European Journal of Haematology 96 (285–290)

ORIGINAL ARTICLE

Risk factors for non-melanoma skin cancer in patients with essential thrombocythemia and polycythemia vera

mez1, Vicent Guillem1, Arturo Pereira2, Francisca Ferrer-Mar
ın3, Alberto Alvarez-Larr
n4 , Montse Go a 5 6 7 4 1 lia Estrada , Joaqu
ın Mart
ınez-Lo
pez , Anna Angona , Paula Amat , Ana Kerguelen , Nata
ndez-Boluda1 Blanca Navarro1, Carles Besses4, Juan-Carlos Herna Hematology and Medical Oncology Department, Hospital Cl
ınico Universitario, INCLIVA, Valencia; 2Hemotherapy and Hemostasis Department, Hospital Cl
ınic, Barcelona; 3Hematology Department, Hospital Morales Messeguer, Murcia; 4Hematology Department, Hospital del Mar-IMIM, Barcelona; 5Hematology Department, Hospital La Paz, Madrid; 6Hematology Department, Institut Catal a d’Oncologia-Hospital Germans Trias i Pujol, Badalona; 7Hematology Department, Hospital 12 de Octubre, Madrid, Spain

1

Abstract Objectives: Population-based studies have reported an increased incidence of skin cancer in patients with essential thrombocythemia (ET) and polycythemia vera (PV). We have examined the risk factors for nonmelanoma skin cancer (NMSC) in patients diagnosed with ET or PV during 1973–2012. Methods: A case– control study was performed to compare the clinical and treatment-related data of 51 ET/PV patients who had NMSC with that of 401 patients who did not. We also evaluated whether polymorphisms in 12 genes involved in DNA integrity predisposed to NMSC. Results: By multivariate logistic regression analysis, risk factors for NMSC were older age (OR: 1.7, 95% CI: 1.3–2.1, P < 0.001), male sex (OR: 2.1, 95% CI: 1.1– 3.8, P = 0.023), higher cumulated hydroxycarbamide dose (OR: 1.3, 95% CI: 1.1–1.7, P = 0.017), and busulphan exposure (OR: 3.2, 95% CI: 1.05–10.0, P = 0.041). On the time-to-event prognostic model, factors independently associated with increased cumulative incidence of NMSC were age (5% increased risk per year; P < 0.001), male sex (91% increased risk; P = 0.022), and hydroxycarbamide exposure (22% increased risk; P = 0.065). No susceptibility gene variant was identified. Conclusions: These findings suggest that the risk to develop NMSC in ET/PV patients results from the combined effect of common risk factors (age, male sex) together with cytoreductive treatment. Key words essential thrombocythemia; polycythemia vera; non-melanoma skin cancer; genetic polymorphisms; hydroxycarbamide Correspondence Juan-Carlos Hern
andez-Boluda, Hematology Department, Hospital Cl
ınico Universitario, INCLIVA, Avd. Blasco ~ez 17, Valencia 46010, Spain. Tel: +34 646584918; Fax: +34 96 3987820; e-mail: [email protected] Ib
an Accepted for publication 16 May 2015

Essential thrombocythemia (ET) and polycythemia vera (PV) are myeloproliferative neoplasms (MPNs) characterized by an indolent clinical course, tendency to vascular complications, and a risk of transformation into myelofibrosis and acute leukemia (1). Recently, the incidence of second nonmyeloid malignancies in patients with MPN has been reported to be higher than that of general population (2–4). The reason for such association remains unclear, although a pathogenic role of the cytoreductive drugs used to treat these conditions can be entertained (5, 6). Although population-based studies have shown that nonmelanoma skin cancer (NMSC) is more frequent in patients with MPN (2, 3), studies on the risk factors for NMSC in

© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

doi:10.1111/ejh.12588

these patients are lacking. In general, exposure to ultraviolet (UV) radiation is recognized as a major etiological factor, by inducing photoproducts that are mutagenic to the DNA (7, 8). However, there is growing recognition that this environmental challenge interacts with individual genetic factors in the process of cancer causation (9). Thus, susceptibility to develop skin cancer could also depend on the mechanisms involved in maintenance of DNA integrity, whose efficiency has been shown to differ significantly within a population (10, 11). This interindividual variation derives, at least in part, from inherited low-penetrance genetic variants, such as single nucleotide polymorphisms (SNPs) (12). To this respect, published data support the existence of predisposi-

285


mez et al. Go

Skin cancer in MPNs

tion variants in genes involved in DNA repair (13–16) or telomere length maintenance (17, 18) that could influence the risk of NMSC in the general population. The aim of the present case–control study was to compare the clinical and treatment-related characteristics of patients with ET and PV who developed NMSC with that of those who did not. We further investigated whether SNPs in genes involved in the preservation of DNA integrity contribute to the risk of NMSC in this population. Patients and methods Study subjects

The study population comprised 452 patients diagnosed with ET or PV. Fifty-one patients (11%) developed histologically confirmed squamous cell carcinoma (SCC) or basal cell carcinoma (BCC) of the skin any time since the time of diagnosis. All patients were diagnosed with ET or PV between 1973 and 2012 according to the criteria in use at the time of first observation (19). Cases were accrued from several Spanish hospitals on the basis of the availability of DNA samples for genotyping and were followed up for a median time of 8.6 yr (range, 1–37). Clinical and hematologic features at MPN diagnosis, as well as treatment-related data and the incidence of NMSC, were obtained from the clinical records. The study was approved by a reference ethics committee and conducted according to the Declaration of Helsinki. Molecular analysis

An analysis of 16 SNPs located in 12 genes involved in DNA repair or maintenance of telomere length was performed. SNPs were selected among those with high heterozygosity and previous scientific evidence for their potential association with susceptibility to human cancer. The candidate SNPs were the following: XPD (rs13181), ERCC3 (rs4662717), ERCC4 (rs3136155), ERCC6 (rs3793784), ERCC8 (rs3117), RPA1 (rs2287321), RPA2 (rs7356), RPA3 (rs6945447), XPA (rs2808668), XPC (rs2228000, rs2228001), XRCC1 (rs25487), TERT (rs2853669, rs2736098, rs2853677, rs2736100) (see Table S1 for details). Genotyping analysis of the SNPs was performed by realtime PCR, using the TaqManâ SNP Genotyping on demand Assays, which are commercially supplied by Applied Biosystems (Life Technologies S.A, Alcobendas, Spain). Briefly, each sample reaction was composed of 2.5 lL of TaqMan Genotyping Master Mix, 0.12 lL of TaqMan probe assay 40X, and 2.5 ll of DNA sample at 5 lg/mL. Thermal cycling and detection was performed in a Fast Real-time PCR system 7900HT from Applied Biosystems (Life Technologies S.A). Thermal cycler conditions were as follows: a first stage of 50°C for 2 min, a second stage of 95°C for

286

10 min, and a third stage consisting 45 cycles of 95°C for 15 s, 60°C for 1 min. Statistical analysis

The statistical tools for genotype analysis of SNPs (Hardy– Weinberg equilibrium, allele and genotype distributions, and association tests) were provided by SNPStats (20). This web-based application generates odds ratios (OR), confidence intervals (CI), and P values for multiple inheritance models. Akaike’s information criterion (AIC) and Bayesian information criterion (BIC) were calculated to select the best inheritance model for each specific polymorphism, with the preferred model being the one with the lowest AIC and BIC values. The association between clinical and genetic variables with the risk of NMSC, adjusted for other potentially predisposing factors, was assessed by logistic regression analysis. Variables included in the analyses were age (categorized by quintiles), sex, exposure to sunlight (considering two geographic areas separated by latitude 40° North), type of MPN (ET/PV), hematologic values at MPN diagnosis, JAK2 mutational status, JAK2 SNP genotype (rs12340895), 16 candidate gene SNPs, and exposure to hydroxycarbamide (HC) and/or other cytoreductive agents. Potential synergisms among risk factors (e.g., SNPs and exposure to cytoreductive agents) were investigated by including the corresponding interaction terms in the regression models. The cumulative incidence of NMSC was calculated by the Fine and Gray’s multivariate analysis, taking death as a competing risk. Statistical analyses were performed using the SPSS 15.0 package (SPSS, Chicago, IL, USA) and the Stata 11.1 software (www.stata.com). Results Characteristics of patients with NMSC

The case series comprised 51 ET/PV patients who developed a total of 66 cutaneous tumors, of which 36 were BCC and 30 SCC (BCC/SCC ratio: 1.2). Their median age at the time of first NMSC was 76 yr (range, 45–91). Median time from MPN diagnosis to first NMSC was 6.2 yr (range, 0.1–23). Twenty-six patients had BCC only, 19 had SCC only, and 6 had both histological subtypes. A total of 10 patients developed multiple skin tumors during the observation period. The anatomical distribution of first tumors was the face/scalp (70%), ear/neck (9%), arms (6%), legs (6%), and trunk (9%). Analysis of factors associated with the risk of NMSC

The 51 case patients who developed NMSC after diagnosis of ET or PV were compared with the 401 controls who did not. The characteristics of both groups are summarized in

© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd


mez et al. Go

Skin cancer in MPNs

Table 1. As can be seen, case patients were more often males and older than control patients, predominantly lived below latitude 40° North, had a higher Hb level at MPN diagnosis, and displayed more frequently a JAK2-mutated gene. A higher proportion of patients with PV were seen in the case group as compared to that of the control group. For the purpose of this study, patients who received interferon or anagrelide as the only cytoreductive treatment were included into the ‘no exposure’ group, as these drugs are regarded as non-carcinogenic. Overall, case patients were more exposed to cytoreductive agents. Moreover, they accumulated a significantly higher dose of HC than control patients, mainly due to the longer treatment duration on this drug. All candidate SNPs were successfully genotyped in more than 95% of the study samples. Genotypic distribution of the SNPs was found to be in Hardy–Weinberg equilibrium. Of note, none of the SNPs in genes involved in the mainte-

Table 1 Clinical characteristics of case patients with ET and PV who developed non-melanoma skin cancer and control patients who did not

Characteristic

nance of DNA integrity was associated with the occurrence of NMSC (Table S2). JAK2 SNP (rs12340895) was also tested regarding its putative influence on the NMSC risk. In univariate analysis, homozygous subjects for the minor allele (G) had a higher incidence of NMSC, but this association disappeared when it was adjusted for the type of MPN. At the multivariate logistic regression analysis of the predictive factors listed in Table 1, the only features independently associated with NMSC were age, categorized by quintiles (OR: 1.7, 95% CI: 1.3–2.1, P < 0.001), male sex (OR: 2.1, 95% CI: 1.1–3.8, P = 0.023), higher cumulated dose of HC (OR: 1.3, 95% CI: 1.1–1.7, P = 0.017), and exposure to busulphan (OR: 3.2, 95% CI: 1.05–10.0, P = 0.041). As patients with longer follow-up were more likely to receive a diagnosis of NMSC and also to accumulate larger doses of cytoreductive agents, we investigated whether these features and other plausibly carcinogenic factors were associ-

Cases (n = 51)

ET/PV 29 (57%)/22 (43%) 70 (29–86) Age at ET/PV diagnosis, years1 Sex, male/female 27 (53%)/24 (47%) Residence below latitude 27 (53%) 40° North Hb, g/L1 157 (98–214) Leukocytes, 9109/L1 10.3 (4.6–24.7) Platelets, 9109/L1 663 (137–1405) JAK2 mutation 36/45 (80%) JAK2 genotype (rs12340895) CC/CG 34/48 (71%) GG 14/48 (29%) Exposure to cytoreductive agents2 No exposure 1 (2%) HC only 43 (84%) Radioactive phosphorus 3 (6%) Busulphan 5 (10%) Cumulated exposure to cytoreductive agents HC cumulated dose (g)1 2415 (22.5–6120) Months on HC1 66 (1–204) Months on alkylating 132 (11–139) agents1 Cumulative dose of HC (in quintiles) 1st quintile (no HC) 1 (2%) 2nd quintile (15–750 g) 18 (35%) 3rd quintile (780–1680 g) 10 (21%) 4th quintile (1687–2940 g) 11 (21.5%) 5th quintile 11 (21.5%) (2955–11 700 g)

Controls (n = 401)

P

286 (71%)/115 (29%) 60 (13–93)

0.03 0.0001

137 (34%)/246 (66%) 24 (47%)

0.01 0.015

145 9.3 719 242/375

(82–238) (4.7–29) (129–2600) (64%)

309/364 (85%) 55/364 (15%)

0.008 0.3 0.2 0.04 0.022

(28%) (67%) (7%) (4%)