Anal Cancer Incidence in the United States, 1977–2011: Distinct Patterns by Histology and Behavior

With only 7,270 cases estimated to occur in the United States during 2015, anal cancer is a relatively rare malignancy (1). However, anal cancer rates have increased steadily for decades in the United States and internationally (2–5). Although the cause of these rising rates is unclear, it has been hypothesized that changing sexual practices, leading to increased prevalence of anal infection with carcinogenic human papillomavirus (HPV), and an increasing number of individuals living with HIV, a group known to have elevated anal cancer risk, have contributed to this increase (6).

The majority of anal cancers are squamous cell carcinomas (SCC), but adenocarcinomas (ADC) make up 9% to 14% of diagnosed cases in the United States, although these proportions may vary considerably internationally (6–9). Although SCC and ADC are usually combined in cancer surveillance, their etiologies may differ. Ninety percent of anal SCCs are caused by infection with oncogenic types of HPV, primarily HPV-16, whereas HPV has been detected in a smaller fraction of ADCs (8, 10, 11). Because of the rarity of anal ADC, little is known about its etiology. Furthermore, it may be difficult to determine whether the primary site of some tumors is the lower rectum or anus; SCCs and ADCs arising across these two sites may have shared etiologies and be prone to misclassification of their primary location.

In the current analysis, we used data from the National Cancer Institute's (NCI) Surveillance, Epidemiology and End Results (SEER) Program to provide a detailed description of the demographic characteristics and temporal trends in incidence rates (IR) of anal cancer by histology (i.e., SCC and ADC) and behavior (i.e., invasive vs. in situ) in the United States. We also compare patterns to rectal cancer to assess similarities across these two sites according to histology.

We used data from SEER to assess recent IRs in the 18 registries with data for cases diagnosed during 2000–2011 (i.e., SEER 18, including approximately 28% of the U.S. population; ref. 12), and to assess temporal trends in the registries that participated in SEER 9 (1977–2011) and SEER 13 (1992–2011; refs. 13, 14).

Microscopically confirmed anal cancers were ascertained using SEER site recode, based on the International Classification of Diseases for Oncology, 3rd edn. (ICD-O-3; site code: C21.0–C21.2, C21.8, excluding histology codes 9050–9055, 9140, 9590–9992; ref. 15). Anal cancers were classified by site (anal, not otherwise specified [NOS; C21.0]; anal canal [C21.1], cloacogenic zone [C21.2], and overlapping lesion of rectum, anus, and anal canal [C21.8]). Invasive cases (i.e., behavior = 3) were further classified by histology (SCC: 8050–8076, 8083–8084, 8123–8124; ADC: 8140–8145, 8190–8231, 8260–8263, 8310, 8401, 8480–8490, 8570–8574; melanoma: 8720–8727, 8730–8743, 8745–8790; and other types; based on a modified version of ref. 16). In situ cases (i.e., behavior = 2) were restricted to SCCs (8050–8077, 8081, 8083–8084, 8123–8124) or carcinoma in situ (CIS), NOS (8010), as these were likely SCCs. For comparison, rates for invasive SCC and ADCs of the rectum (site code: C20.9) were also calculated to assess similarities across sites.

Using data from SEER 18, we estimated anal cancer IRs per 1,000,000 person-years, age-adjusted to the 2000 U.S. population across behavior/histology by age group (<30, 30–49, 50–59, 60–69, 70+ years), race/ethnicity (non-Hispanic white, black, Hispanic white, Asian/Pacific Islander, American Indian/Alaskan Native, and other/unknown), cancer registry, and anatomic subsite. IRs for American Indian/Alaskan Natives were restricted to Contract Health Service Delivery Areas (CHSDA). Rate ratios (RR) and 95% confidence intervals (CI) were estimated by sex (women vs. men) and race/ethnicity (blacks and Hispanic whites vs. non-Hispanic whites). Data from SEER 18 were also used to assess patterns across age groups by sex, race/ethnicity, and histology/behavior.

Information on Hispanic ethnicity is not available in SEER <1992. Therefore, trends in age-standardized IRs of anal cancer over calendar time (presented in 5-year calendar periods) were estimated using SEER 9 for whites and blacks (1977–1991) and SEER 13 for non-Hispanic whites, blacks, and Hispanic whites (1992–2011), stratified by sex, race/ethnicity, and histology/behavior. Trends for Asian/Pacific Islanders and American Indians/Alaska Natives were not shown due to scarcity of the data. For comparison, calendar trends (1977–2011) were also assessed for rectal cancers, restricted to whites/non-Hispanic whites, as the data in other racial/ethnic groups were too sparse for analysis. Annual percent changes (APC) in IRs during 1992–2011 were assessed with least squares regression of the natural logarithm of the rate (17). IRs were suppressed if based on <16 cases. All analyses were carried out with SEER*Stat 8.1.5 (18).

During 2000–2011 in the 18 SEER registries, a total of 16,141 cases of invasive anal cancer were diagnosed during 993,234,972 person-years of follow-up (IR = 16.3/1,000,000; Table 1). Cases were primarily SCC (n = 13,274, 82.2%), followed by ADC (n = 1,992; 12.3%), and melanoma (n = 254; 1.6%). There were 621 invasive anal cancers with other/poorly specified histologies (3.8%). An additional 6,686 cases of anal squamous CIS were diagnosed during the same time period. We excluded from analysis the 131 invasive anal cancers that were not microscopically confirmed and the 131 anal CIS with non-SCC or CIS, NOS histologies.

Table 1 presents IRs by behavior/histology across categories of age, sex, race/ethnicity, registry, and anal site. IRs increased across age groups for all invasive cancers. For squamous CIS, the highest IR occurred among 40 to 49 year-olds. The IR for anal SCC was greater among women than among men. In contrast, rates of both anal ADC and squamous CIS were greater in men than in women. The SCC IR was highest among non-Hispanic whites, while the ADC IR was highest among blacks. The IRs varied notably across registries. The total invasive anal cancer rate was highest among Alaska Natives, but the small number of cases precluded estimation of the type-specific rates. The SCC IRs exceeded 15/million in Atlanta, San Francisco-Oakland, Kentucky, and Seattle and were <10/million in Utah and Hawaii. Rates of anal CIS were dramatically higher in San Francisco-Oakland (35.2/million), about three times the next highest rate of 11.9 in Seattle. Eighty-three percent of SCC developed in the anal canal or anus, NOS. In contrast, nearly half of the ADCs had the primary site designated as overlapping lesion of the rectum, anus, and anal canal.

The different age-specific incidence patterns for SCC, ADC, and anal CIS are shown in Fig. 1. Across race/ethnicities, SCC IRs generally increased steeply until ages 50 or 60 years, and then increased more gradually at older ages, except among black males. Among non-Hispanic whites and blacks, anal ADC IRs increased steadily with advancing age, while rates among Hispanic whites were too sparse to evaluate. In contrast, squamous CIS IRs were highest among 40- to 49-year-old men and 50- to 59-year-old women who were non-Hispanic white or black. Among Hispanic whites, CIS rates decreased with age among men and increased among women.

The IR for invasive anal cancer was higher among women compared with men (overall RR, 1.29; 95% CI, 1.25–1.33) with the excess particularly notable among those of ages ≥50 years (RR, 1.32–1.53) and among cases arising in the cloacogenic zone (RR, 2.29; 95% CI, 1.91–2.76; Table 2). There were a few notable exceptions, however, where anal cancer IRs were higher among men; including <40 year-olds (RR, 0.41 and 0.62) and blacks (RR, 0.84; 95% CI, 0.76–0.93). Consistent with data presented in Table 1, IRs of ADC (RR, 0.68; 95% CI, 0.62–0.74) and squamous CIS (RR, 0.36; 95% CI, 0.34–0.38) were also higher in men.

Compared with non-Hispanic whites, Hispanic whites (overall RR, 0.64; 95% CI, 0.60–0.68) and Asian/Pacific Islanders (overall RR, 0.27; 95% CI, 0.25–0.30) had lower rates of invasive anal cancer across sexes, age groups, primary sites, and histology, with the exception of melanoma (Table 3). Compared with non-Hispanic whites, blacks also had somewhat lower rates of invasive anal cancer overall (RR, 0.91; 95% CI, 0.86–0.95), and lower rates among women (RR, 0.72; 95% CI, 0.67–0.77) and 60+ year-olds (RR, 0.72 and 0.80). Rates of anal cancer occurring in the anal canal (RR, 0.81; 95% CI, 0.74–0.88) or cloacogenic zone (RR, 0.62; 95% CI, 0.44–0.85), and SCC (RR, 0.82; 95% CI, 0.77–0.87) were also lower among blacks. In contrast, consistent with Table 2, invasive anal cancer rates were higher among black men overall (RR, 1.19; 95% CI, 1.10–1.28), and particularly among men of ages <40 (RR, 4.56 and1.73), and for ADCs (RR, 1.48; 95% CI, 1.29–1.70).

During 1992–2011 in the 13 SEER registries, invasive anal cancer IRs increased 2.2%/year overall (95% CI, 1.8–2.6) and among men (95% CI, 1.6–2.8), and 2.3%/year (95% CI, 1.9–2.8) among women (Supplementary Table S1 and Fig. 2). The increases were limited to SCC (overall APC, 2.9%; 95% CI, 2.5–3.3), with significant increases among non-Hispanic white and black men and women, and nonsignificant increases among Hispanic whites. Among non-Hispanic whites, APCs for SCC increased for both men and women across age groups with the exception of 0 to 39 year-olds (Supplementary Fig. S1). Of note, SCC IRs have been consistently higher among women compared with men over time among non-Hispanic and Hispanic whites; however, rates have been higher among black men compared with black women since 1992–1996. Since 1992, ADC IRs have been stable among men (APC, −0.6; 95% CI, −1.9–0.8) and declined among women (APC, −1.8; 95% CI, −3.5–0.0). Squamous CIS IRs increased 14.2%/year overall (95% CI, 12.6–15.9), with more rapid increases among men (APC, 15.5%; 95% CI, 13.5–17.6) than women (APC, 10.1%; 95% CI, 8.7–11.6); these patterns are apparent in each racial/ethnic group. When the San Francisco-Oakland registry was excluded (i.e., the registry with the highest rates of squamous CIS), a steep incline in rates remained (APCs: overall, 13.4%; 95% CI, 12.0–14.8%; men: 15.2%, 95% CI, 13.4–17.1%; women: 9.3%, 95% CI, 7.7–11.0%).

Because of the juxtaposition of the anus and rectum and the potential for misclassification of site of origin, we additionally examined trends in rectal cancer IRs by histology among whites/non-Hispanic whites during 1977–2011 (Fig. 3). The trends over time were quite similar for rectal and anal SCCs and, to a lesser extent, for rectal and anal ADCs. Like anal cancer, rectal SCC IRs have been consistently higher among women, and rectal ADC IRs have been consistently higher among men. During 1992–2011, similar to the anal cancer APCs, among women, rectal SCC IRs increased 3.9%/year (95% CI, 2.6–5.3) and ADC IRs declined 1.3%/year (95% CI, −1.8, −0.9). In contrast to the significant increase in anal SCC IRs and stable anal ADC IRs among men, rectal SCC IRs have been relatively stable since 1992 (APC, 0.3; 95% CI, −1.7–2.4), and ADC IRs significantly declined (APC, −1.4; 95% CI, −1.9 to −0.8). When rectal and anal cancers were combined, during 2007–2011, ADC rates were 98.3/1,000,000 among men and 56.7/1,000,000 among women, while SCC rates were 13.6/1,000,000 among men and 21.3/1,000,000 among women.

Using population-based data from the SEER program, we have shown that the demographic characteristics and temporal trends in anal cancer IRs differ dramatically by histologic type. SCCs are more common among women and occur at younger ages than ADCs. Furthermore, during 1992–2011, anal SCC IRs increased significantly, whereas ADC IRs declined, at least among women. In addition, anal squamous CIS IRs increased steeply over time, with rates highest among men and in the San Francisco-Oakland registry.

Eighty-two percent of all anal cancers in the United States are SCCs, though ADCs may comprise a larger fraction of cases in some countries (9). Anal SCCs are largely caused by infection with oncogenic HPV, primarily HPV-16 (10, 11, 19). In contrast, there are very limited data on the cause of anal ADC. In a meta-analysis, based on only seven anal ADCs, three were positive for HPV infection (10). Many of the risk factors that have been described for anal cancer reflect infection with or persistence of anal HPV infection, including lifetime number of sexual partners, receptive anal intercourse, and HIV infection (8, 20). Given the predominance of SCCs among anal cancers, these risk factors likely reflect the causal association between HPV infection and anal SCC. In contrast to SCC, little is known about risk factors for anal ADC. Descriptively, anal ADC differs from SCC in case demographics and temporal trends, with higher rates among males, steep increases in rates with age and stable or declining rates over time, providing evidence that ADC is etiologically distinct from SCC.

Anal SCC IRs are higher among women than men, and consistently lower among Hispanic whites and Asian/Pacific Islanders than among non-Hispanic whites. These differences are likely driven by a higher prevalence of anal infection with carcinogenic HPV among women and non-Hispanic whites. Compared with men, women are more likely to engage in anal receptive intercourse, which is known to be associated with anal HPV infection (21). In addition, transmission of cervical HPV infection to the anus may occur independent of anal intercourse, as those with cervical HPV infection have three times the risk of anal HPV infection (22). Although not established in the literature, it is possible that Hispanics and Asian/Pacific Islanders may also have a lower rate of anal HPV infection than non-Hispanic whites. In the Hawaii Multiethnic Cohort Study, anal and cervical HPV infections (either detected singularly or with infection at both anatomic sites) tended to be more common among Caucasian women compared with women who were Japanese, Hawaiian, or Filipino (22). Increasing rates over time across sexes and race/ethnicities are likely driven by changes in sexual behavior and are consistent with the rise of other HPV-associated cancers (23).

Unlike non-Hispanic whites, anal SCC IRs have been higher in black men than black women in recent years. In addition, among those <40 years old, IRs are higher in men than women, particularly among young, black men. These patterns are consistent with the impact of HIV on anal cancer rates. HIV-infected individuals have a 30-fold increased risk of anal cancer compared with the general population, due to an elevated prevalence of anal intercourse among men who have sex with men and other sexual behaviors leading to increased HPV acquisition and a role of immunosuppression (24, 25). Approximately 28% of male and 1% of female anal cancers in the United States occur among people with HIV (6). The HIV prevalence is particularly high (84%) among anal cancer cases occurring in young, black men (6). In addition, HIV-infected anal cancer cases have contributed strongly to the rising anal cancer rates in men but not women (6). The HIV status of anal cancer cases is not collected by SEER; thus, we could not address HIV directly in the current analysis.

Although anal SCC and ADC differ from each other, they resemble rectal SCC and ADC, respectively, particularly among women. For example, both rectal and anal SCCs are more common among women, and the rates among women have increased for both sites over time. Similarly, both rectal and anal ADCs are more common in men, and rates for both malignancies have declined significantly over time among women. These similarities within histologic types could have multiple explanations, including distinct malignancies arising in the anus and the rectum with similar etiologies and difficulties determining the primary site.

Although rare, rectal SCC and anal ADC may be distinct cancers that share etiologic risk factors with anal SCC and rectal ADC, respectively. There have been case reports of HPV detected in rectal SCC tumors (26, 27), and one small study found 77% of rectal SCCs to be HPV positive (8). Another study showed an increased risk of rectal SCC in HIV-infected individuals, similar to what is observed for anal SCC (28), implicating a potential role of HPV in the development of rectal SCC. In addition, there is limited evidence that Crohn's disease and chronic inflammation may be associated with anal ADC risk, similar to colorectal cancer (29).

Given the juxtapostion of the anus and the rectum, it is also likely that some tumors arising in the anus or the lower part of the rectum would overlap these two anatomic sites. Therefore, anal SCCs may be misclassified as rectal SCCs and the rarer primary anal ADCs may be misclassified rectal ADCs. For example, as nearly half of the ADCs had the primary site designated as overlapping lesion of the rectum, anus, and anal canal, it is possible that many anal ADCs actually arose in the rectum. The anal canal is slightly shorter in women than men, perhaps leading to a greater degree of misclassification consistent with the data in this study (30).

The anal squamous CIS IRs increased dramatically during 1992–2011 across sexes and racial groups and were highest in middle age, likely reflecting the uptake of anal cancer screening in certain geographic areas. The consistency across racial/ethnic groups suggests a lack of racial/ethnic disparities in anal cancer screening, and the higher rates among men than women likely reflect increased anal cancer screening, particularly among men who have sex with men. Several organizations now recommend anal cancer screening for HIV-infected individuals (31). Notably, the anal squamous CIS IR in San Francisco is three to 17 times higher than other registry areas, likely due to the establishment of an anal neoplasia clinic at the University of California San Francisco in 1990 (32). It is unclear whether anal cancer screening and subsequent treatment prevents the development of anal cancer; however, the Anal Cancer/HSIL Outcomes Research (ANCHOR) Study (NIH clinical trials identification number: NCT02135419), has been undertaken to directly assess the benefits of treating precancerous anal lesions. Given the rapid increase in anal SCC rates over time, there is an urgent need for prevention and early detection strategies, particularly in high-risk populations. Although HPV vaccination has been shown to protect against anal HPV infection (33, 34), uptake remains suboptimal among adolescents, with only 57% of females and 35% of males receiving at least one dose in 2013 (35). Furthermore, the reduction of anal cancer due to vaccination will not be seen for decades.

The use of high-quality, population-based SEER cancer registry data is the main strength of this analysis. We provide a detailed analysis of rates and trends by primary site, histology and behavior, and, for the first time, provide comparisons between anal and rectal cancer rates by histology. We separate Hispanic whites from non-Hispanic whites during 1992–2011, which is important as the fraction of whites who were of Hispanic ethnicity in SEER 13 increased from 21% in 1992–96 to 28% in 2007–11. In addition, due to the expansion of the SEER Program over time, we used data from both SEER 9 and 13 to assess temporal trends. Although rates were similar in SEER 9 and 13 (Supplementary Fig. S2), changes in IRs between 1977–1991 and 1992–2011 reflect small changes attributable to the inclusions of additional registries with somewhat different population compositions, in addition to changes due to the separation of Hispanic whites and non-Hispanic whites. During 1992–1996, total invasive anal cancer rates among whites were 11.8/million in SEER 9 and 12.3/million in SEER 13, where rates were higher among non-Hispanic whites (12.7/million) compared with Hispanic whites (9.4/million). Rates among blacks were also similar in SEER 9 (14.2/million) and SEER 13 (14.5/million). The main limitation of this analysis is the lack of direct information on HPV or HIV infection and anal cancer screening, important factors that have likely influenced temporal trends. Furthermore, although anal squamous CIS is reportable to SEER registries, because it is primarily detected via screening, it is possible that it is not completely ascertained. Thus, rates of CIS may be underestimated and should be interpreted with caution. Only in San Francisco was the CIS rate higher than the invasive SCC rate, reflecting active screening programs in San Francisco, which are not as prominent in other registry areas.

In an analysis of several decades of U.S. data on anal cancer, we have shown that anal SCC rates have increased rapidly over time. Importantly, we have highlighted the differences in anal SCC and ADC, which likely have different etiologies, and reveal the importance of analyzing these histologic types separately. Furthermore, we have shown the similarity of rectal SCC to anal SCC, which may call for the reclassification of anal and rectal cancer tumors based on their histologies. Finally, we have shown dramatic increases in the detection of anal squamous CIS, presumably reflecting anal cancer screening in certain geographic locations in the United States.

T.M. Darragh has received speakers bureau honoraria from Roche and Ventana Roche, is a consultant/advisory board member for TheVax, and has provided expert testimony for Hologic. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.S. Shiels, A.R. Kreimer, S.S. Devesa

Development of methodology: M.S. Shiels, T.M. Darragh, S.S. Devesa

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.S. Shiels, A.R. Kreimer, A.E. Coghill, T.M. Darragh, S.S. Devesa

Writing, review, and/or revision of the manuscript: M.S. Shiels, A.R. Kreimer, A.E. Coghill, T.M. Darragh, S.S. Devesa

Study supervision: S.S. Devesa

This study was supported by the Intramural Research Program of the National Cancer Institute (to M.S. Shiels, A.R. Kreimer, A.E. Coghill, and S.S. Devesa).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.