Epidemiology of keratinocyte carcinomas after organ transplantation

Keratinocyte carcinoma (KC) is the most common type of cancer among white populations, but it is even more common among solid organ transplant recipients (OTRs). The most frequent histological type of KC among OTRs is cutaneous squamous cell carcinoma (cSCC), followed by basal cell carcinoma, although the reverse is seen in the general population. Metastatic cSCCs are more frequent, and mortality is increased compared with immunocompetent populations. There is strong evidence that the risk of KC among OTRs rises with increasing time after transplantation and older age at transplantation, and that KC is enhanced in those with sun‐damaged skin. This evidence suggests that accelerated accumulation of genetic damage from several sources leads to excess KC in OTRs. We describe international variation in KC and focus on trends in immunosuppressive regimens, the role of ultraviolet susceptibility and exposure, and the contribution of genetics to tumour development. Further epidemiological studies are needed to address gaps in our understanding of the mediation of excess KC by immunosuppressive drugs, viral infection, genetic susceptibility, timing of relevant ultraviolet exposure or some combination of these factors.

dependence on immunosuppression to maintain engraftment and an increase in cancer rates among organ transplant recipients (OTRs). Specifically, there is an approximately two-to sixfold increased risk of all types of malignancy compared with cancer rates in the general population. [1][2][3] In this systematic review we explore the epidemiology of keratinocyte carcinoma (KC), the most common malignancy that occurs among transplant recipients in light-skinned populations.
KC comprises two main histological types, cutaneous squamous cell carcinoma (cSCC) and basal cell carcinoma (BCC), which together account for 90% of post-transplantation KC. 4 In OTRs, as with KC in the general population, both histological types are caused mainly by excessive exposure to ultraviolet (UV) radiation and are found most often on sun-exposed skin. They also occur more often in men than in women and are more frequent with increasing age. Despite these commonalities with KC in the general population, KCs among OTRs are characterized by higher rates of cSCC than BCC, while the reverse is seen among immunocompetent populations. 5 This excess of cSCC after transplantation suggests that unique factors associated with transplantation play an important role in promoting cSCC development among OTRs.
In this review we discuss epidemiological studies of demographic factors, transplant-related factors, UV susceptibility and exposure measures, viral aetiology and genetic variation associated with development of KC after organ transplantation. We also point out gaps in our understanding that can be addressed in future epidemiological studies of risk factors and prognosis of KC in OTRs.

Demography and demographic trends Incidence
Studies conducted in the last 10 years reporting standardized incidence ratios (SIRs, which measure incidence relative to the general population) are summarized in Table 1, and cumulative incidences over time are summarized in Table 2. SIRs of 82 [95% confidence interval (CI) 70-95] for cSCC and 6Á4 (95% CI 5Á5-7Á5) for BCC have been reported for all transplant recipients combined in Denmark, 6 although incidence rates differ in other studies by organ transplanted and study location ( Table 1). The SIRs for cSCC strongly increased with longer time since transplantation, which supports a role for duration of immunosuppression in cSCC development. As KC risks increase with time after transplantation for all solid organ types, cumulative incidence rates may be more clinically relevant than SIRs for planning treatment and surveillance posttransplantation (Table 2).
Cumulative incidence estimates also vary somewhat by geographical location, study period and centre, but all support an excess of BCC and cSCC regardless of the type of organ transplanted. A long-term follow-up of a cohort in Sweden 3 reported that 19% (95% CI 14-25%) of heart/lung, 18% (95% CI 12-29%) of liver and 9% (95% CI 8Á2-9Á9%) of kidney recipients had a cSCC by 20 years post-transplantation. Another large cohort study in the U.K. followed over 900 transplant recipients, for an average of approximately 10 years; the investigators reported a 30-year cumulative incidence of 65Á4% for cSCC and 46Á5% BCC. 7 These long-term studies in countries with population-based data suggest that the majority of light-skinned OTRs will eventually develop cSCC, providing key information for surveillance guidelines.

Mortality
A recent review 8 reports that mortality rates for KC in the general population, dominated by cSCC, are 0Á5 and 0Á4 per 100 000 in developed and developing countries, respectively. Much higher mortality rates were reported in a study of transplant recipients in the U.S.A., with a cSCC-specific mortality rate of 4Á9 per 100 000. 9 KC-specific mortality among OTRs in the U.S.A. was reported as 11% for BCC, 18% for cSCC and 15% for both. 10 A U.K. study of all OTRs reported that seven of 26 OTRs (27%) with more than 10 diagnosed skin cancers died of their disease. 7 In a US registry-based study of nearly 500 000 OTRs (all organ types), patients aged ≥ 50 years at the time of transplantation were at higher risk of dying from cSCC than those aged < 50 years at transplantation (hazard ratio 2Á8, 95% CI 1Á9-4Á0). An international study of OTRs with metastatic skin cancer followed 68 patients treated between 1989 and 2001, among whom the majority (n = 62) had metastatic cSCC. The study reported that among the 68 OTRs with metastatic skin cancers, 34 died, 23 with metastatic skin cancer, compared with 34 alive at the end of the study, 14 with disease. 11 In a large single-centre study in France that followed 188 OTRs, 10 developed metastatic cSCC and seven died of disease. 12 The exact reason for the comparatively high rate of mortality from cSCC in OTRs compared with the general population is unknown. It is likely that aggressive cSCC in OTRs is attributable to a combination of factors such as increased frequency of cSCC leading to genetic instability, carcinogenic immunotherapies, prior UV damage and a blunted immune response to tumour antigens due to underlying disease, immunosuppression and older age.

Geography
Proximity to the equator and susceptibility to cumulative sun damage explain part of the worldwide variability in KC incidence among OTRs. Thus, higher rates of KC are reported in white OTRs in low-latitude Australia compared with Europe or the U.S.A. The lowest rates are reported for OTRs from Africa, East Asia and India. Taken together, these data point to the protective effect of darker skin pigmentation 13 (also see 'UV susceptibility' below). This is illustrated by the much less significant burden of KC among Asian OTRs. A Japanese study of 1744 kidney transplant recipients (KTRs) with 12 982 person-years of observation reported only one cSCC. 14 Similarly, a review of cancer registry data of 283 Taiwanese KTRs from 1981 to 2002 and a retrospective review of 569 Chinese liver transplant recipients failed to reveal even a single KC. 15,16 A low incidence of KC has also been reported in liver transplant recipients in Korea, with only two skin cancers out of 44 de novo cancers. 17

Second keratinocyte carcinoma tumours
In the general population, the risk of developing second KC tumours was 50% at 5 years after a first KC. 18 This high rate of second KC primaries is even more elevated among OTRs. In the U.K. long-term cohort study of OTRs, 7 66% of those with KC developed multiple tumours. In a US study of 166 lung transplant recipients, 80% developed a second cSCC an average of 4 years after the initial cSCC. 19 A study in the Netherlands reported that 72% of KTRs developed a second KC 5 years after transplantation, 20 with similarly high rates of second cSCCs in other studies of KTRs. 12,21 Age at transplantation Age at transplantation is an independent risk factor in the development of KC. 7,9 Many studies have reported that older

Transplantation-related factors Maintenance immunosuppression
Immunosuppression is a key causal cofactor for KC among those OTRs susceptible to UV-induced skin damage. Long duration of induced immunosuppression post-transplantation to maintain grafts results in not just years of a depressed immune response but also a longer duration of exposure to potentially carcinogenic medications. 25,26 Azathioprine, one of the first drugs used to maintain graft function, interrupts rapidly dividing cells, including synthesis of immune cells, and is most often associated with increased risk of KC. 27 In the setting of significant ultraviolet A exposure, azathioprine has been found to be mutagenic. 28 Ciclosporin, a calcineurin inhibitor, is another medication used to prevent graft rejection post-transplantation. It was introduced in the 1980s to supplement azathioprine-based regimens, and can disrupt nucleotide excision repair and increase sensitization to UV damage. 26,29 Through its action on the transforming growth factor-b pathway, ciclosporin has been found both to increase phenotypic changes in cells and to increase tumour invasiveness. 30 The KC risk profile of drugs introduced to post-transplantation medication regimens in more recent years has not yet been well characterized. Mycophenolate mofetil and mycophenolate sodium are nucleotide inhibitors introduced in 1995 and 2004, respectively, to replace azathioprine. Limited data provide evidence for a role of these mycophenolic acid preparations in reducing KC risk compared with other immunosuppressive medications. 31 Likewise, tacrolimus is a calcineurin inhibitor introduced to replace ciclosporin in 1994. Data from a meta-analysis of randomized trials comparing overall cancer rates according to choice of calcineurin inhibitor demonstrated no difference between tacrolimus and ciclosporin, 32 but more recent data suggest lower skin cancer risk associated with tacrolimus. 25 Changing clinical practice will require pooling data from longitudinal projects that keep track of major changes in the maintenance regimen over time post-transplantation.

Newer drugs
Mammalian target of rapamycin pathway inhibitors like sirolimus and everolimus may have anticancer properties, and have been studied in OTRs, [33][34][35][36] although side-effects have discouraged wide-scale adoption. 37 A multicentre trial that switched individuals using calcineurin inhibitors to sirolimus observed a decreased risk of second cSCC compared with the usualtreatment calcineurin inhibitor group in a trial among KTRs. 34 A single-institution study found that sirolimus did not reduce risk of skin cancers after lung transplantation in a US study. 19 Immunosuppression by type of organ transplanted The average level of immunosuppression post-transplantation varies by the type of organ transplanted. The highest levels of long-term immunosuppression are required for heart/lung transplants, followed by kidney, and lowest for liver. Also, lower levels of immunosuppression are reported with organs from living compared with deceased donors, and by level of mismatch between donors and recipients. 38 These relative levels of immunosuppression correlate with rates of KC by organ type, for both incidence and mortality. 39 Higher levels of immunosuppression are associated with higher risk of KC, and therefore immunosuppression is lowered when possible as part of treatment for aggressive KC. 40 Ultraviolet susceptibility and exposure

Measures of ultraviolet susceptibility
Studies in the last 5 years have used a variety of measures of UV susceptibility in KC among OTRs ( Table 3). The most common measure of skin susceptibility to the harmful effects of UV has been a self-reported tendency to sunburn or develop a tan after acute sun exposure, classified according to the Fitzpatrick sixpoint phototype scale. The Fitzpatrick scale ranges from I, skin that always burns without tanning, to VI, black skin. 25 An investigation at Northwestern University, Illinois, of 63 liver recipients and 620 kidney recipients reported that Fitzpatrick skin types I and II were independently associated with the development of KC. 41 They also found that the duration of time until an incident KC was greater in patients with higher Fitzpatrick skin types. Similarly, an investigation of US transplant patients demonstrated a cSCC accrual of 14% among nonwhite OTRs (eight of 58) compared with 65% among white OTRs (61 of 94). 42 BCC was also diagnosed more frequently among white vs. nonwhite OTRs: 28% (26 of 94) vs. 2% (one of 58). Indeed, all recent studies show higher skin cancer and cSCC risk in OTRs with fair skin (phototypes I-II), or have shown a tendency towards higher numbers of new cSCCs compared with those with dark skin phototypes, irrespectively of the organ type (Table 3). 43 Positive associations were also seen with light hair 44 and light eye colour. 45

Measures of ultraviolet exposure
Measures of cumulative UV exposure have ranged from a focus on geographical location to indices of occupational sun exposure. Other exposure variables used include residential history, sun avoidance, use of sunscreen and clinical elastosis of the neck (a sign of chronic photodamage).
The relationship of UV exposure to cSCC in OTRs is not as clear as for general populations, suggesting a need for further research into UV exposure among OTRs. In one Australian cross-sectional study, 46 native-born OTRs had twice the prevalence of cSCC of those born elsewhere, while clinical elastosis of the neck was not convincingly associated with cSCC.  Occupational sun exposure was inversely associated in some studies, 43,47 but positively 45 in another (Table 3). In several of these studies, potential confounding factors were not assessed, and thus confounding could explain differences in results.
Most of the more recently published studies of KC risk in relation to UV exposure have been retrospective cohort studies in KTRs (Table 3). Also, UV exposure has usually been solar as opposed to artificial UV (indoor tanning). Most outcomes studied were 'all KCs combined', with a minority looking at cSCC alone. Reviews that summarized KC studies in OTRs in the previous decade consistently found fair skin and high sun exposure to be risk factors for the development of KC along with a variety of secondary factors. 48,49 Future studies of patterns of UV exposure and risk of KC must be specific about the histological type(s) of KC that develop after transplantation.

Viral aetiology and replication
In addition to its intended beneficial effects, immunosuppressive therapy is also associated with increased susceptibility to virus-associated cancers in OTRs, including non-Hodgkin lymphoma (sevenfold SIR, associated with Epstein-Barr virus infection), Kaposi sarcoma (61-fold SIR, associated with human herpesvirus 8) and liver cancer (11-fold SIR, commonly caused by persistent infection with hepatitis B or C virus). 1 Given that all OTR studies have reported greatly increased SIRs for KC (Table 1), a potential viral aetiology for KC is described in more detail in the accompanying review on pathogenesis. 50 The leading candidate viruses in prior studies of OTRs are cutaneous types of human papillomavirus (HPV). Despite rigorous study designs and occasional significant findings for HPV types, studies of associations between HPV and KC among OTRs are so far inconclusive. Larger studies with pooled samples and harmonized laboratory techniques may increase the specificity of the associations. Alternatively, undetected HPV types, or indeed other viruses, may account for the associations. [51][52][53] Taken together, the epidemiological data support a possible but not definitive role for a cutaneous HPV virus in the aetiology of KC, supporting the need for further studies.

Genetic variation: germline genetic risk factors
A limited number of genetic studies have been performed in OTRs, and most studies were small in size, which may decrease reproducibility as the effect size for most confirmed genetic associations is < 1Á5. As in the general population, variants in pigment-related genes, such as MC1R, were associated with an approximately twofold increased risk for cSCC. 54 A second study evaluated the impact of genetic variants in eight pigmentation-related genes on time to first cSCC posttransplant. That study reported that the rs12203592 T allele in IRF4, a gene important in melanin synthesis, was associated with an increased hazard ratio for time to first cSCC, and the rs16891982 C allele in SLC45A2, which encodes a melanosome transport-associated protein, was associated with decreased risk of cSCC. 55 These findings persisted after adjusting for skin, hair and eye colour, suggesting that the effects of the melanin genes extend beyond phototype.
Other candidate genes were hypothesized to be important in KC risk among OTRs. Due to the link between immune system suppression and increased KC risk, human leucocyte antigen alleles that activate acquired T-cell responses were assessed for their potential impact. 56,57 Additional candidate gene polymorphism studies have not yet yielded convincing associations. These include KC risk linked to TP53 polymorphisms associated with UV damage, 58 variants in detoxifying enzymes such as glutathione S-transferase, 59 and CYP1A1 polymorphisms associated with drug metabolism. In short, larger studies are needed to clarify the genetic associations and to allow for adjustment for potential confounding effects.
Future studies may be warranted to examine drug-gene interactions for OTRs. For example, a recent report suggests that the excess risk of cSCC observed in individuals with lung transplants who take voriconazole may be related to having two copies of a CYP2C19 rapid-metabolizing allele. 60 This finding adds to the data supporting excess risk in these patients being due to the photosensitizing properties of the drug. 61 The implications of such drug-gene interaction studies may be particularly important for highly medicated OTRs.

Future studies
The epidemiological evidence in OTRs strongly supports an excess of cSCC compared with BCC, and higher mortality associated with cSCC after transplantation, but the aetiology that drives these differences is unclear. Future studies may comprehensively assess the direct impact of specific immunosuppressive drugs, patterns in UV exposure before and after transplantation, or a role for cutaneous viruses, but these remain areas that require careful and collaborative epidemiological study.
Challenges for such future studies include collecting data in a standardized way across studies in different geographical locations, preferably in longitudinal cohort settings with harmonized data collection instruments. Such studies need to evaluate best approaches to measuring UV exposure, and changes over time in immunosuppressive medication use. Studies of cutaneous viruses, such as HPV, will need to standardize their approach to DNA and serological testing of the same samples across laboratories. As multiple KC tumours and the potential for metastatic disease have been noted, an important aspect of future studies will be continued surveillance over time to assess risk factors for aggressive KC after transplantation.
Key ongoing large surveillance studies are methodologically challenging but important for assessing the epidemiology of multiple primary KCs. Consortia that include studies of OTRs and immunocompetent individuals will facilitate sharing data across research groups so that future epidemiological studies are well positioned to contribute to a more definitive understanding of the excess risk of KC among OTRs. Such findings may enhance surveillance and provide new treatment targets for KC.

Conclusions
This review highlights the importance of assessing KC risk by histological type, and highlights that the common cofactors, such as UV exposure and level of immunosuppression, cannot be overlooked. We find that the risk of KC after organ transplantation is generally higher among those whose skin is susceptible to photodamage and who have a prolonged immunosuppression history. Thus, OTRs with the highest reported history of sun damage before transplantation have the most KCs. In addition, older age at transplantation, time since transplantation, and type of organ transplanted are likely surrogates for accumulated UV damage and the level or duration of immunosuppression.