Materials and Methods

Study population

The PUVA Follow-up Study has prospectively studied 1380 patients with psoriasis first treated with PUVA in 1975 to 1976 at 16 university centers in the USA. At time of enrollment all patients provided written informed consent. From 1975 to 2001, we have performed 19 cycles of follow-up interviews. Structured questions concerning health events, including skin cancers and their use of psoriasis treatments are included in each follow-up. The Beth Israel Deaconess Medical Center Committee on Clinical Investigation has approved the study.

Tumor risk

We determined the incidence of SCC and BCC counting all tumors in the incidence calculation. The values for all variables were determined for each patient for each calendar year and hence were time dependent. In each year, we classified each patient according to PUVA exposure, treatment exposures, and their disease and demographic attributes, including age. For each type of tumor, the analysis included 27,840 observations (i.e., one for each year for each patient). Only biopsy confirmed SCC (not including SCC in situ or keratoacanthoma) or BCC were considered. We defined PUVA exposure according to total number of treatments received up to that year. As previously, high ultraviolet (UV)B and tar exposure were defined as 300 or more UVB treatments and/or 45 mo or more of tar use (^{Stern et al, 1979}) and high methotrexate exposure was defined as 36 mo or more of therapy (^{Stern et al, 1984}). We considered calendar years with 10 PUVA treatments to be PUVA-exposed years. Those with 10 or fewer treatments to be non-exposed years. In calculating years since last PUVA therapy we excluded the first calendar year after 10 or more treatments as a year of no PUVA exposure. All analyses were done separately for BCC and SCC.

Statistical methods

We used the Chi-square test to determine the statistical significance of difference in the distribution of categorical variables. Initially, we applied Poisson regression models to our data to relate risk factors to skin cancer incidence; however, when we tested for goodness-of-fit of the Poisson model to our data, the fit was poor. Therefore, we applied a negative binomial regression model to calculate the incidence rate ratios (IRR) and calculate 95% confidence intervals (CI) (^{Glynn and Buring, 1996}). We included all factors significantly associated with incidence in the univariate analysis in a multivariate negative binomial regression model. This model also included calendar year. The multivariate negative binomial regression model estimated the IRR and 95% CI for each factor after adjusting for all other significant predictors of tumor incidence.

To quantify the risk of a first tumor with the passage of time since first exposure to PUVA, we utilized proportional hazards models to estimate the hazard ratios (HR) and 95% CI associated with risk of first tumor for levels of PUVA exposure, age, and gender. In the multivariate analysis, the HR and 95% CI for PUVA dose was adjusted for age and gender, both significant predictors of the risk of a first tumor in the univariate analysis. Kaplan–Meier and predicted survival functions were generated using standard techniques.

All these statistical models incorporated time-dependent variables. All statistical tests were two sided and all except Chi-square tests were performed with Stata 7.0 (Stata Corp., College Station, Texas). Crunch version 4 (Crunch Software Corp., California) was used to calculate the chi-square tests.

Results

We have prospectively followed the 1380 members since 1975/1976 with an average of more than 20 y per person. At enrollment 64% were male and the mean age was 46. The status of cohort patients at representative times during the study are presented in Table I. From first treatment to most recent follow-up this cohort has been observed for 27,840 person-years of which 19,138 (68.7%) person-years were years with less than 10 PUVA treatments reported (nonexposed years). For 16,318 (85.3%) of these 19,138 person-years, no use of PUVA was recorded. The median number of years of follow up after calendar year with PUVA exposure reported is 10. At time of enrollment, 127 patients of the inception cohort (9.2%) reported high-dose exposure to methotrexate and 250 patients high exposure to tar and/or UVB (18.1%). In Table II, we present the number of person-years of exposure to each level of each therapy. In addition, we present the distribution of patient's levels of exposure at their last follow up.

Table I - Status of 1380 patients originally enrolled in the PUVA follow-up status by the end of the year.

Full table

Table II - The number of person-years of exposure to each level for each variable and the distribution of patient's characteristics at final follow up in the PUVA Follow-Up Study.

Full table

Tumor risk

SCC

Between 1975 and 2001, a total of 2147 invasive SCC were documented in 303 persons (21.9% of the inception cohort) with a median of 3.0 SCC per person with one or more SCC (25th percentile=1.0 and 75th percentile=7.0). The average incidence of SCC is 7700 per 100,000 person-years. The incidence of SCC has increased dramatically over the 25 y of the study (Figure 1) and now is nearly 20,000 per 100,000 person-years.

Figure 1.

The age-adjusted incidence rate of SCC and BCC by calendar year in the PUVA cohort adjusted for age based on a negative binominal regression model.

Full figure and legend (13K)

In the univariate analysis, the incidence of SCC is 30 times higher in the highest PUVA dose group compared with patients with low-dose exposure (Table III). Other factors were much more modestly associated with SCC risk. The results of the multivariate analysis, which adjust each risk estimate for all significant predictors of SCC risk, are presented in Table IV. Level of PUVA exposure is the single most strongly factor associated with SCC risk (Table IV, Figure 2). Patients with high PUVA exposure (>500 treatments) were 19 times more likely to develop SCC compared with patients with low-dose exposure. Even 15 y after stopping PUVA, the risk of SCC was not lower than the risk observed while still utilizing PUVA (IRR=0.93; 95% CI=0.58–1.52). Early intense therapy with PUVA (defined as 200 PUVA treatments in the first 5 y of study) was not significantly related to SCC risk. For most other significant predictors of risk for SCC, including methotrexate, X-ray, UVB, and tar, the risk estimates from the multivariate analysis were somewhat lower than the univariate estimates (Table III and Table IV).

Figure 2.

The multivariate estimates (Table III) of SCC and BCC risk for PUVA dose with one-sided CI in PUVA follow-up Study.

Full figure and legend (12K)

Table III - Univariate analysis of potential risk factors associated with development of SCC and BCC in patients in the PUVA Follow-up Study (n=1380).

Full table

Table IV - Multivariate estimates of IRR and 95% CI for SCC and BCC adjusted for all other significant risk factors and for study year in members of the PUVA Follow-up Study (n=1380).

Full table

BCC

Between 1975 and 2001, we documented 1363 BCC in 294 persons (21.3% of the inception cohort) with a median of 2.0 SCC per person with one or more BCC (25th percentile=1.0 and 75th percentile=5.0). The average BCC incidence is 4400 per 100,000 person-years, but has increased substantially over the last 10 y of this study and now is more than 12,500 per 100,000 person-years (Figure 1).

The results of the univariate analysis are summarized in Table III. BCC risk is significantly associated with increasing PUVA use, the years since first PUVA treatment and age. In the multivariate analysis, which adjusted for all significant predictors in the univariate analysis, BCC risk increased greatly with very high levels of PUVA exposure (500 PUVA treatments; IRR=12.69, 95% CI=6.34–25.40) (Figure 2). For most significant predictors of risk for BCC, including years since stopping PUVA, there was little difference in the estimates obtained in the univariate and multivariate analysis (Table III and Table IV). The risk of BCC increases significantly with the passage of time since stopping PUVA and is about three times higher 10 y after stopping PUVA than during treatment (Table IV).

Risk of first tumor

SCC

Total PUVA exposure was significantly associated with an individual's risk of a first SCC (Table V and Table VI). Males had a modestly, but significantly higher risk of developing a first SCC. Surprisingly, the risk of a first SCC and age were not significantly associated.

Table V - Univariate estimates of HR and 95% CI for a first SCC (n=303) and a first BCC (n=294) in members of the PUVA Follow-up Study.

Full table

Table VI - Multivariate estimates of HR and 95% CI for a first SCC (n=303) and a first BCC (n=294) in members of the PUVA Follow-up Study.

Full table

In 25 y, patients with fewer than 200 PUVA treatments have about 7% risk at least one SCC (Figure 3). After 25 y, more than half of the patients with 400 or more treatments develop at least one SCC.

Figure 3.

The risk of a first SCC over time according to PUVA dose.

Full figure and legend (11K)

BCC

More than 100 PUVA treatments was significantly associated with an individual's increased risk of a first BCC (Table V and Table VI). Risk of developing at least one BCC was higher in males. Surprisingly, in the multivariate analysis the risk of a first BCC was lower among patients older than 44 than younger patients (Table VI). Risk of developing at least one BCC did not increase as sharply with increasing PUVA dose as did the risk of SCC. After 25 y, almost one-third of the patients with 200 treatments develop at least one BCC (Figure 4).

Figure 4.

The risk of a first BCC over time according to PUVA dose.

Full figure and legend (10K)

Discussion

From 1975 to 2001, our cohort has developed of 1380 PUVA-treated patients. The study continues. This experience enables us to determine the magnitude and persistence of carcinogenic effects of PUVA. Our cohort has developed more than 2100 invasive SCC, 1400 BCC as well as over 800 keratoacanthoma and SCC in situ (Bowen's disease).

We have demonstrated that the incidence of both SCC and BCC continues to increase dramatically over time. In contrast to the general population (^{Karagas et al, 1999;Harris et al, 2001}) the SCC incidence is higher than the BCC incidence. Risk of both SCC and BCC are now extremely high for patients with very high levels of PUVA exposure. At least for 15 y, stopping PUVA in psoriasis patients does not reduce the risk of developing a SCC. Among patients with 200 or more PUVA treatments, about half will develop at least one SCC and almost a third at least one BCC in 25 y after achieving this dose level.

Tumor risk

SCC

SCC incidence has increased dramatically and is now about 20,000 SCC per 100,000 person-years. This rate is almost 25 times the SCC risk of a comparably aged population in Arizona (^{Harris et al, 2001}). The SCC risk of patients with high-dose PUVA (400 treatments), however, is 250-fold the risk of the comparably aged Arizona population. Patients with more than 200 treatments appear to be at higher risk of developing skin cancer than those exposed to other known carcinogenic treatments such as systemic immunosuppression in organ transplant recipients and therapeutic radiation therapy (^{Bouwes-Bavinck et al, 1991;London et al, 1995;Karagas et al, 1996;Caforio et al, 2000;Shore et al, 2002}).

Our earlier data demonstrated a persistence of risk over shorter periods after stopping PUVA (^{Stern et al, 1998}) and that discontinuation of PUVA treatment after a first SCC was not significantly related to the risk of a first tumor (^{Katz et al, 2002}). This analysis demonstrates that even after 15 y after stopping therapy the risk of SCC is not significantly decreased compared with the years when PUVA was utilized. We detected no association between the intensity of PUVA treatment in the first 5 y of treatment and tumor risk. Although many factors not considered in this analysis could explain the differences between European and American results, this finding argues against the hypothesis that the difference in tumor incidence between European and American studies was due to different PUVA regimens (^{Henseler et al, 1987}). The stabilization of SCC incidence after about 20 y of follow-up may reflect the death and/or loss of follow-up of highly susceptible patients or, at last, a stabilization of the SCC risk associated with PUVA therapy with the additional passage of time.

BCC

In the last decade BCC incidence has increased substantially in our cohort (Figure 1). Compared with the risk in Arizona the BCC risk in our population is about a 4-fold higher during the last 5 y (^{Harris et al, 2001}). Patients with 500 treatments, however, have almost a 50-fold increase of BCC risk compared with Arizonans of comparable age. Low-dose groups risks of BCC are only modestly elevated compared with Arizonans. Previously, we, and others, have reported a modest elevation of BCC risk in PUVA-treated patients (^{Stern and Lange, 1988};Forman et al, 1989;^{Bruynzeel et al, 1991};^{Stern and Laird, 1994;Stern et al, 1998;Lindelof et al, 1999}). We now observe an increase in BCC risk with increasing PUVA exposure that we had previously only observed for SCC suggests the effect of high-dose PUVA on BCC risk is manifest only after a long latency period. Incidences of both tumor types increased significantly with the passage of time. In the multivariate analysis, however, BCC risk increased significantly after stopping PUVA. We do not have an explanation for this finding.

Risk of first tumor

SCC

The morbidity of each tumor whether a first or subsequent one is probably comparable. Therefore, our past analyses have emphasized absolute risk. The odds of a PUVA exposed patient remaining tumor free over time, however, is also of clinical interest. The risk of ever developing a first SCC is much higher among patients who have used at least 200 treatments (Figure 3).^{Grodstein et al (1995)} estimated that 0.2% of middle-aged white females develop a first SCC over 8 y. In a comparable period, we estimated about 50 times more of our patients with 200 PUVA treatments will develop a first SCC over this time. The risks of cutaneous SCC for patients immunosuppressed for 20 y for solid organ transplantation (^{London et al, 1995}) and PUVA patients with at least 200 treatments appear to be roughly comparable.

BCC

Patients with 100 PUVA treatments have a significantly increased risk of developing at least one BCC (Figure 4). About 13% of male caucasian US health professionals of 40 y to 75 y self-reported a first BCC during 8 y of follow-up (^{Van Dam et al, 1999}). In a comparable period, our patients were about as likely to develop a first BCC after adjusting for age and gender. In up to 50 y of follow-up after ionizing radiation therapy almost every patient with radiodermatitis of the face developed a BCC (^{Davis et al, 1988}) and about 40% of those treated on the scalp (^{Shore et al, 2002}). When adjusted for length of follow-up the risks of ionizing radiation and those of substantial PUVA exposure appear to be highly comparable.

Translation to basic research

The carcinogenic effects of PUVA increase over time, are independent of the intensity of the initial therapeutic regimen and persist for many years after stopping treatment. This suggests that PUVA mutagenic effects are likely to be important in skin carcinogenesis (^{Nataraj et al, 1997;Kreimer-Erlacher et al, 2001;Seidl et al, 2001}); however, the PUVA-induced mutation spectrum suggests that factor other than DNA photoadduct formation may be responsible of inducing skin cancer (^{Peritz and Gasparro, 1999}). The persistent tumor risk after stopping PUVA and an increasing absolute risk over time suggest that accurate documentation of carcinogenic effects of photosensitizers both in human and experimental systems require substantial postexposure observation. The carcinogenic risks of ionizing radiation, UVB therapy, tar, and methotrexate in our cohort are modest and remarkably consistent over time (^{Stern and Lange, 1988;Stern and Laird, 1994;Stern et al, 1998}).

Our findings are most consistent with additive effects of various carcinogenic exposures and argue strongly against PUVA acting primarily to enhance tumor progression by other epigenetic effects (^{Granstein et al, 1992;Gasparro et al, 1998;Lindelof et al, 1999}). Many immunosuppressive agents, except methotrexate, substantially increase tumor risk in UV-exposed mice (^{Daynes et al, 1979;Kelly et al, 1987;Servilla et al, 1987}). This finding in mice and our findings are consistent (^{Stern and Lange, 1988;Stern et al, 1998;Lindelof et al, 1999;Marcil and Stern, 2001}). Methotrexate appears to be at most a modest risk factor for developing NMSC in contrast to the substantial enhancing effects of cyclosporine on tumor risk (^{Marcil and Stern, 2001}) in PUVA-treated patients. Different immunosuppressive regimens in solid organ transplant recipients also have a different magnitude of carcinogenic effect (^{Jensen et al, 1999;Caforio et al, 2000}). The mechanism and magnitude of interaction between ionizing radiation and UV is not well understood (^{Mohankumar et al, 2000;Shore et al, 2002}). The carcinogenic effect of UVB has been well established, but it does not increase tumor risk in mice when simultaneously administered with PUVA compared with PUVA alone (^{Granstein et al, 1992}). High levels of UVB exposure is a modest, but significant, risk factor for SCC and BCC among PUVA-treated patients.

Methodologic aspects

Among the strengths of this study are its prospective nature, duration, completeness of follow-up, and pathologic confirmation of endpoints. The large number of tumors documented and the length of the follow-up increases our statistical power and precision. Our analyses used multivariate models to estimate the risks of developing a tumor or a first tumor which adjust for other significantly associated risk factors in the univariate analysis; however, not all risk factors such as exposure to sunlight were accounted for in this model.

Both observational and selection bias may have affected the findings of our tumor risk analysis because patients with high doses of PUVA and those with a prior tumor were more likely to be closely followed and/or to discontinue PUVA treatment. Moreover, only 54 of the 303 patients who reported at least one SCC developed a first SCC within 2 y prior to stopping PUVA (data not shown). In the risk-of-first-tumor analysis we adjusted for age and gender. To correct for possible misclassification, we only considered years with 10 or more treatments and the first year after PUVA exposure as years of exposure. We may therefore have underestimated the tumor incidence in the period immediately after stopping PUVA. Moreover, the results from a preliminary analysis that considered years with one to 10 treatments to be years with PUVA exposure did not differ substantially from the results reported here (data not shown).

Conclusions

After more than 20 y, substantial exposure to PUVA dramatically increases the risk of NMSC. Although alternative treatments such as narrow-band UVB are now available prior exposure to PUVA remains an important issue in the management of patients. The risk of SCC associated with PUVA appears to persist for more than 15 y after stopping treatment and the risk of BCC continues to increase after stopping PUVA. Past use of PUVA may effect the risk of subsequent therapies, particularly newer immunosuppressive and immunomodulatory agents that could potentially unmask and/or induce skin carcinogenesis in these high-risk patients (^{Marcil and Stern, 2001}). We recommend that PUVA-treated patients be rigorously followed on a long-term basis.

Physicians and patients should carefully balance the risks of PUVA with the benefits it affords for the management of severe psoriasis and other skin diseases, and consider past use of PUVA in determining the safety of alternative therapies for an individual patient.

References

1. Bavinck, JNM, Vermeer, BJ, van der Woude, FJ, et al: Relation between skin cancer and HLA antigens in renal transplant recipients. N Engl J Med 1991 0: 843–848,
2. Bruynzeel, I, Bergman, W, Hartevelt, HM, et al: "High single-dose" European PUVA regimen also causes an excess of non-melanoma skin cancer. Br J Dermatol 1991 124: 49–55,  | Article | PubMed | ISI | ChemPort |
3. Caforio, ALP, Fortina, AB, Piaserico, S, et al: Skin cancer in heart transplant recipients: Risk factor analysis and relevance of immunosuppressive therapy. Circulation 2000 102 (Suppl. III) III-222–III-227,
4. Chuang, TY, Heinrich, LA, Schultz, MD, Reizner, GT, Kumm, RC, Cripps, DJ: PUVA and skin cancer. A historical cohort study on 492 patients. J Am Acad Dermatol 1992 26: 173–177,  | PubMed | ISI | ChemPort |
5. Davis, MM, Hanke, W, Zollinger, TW, Montebello, JF, Hornback, NB, Norins, AL: Skin cancer in patients with chronic radiation dermatitis. J Am Acad Dermatol 1988 20: 608–616,  | ISI |
6. Daynes, RA, Harris, CC, Conner, RJ, Eichwald, EJ: Skin cancer development in mice exposed chronically to immunosuppressive agents. J Natl Cancer Inst 1979 62: 1075–1081,  | PubMed | ChemPort |
7. Dunnick, JK, Forbes, PD, Eustis, SL, Hardisty, JF, Goodman, DG: Tumors of the skin in the HRA/Skh mouse after treatment with 8-methoxypsoralen and UVA radiation. Fundam Appl Toxicol 1991 16: 92–102,  | Article | PubMed | ISI | ChemPort |
8. Forman, AB, Roenigk, HH Jr, Caro, WA, Magid, ML: Long-term follow-up of skin cancer in the PUVA-48 cooperative study. Arch Dermatol 1987 16: 108–116,
9. Gasparro, FP, Liao, B, Foley, PJ, Wang, XM, McNiff, JM: Psoralen photochemotherapy, clinical efficacy, and photomutagenicity: The role of molecular epidemiology in minimizing risk. Environ Mol Mutagen 1998 31: 105–112,  | Article | PubMed | ISI | ChemPort |
10. Glynn, RJ, Buring, JE: Ways of measuring rates of recurrent events. Br Med J 1996 312: 364–367,  | ISI | ChemPort |
11. Granstein, RD, Morison, WL, Kripke, ML: Carcinogenicity of combined ultraviolet B radiation and psoralen plus ultraviolet A irradiation treatment of mice. Photodermatol Photoimmunol Photomed 1992 9: 198–202,  | PubMed | ISI | ChemPort |
12. Grodstein, F, Speizer, FE, Hunter, DJ: A prospective study of incident squamous cell carcinoma of the skin in the nurses' health study. J Natl Cancer Inst 1995 87: 1061–1066,  | Article | PubMed | ChemPort |
13. Hannuksela-Svahn, A, Pukkala, E, Laara, E, Poikolainen, K, Karvonen, J: Psoriasis, its treatment, and cancer in a cohort of Finnish patients. J Invest Dermatol 2000 114: 587–590,  | Article | PubMed | ISI | ChemPort |
14. Harris, RB, Griffith, K, Moon, TE: Trends in the incidence of nonmelanoma skin cancers in southeastern Arizona, 1985–1996. J Am Acad Dermatol 2001 45: 528–536,  | Article | PubMed | ISI | ChemPort |
15. Henseler, T, Christophers, E, Honigsmann, H, Wolff, K: Skin tumors in the European PUVA Study. Eight-year follow-up of 1643 patients treated with PUVA for psoriasis. J Am Acad Dermatol 1987 16: 108–116,  | PubMed | ISI | ChemPort |
16. Jensen, P, Hansen, S, Moller, B, et al: Skin cancer in kidney and heart transplant recipients and different long-term immunosuppressive therapy regimens. J Am Acad Dermatol 1999 40: 177–186,  | Article | PubMed | ISI | ChemPort |
17. Karagas, MR, Greenberg, ER, Spencer, SK, Stukel, TA, Mott, LA: Increase in incidence rates of basal cell and squamous cell skin cancer in New Hampshire, USA: New Hampshire Skin Cancer Study Group. Int J Cancer 1999 81: 555–559,  | Article | PubMed | ISI | ChemPort |
18. Karagas, MR, McDonald, JA, Greenberg, ER, Stukel, TA, Weiss, JE, Baron, JA, Stevens, MM. Risk of basal cell and squamous cell skin cancers after ionizing radiation therapy. For The Skin Cancer Prevention Study Group. J Natl Cancer Inst 1996 88: 1848–1853,  | Article | PubMed | ChemPort |
19. Katz, KA, Marcil, I, Stern, RS: Incidence and risk factors associated with a second SCC or BCC in PUVA-treated psoriasis patients. J Invest Dermatol 2002 118: 1038–1043,  | Article | PubMed | ISI | ChemPort |
20. Kelly, GE, Meikle, W, Ross Sheil, AG: Effects of immunosuppressive therapy on the induction of skin tumors by ultraviolet irradiation in hairless mice. Transplantation 1987 44: 429–434,  | PubMed | ISI | ChemPort |
21. Kreimer-Erlacher, H, Seidl, H, Back, B, Kerl, H, Wolf, P: High mutation frequency at Ha-ras exons 1–4 in squamous cell carcinomas from PUVA-treated psoriasis patients. Photochem Photobiol 2001 74: 323–330,  | Article | PubMed | ISI | ChemPort |
22. Lindelof, B, Sigugeirsson, E, Tegner, E, et al: PUVA and cancer: A large scale epidemiological study. Lancet 1991 338: 91–93,  | Article | PubMed | ISI | ChemPort |
23. Lindelof, B, Sigurgeirson, B, Tegner, E, et al: PUVA and cancer risk: The Swedish follow-up study. Br J Dermatol 1999 141: 108–112,  | Article | PubMed | ISI | ChemPort |
24. London, NJ, Farmery, SM, Will, EJ, Davison, AM, Lodge, JPA: Risk of neoplasia in renal transplant patients. Lancet 1995 346: 403–406,  | Article | PubMed | ISI | ChemPort |
25. Maier, H, Schemper, M, Ortel, B, Binder, M, Honingsmann, H: Skin tumors in photochemotherapy for psoriasis. A single-center follow-up of 496 patients. Dermatology 1996 193: 185–191,  | PubMed | ISI | ChemPort |
26. Marcil, I, Stern, RS: Squamous-cell cancer of the skin in patients given PUVA and cyclosporin: Nested cohort crossover study. Lancet 2001 358: 1042–1045,  | Article | PubMed | ISI | ChemPort |
27. Mohankumar, MN, Venkatachalam, P, Prabhu, BK, Jeevanram, RK. Comparison of UV-induced unscheduled DNA synthesis in lymphocytes exposed to low doses of ionising radiation in vivo and in vitro. Mutat Res 2000 447: 199–207,  | PubMed | ISI | ChemPort |
28. Nataraj, AJ, Wolf, P, Cerroni, L, Ananthaswamy, HN. p53 mutation in squamous cell carcinomas from psoriasis patients treated with psoralen + UVA (PUVA). J Invest Dermatol 1997 109: 238–243,  | Article | PubMed | ISI | ChemPort |
29. Parrish, JA, Fitzpatrick, TB, Tanenbaum, L, Pathak, MA. Photochemotherapy of psoriasis with oral methoxsalen and longwave ultraviolet light. N Engl J Med 1974 29: 1207–1211,
30. Peritz, AE, Gasparro, FP: Psoriasis, PUVA, and skin cancer-molecular epidemiology: The curious question of TA transversions. J Invest Dermatol Symp Proc 1999 4: 11–16,  | ISI | ChemPort |
31. Seidl, H, Kreimer-Erlacher, H, Back, B, Soyer, HP, Hofler, G, Kerl, H, Wolf, P: Ultraviolet exposure as main initiator of p53 mutations in basal cell carcinomas from psoralen and ultraviolet A-treated patients with psoriasis. J Invest Dermatol 2001 117: 365–370,  | Article | PubMed | ISI | ChemPort |
32. Servilla, KS, Burnham, DK, Daynes, RA: Ability of cyclosporine to promote the growth of transplanted ultraviolet radiation-induced tumors in mice. Transplantation 1987 44: 291–295,  | PubMed | ISI | ChemPort |
33. Shore, RE, Moseson, M, Xue, X, Tse, Y, Harley, N, Pasternack, BS. Skin cancer after X-ray treatment for scalp ringworm. Radiat Res 2002 157: 410–418,  | Article | PubMed | ISI | ChemPort |
34. Stern, RS, Lange, R: Non-melanoma skin cancer occurring in patients treated with PUVA five to ten years after first treatment. J Invest Dermatol 1988 91: 120–124,  | Article | PubMed | ISI | ChemPort |
35. Stern, RS, Laird, N: The carcinogenic risk of treatments for severe psoriasis. Cancer 1994 73: 2759–2764,  | Article | PubMed | ISI | ChemPort |
36. Stern, RS, Thibodeau, LA, Kleinermann, RA, Parrish, JA, Fitzpatrick, TB: Risk of cutaneous carcinoma in patients treated with oral methoxsalen photochemotherapy for psoriasis. N Engl J Med 1979 300: 809–813,  | PubMed | ISI | ChemPort |
37. Stern, RS, Laird, N, Melski, J, Parrish, JA, Fitzpatrick, TB, Bleich, HL: Cutaneous squamous-cell carcinoma in patients treated with PUVA. N Engl J Med 1984 310: 1156–1161,  | PubMed | ISI | ChemPort |
38. Stern, RS, Liebman, EJ, Vakeva, L: Oral psoralen and ultraviolet-A light (PUVA) treatment of psoriasis and persistent risk of non-melanoma skin cancer. PUVA Follow-up Study. J Natl Cancer Inst 1998 90: 1278–1284,  | Article | PubMed | ChemPort |
39. Van Dam, RM, Huang, Z, Rimm, EB, et al: Risk factors for basal cell carcinoma of the skin in men: Results from the health professionals follow-up study. Am J Epidemiol 1999 150: 459–468,  | PubMed | ISI | ChemPort |

Acknowledgments

Supported by Public Health Service grant N01-AR-0-2246 from the National Health Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Department of Health and Human Services. T.N. was also supported by a grant from the Fund for Scientific Research-Flanders, Belgium (F.W.O.-Vlaanderen)