.4), 14.3 (95% CI 9.0–20.9), and 1.8 (95% CI 0.2–5.4) for women with an initial invasive cervical, vulvar, and vaginal cancer diagnosis, respectively. The standardized incidence ratio for anal cancer was 14.3 (95% CI 12.0–17.2), 16.1 (95% CI 11.4–21.6), and 6.1 (95% CI 1.5–13.6) for in situ cervical, vulvar, and vaginal neoplasm, respectively.
Kaplan-Meier curves characterizing the time to anal cancer diagnosis for each gynecologic malignancy were calculated. The mean time interval between the incidence of primary gynecologic malignancy and the diagnosis of a second primary anal cancer was longest in women with in situ cervical cancer (15.7 years). In fact, the interval between diagnoses was longer for in situ compared with invasive cancers for all gynecologic neoplasm, although the small number of anal cancers observed among the vaginal cancer cases limits this conclusion (Table 4).
Radiation therapy was reported as a treatment used in 23,884 (55.6%) women with invasive cervical cancer, 2,215 (22.7%) women with invasive vulvar cancer, and 2,008 (63.6%) women with invasive vaginal cancer. The risk of anal cancer in women with a previous cervical cancer diagnosis was similar among women who did not receive radiotherapy (standardized incidence ratio=3.1, 95% CI 1.6–4.9) compared with those who did (standardized incidence ratio=2.9, 95% CI 1.5–4.6) (Table 5). The data for vulvar and vaginal cancer were inconclusive.
We identified a significant association between gynecologic neoplasm and anal cancer for both in situ and invasive cancers of the cervix and vulva and in situ neoplasm of the vagina. The highest risk for anal cancer was identified in those women with evidence of either in situ or invasive squamous cell cancer of the vulva. These data indicate that women with both in situ and invasive cancers of the cervix and vulva are at higher risk for developing anal cancer than the general population and may benefit from close observation and anal cancer screening.
The pathway of human papillomavirus–related malignant transformation for cervical cancer has been well established and has led to effective prevention strategies. The National Cancer Institute recommends cervical cancer screening with Papanicolaou smear at least once every 3 years, starting within 3 years of the initiation of sexual intercourse, but no later than age 21.15 Although randomized controlled trials demonstrating a benefit to cervical cancer screening have not been conducted, observational studies from the International Agency for Research on Cancer have shown a 91–94% reduction in cervical cancer incidence with screening.16 The incidence of cervical cancer decreased from 32 cases per 100,000 women in the 1940s to 8.3 cases per 100,000 women in the 1980s in the United States, and is largely attributed to increased use of cervical cytology to detect precursor lesions.17
Although it would be impractical and not cost-effective to implement a policy of routine anal cancer screening in the United States for all sexually active adults, screening has been proposed for individuals at elevated risk for anal cancer. Given that routine Papanicolaou smear screening has significantly reduced cervical cancer incidence and mortality across many populations, it is expected that similar screening of populations at high risk for anal cancer precursor lesions will also reduce the burden of this disease.18 To accomplish this goal, high-risk groups must be identified to properly target those individuals who may benefit. Our data reveal that women with a previous gynecologic neoplasm represent a high-risk group, particularly those women with vulvar neoplasm.
Epidemiologic data reveal that as in cervical cancer, the majority of patients with anal squamous cell cancer have had associated infection with a similar HPV subtypes. In particular, HPV-16 is the most commonly isolated subtype from anal cancer tissue specimens.3 Given the common precursor for anal and gynecologic squamous cell cancers, a common pathway for cancer transformation in the genitalia and anus has been proposed, thereby placing women with genital cancers at high risk for developing anal cancer. Previous data have shown that women with cervical, vulvar, and vaginal cancers are at higher risk for anal human papillomavirus infection and anal intraepithelial neoplasia.19–22 In addition, population-based studies reveal that the risk of anal cancer is significantly greater in women with in situ or invasive cervical cancer.8 Our study using population-based data from the United States reports a standardized incidence ratio of 16.3 and 6.2 for women with in situ and invasive cervical cancer, respectively. The increased risk among the in situ group may be due to differences in treatment, surveillance, strain of human papillomavirus, or other host defenses. Not surprisingly, because of the proximity, the highest risk of anal cancer was noted in women with in situ vulvar neoplasm. Length of follow-up is unlikely to account for the increased risk with in situ neoplasia as our study did adjust for patient follow-up. Other possible explanations for the increased risk of anal cancer with in situ neoplasm may be related to differences in treatment, surveillance, strain of human papillomavirus, or other host defenses.
In addition to the increased risk of anal cancer with in situ neoplasm, we also evaluated the role of radiation in modifying the association between gynecologic neoplasm and anal cancer. We hypothesized that exposure to irradiation would prevent development of anal cancer, given the therapeutic role of irradiation as a component of the Nigro protocol, consisting of 5-fluorouracil, mitomycin C, and 30 Gy of radiation therapy.23 Chaturvedi et al similarly examined the incidence of secondary cancers after cervical cancer and found that all women had a higher risk of developing anal cancer compared with the general population, regardless of radiation therapy status.24 Consistent with their result, we found that women with invasive gynecologic malignancy who were exposed to radiotherapy had a similar risk of anal cancer compared with those who did not receive radiotherapy. Our findings are not compatible with our hypothesis that radiation therapy sterilizes dysplasia in those areas included in the field of treatment, including the anus.
Our study has limitations related to the nature of the data used in our analyses. First, although the data are collected prospectively, we analyzed the data retrospectively, which might lead to selection or information bias. Second, migration of participants out of the states included in the SEER registry may have limited the capture of second primary tumors. Migration status is not recorded in our dataset25 and might lead to underreporting of second primary cancers, resulting in a conservative estimate of subsequent cancer risk. Additionally, we were unable to explore the role of human papillomavirus–positive status because the tumor registry does not record this information. However, it is acknowledged that human papillomavirus infection is present in almost all cases of cervical cancer,26 and in the vast proportion of vulvar and vaginal cancers.27 Fourth, we were unable to identify the role of HIV on the overall effect toward anal cancer development. It is clear that patients with HIV are at increased risk for squamous cell cancer of the anus, and that infection with this virus rose substantially during the study period. A fifth limitation is that cancer patients are often under more intense medical surveillance than the general population, leading to ascertainment bias11 and a possible inflation of the standardized incidence ratio. Notwithstanding the potential biases and caveats described, the SEER program is an excellent population-based tool that broadly characterizes cancer epidemiology and treatment, avoiding potential biases that can be introduced in hospital-based studies as a result of area referral patterns and incomplete follow-up.25
In conclusion, our study has demonstrated that women with a wide range of primary human papillomavirus–related gynecologic neoplasm are at higher risk for developing anal cancer compared with the general population. On average, anal cancer developed between 4 and 16 years after diagnosis and thus, women with gynecologic neoplasm may benefit from early anal cancer screening. According to our data, those treated with irradiation are not protected from the development of incident anal cancer and should still be considered high risk. At this time, cost-effectiveness analyses are needed to determine the value of anal cancer screening in women with primary human papillomavirus–related gynecologic neoplasm.
2. Maggard MA, Beanes SR, Ko CY. Anal canal cancer: a population-based reappraisal. Dis Colon Rectum 2003;46:1517–24; discussion 1523–4; author reply 1524.
3. Frisch M, Glimelius B, van den Brule AJ, Wohlfahrt J, Meijer CJ, Walboomers JM, et al. Sexually transmitted infection as a cause of anal cancer. N Engl J Med 1997;337:1350–8.
4. Daling JR, Madeleine MM, Johnson LG, Schwartz SM, Shera KA, Wurscher MA, et al. Human papillomavirus, smoking, and sexual practices in the etiology of anal cancer. Cancer 2004;101:270–80.
5. Bjørge T, Engeland A, Luostarinen T, Mork J, Gislefoss RE, Jellum E, et al. Human papillomavirus infection as a risk factor for anal and perianal skin cancer in a prospective study. Br J Cancer 2002;87:61–4.
6. Carter JJ, Madeleine MM, Shera K, Schwartz SM, Cushing-Haugen KL, Wipf GC, et al. Human papillomavirus 16 and 18 L1 serology compared across anogenital cancer sites. Cancer Res 2001;61:1934–40.
8. Edgren G, Sparen P. Risk of anogenital cancer after diagnosis of cervical intraepithelial neoplasia: a prospective population-based study. Lancet Oncol 2007;8:311–6.
9. Hemminki K, Dong C, Vaittinen P. Second primary cancer after in situ and invasive cervical cancer. Epidemiology 2000;11:457–61.
10. Jiménez W, Paszat L, Kupets R, Wilton A, Tinmouth J. Presumed previous human papillomavirus (HPV) related gynecological cancer in women diagnosed with anal cancer in the province of Ontario. Gynecol Oncol 2009;114:395–8.
11. Curtis RE, Freedman DM, Ron E, Ries LAG, Hacker DG, Edwards BK, et al (eds). New malignancies among cancer survivors: SEER cancer registries, 1973–2000. NIH Publication No. 05-5302. Bethesda (MD): National Cancer Institute; 2006.
13. National Cancer Institute's Surveillance, Epidemiology and End Results: SEER*Stat Software, Version 6.62. Available from: http://seer.cancer.gov/seerstat/
. Retrieved December 2009.
14. Vandenbroucke JP. A shortcut method for calculating the 95 per cent confidence interval of the standardized mortality ratio [letter]. Am J Epidemiol 1982;115:303–4.
16. Lynge E. Screening for cancer of the cervix uteri. World J Surg 1989;13:71–8.
17. Devesa SS, Silverman DT, Young JL Jr, Pollack ES, Brown CC, Horm JW, et al. Cancer incidence and mortality trends among whites in the United States, 1947–84. J Natl Cancer Inst 1987;79:701–70.
18. Arain S, Walts AE, Thomas P, Bose S, et al. The Anal Pap Smear: cytomorphology of squamous intraepithelial lesions. Cytojournal 2005;2:4.
19. Park IU, Ogilvie JW Jr, Anderson KE, Li ZZ, Darrah L, Madoff R, et al. Anal human papillomavirus infection and abnormal anal cytology in women with genital neoplasia. Gynecol Oncol 2009;114:399–403.
20. Ogunbiyi OA, Scholefield JH, Robertson G, Smith JH, Sharp F, Rogers K. Anal human papillomavirus infection and squamous neoplasia in patients with invasive vulvar cancer. Obstet Gynecol 1994;83:212–6.
21. Scholefield JH, Sonnex C, Talbot IC, Palmer JG, Whatrup C, Mindel A, et al. Anal and cervical intraepithelial neoplasia: possible parallel. Lancet 1989;2:765–9.
22. Véo CA, Saad SS, Nicolau SM, Melani AG, Denadai MV. Study on the prevalence of human papillomavirus in the anal canal of women with cervical intraepithelial neoplasia grade III. Eur J Obstet Gynecol Reprod Biol 2008;140:103–7.
23. Nigro ND. Multidisciplinary management of cancer of the anus. World J Surg 1987;11:446–51.
24. Chaturvedi AK, Engels EA, Gilbert ES, Chen BE, Storm H, Lynch CF, et al. Second cancers among 104,760 survivors of cervical cancer: evaluation of long-term risk. J Natl Cancer Inst 2007;99:1634–43.
25. Yu JB, Gross CP, Wilson LD, Smith BD. NCI SEER public-use data: applications and limitations in oncology research. Oncology (Williston Park) 2009;23:288–95.
26. Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189:12–9.
27. De Vuyst H, Clifford GM, Nascimento MC, Madeleine MM, Franceschi S. Prevalence and type distribution of human papillomavirus in carcinoma and intraepithelial neoplasia of the vulva, vagina and anus: a meta-analysis. Int J Cancer 2009;124:1626–36.