- Email: [email protected]
Trends in incidence and survival for anal cancer in New South Wales, Australia, 1972–2009
Cancer Epidemiology 39 (2015) 842–847
Contents lists available at ScienceDirect
Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention journal homepage: www.cancerepidemiology.net
Trends in incidence and survival for anal cancer in New South Wales, Australia, 1972–2009 Matthew J. Soeberga,* , Kris Rogersa , David C. Currowb , Jane M. Younga,c a b c
Cancer Epidemiology and Services Research (CESR), Sydney School of Public Health, Sydney Medical School, University of Sydney, NSW, Australia Cancer Institute NSW, Australia Surgical Outcomes Research Centre (SOuRCe), Sydney Local Health District and University of Sydney, NSW, Australia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 25 May 2015 Received in revised form 1 October 2015 Accepted 7 October 2015 Available online 24 October 2015
Introduction: Little is known about the incidence and survival of anal cancer in New South Wales (NSW), Australia, as anal cancer cases are often grouped together with other colorectal cancers in descriptive epidemiological analyses. Methods: We studied patterns and trends in the incidence and survival of people diagnosed with anal cancer in NSW, Australia, 1972–2009 (n = 2724). We also predicted anal cancer incidence in NSW during 2010–2032. Given the human papilloma virus-associated aetiology for most anal cancers, we quantified these changes over time in incidence and survival by histological subtype: anal squamous cell carcinoma (ASCC); and anal adenocarcinoma (AAC). Results: There was a linear increase in incident anal cancer cases in NSW with an average annual percentage change (AAPC) of 1.6 (95% CI 1.1–2.0) such that, in combination with age-period-cohort modelling, we predict there will be 198 cases of anal cancer in the 2032 calendar year (95% CI 169–236). Almost all of these anal cancer cases are projected to be ASCC (94%). Survival improved over time regardless of histological subtype. However, five-year relative survival was substantially higher for people with ASCC (70% (95% CI 66–74%)) compared to AAC (51% (95% CI 43–59%)), a 37% difference. Survival was also greater for women (69% (95% CI 64–73%)) with ASCC compared to men (55% (95% CI 50– 60%)). It was not possible to estimate survival by stage at diagnosis particularly given that 8% of all cases were recorded as having distant stage and 22% had missing stage data. Interpretation: Aetiological explanations, namely exposure to oncogenic types of human papillomavirus, along with demographic changes most likely explain the actual and projected increase in ASCC case numbers. Survival differences by gender and histological subtype point to areas where further research is warranted to improve treatment and outcomes for all anal cancer patients. ã 2015 Elsevier Ltd. All rights reserved.
Keywords: Anal cancer Australia Squamous cell carcinoma Adenocarcinoma Incidence Relative survival
1. Introduction The incidence of anal cancer, a rare digestive tract malignant tumour located primarily in the anal canal, is strongly associated with exposure to oncogenic types of human papillomavirus [1]. Reported increases in incidence worldwide have led to greater clarity on the role that HPV plays in anal cancer aetiology including population subgroups at increased risk of disease [2–4], specifically that the aetiology of anal cancer is closer to that of other genital cancers compared to gastrointestinal malignancies. The most common histological subtype of anal cancer is anal squamous
* Corresponding author at: Cancer Epidemiology and Services Research (CESR), Level 6, The Lifehouse (C39Z), RPA Hospital, The University of Sydney, NSW 2006 Australia. Fax: +61 2 9515 3222. E-mail address: [email protected] (M.J. Soeberg). http://dx.doi.org/10.1016/j.canep.2015.10.008 1877-7821/ ã 2015 Elsevier Ltd. All rights reserved.
cell carcinoma (ASCC), estimated to comprise greater than 70% of all anal cancer cases [5]. ASCC is known to have substantially increased in a number of countries over the last fifty years [6–10]. Few studies have simultaneously published incidence data on ASCC and anal adenocarcinoma (AAC). An Australian study of anal cancer for people diagnosed during the 1987–2005 period show increases over time in both ASCC and AAC age-standardised incidence rates per 100,000 person-years [6]. However, the investigators of this study did not estimate the magnitude of any future increase in incidence over time or investigate survival differences by histological subtype. Compared to cancers of the colon and rectum [11,12], little is understood about the survival outcomes of anal cancer in New South Wales and in other jurisdictions. This is primarily due to the fact that anal cancer is a rare outcome making trend measurement and multivariate analyses computationally difficult with sparse
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
data. For example, the investigators of a study of relative survival for anorectal cancers in England and Wales combined data for cancers of the rectum and anus allowing for larger datasets to be analysed in a more complex manner [13]. There is the potential to mask important differences in survival outcomes when, in this example, data for two anatomical sites were aggregated. Given improvements in anal cancer treatment over the last 20 years [5], it is timely to measure incidence and survival for anal cancer in NSW including by histological subtype. The aim of this study was to update incidence trends using more up-to-date cancer registration data, to predict the number of anal cancer cases in New South Wales (NSW) up to 2032, and to compare 5-year relative survival by histological subtype. We hypothesise that incidence for ASCC and AAC have increased over time. Similarly, 5-year survival has improved but that differences exist by sex and histological subtype. 2. Materials and methods 2.1. Data sources De-identified unit records for all unique cases of anal cancer diagnosed in NSW residents between 1972 and 2009 and notified to the NSW Cancer Registry were included. Operational details of the Registry have been published elsewhere [14]. Briefly, notifications to the Registry of invasive cancers are mandated under the NSW Public Health Act 2010. The study cohort was defined using International Classification of Diseases for Oncology (ICD-O-3), 3rd edition, topography and morphology codes. All records for the ICD-O-3 topography code C21, regardless of morphology code, were extracted. Records for ICDO-3 topography code C20 where the ICD-O-3 morphology codes were 8050–8085 or 8120–8131 were also included, consistent with prior knowledge [15], where the majority were squamous cell carcinomas (Supplementary Table 1). Cases of anal cancer were categorised into three histological subtypes: ASCC; AAC; and other anal tumours. Morphology codes used to define each category were adapted from published Australian data [6] (Table 1). Incidence count and survival data were analysed using all anal cancer data combined and for each histological subtype. Due to sparse data, information about people with other anal tumours (7% of all cases) was excluded from time trend incidence analyses. Data presented here are for all anal cancer data combined and for ASCC and AAC. 2.2. Statistical analyses Incidence was calculated as both a count and rate measure, with the age-standardised incidence rate at 100,000 person-years standardised to the Australia population as at 30 June 2001. Incidence was calculated for each combination of calendar year and histological subtype. Changes over time in incidence during 1972–2009 were evaluated by fitting piece-wise linear regression segments [16]. Permutation tests, a method for selecting the optimal number of data segments, were performed. For sensitivity analyses, we fitted the regressions for males and females separately; there was little difference in the magnitude of the
843
slope for case numbers and a faster increase over time for males when assessing rates. However, due to sparse data, we only present data here for males and females combined. Using the Nordpred method [17], estimation of future incidence for 2010–2032 was calculated by combining age, period and cohort effects with estimated actual and projected resident population structures in NSW by 5-year age groups and sex [18,19]. Relative survival analysis was restricted to adults aged 15– 99 years and diagnosed with anal cancer between 1972 and 2006. Complete follow-up data were available up to 31 December 2007. We estimated relative survival for people diagnosed up to the end of 2001 using the cohort approach. All remaining cases were estimated using the period approach [20,21], commonly used for more recently diagnosed cases where full five-year follow-up data are not available. Relative survival was calculated for each combination of collapsed age groups (15–54 years, 55–64 years, 65–74 years, and 75 years and older), sex, calendar period (perdecade groupings), and histological subtype. Data were too sparse to estimate relative survival by recorded stage at diagnosis. The Ederer II method [22] was used to calculate expected survival using life tables constructed from raw population mortality data for each combination of calendar year, sex, and single year of age. Incidence data were analysed using SAS version 9.3 [23], with trend analysis performed in JoinPoint software [24], and projections undertaken in Nordpred software [25]. Relative survival was calculated using the strs program [26] in Stata Statistical Software: 12.0, StataCorp., Lp, College Station, Texas. 3. Results 3.1. Characteristics of persons with ASCC and AAC Overall, there were 2724 newly diagnosed with anal cancer recorded in the NSW Cancer Registry between 1972 and 2009 (Table 2). The majority of these people (72%) were diagnosed with ASCC. There was an even distribution of the number of cases by age group for all anal cancers combined and for ASCC. People with AAC tended to be older at diagnosis, with 40% of these being diagnosed in people aged 75 years and older. More women than men were diagnosed with all anal cancers combined and ASCC, with more men diagnosed with AAC. For all anal cancer cases combined, there was an apparent increase in the number of cases in the most recent decades with a notable peak in the number AAC in the 1990s (see Section 3.2). Almost all cases of anal cancer were confirmed on the NSW Cancer Registry through histopathology information. Approximately 40% of all anal cancer cases, regardless of histological subtype, had local extent of disease within the first four months of diagnosis, with around 8% of cases having distant disease. Extent of diagnosis was missing for about 20% of all cases. For people who were recorded as having died, 41% of ASCC cases had died of their anal cancer compared to 29% for people with AAC. 3.2. Actual and projected incidence of ASCC and AAC For all histological subtypes combined, there was a linear increase in the incidence of anal cancer (Fig. 1a). In 2009, the age-
Table 1 ICD-O-3 morphology codes using to define anal squamous cell carcinoma (ASCC), anal adenocarcinoma (AAC), and other anal tumours. Histological subtype
ICD-O-3 Morphology codes used to define
Anal squamous cell carcinoma (ASCC) Anal adenocarcinoma (AAC) Other anal tumour
8050–8051; 8052; 8070–8073; 8076; 8083; 8120; 8123–8124; and 8130 8140; 8143–8144; 8120–8211; 8215; 8260–8261; 8263; 8430; 8480–8481; 8490; and 8542 8000; 8010; 8012–8013; 8020–8021; 8033; 8041; 8090; 8094; 8231; 8240; 8246; 8560; 8720–8721; 8743; 8890; 8900; 8936; 9140; 9540; and 9680
844
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Table 2 Number and percent distribution of people with anal squamous cell carcinoma and anal adenocarcinoma in New South Wales, Australia, 1972–2009, by age group, sex, calendar period, basis of diagnosis, stage within four months of diagnosis, and recorded cause of death. All malignant anal tumours
Squamous cell carcinoma
Adenocarcinoma
Number
Percent distribution
Number
Percent distribution
Number
Percent distribution
Age group 15–54 years 55–64 years 65–74 years 75+ years All age groups
695 616 678 735 2724
(25) (23) (25) (27) (100)
563 478 489 426 1956
(29) (24) (25) (22) (100)
90 110 147 235 582
(16) (19) (25) (40) (100)
Sex Male Female
1219 1505
(45) (55)
812 1144
(42) (58)
316 266
(54) (46)
Calendar period 1972–1981 1982–1991 1992–2001 2002–2009
387 540 928 869
(14) (20) (34) (32)
287 361 627 681
(15) (18) (32) (35)
78 131 247 126
(13) (23) (42) (22)
Best basis of diagnosis Histopathology Clinical Other
2586 108 30
(95) (4) (1)
1863 73 0
(95) (4) (1)
571 9 <6
(98) (2) (<1)
Stage at diagnosisa Local Regional Distant Missing
1208 690 216 610
(44) (26) (8) (22)
905 464 133 454
(46) (24) (7) (23)
246 183 55 98
(43) (31) (9) (17)
Cause of death Anal cancer Other causes Total
587 1030 1617
(36) (64) (100)
433 626 1059
(41) (59) (100)
120 289 409
(29) (71) (100)
a These categories represent information collected by the NSW Cancer Registry for the degree of disease spread within four months of diagnosis. These data do not include clinical information such as values for tumour, node or metastasis.
standardised incidence rate for all anal cancer histology subtypes combined was 1.45 per 100,000 (95% 1.15, 1.75). There were no differences in slope between men and women with an average annual percentage change of 1.7 (95% CI 1.1, 2.3) in men and 1.6 (95% CI 1.0, 2.2) in women (Supplementary Fig. 1). Most of the change over time in incidence has been due to increases in ASCC (Fig. 1c), with the incidence of AAC declining over time (Fig. 1b). For example, the age-standardised rate per 100,000 for AAC in 2009 was 0.19 (95% CI 0.09, 0.28), a substantial decrease from 0.37 (95% CI 0.22, 0.56) in 1998. The number of anal cancer cases has been projected to continue to increase over time (Supplementary Fig. 2) with the number of cases reaching a maximum of 198 per year (95% CI 169–236) by 2032. The incidence of AAC is projected to remain at the low current levels, which means that the increases in incidence and cases into the future will be from cases of ASCC (94% of cases projected in 2030–2035). 3.3. Relative survival for persons with ASCC and AAC Five-year relative survival estimates by age group and sex are first presented for people with anal cancer diagnosed in the most recent calendar period (1997–2006) (Table 3). Five-year relative survival in this calendar period across all anal cancers combined was 63% (95% CI 59–66%). This varied by histological subtype and gender. Survival was 37% greater for people with ASCC compared to people with AAC; 70% (95% CI 66–73%) and 51% (95% CI 43–59%) respectively. Women diagnosed with anal cancer experienced better survival than men, 69% (95% CI 64–73%) and 55% (95% CI 50– 60%) respectively—a 20% survival advantage for women with anal cancer compared to men. These gender differences in survival 5-
years after diagnosis were consistent in pattern across histological subtypes. People diagnosed at a younger age also had better survival compared to people diagnosed at older ages. For all histological subtypes combined, 5-year relative survival for people diagnosed 1997–2006 and aged 15–54 years was 70% (95% CI 64–75%) compared to 50% (95% CI 41–58%) for people aged 75 years or more. Similar patterns in age group differences in 5-year relative survival can be seen across histological subtypes. Estimates for people diagnosed 1972–1981, 1982–1991, 1992– 2001, and 2002–2006 are shown in Table 3. Survival improved over time for people diagnosed with anal cancer in NSW. Five-year relative survival for people with anal cancer diagnosed between 2002 and 2006 was 64% (95% CI 58–69%), an increase of 6% from 1992 to 2001. A step-wise increase over time in 5-year relative survival occurred for people diagnosed with ASCC. Five-year relative survival also increased over time for people diagnosed with AAC. 4. Discussion Our study was based on high-quality anal cancer data over a 38year period from the NSW Cancer Registry. We observed a significant, linear increase (AAPC of 1.6%) in incident cases of ASCC and a decline in the number of AAC cases. The incidence of ASCC is expected to steadily increase and make up the majority of cases of anal cancer in the future. We estimate there will be 198 cases of anal cancer in 2032. We also demonstrated a rise in 5-year relative survival for all histological subtypes and differences in survival by sex and histology.
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Fig. 1. Actual (1972–2009) and predicted (2010–2032) age-standardised incidence rates per 100,000 person-years of anal cancer in New South Wales, Australia. Estimates are shown separately for all anal cancer cases combined (A), anal squamous cell carcinomas (B), and anal adenocarcinomas (C).
Interpretation of our findings using data from other jurisdiction is more limited due to inconsistent approaches in cohort definitions and statistical analyses across studies. We consider that a more standard approach to defining malignant anal disease is warranted, including a consistent method for defining topography and morphology codes. Investigation of anal cancer incidence using national Australian cancer registration data as reported by Jin et al. showed AAPCs of 1.9 and 3.4 for men and women with ASCC
845
respectively, a statistically significant finding [6]. We found similar AAPC values when stratified by sex. Jin et al. did not report an AAPC for males and females combined for either ASCC or AAC. However, given consistency with the sex-specific AAPC estimates, we assume that our AAPC for age-standardised rates would be similar to the work undertaken by these investigators. Data from the United States (US) provides an international context. Melbye et al. studied changing incidence of anal cancer in the US by sex, ethnicity, and marital status between 1973 and 1989 [8]. Although APCs were not calculated, a significant increase in the incidence of anal cancer occurred over the study period. In particular, Melbye et al. note the increase over time that occurred among never married men and men living in San Francisco in the most recent calendar period in their study. They speculated whether part of the increase in this sub-population group could be attributed to the AIDS epidemic. It is likely that behaviour changes, including condom use, will have had an impact on anal cancer incidence trajectories over time but this is difficult to quantify. An increase over time in anal cancer incidence, particularly among men, was reported by Johnson et al. [7]. A more recent study from the US by Nelson et al. analysed changes over time, 1973–2009, in noninvasive anal disease (half of all cases analysed) and anal cancer where the AAPC was 16.7%, skewing the combined non-invasive and anal cancer AAPC estimates [9]. The Nelson et al. study did not assess changes over time for AAC. Differences in aetiology and age at diagnosis between ASSC and AAC will partially explain the different incident trajectories reported in this study. There is an increasing body of evidence that the strength of association between human papillomavirus and ASCC is much stronger compared to the association with AAC [27]. In addition, the age-specific rate for ASCC peaked at a much lower age (around 65 years of age) compared with AAC where the rate was highest in people aged 85 years and over. Through the use of age-period-cohort modelling in this study [17], we were able to simultaneously allow for these differences between ASCC and AAC disease trajectories. The impact of the introduction of prophylactic human papillomavirus (HPV) vaccines, including the qHPV vaccine offered in Australia to all secondary school students, on the future incidence of anal cancer is not yet quantified. A recent study by Hillman et al., where biopsy material from Australian people with anal cancer was analysed suggests that the HPV vaccine is unlikely to have a significant impact on anal cancer incidence for several decades [28]. However, this effect on the future incidence of anal cancer in Australia is likely to vary by gender due to differences in the introduction of the vaccination for males and females, with the male vaccination program only starting in 2013. Our study is limited in what we can say about the public health implications of improvements in anal cancer survival. For instance, without clinical TNM staging and hospital episode data it is difficult to quantify whether these survival improvements are due to earlier diagnosis or improvements in treatment or a combination of both factors. Survival following an anal cancer diagnosis is amenable to improvement through earlier detection, chemotherapy, radiotherapy, and, where appropriate surgical resection [5,29– 31]. Recent studies of colorectal treatment and outcomes during 2000–2008 have been conducted in NSW using linked administrative data such that it is possible to identify predictors of treatment receipt and factors associated with good prognosis [32– 34]. Use of this linked data for investigating anal cancer is feasible but not without limitations. For example, data on approximately 700 people with ASCC are recorded in the NSW Cancer Registry for the 2000–2008 period, or about 70 cases per year, leading to instability of estimates following stratification by covariates and limited interpretation of results. Further, linked cancer registry, administrative hospital episode and cancer treatment data
846
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Table 3 Five-year relative survival: adults (15–99 years) diagnosed with anal cancer in New South Wales, 1972–2006 with follow up to the end of 2007, by age group, sex and calendar period. Age group and sex estimates were calculated for people diagnosed 1997–2006. Survival estimates for people diagnosed 1997–2006 and 2002–2006 were calculated using the period approach; all other estimates were calculated using the cohort approach. All malignant anal cancer tumours
Squamous cell carcinoma
5-year relative survival
95% CI
5-year relative survival
95% CI
5-year relative survival
95% CI
Age group 15–54 years 55–64 years 65–74 years 75+ years
70 66 63 50
(64–75) (59–73) (55–70) (41–58)
73 71 69 65
(66–88) (63–78) (60–78) (53–76)
63 54 58 38
(43–78) (36–70) (41–72) (24–54)
Sex Male Female Total
55 69 63
(50–60) (64–73) (59–66)
64 75 70
(57–70) (69–80) (66–74)
47 60 51
(35–58) (43–68) (43–59)
Calendar period 1972–1981 1982–1991 1992–2001 2002–2006
46 55 60 64
(40–52) (50–60) (57–64) (58–69)
48 63 67 70
(41–55) (57–69) (62–71) (63–76)
45 37 50 58
(32–59) (28–46) (43–58) (41–74)
currently available in NSW do not yet have adequate information on receipt of radiation therapy, a primary treatment modality for recently diagnosed anal cancer cases, and the data on clinical TNM staging is yet to be validated. A number of population groups appear to be at higher risk of developing anal cancer, namely men who have sex with men [35,36], women with a history of previous HPV-associated cervical dysplasia or malignancy [37,38], and people whose immune response is limited in controlling HPV infection [39]. The data in this study do not allow for incidence and survival in these at-risk groups to be explored. Such data are value such as the recent study from Marcus et al. who identified that survival 5-years following a cancer diagnosis was poorer for people who were HIV positive compared to those who were HIV negative [40]. Future use of linked administrative datasets in NSW may provide useful insights. We found substantial survival differences by gender, with women having better survival. We also showed a 21% difference in 5-year relative survival difference between people with ASCC compared to people with AAC for the 1997–2006 period. These findings are consistent with previously published data and have the potential to provide more clarity on histological differences in anal cancer survival. In a study of anal cancer incidence and survival in the United States, Johnson et al. reported a 10% sexspecific difference in 5-year relative survival for people diagnosed in the 1973–2000 period with females have a greater survival outcome [7]. In their investigation of survival for ASSC in Australia, 1982–2004, Jin et al. reported a 20% difference between males and females in 5-year relative survival (74% for females and 61% for males) [6]. This study did not report relative survival by histological subtype. A more recent French study measured population-based survival for people diagnosed with anal cancer during 1982–2004 followed-up to the start of 2008. Bouvier et al. found that, regardless of age at diagnosis, 5-year net survival was greater for females compared to males [41]. With their larger sample size, the investigators were also able to model relative survival where they reported that excess mortality rates were highest in the first two years since diagnosis and then steadily declined up to 10 years post diagnosis. Data from England and Wales reported by Jeffreys et al., where rectal and anal cancers were combined in the analyses, showed differences in histologyspecific relative survival differences [13]. Five-year relative survival in the most recent period, 1991–1996, was 59% for epidermoid carcinomas (accounting for over half of all anal cancers) and 52% for adenocarcinomas. This translates into a 14%
Adenocarcinoma
difference. Jeffreys et al. also performed excess mortality rate modelling to compare differences by patient and tumour characteristics. Interestingly, the results of these analyses show that relative excess mortality rate differences between these two histological subtypes reduced to the null. Age at diagnosis and differences in tumour behaviour are likely to explain histologyspecific differences in relative survival. The age group distribution for adenocarcinoma is skewed to older age groups, with older age widely recognised as a factor associated with poorer relative survival [26,42]. Explanations for sex-specific differences in relative survival on the magnitude reported here are less clear but are possibly related to differences in age at diagnosis [33,34]. To our knowledge, no other study has predicted the number of future anal cancer cases and age-standardised incidence rates. It seems likely that the anal cancer incidence increases we have reported are due to changes over time in aetiology and demographic factors. We also confirmed differences in survival 5-years following an anal cancer diagnosis by sex and histological group but due to sparse data and limited staging data were unable to directly test whether these differences can be explained by stage at diagnosis or treatment. Nevertheless, these differences are marked enough to warrant further investigation and possible health system interventions to reduce these gaps in anal cancer survival. Conflict of interest None. Authors’ contribution J.Y., K.R., and M.S. designed the study. K.R. and M.S. analysed the data. M.S. wrote the manuscript. All authors revised the manuscript for intellectual content. Sources of funding This research was funded by a Cancer Epidemiology Linkage Grant from the Cancer Institute NSW. MS and JY were also supported by the Academic Leader, Cancer Epidemiology, grant awarded to The University of Sydney from the Cancer Institute NSW.
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. canep.2015.10.008. References [1] H. De Vuyst, G.M. Clifford, M.C. Nascimento, M.M. Madeleine, S. Franceschi, Prevalence and type distribution of human papillomavirus in carcinoma and intraepithelial neoplasia of the vulva, vagina and anus: a meta-analysis, Int. J. Cancer 124 (7) (2009) 1626–1636. [2] N. Ouhoummane, M. Steben, F. Coutlee, T. Vuong, P. Forest, C. Rodier, et al., Squamous anal cancer: patient characteristics and HPV type distribution, Cancer Epidemiol. 37 (6) (2013) 807–812. [3] A.M. Saleem, J.K. Paulus, A.P. Shapter, N.N. Baxter, P.L. Roberts, R. Ricciardi, Risk of anal cancer in a cohort with human papillomavirus-related gynecologic neoplasm, Obstet. Gynecol. 117 (3) (2011) 643–649. [4] C.M. Vajdic, M.T. van Leeuwen, F. Jin, G. Prestage, G. Medley, R.J. Hillman, et al., Anal human papillomavirus genotype diversity and co-infection in a community-based sample of homosexual men, Sex. Transm. Infect. 85 (5) (2009) 330–335. [5] R. Glynne-Jones, J.M. Northover, A. Cervantes, Anal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up, Ann. Oncol. 21 (Suppl. (5)) (2010) v87–v92. [6] F. Jin, A.N. Stein, E.L. Conway, D.G. Regan, M. Law, J.M. Brotherton, et al., Trends in anal cancer in Australia, 1982–2005, Vaccine 29 (12) (2011) 2322–2327. [7] L.G. Johnson, M.M. Madeleine, L.M. Newcomer, S.M. Schwartz, J.R. Daling, Anal cancer incidence and survival: the surveillance, epidemiology, and end results experience, 1973–2000, Cancer 101 (2) (2004) 281–288. [8] M. Melbye, C. Rabkin, M. Frisch, R.J. Biggar, Changing patterns of anal cancer incidence in the United States, 1940–1989, Am. J. Epidemiol. 139 (8) (1994) 772–780. [9] R.A. Nelson, A.M. Levine, L. Bernstein, D.D. Smith, L.L. Lai, Changing patterns of anal canal carcinoma in the United States, J. Clin. Oncol. 31 (12) (2013) 1569–1575. [10] S. Simpson, R. Turner, Four decades of anal cancer in Tasmania, Australia: what do the case data tell us? Sex. Health 9 (3) (2012) 213–219. [11] C. Allemani, H.K. Weir, H. Carreira, R. Harewood, D. Spika, X.S. Wang, et al., Global surveillance of cancer survival 1995–2009: analysis of individual data for 25,676,887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet 385 (March (9972)) (2015) 977–1010. [12] M.P. Coleman, D. Forman, H. Bryant, J. Butler, B. Rachet, C. Maringe, et al., Cancer survival in Australia, Canada, Denmark, Norway, Sweden, and the UK, 1995–2007 (the International Cancer Benchmarking Partnership): an analysis of population-based cancer registry data, Lancet 377 (9760) (2011) 127–138. [13] M. Jeffreys, B. Rachet, S. McDowell, A.G. Habib, C. Lepage, M.P. Coleman, Survival from rectal and anal cancers in England and Wales, 1986–2001, Eur. J. Cancer 42 (10) (2006) 1434–1440. [14] D. Currow, W. Thompson, Cancer in NSW: Incidence Report, Cancer Institute NSW, Sydney, 2014, doi:http://dx.doi.org/10.1097/CEJ.0000000000000163. [15] M. Frisch, M. Melbye, Anal Cancer, in: D. Schottenfeld, J. Fraumeni Jr. (Eds.), Cancer Epidemiology and Prevention, 3rd ed., Oxford University Press, USA, 2006. [16] H.J. Kim, M.P. Fay, E.J. Feuer, D.N. Midthune, Permutation tests for joinpoint regression with applications to cancer rates, Stat. Med. 19 (3) (2000) 335–351. [17] B. Moller, H. Fekjaer, T. Hakulinen, H. Sigvaldason, H.H. Storm, M. Talback, et al., Prediction of cancer incidence in the Nordic countries: empirical comparison of different approaches, Stat. Med. 22 (17) (2003) 2751–2766. [18] Department of Planning and Infrastructure, New South Wales in the future: Preliminary 2013 population projections. Preliminary release of NSW state and local government area population projections. Sydney NSW Department of Planning & Infrastructure, 2013.
847
[19] Australian Bureau of Statistics. 3235.0—Population by Age and Sex, Regions of Australia. Canberra: Australian Bureau of Statistics, 2013. [20] H. Brenner, V. Arndt, Recent increase in cancer survival according to age: higher survival in all age groups, but widening age gradient, CCC 15 (9) (2004) 903–910. [21] H. Brenner, O. Gefeller, An alternative approach to monitoring cancer patient survival, Cancer 78 (9) (1996) 2004–2010. [22] F. Ederer, L.M. Axtell, S.J. Cutler, The relative survival rate: a statistical methodology, Natl. Cancer Inst. Monogr. 6 (1961) 101–121. [23] SAS Institute Inc., 2011. Cary, North Carolina, USASAS 9.3. [24] National Cancer Institute, Joinpoint Regression Program 4.1.1 ed. Bethesda, MD, USA, 2014. [25] Cancer Registry of Norway, NordPred 2.8 ed. Norway, 2003. [26] P.W. Dickman, A. Sloggett, M. Hills, T. Hakulinen, Regression models for relative survival, Stat. Med. 23 (1) (2004) 51–64. [27] J.R. Daling, M.M. Madeleine, L.G. Johnson, S.M. Schwartz, K.A. Shera, M.A. Wurscher, et al., Human papillomavirus, smoking, and sexual practices in the etiology of anal cancer, Cancer 101 (2) (2004) 270–280. [28] R.J. Hillman, S.M. Garland, M.P. Gunathilake, M. Stevens, N. Kumaradevan, C. Lemech, et al., Human papillomavirus (HPV) genotypes in an Australian sample of anal cancers, Int. J. Cancer 135 (4) (2014) 996–1001. [29] P.F. Engstrom, J.P. Arnoletti, A.B. Benson 3rd, J.D. Berlin, J.M. Berry, Y.J. Chen, et al., NCCN clinical practice guidelines in oncology. Anal carcinoma, J. Natl. Compr. Cancer Netw. 8 (1) (2010) 106–120. [30] F. Lim, R. Glynne-Jones, Chemotherapy/chemoradiation in anal cancer: a systematic review, Cancer Treat. Rev. 37 (7) (2011) 520–532. [31] R.J. Myerson, M.C. Garofalo, I. El Naqa, R.A. Abrams, A. Apte, W.R. Bosch, et al., Elective clinical target volumes for conformal therapy in anorectal cancer: a radiation therapy oncology group consensus panel contouring atlas, Int. J. Radiat. Oncol. Biol. Phys. 74 (3) (2009) 824–830. [32] M.L. Jorgensen, J.M. Young, T.A. Dobbins, M.J. Solomon, Assessment of abdominoperineal resection rate as a surrogate marker of hospital quality in rectal cancer surgery, Br. J. Surg. 100 (12) (2013) 1655–1663. [33] M.L. Jorgensen, J.M. Young, T.A. Dobbins, M.J. Solomon, Predictors of variation in colorectal cancer care and outcomes in New South Wales: a populationbased health data linkage study, Med. J. Aust. 200 (7) (2014) 403–407. [34] M.L. Jorgensen, J.M. Young, M.J. Solomon, Older patients and adjuvant therapy for colorectal cancer: surgeon knowledge, opinions, and practice, Dis. Colon Rectum 54 (3) (2011) 335–341. [35] J.R. Daling, N.S. Weiss, T.G. Hislop, C. Maden, R.J. Coates, K.J. Sherman, et al., Sexual practices, sexually transmitted diseases, and the incidence of anal cancer, N. Engl. J. Med. 317 (16) (1987) 973–977. [36] G. D'Souza, D.J. Wiley, X. Li, J.S. Chmiel, J.B. Margolick, R.D. Cranston, et al., Incidence and epidemiology of anal cancer in the multicenter AIDS cohort study, J. Acquir. Immune Defic. Syndr. 48 (4) (2008) 491–499. [37] M.T. Goodman, Y.B. Shvetsov, K. McDuffie, L.R. Wilkens, X. Zhu, P.J. Thompson, et al., Sequential acquisition of human papillomavirus (HPV) infection of the anus and cervix: the Hawaii HPV Cohort Study, J. Infect. Dis. 201 (9) (2010) 1331–1339. [38] D.P. Ryan, C.C. Compton, R.J. Mayer, Carcinoma of the anal canal, N. Engl. J. Med. 342 (11) (2000) 792–800. [39] A.E. Grulich, M.T. van Leeuwen, M.O. Falster, C.M. Vajdic, Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis, Lancet 370 (9581) (2007) 59–67. [40] J.L. Marcus, C. Chao, W.A. Leyden, L. Xu, J. Yu, M.A. Horberg, et al., Survival among HIV-infected and HIV-uninfected individuals with common non-AIDSdefining cancers, Cancer Epidemiol. Biomarkers Prev. 24 (8) (2015) 1167–1173. [41] A.M. Bouvier, A. Belot, S. Manfredi, V. Jooste, Z. Uhry, J. Faivre, et al., Trends of incidence and survival in squamous-cell carcinoma of the anal canal in France: a population-based study, Eur. J. Cancer Prev. (2015) , doi:http://dx.doi.org/ 10.1097/CEJ.0000000000000163. [42] H. Charvat, N. Bossard, L. Daubisse, F. Binder, A. Belot, L. Remontet, Probabilities of dying from cancer and other causes in French cancer patients based on an unbiased estimator of net survival: a study of five common cancers, Cancer Epidemiol. 37 (6) (2013) 857–863.
Contents lists available at ScienceDirect
Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention journal homepage: www.cancerepidemiology.net
Trends in incidence and survival for anal cancer in New South Wales, Australia, 1972–2009 Matthew J. Soeberga,* , Kris Rogersa , David C. Currowb , Jane M. Younga,c a b c
Cancer Epidemiology and Services Research (CESR), Sydney School of Public Health, Sydney Medical School, University of Sydney, NSW, Australia Cancer Institute NSW, Australia Surgical Outcomes Research Centre (SOuRCe), Sydney Local Health District and University of Sydney, NSW, Australia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 25 May 2015 Received in revised form 1 October 2015 Accepted 7 October 2015 Available online 24 October 2015
Introduction: Little is known about the incidence and survival of anal cancer in New South Wales (NSW), Australia, as anal cancer cases are often grouped together with other colorectal cancers in descriptive epidemiological analyses. Methods: We studied patterns and trends in the incidence and survival of people diagnosed with anal cancer in NSW, Australia, 1972–2009 (n = 2724). We also predicted anal cancer incidence in NSW during 2010–2032. Given the human papilloma virus-associated aetiology for most anal cancers, we quantified these changes over time in incidence and survival by histological subtype: anal squamous cell carcinoma (ASCC); and anal adenocarcinoma (AAC). Results: There was a linear increase in incident anal cancer cases in NSW with an average annual percentage change (AAPC) of 1.6 (95% CI 1.1–2.0) such that, in combination with age-period-cohort modelling, we predict there will be 198 cases of anal cancer in the 2032 calendar year (95% CI 169–236). Almost all of these anal cancer cases are projected to be ASCC (94%). Survival improved over time regardless of histological subtype. However, five-year relative survival was substantially higher for people with ASCC (70% (95% CI 66–74%)) compared to AAC (51% (95% CI 43–59%)), a 37% difference. Survival was also greater for women (69% (95% CI 64–73%)) with ASCC compared to men (55% (95% CI 50– 60%)). It was not possible to estimate survival by stage at diagnosis particularly given that 8% of all cases were recorded as having distant stage and 22% had missing stage data. Interpretation: Aetiological explanations, namely exposure to oncogenic types of human papillomavirus, along with demographic changes most likely explain the actual and projected increase in ASCC case numbers. Survival differences by gender and histological subtype point to areas where further research is warranted to improve treatment and outcomes for all anal cancer patients. ã 2015 Elsevier Ltd. All rights reserved.
Keywords: Anal cancer Australia Squamous cell carcinoma Adenocarcinoma Incidence Relative survival
1. Introduction The incidence of anal cancer, a rare digestive tract malignant tumour located primarily in the anal canal, is strongly associated with exposure to oncogenic types of human papillomavirus [1]. Reported increases in incidence worldwide have led to greater clarity on the role that HPV plays in anal cancer aetiology including population subgroups at increased risk of disease [2–4], specifically that the aetiology of anal cancer is closer to that of other genital cancers compared to gastrointestinal malignancies. The most common histological subtype of anal cancer is anal squamous
* Corresponding author at: Cancer Epidemiology and Services Research (CESR), Level 6, The Lifehouse (C39Z), RPA Hospital, The University of Sydney, NSW 2006 Australia. Fax: +61 2 9515 3222. E-mail address: [email protected] (M.J. Soeberg). http://dx.doi.org/10.1016/j.canep.2015.10.008 1877-7821/ ã 2015 Elsevier Ltd. All rights reserved.
cell carcinoma (ASCC), estimated to comprise greater than 70% of all anal cancer cases [5]. ASCC is known to have substantially increased in a number of countries over the last fifty years [6–10]. Few studies have simultaneously published incidence data on ASCC and anal adenocarcinoma (AAC). An Australian study of anal cancer for people diagnosed during the 1987–2005 period show increases over time in both ASCC and AAC age-standardised incidence rates per 100,000 person-years [6]. However, the investigators of this study did not estimate the magnitude of any future increase in incidence over time or investigate survival differences by histological subtype. Compared to cancers of the colon and rectum [11,12], little is understood about the survival outcomes of anal cancer in New South Wales and in other jurisdictions. This is primarily due to the fact that anal cancer is a rare outcome making trend measurement and multivariate analyses computationally difficult with sparse
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
data. For example, the investigators of a study of relative survival for anorectal cancers in England and Wales combined data for cancers of the rectum and anus allowing for larger datasets to be analysed in a more complex manner [13]. There is the potential to mask important differences in survival outcomes when, in this example, data for two anatomical sites were aggregated. Given improvements in anal cancer treatment over the last 20 years [5], it is timely to measure incidence and survival for anal cancer in NSW including by histological subtype. The aim of this study was to update incidence trends using more up-to-date cancer registration data, to predict the number of anal cancer cases in New South Wales (NSW) up to 2032, and to compare 5-year relative survival by histological subtype. We hypothesise that incidence for ASCC and AAC have increased over time. Similarly, 5-year survival has improved but that differences exist by sex and histological subtype. 2. Materials and methods 2.1. Data sources De-identified unit records for all unique cases of anal cancer diagnosed in NSW residents between 1972 and 2009 and notified to the NSW Cancer Registry were included. Operational details of the Registry have been published elsewhere [14]. Briefly, notifications to the Registry of invasive cancers are mandated under the NSW Public Health Act 2010. The study cohort was defined using International Classification of Diseases for Oncology (ICD-O-3), 3rd edition, topography and morphology codes. All records for the ICD-O-3 topography code C21, regardless of morphology code, were extracted. Records for ICDO-3 topography code C20 where the ICD-O-3 morphology codes were 8050–8085 or 8120–8131 were also included, consistent with prior knowledge [15], where the majority were squamous cell carcinomas (Supplementary Table 1). Cases of anal cancer were categorised into three histological subtypes: ASCC; AAC; and other anal tumours. Morphology codes used to define each category were adapted from published Australian data [6] (Table 1). Incidence count and survival data were analysed using all anal cancer data combined and for each histological subtype. Due to sparse data, information about people with other anal tumours (7% of all cases) was excluded from time trend incidence analyses. Data presented here are for all anal cancer data combined and for ASCC and AAC. 2.2. Statistical analyses Incidence was calculated as both a count and rate measure, with the age-standardised incidence rate at 100,000 person-years standardised to the Australia population as at 30 June 2001. Incidence was calculated for each combination of calendar year and histological subtype. Changes over time in incidence during 1972–2009 were evaluated by fitting piece-wise linear regression segments [16]. Permutation tests, a method for selecting the optimal number of data segments, were performed. For sensitivity analyses, we fitted the regressions for males and females separately; there was little difference in the magnitude of the
843
slope for case numbers and a faster increase over time for males when assessing rates. However, due to sparse data, we only present data here for males and females combined. Using the Nordpred method [17], estimation of future incidence for 2010–2032 was calculated by combining age, period and cohort effects with estimated actual and projected resident population structures in NSW by 5-year age groups and sex [18,19]. Relative survival analysis was restricted to adults aged 15– 99 years and diagnosed with anal cancer between 1972 and 2006. Complete follow-up data were available up to 31 December 2007. We estimated relative survival for people diagnosed up to the end of 2001 using the cohort approach. All remaining cases were estimated using the period approach [20,21], commonly used for more recently diagnosed cases where full five-year follow-up data are not available. Relative survival was calculated for each combination of collapsed age groups (15–54 years, 55–64 years, 65–74 years, and 75 years and older), sex, calendar period (perdecade groupings), and histological subtype. Data were too sparse to estimate relative survival by recorded stage at diagnosis. The Ederer II method [22] was used to calculate expected survival using life tables constructed from raw population mortality data for each combination of calendar year, sex, and single year of age. Incidence data were analysed using SAS version 9.3 [23], with trend analysis performed in JoinPoint software [24], and projections undertaken in Nordpred software [25]. Relative survival was calculated using the strs program [26] in Stata Statistical Software: 12.0, StataCorp., Lp, College Station, Texas. 3. Results 3.1. Characteristics of persons with ASCC and AAC Overall, there were 2724 newly diagnosed with anal cancer recorded in the NSW Cancer Registry between 1972 and 2009 (Table 2). The majority of these people (72%) were diagnosed with ASCC. There was an even distribution of the number of cases by age group for all anal cancers combined and for ASCC. People with AAC tended to be older at diagnosis, with 40% of these being diagnosed in people aged 75 years and older. More women than men were diagnosed with all anal cancers combined and ASCC, with more men diagnosed with AAC. For all anal cancer cases combined, there was an apparent increase in the number of cases in the most recent decades with a notable peak in the number AAC in the 1990s (see Section 3.2). Almost all cases of anal cancer were confirmed on the NSW Cancer Registry through histopathology information. Approximately 40% of all anal cancer cases, regardless of histological subtype, had local extent of disease within the first four months of diagnosis, with around 8% of cases having distant disease. Extent of diagnosis was missing for about 20% of all cases. For people who were recorded as having died, 41% of ASCC cases had died of their anal cancer compared to 29% for people with AAC. 3.2. Actual and projected incidence of ASCC and AAC For all histological subtypes combined, there was a linear increase in the incidence of anal cancer (Fig. 1a). In 2009, the age-
Table 1 ICD-O-3 morphology codes using to define anal squamous cell carcinoma (ASCC), anal adenocarcinoma (AAC), and other anal tumours. Histological subtype
ICD-O-3 Morphology codes used to define
Anal squamous cell carcinoma (ASCC) Anal adenocarcinoma (AAC) Other anal tumour
8050–8051; 8052; 8070–8073; 8076; 8083; 8120; 8123–8124; and 8130 8140; 8143–8144; 8120–8211; 8215; 8260–8261; 8263; 8430; 8480–8481; 8490; and 8542 8000; 8010; 8012–8013; 8020–8021; 8033; 8041; 8090; 8094; 8231; 8240; 8246; 8560; 8720–8721; 8743; 8890; 8900; 8936; 9140; 9540; and 9680
844
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Table 2 Number and percent distribution of people with anal squamous cell carcinoma and anal adenocarcinoma in New South Wales, Australia, 1972–2009, by age group, sex, calendar period, basis of diagnosis, stage within four months of diagnosis, and recorded cause of death. All malignant anal tumours
Squamous cell carcinoma
Adenocarcinoma
Number
Percent distribution
Number
Percent distribution
Number
Percent distribution
Age group 15–54 years 55–64 years 65–74 years 75+ years All age groups
695 616 678 735 2724
(25) (23) (25) (27) (100)
563 478 489 426 1956
(29) (24) (25) (22) (100)
90 110 147 235 582
(16) (19) (25) (40) (100)
Sex Male Female
1219 1505
(45) (55)
812 1144
(42) (58)
316 266
(54) (46)
Calendar period 1972–1981 1982–1991 1992–2001 2002–2009
387 540 928 869
(14) (20) (34) (32)
287 361 627 681
(15) (18) (32) (35)
78 131 247 126
(13) (23) (42) (22)
Best basis of diagnosis Histopathology Clinical Other
2586 108 30
(95) (4) (1)
1863 73 0
(95) (4) (1)
571 9 <6
(98) (2) (<1)
Stage at diagnosisa Local Regional Distant Missing
1208 690 216 610
(44) (26) (8) (22)
905 464 133 454
(46) (24) (7) (23)
246 183 55 98
(43) (31) (9) (17)
Cause of death Anal cancer Other causes Total
587 1030 1617
(36) (64) (100)
433 626 1059
(41) (59) (100)
120 289 409
(29) (71) (100)
a These categories represent information collected by the NSW Cancer Registry for the degree of disease spread within four months of diagnosis. These data do not include clinical information such as values for tumour, node or metastasis.
standardised incidence rate for all anal cancer histology subtypes combined was 1.45 per 100,000 (95% 1.15, 1.75). There were no differences in slope between men and women with an average annual percentage change of 1.7 (95% CI 1.1, 2.3) in men and 1.6 (95% CI 1.0, 2.2) in women (Supplementary Fig. 1). Most of the change over time in incidence has been due to increases in ASCC (Fig. 1c), with the incidence of AAC declining over time (Fig. 1b). For example, the age-standardised rate per 100,000 for AAC in 2009 was 0.19 (95% CI 0.09, 0.28), a substantial decrease from 0.37 (95% CI 0.22, 0.56) in 1998. The number of anal cancer cases has been projected to continue to increase over time (Supplementary Fig. 2) with the number of cases reaching a maximum of 198 per year (95% CI 169–236) by 2032. The incidence of AAC is projected to remain at the low current levels, which means that the increases in incidence and cases into the future will be from cases of ASCC (94% of cases projected in 2030–2035). 3.3. Relative survival for persons with ASCC and AAC Five-year relative survival estimates by age group and sex are first presented for people with anal cancer diagnosed in the most recent calendar period (1997–2006) (Table 3). Five-year relative survival in this calendar period across all anal cancers combined was 63% (95% CI 59–66%). This varied by histological subtype and gender. Survival was 37% greater for people with ASCC compared to people with AAC; 70% (95% CI 66–73%) and 51% (95% CI 43–59%) respectively. Women diagnosed with anal cancer experienced better survival than men, 69% (95% CI 64–73%) and 55% (95% CI 50– 60%) respectively—a 20% survival advantage for women with anal cancer compared to men. These gender differences in survival 5-
years after diagnosis were consistent in pattern across histological subtypes. People diagnosed at a younger age also had better survival compared to people diagnosed at older ages. For all histological subtypes combined, 5-year relative survival for people diagnosed 1997–2006 and aged 15–54 years was 70% (95% CI 64–75%) compared to 50% (95% CI 41–58%) for people aged 75 years or more. Similar patterns in age group differences in 5-year relative survival can be seen across histological subtypes. Estimates for people diagnosed 1972–1981, 1982–1991, 1992– 2001, and 2002–2006 are shown in Table 3. Survival improved over time for people diagnosed with anal cancer in NSW. Five-year relative survival for people with anal cancer diagnosed between 2002 and 2006 was 64% (95% CI 58–69%), an increase of 6% from 1992 to 2001. A step-wise increase over time in 5-year relative survival occurred for people diagnosed with ASCC. Five-year relative survival also increased over time for people diagnosed with AAC. 4. Discussion Our study was based on high-quality anal cancer data over a 38year period from the NSW Cancer Registry. We observed a significant, linear increase (AAPC of 1.6%) in incident cases of ASCC and a decline in the number of AAC cases. The incidence of ASCC is expected to steadily increase and make up the majority of cases of anal cancer in the future. We estimate there will be 198 cases of anal cancer in 2032. We also demonstrated a rise in 5-year relative survival for all histological subtypes and differences in survival by sex and histology.
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Fig. 1. Actual (1972–2009) and predicted (2010–2032) age-standardised incidence rates per 100,000 person-years of anal cancer in New South Wales, Australia. Estimates are shown separately for all anal cancer cases combined (A), anal squamous cell carcinomas (B), and anal adenocarcinomas (C).
Interpretation of our findings using data from other jurisdiction is more limited due to inconsistent approaches in cohort definitions and statistical analyses across studies. We consider that a more standard approach to defining malignant anal disease is warranted, including a consistent method for defining topography and morphology codes. Investigation of anal cancer incidence using national Australian cancer registration data as reported by Jin et al. showed AAPCs of 1.9 and 3.4 for men and women with ASCC
845
respectively, a statistically significant finding [6]. We found similar AAPC values when stratified by sex. Jin et al. did not report an AAPC for males and females combined for either ASCC or AAC. However, given consistency with the sex-specific AAPC estimates, we assume that our AAPC for age-standardised rates would be similar to the work undertaken by these investigators. Data from the United States (US) provides an international context. Melbye et al. studied changing incidence of anal cancer in the US by sex, ethnicity, and marital status between 1973 and 1989 [8]. Although APCs were not calculated, a significant increase in the incidence of anal cancer occurred over the study period. In particular, Melbye et al. note the increase over time that occurred among never married men and men living in San Francisco in the most recent calendar period in their study. They speculated whether part of the increase in this sub-population group could be attributed to the AIDS epidemic. It is likely that behaviour changes, including condom use, will have had an impact on anal cancer incidence trajectories over time but this is difficult to quantify. An increase over time in anal cancer incidence, particularly among men, was reported by Johnson et al. [7]. A more recent study from the US by Nelson et al. analysed changes over time, 1973–2009, in noninvasive anal disease (half of all cases analysed) and anal cancer where the AAPC was 16.7%, skewing the combined non-invasive and anal cancer AAPC estimates [9]. The Nelson et al. study did not assess changes over time for AAC. Differences in aetiology and age at diagnosis between ASSC and AAC will partially explain the different incident trajectories reported in this study. There is an increasing body of evidence that the strength of association between human papillomavirus and ASCC is much stronger compared to the association with AAC [27]. In addition, the age-specific rate for ASCC peaked at a much lower age (around 65 years of age) compared with AAC where the rate was highest in people aged 85 years and over. Through the use of age-period-cohort modelling in this study [17], we were able to simultaneously allow for these differences between ASCC and AAC disease trajectories. The impact of the introduction of prophylactic human papillomavirus (HPV) vaccines, including the qHPV vaccine offered in Australia to all secondary school students, on the future incidence of anal cancer is not yet quantified. A recent study by Hillman et al., where biopsy material from Australian people with anal cancer was analysed suggests that the HPV vaccine is unlikely to have a significant impact on anal cancer incidence for several decades [28]. However, this effect on the future incidence of anal cancer in Australia is likely to vary by gender due to differences in the introduction of the vaccination for males and females, with the male vaccination program only starting in 2013. Our study is limited in what we can say about the public health implications of improvements in anal cancer survival. For instance, without clinical TNM staging and hospital episode data it is difficult to quantify whether these survival improvements are due to earlier diagnosis or improvements in treatment or a combination of both factors. Survival following an anal cancer diagnosis is amenable to improvement through earlier detection, chemotherapy, radiotherapy, and, where appropriate surgical resection [5,29– 31]. Recent studies of colorectal treatment and outcomes during 2000–2008 have been conducted in NSW using linked administrative data such that it is possible to identify predictors of treatment receipt and factors associated with good prognosis [32– 34]. Use of this linked data for investigating anal cancer is feasible but not without limitations. For example, data on approximately 700 people with ASCC are recorded in the NSW Cancer Registry for the 2000–2008 period, or about 70 cases per year, leading to instability of estimates following stratification by covariates and limited interpretation of results. Further, linked cancer registry, administrative hospital episode and cancer treatment data
846
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Table 3 Five-year relative survival: adults (15–99 years) diagnosed with anal cancer in New South Wales, 1972–2006 with follow up to the end of 2007, by age group, sex and calendar period. Age group and sex estimates were calculated for people diagnosed 1997–2006. Survival estimates for people diagnosed 1997–2006 and 2002–2006 were calculated using the period approach; all other estimates were calculated using the cohort approach. All malignant anal cancer tumours
Squamous cell carcinoma
5-year relative survival
95% CI
5-year relative survival
95% CI
5-year relative survival
95% CI
Age group 15–54 years 55–64 years 65–74 years 75+ years
70 66 63 50
(64–75) (59–73) (55–70) (41–58)
73 71 69 65
(66–88) (63–78) (60–78) (53–76)
63 54 58 38
(43–78) (36–70) (41–72) (24–54)
Sex Male Female Total
55 69 63
(50–60) (64–73) (59–66)
64 75 70
(57–70) (69–80) (66–74)
47 60 51
(35–58) (43–68) (43–59)
Calendar period 1972–1981 1982–1991 1992–2001 2002–2006
46 55 60 64
(40–52) (50–60) (57–64) (58–69)
48 63 67 70
(41–55) (57–69) (62–71) (63–76)
45 37 50 58
(32–59) (28–46) (43–58) (41–74)
currently available in NSW do not yet have adequate information on receipt of radiation therapy, a primary treatment modality for recently diagnosed anal cancer cases, and the data on clinical TNM staging is yet to be validated. A number of population groups appear to be at higher risk of developing anal cancer, namely men who have sex with men [35,36], women with a history of previous HPV-associated cervical dysplasia or malignancy [37,38], and people whose immune response is limited in controlling HPV infection [39]. The data in this study do not allow for incidence and survival in these at-risk groups to be explored. Such data are value such as the recent study from Marcus et al. who identified that survival 5-years following a cancer diagnosis was poorer for people who were HIV positive compared to those who were HIV negative [40]. Future use of linked administrative datasets in NSW may provide useful insights. We found substantial survival differences by gender, with women having better survival. We also showed a 21% difference in 5-year relative survival difference between people with ASCC compared to people with AAC for the 1997–2006 period. These findings are consistent with previously published data and have the potential to provide more clarity on histological differences in anal cancer survival. In a study of anal cancer incidence and survival in the United States, Johnson et al. reported a 10% sexspecific difference in 5-year relative survival for people diagnosed in the 1973–2000 period with females have a greater survival outcome [7]. In their investigation of survival for ASSC in Australia, 1982–2004, Jin et al. reported a 20% difference between males and females in 5-year relative survival (74% for females and 61% for males) [6]. This study did not report relative survival by histological subtype. A more recent French study measured population-based survival for people diagnosed with anal cancer during 1982–2004 followed-up to the start of 2008. Bouvier et al. found that, regardless of age at diagnosis, 5-year net survival was greater for females compared to males [41]. With their larger sample size, the investigators were also able to model relative survival where they reported that excess mortality rates were highest in the first two years since diagnosis and then steadily declined up to 10 years post diagnosis. Data from England and Wales reported by Jeffreys et al., where rectal and anal cancers were combined in the analyses, showed differences in histologyspecific relative survival differences [13]. Five-year relative survival in the most recent period, 1991–1996, was 59% for epidermoid carcinomas (accounting for over half of all anal cancers) and 52% for adenocarcinomas. This translates into a 14%
Adenocarcinoma
difference. Jeffreys et al. also performed excess mortality rate modelling to compare differences by patient and tumour characteristics. Interestingly, the results of these analyses show that relative excess mortality rate differences between these two histological subtypes reduced to the null. Age at diagnosis and differences in tumour behaviour are likely to explain histologyspecific differences in relative survival. The age group distribution for adenocarcinoma is skewed to older age groups, with older age widely recognised as a factor associated with poorer relative survival [26,42]. Explanations for sex-specific differences in relative survival on the magnitude reported here are less clear but are possibly related to differences in age at diagnosis [33,34]. To our knowledge, no other study has predicted the number of future anal cancer cases and age-standardised incidence rates. It seems likely that the anal cancer incidence increases we have reported are due to changes over time in aetiology and demographic factors. We also confirmed differences in survival 5-years following an anal cancer diagnosis by sex and histological group but due to sparse data and limited staging data were unable to directly test whether these differences can be explained by stage at diagnosis or treatment. Nevertheless, these differences are marked enough to warrant further investigation and possible health system interventions to reduce these gaps in anal cancer survival. Conflict of interest None. Authors’ contribution J.Y., K.R., and M.S. designed the study. K.R. and M.S. analysed the data. M.S. wrote the manuscript. All authors revised the manuscript for intellectual content. Sources of funding This research was funded by a Cancer Epidemiology Linkage Grant from the Cancer Institute NSW. MS and JY were also supported by the Academic Leader, Cancer Epidemiology, grant awarded to The University of Sydney from the Cancer Institute NSW.
M.J. Soeberg et al. / Cancer Epidemiology 39 (2015) 842–847
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. canep.2015.10.008. References [1] H. De Vuyst, G.M. Clifford, M.C. Nascimento, M.M. Madeleine, S. Franceschi, Prevalence and type distribution of human papillomavirus in carcinoma and intraepithelial neoplasia of the vulva, vagina and anus: a meta-analysis, Int. J. Cancer 124 (7) (2009) 1626–1636. [2] N. Ouhoummane, M. Steben, F. Coutlee, T. Vuong, P. Forest, C. Rodier, et al., Squamous anal cancer: patient characteristics and HPV type distribution, Cancer Epidemiol. 37 (6) (2013) 807–812. [3] A.M. Saleem, J.K. Paulus, A.P. Shapter, N.N. Baxter, P.L. Roberts, R. Ricciardi, Risk of anal cancer in a cohort with human papillomavirus-related gynecologic neoplasm, Obstet. Gynecol. 117 (3) (2011) 643–649. [4] C.M. Vajdic, M.T. van Leeuwen, F. Jin, G. Prestage, G. Medley, R.J. Hillman, et al., Anal human papillomavirus genotype diversity and co-infection in a community-based sample of homosexual men, Sex. Transm. Infect. 85 (5) (2009) 330–335. [5] R. Glynne-Jones, J.M. Northover, A. Cervantes, Anal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up, Ann. Oncol. 21 (Suppl. (5)) (2010) v87–v92. [6] F. Jin, A.N. Stein, E.L. Conway, D.G. Regan, M. Law, J.M. Brotherton, et al., Trends in anal cancer in Australia, 1982–2005, Vaccine 29 (12) (2011) 2322–2327. [7] L.G. Johnson, M.M. Madeleine, L.M. Newcomer, S.M. Schwartz, J.R. Daling, Anal cancer incidence and survival: the surveillance, epidemiology, and end results experience, 1973–2000, Cancer 101 (2) (2004) 281–288. [8] M. Melbye, C. Rabkin, M. Frisch, R.J. Biggar, Changing patterns of anal cancer incidence in the United States, 1940–1989, Am. J. Epidemiol. 139 (8) (1994) 772–780. [9] R.A. Nelson, A.M. Levine, L. Bernstein, D.D. Smith, L.L. Lai, Changing patterns of anal canal carcinoma in the United States, J. Clin. Oncol. 31 (12) (2013) 1569–1575. [10] S. Simpson, R. Turner, Four decades of anal cancer in Tasmania, Australia: what do the case data tell us? Sex. Health 9 (3) (2012) 213–219. [11] C. Allemani, H.K. Weir, H. Carreira, R. Harewood, D. Spika, X.S. Wang, et al., Global surveillance of cancer survival 1995–2009: analysis of individual data for 25,676,887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet 385 (March (9972)) (2015) 977–1010. [12] M.P. Coleman, D. Forman, H. Bryant, J. Butler, B. Rachet, C. Maringe, et al., Cancer survival in Australia, Canada, Denmark, Norway, Sweden, and the UK, 1995–2007 (the International Cancer Benchmarking Partnership): an analysis of population-based cancer registry data, Lancet 377 (9760) (2011) 127–138. [13] M. Jeffreys, B. Rachet, S. McDowell, A.G. Habib, C. Lepage, M.P. Coleman, Survival from rectal and anal cancers in England and Wales, 1986–2001, Eur. J. Cancer 42 (10) (2006) 1434–1440. [14] D. Currow, W. Thompson, Cancer in NSW: Incidence Report, Cancer Institute NSW, Sydney, 2014, doi:http://dx.doi.org/10.1097/CEJ.0000000000000163. [15] M. Frisch, M. Melbye, Anal Cancer, in: D. Schottenfeld, J. Fraumeni Jr. (Eds.), Cancer Epidemiology and Prevention, 3rd ed., Oxford University Press, USA, 2006. [16] H.J. Kim, M.P. Fay, E.J. Feuer, D.N. Midthune, Permutation tests for joinpoint regression with applications to cancer rates, Stat. Med. 19 (3) (2000) 335–351. [17] B. Moller, H. Fekjaer, T. Hakulinen, H. Sigvaldason, H.H. Storm, M. Talback, et al., Prediction of cancer incidence in the Nordic countries: empirical comparison of different approaches, Stat. Med. 22 (17) (2003) 2751–2766. [18] Department of Planning and Infrastructure, New South Wales in the future: Preliminary 2013 population projections. Preliminary release of NSW state and local government area population projections. Sydney NSW Department of Planning & Infrastructure, 2013.
847
[19] Australian Bureau of Statistics. 3235.0—Population by Age and Sex, Regions of Australia. Canberra: Australian Bureau of Statistics, 2013. [20] H. Brenner, V. Arndt, Recent increase in cancer survival according to age: higher survival in all age groups, but widening age gradient, CCC 15 (9) (2004) 903–910. [21] H. Brenner, O. Gefeller, An alternative approach to monitoring cancer patient survival, Cancer 78 (9) (1996) 2004–2010. [22] F. Ederer, L.M. Axtell, S.J. Cutler, The relative survival rate: a statistical methodology, Natl. Cancer Inst. Monogr. 6 (1961) 101–121. [23] SAS Institute Inc., 2011. Cary, North Carolina, USASAS 9.3. [24] National Cancer Institute, Joinpoint Regression Program 4.1.1 ed. Bethesda, MD, USA, 2014. [25] Cancer Registry of Norway, NordPred 2.8 ed. Norway, 2003. [26] P.W. Dickman, A. Sloggett, M. Hills, T. Hakulinen, Regression models for relative survival, Stat. Med. 23 (1) (2004) 51–64. [27] J.R. Daling, M.M. Madeleine, L.G. Johnson, S.M. Schwartz, K.A. Shera, M.A. Wurscher, et al., Human papillomavirus, smoking, and sexual practices in the etiology of anal cancer, Cancer 101 (2) (2004) 270–280. [28] R.J. Hillman, S.M. Garland, M.P. Gunathilake, M. Stevens, N. Kumaradevan, C. Lemech, et al., Human papillomavirus (HPV) genotypes in an Australian sample of anal cancers, Int. J. Cancer 135 (4) (2014) 996–1001. [29] P.F. Engstrom, J.P. Arnoletti, A.B. Benson 3rd, J.D. Berlin, J.M. Berry, Y.J. Chen, et al., NCCN clinical practice guidelines in oncology. Anal carcinoma, J. Natl. Compr. Cancer Netw. 8 (1) (2010) 106–120. [30] F. Lim, R. Glynne-Jones, Chemotherapy/chemoradiation in anal cancer: a systematic review, Cancer Treat. Rev. 37 (7) (2011) 520–532. [31] R.J. Myerson, M.C. Garofalo, I. El Naqa, R.A. Abrams, A. Apte, W.R. Bosch, et al., Elective clinical target volumes for conformal therapy in anorectal cancer: a radiation therapy oncology group consensus panel contouring atlas, Int. J. Radiat. Oncol. Biol. Phys. 74 (3) (2009) 824–830. [32] M.L. Jorgensen, J.M. Young, T.A. Dobbins, M.J. Solomon, Assessment of abdominoperineal resection rate as a surrogate marker of hospital quality in rectal cancer surgery, Br. J. Surg. 100 (12) (2013) 1655–1663. [33] M.L. Jorgensen, J.M. Young, T.A. Dobbins, M.J. Solomon, Predictors of variation in colorectal cancer care and outcomes in New South Wales: a populationbased health data linkage study, Med. J. Aust. 200 (7) (2014) 403–407. [34] M.L. Jorgensen, J.M. Young, M.J. Solomon, Older patients and adjuvant therapy for colorectal cancer: surgeon knowledge, opinions, and practice, Dis. Colon Rectum 54 (3) (2011) 335–341. [35] J.R. Daling, N.S. Weiss, T.G. Hislop, C. Maden, R.J. Coates, K.J. Sherman, et al., Sexual practices, sexually transmitted diseases, and the incidence of anal cancer, N. Engl. J. Med. 317 (16) (1987) 973–977. [36] G. D'Souza, D.J. Wiley, X. Li, J.S. Chmiel, J.B. Margolick, R.D. Cranston, et al., Incidence and epidemiology of anal cancer in the multicenter AIDS cohort study, J. Acquir. Immune Defic. Syndr. 48 (4) (2008) 491–499. [37] M.T. Goodman, Y.B. Shvetsov, K. McDuffie, L.R. Wilkens, X. Zhu, P.J. Thompson, et al., Sequential acquisition of human papillomavirus (HPV) infection of the anus and cervix: the Hawaii HPV Cohort Study, J. Infect. Dis. 201 (9) (2010) 1331–1339. [38] D.P. Ryan, C.C. Compton, R.J. Mayer, Carcinoma of the anal canal, N. Engl. J. Med. 342 (11) (2000) 792–800. [39] A.E. Grulich, M.T. van Leeuwen, M.O. Falster, C.M. Vajdic, Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis, Lancet 370 (9581) (2007) 59–67. [40] J.L. Marcus, C. Chao, W.A. Leyden, L. Xu, J. Yu, M.A. Horberg, et al., Survival among HIV-infected and HIV-uninfected individuals with common non-AIDSdefining cancers, Cancer Epidemiol. Biomarkers Prev. 24 (8) (2015) 1167–1173. [41] A.M. Bouvier, A. Belot, S. Manfredi, V. Jooste, Z. Uhry, J. Faivre, et al., Trends of incidence and survival in squamous-cell carcinoma of the anal canal in France: a population-based study, Eur. J. Cancer Prev. (2015) , doi:http://dx.doi.org/ 10.1097/CEJ.0000000000000163. [42] H. Charvat, N. Bossard, L. Daubisse, F. Binder, A. Belot, L. Remontet, Probabilities of dying from cancer and other causes in French cancer patients based on an unbiased estimator of net survival: a study of five common cancers, Cancer Epidemiol. 37 (6) (2013) 857–863.