Asymmetric lateral distribution of melanoma and Merkel cell carcinoma in the United States

Transcription

1 Asymmetric lateral istribution of melanoma an Merkel cell carcinoma in the Unite States KellyG.Paulson,PhD,JayasriG.Iyer,MD,anPaulNghiem,MD,PhD Seattle, Washington Backgroun: A recent report suggeste a tren towar more ultraviolet-linke skin cancers arising on the left rather than the right sie of the boy in the Unite States. Objective: We sought to test whether the reporte inciences of two ultraviolet-linke skin cancers, malignant melanoma (MM) an Merkel cell carcinoma (MCC), are statistically significantly greater on the left than the right in the Unite States. Methos: MMs (n = 82,587) an MCCs (n = 2384) occurring on the left or right sie of the face, arm, or leg that were reporte in the Surveillance, Epiemiology, an En Results registry between 1986 an 2006 were inclue for analysis. Results: MM an MCC were significantly more likely to present on the left than the right (P \.01 for both MM an MCC). In all, 53% of arm melanomas, 51% of facial melanomas, an 52% of leg melanomas presente on the left. A total of 55% of arm MCCs an 52% of facial MCCs presente on the left, whereas leg MCCs were equally istribute. Limitations: National registry ata i not provie information regaring sun exposure or riving habits. No equivalent registry ata were available for basal or squamous cell carcinoma. Conclusions: Both melanoma an MCC are significantly more likely to arise on the left than the right, an this effect was most prominent on the arm. Driver-sie automobile ultraviolet exposure (approximately 5-fol stronger on the left than right arm) is a likely contributing factor. It may be pruent to remin iniviuals prone to skin cancer to take appropriate sun precautions when riving in an automobile. ( J Am Aca Dermatol 2011;65:35-9.) Key wors: asymmetric; automobile; melanoma; Merkel cell carcinoma; presentation; ultraviolet. Ultraviolet (UV) light is a well-characterize risk factor for most skin cancers, incluing melanoma 1 an Merkel cell carcinoma (MCC). 2,3 One source of environmental UV exposure is riving or riing in automobiles. Actinic keratoses Abbreviations use: MCC: Merkel cell carcinoma MM: malignant melanoma SEER: Surveillance, Epiemiology, an En Results UV: ultraviolet From the Division of Dermatology, Department of Meicine, University of Washington. Conflicts of interest: None eclare. Fune by the Merkel Cell Carcinoma Patient Gift Fun at the University of Washington, ACS RSG CCE an K24 CA (Dr Nghiem) an National Institutes of Health F30ES (Dr Paulson). Accepte for publication May 20, Reprint requests: Paul Nghiem, MD, PhD, University of Washington, 815 Mercer St, Brotman Blg, Room 242, Seattle, WA pnghiem@uw.eu. Publishe online April 22, /$36.00 ª 2010 by the American Acaemy of Dermatology, Inc. oi: /j.jaa have been reporte to occur more frequently on the right sie of the boy in Australian rivers, where rivers sit on the right sie of the car. 4 Conversely, in a recent case series from the Unite States, where rivers sit on the left sie of the car, a tren towar left-sie bias in skin cancer was observe. 5 In this series of 890 cases, skin cancers (mostly basal cell carcinomas) were more frequently etecte on the left sie of the boy (53% vs 47%), although the tren i not reach statistical significance. A recent stuy by Moehrle et al 6 carefully examine the pattern of UV exposure receive by a river 35

2 36 Paulson, Iyer, an Nghiem JAM ACAD DERMATOL JULY 2011 sitting on the left sie of an automobile. Moehrle et al 6 reporte that approximately 25% to 31% of ambient UV raiation is transmitte through an open car winow (the remainer is blocke by the vehicle boy). In contrast, a close car winow that only transmits UVA rays allows approximately 3% to 4% of environmental UV raiation to be transmitte. 6 For the river, this raiation more strongly affects the left sie of the boy. In particular, a river sitting on the left next to a close car winow receives many times more UV on the left arm an face than the right, with the highest total ose being on the left arm. In contrast, UV exposure to the leg is relatively evenly istribute, with only slightly more exposure on the right than the left leg. As note above, a nonsignificant tren towar leftsie skin cancer bias has recently been suggeste in the US population. 5 To buil CAPSULE SUMMARY on this stuy, we use national registry ata from the Surveillance, Epiemiology, an En Results (SEER) atabase to test in a large number of cases whether there are significantly more left- than right-sie UVassociate cancers reporte in the Unite States. Specifically, we focuse on two skin cancersemalignant melanoma (MM) an MCCebecause these are capture in the SEER atabase. We hypothesize that both MM an MCC woul isplay a left-sie bias. Furthermore, we hypothesize this bias might be a result of riving an thus particularly affect the face an left arm, but not the leg. METHODS Patient inclusion Information was obtaine from the SEER registry atabase. 7 We consiere all melanoma (International Classification of Diseases for Oncology, Thir Eition coe ) an MCC (International Classification of Diseases for Oncology, Thir Eition coe 8247) cases that met the following criteria: iagnose between 1986 an 2006; lateralize cancer (as oppose to miline) with known sie, known age, known gener. Cancers that presente on one of 3 boy sites were consiere for analysis: skin of face an scalp (C44.3, C44.4), upper limb (C44.6), an lower limb (C44.7). A total of 82,587 melanoma an 2384 MCC cases were inclue in the analysis. More ultraviolet-associate skin cancers (melanomas an Merkel cell carcinomas) occur on the left sie of the boy than the right in the Unite States. The left arm is most prominently affecte by this asymmetry for both cancer types. It may be appropriate to remin iniviuals prone to skin cancer that ultraviolet exposure from automobile riving is a common an establishe source of ultraviolet that preferentially affects the left arm. Data extraction Data were extracte using SEERstat software (National Cancer Institute, Bethesa, MD). Statistical analysis Melanoma an MCC were analyze separately. Statistical analysis was performe with GraphPa Prism software (GraphPa Software, La Jolla, CA). The x 2 test was use to test whether the observe leftright istribution iffere significantly from 50% on the left an 50% on the right. RESULTS Malignant melanoma Among 82,587 melanomas analyze, 52.3% presente on the left sie an 47.7% on the right. This left bias was statistically significant (P \.001). Appreciably more tumors were observe on the left sie of the face an the left arm as compare with the right (Fig 1). In all, 51% of face an scalp melanomas (7787 of 15,185, P =.002) an 53% of arm melanomas (19,791 of 37,588, P \.001) presente on the left sie. However, melanomas were also significantly more likely to present on the left leg than the right. A total of 52% of leg tumors (15,627 of 29,814, P\.001) presente on the left sie. Data were split out by age an sex to etermine whether left-right biases were more pronounce in any particular group. Women oler than 65 years trene towar less left-sie bias in the hea an neck as compare with oler men, however, this ifference between men an women was not significant (P =.56) (Table I). Merkel cell carcinoma A total of 2384 MCCs met inclusion criteria an were analyze. Of these, 1256 (52.7%) presente on the left an 1128 (47.3%) on the right (P =.046) (Table II). Face, scalp, an upper limb cases were more likely to present on the left than the right, but cases on the leg were evenly istribute. Left sie presentation occurre in 465 of 845 cases on the upper limb (55%, P =.004), 503 of 968 cases on the face an scalp (52%, P =.22), an 288 of 571 cases on the lower limb (50%, P =.83) (Fig 2). DISCUSSION It has previously been suggeste by Butler an Fosko 5 that skin cancers in the Unite States,

3 JAM ACAD DERMATOL VOLUME 65, NUMBER 1 Paulson, Iyer, an Nghiem 37 Fig 1. Fraction of non-miline melanomas that present on each sie of boy. Percentage of lateralize melanomas presenting on left an right sies are shown for face/scalp, arm, an leg. Statistical significance, **P less than.01. Shae areas receive relatively more ultraviolet (UV). Boy site shaing reflects fraction of total ambient UV that reaches river seate on left with winows an sunroof close, as reporte in Moehrle et al. 6 Open winow increases UV ose by more than 5-fol with continue preferential exposure to left arm. Table I. Percent of melanomas that present on left, broken own by age, sex, an boy site MM: lateralize tumors presenting on left sie Scalp an face Upper limb Lower limb Total Men age [ 65 y 52%* (3080/5920) 53%* (4511/8514) 53%* (1339/2536) 53%* (8930/16,970) Women age [ 65 y 51% (1786/3474) 51% y (3224/6262) 51% y (3065/5972) 51%* (8075/15,708) Men age \ 65 y 50% (1826/3618) 53%* (5855/10,960) 53%* (3012/5703) 53%* (10,693/20,281) Women age \ 65 y 50% (1095/2173) 52%* (6201/11,852) 53%* (8211/15,603) 52%* (15,507/29,628) Total 51%* (7787/15,185) 53%* (19,791/37,588) 52%* (15,627/29,814) 52%* (43,205/82,587) MM, Malignant melanoma. Statistical significance: *P \.01; y P \.05. particularly in men, are more likely to present on the left sie of the boy. Importantly, this contrasts with Australia, where a similarly size stuy fins that actinic keratoses in men are significantly more likely to present on the right sie of the boy. 4 Both sets of authors suggest automobile UV exposure as a possible explanation, an the observe reversal in the asymmetrically involve sie is consistent because rivers in the Unite States sit on the left sie of the car whereas rivers in Australia sit on the right.

4 38 Paulson, Iyer, an Nghiem JAM ACAD DERMATOL JULY 2011 Table II. Percent of Merkel cell carcinomas that present on left, broken own by age, sex, an boy site MCC: lateralize tumors presenting on left sie Scalp an face Upper limb Lower limb Total Men age [ 65 y 54% (266/493) 55%* (236/430) 52% (98/187) 54% y (600/1110) Women age [ 65 y 49% (175/357) 58%* (133/230) 54% (131/244) 53% (439/831) Men age \ 65 y 53% (31/59) 55% (58/106) 43% (43/100) 50% (132/265) Women age \ 65 y 53% (31/59) 48% (38/79) 40% (16/40) 48% (85/178) Total 52% (503/968) 55% y (465/845) 50% (288/571) 53% y (1256/2384) MCC, Merkel cell carcinoma. Statistical significance: *P \.05; y P \.01. Fig 2. Fraction of non-miline Merkel cell carcinomas (MCC) that present on each sie of boy. Percentage of lateralize MCC presenting on left an right sies are shown for face/scalp, arm, an leg. Statistical significance, **P less than.01. Shae areas receive relatively more ultraviolet (UV). Boy site shaing reflects fraction of total ambient UV that reaches river seate on left with winows an sunroof close, as reporte in Moehrle et al. 6 However, Butler an Fosko 5 only looke at a limite number of cases (890), an the observe left-sie tren in the Unite States was nonsignificant except in a few subgroups. To test for presence of left-sie bias in the Unite States, we analyze the laterality of skin cancers reporte in the SEER national registry atabase. MCC an MM were the focus, because these aggressive UV-linke cancers are capture by the SEER program. Inee, both of these cancers showe a highly significant left-sie bias in their primary location that was particularly pronounce on the arms.

5 JAM ACAD DERMATOL VOLUME 65, NUMBER 1 Paulson, Iyer, an Nghiem 39 The left-sie bias of melanoma can be partially but not perfectly explaine by UV exposure while riving. The observe asymmetry for melanoma was most prominent for the left arm, which receives the largest UV exposure from riving an significantly more UV than the right arm (approximately 5-fol ifference with winows close). 6 Significant leftsie asymmetry was also etecte for leg melanomas. For the leg, UV exposure in the automobile is more evenly istribute, with only an approximate 1.5-fol ifference between sies an the expecte UV exposure being greater on the right. Finings on the leg may be a result of alternate sources of asymmetric UV, however, population-wie asymmetric UV exposures that coul account for this observe bias are elusive an no aitional potential exposures were gleane from a PubMe search of asymmetrical ultraviolet. Regarless of cause, this left-sie association was statistically significant at all stuie sites (arm, leg, an scalp/face) an is equivalent to approximately one in 20 melanomas that arise in excess on the left in the Unite States. MCC is an increasingly common skin cancer that has been strongly linke to environmental UV. 2,3,8,9 We observe significantly more MCCs presenting on the left than the right. In particular, the arm emonstrate the most left-sie bias, with 1.2 MCCs occurring on the left for each on the right. This observe pattern is consistent with excess left-sie UV exposure from automobile riving. 6,10 Therefore, automobile UV exposure may represent an appreciable risk factor for MCC. However, we cannot rule out other possible explanations. Our stuy has several limitations. SEER oes not provie information on whether the skin was sun expose or not. Furthermore, we o not have information on iniviual riving or automobile riing habits, an cannot test the association of left bias with time in the river s seat. To etermine the importance of automobile UV exposure in this asymmetric istribution, further stuy is inicate in countries with opposite riving patterns an/or of sets of ata annotate with riving habit information. Despite these limitations, these finings represent the largest stuy of asymmetry in skin cancer in the Unite States to ate. In summary, both MCC an MM are significantly more likely to present on the left sie than the right in the Unite States. Automobile UV exposure is a wellcharacterize source of asymmetric UV exposure an likely contributes to the observe excess of melanoma an MCC presenting on the left arm. These excess left-sie cancers represent one of every 20 upper limb melanomas an a striking one of every 10 upper limb MCCs. Therefore, it may be pruent to recommen sunscreen application or other UV protective measures in iniviuals prone to skin cancer who spen significant time riving automobiles or trucks. REFERENCES 1. Veiero MB, Aami HO, Lun E, Armstrong BK, Weierpass E. Sun an solarium exposure an melanoma risk: effects of age, pigmentary characteristics, an nevi. Cancer Epiemiol Biomarkers Prev 2010;19: Heath M, Jaimes N, Lemos B, Mostaghimi A, Wang LC, Penas PF, et al. Clinical characteristics of Merkel cell carcinoma at iagnosis in 195 patients: the AEIOU features. J Am Aca Dermatol 2008;58: Popp S, Waltering S, Herbst C, Moll I, Boukamp P. UV-B-type mutations an chromosomal imbalances inicate common pathways for the evelopment of Merkel an skin squamous cell carcinomas. Int J Cancer 2002;99: Foley P, Lanzer D, Marks R. Are solar keratoses more common on the river s sie? Br Me J (Clin Res E) 1986;293: Butler ST, Fosko SW. Increase prevalence of left-sie skin cancers. J Am Aca Dermatol oi: /j.jaa Publishe online March 12, Moehrle M, Soballa M, Korn M. UV exposure in cars. Photoermatol Photoimmunol Photome 2003;19: Surveillance, Epiemiology, an En Results (SEER) Program ( SEER*Stat Database. Release April 2009, base on the November 2008 submission. 8. Agelli M, Clegg LX. Epiemiology of primary Merkel cell carcinoma in the Unite States. J Am Aca Dermatol 2003; 49: Luner EJ, Stern RS. Merkel-cell carcinomas in patients treate with methoxsalen an ultraviolet A raiation. N Engl J Me 1998;339: Singer RS, Hamilton TA, Voorhees JJ, Griffiths CE. Association of asymmetrical facial photoamage with automobile riving. Arch Dermatol 1994;130:121-3.