Int. J. Cancer: 67,24-28 (1996) 0 1996 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de I’Union InternationaleContre le Cancel SITE DISTRIBUTION OF DIFFERENT TYPES OF SKIN CANCER: NEW AETIOLOGICAL CLUES Silvia FRANCESCHI’, Fabio L E V I ~Lalao , ~ , RANDIMBISON’ and Carlo LA vECCHIA3 lServizio di Epidemiologia, Centro di Riferimento Oncologico, Aviano, Italy; 2Registre vaudois des tumeurs, Institut universitaire de mCdecine sociale et priventive, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland; and 31stitutodi Ricerche Farmacologiche “Mario Negri; ” and Istituto di Statistica Medica e Biometria, Universita degli Studi di Milano, Milan, Italy. Since the investigation, at an individual level, of lifetime sun exposure remains difficult, the site distribution of different types of skin cancer can be an important source of aetiological clues. The present report deals with I, 149 cases of cutaneous malignant melanoma (CMM), 7,685 of basal-cell carcinoma (BCC) and 3,049 of squamous-cell carcinoma (SCC) reported between 1976 and 1992 to the Vaud Cancer Registry, in Switzerland. Site- and type-specific age-standardized (on the world population) incidence rates per IOO.000 population and per IOO.000 unit surface were computed, together with relative age-standardized incidence rates (i.e., rates per 100,000 surface unit in each anatomical site relative to rates for the body as a whole). The highest rates per unit surface were seen for both genders in the face, thus indicating some role of cumulative sun exposure in all skin-cancer types. Relative to the incidence in the whole body, the excess on the face was, however, more than 20-fold for BCC and SCC, but only 4-fold for CMM. The relative incidence in males was very much higher for SCC than for CMM and BCC in the neck, ears and scalp, a heavily sun-exposed area in men but not in women. Conversely, a substantial lack of SCC was seen in the trunk. In conclusion, site distribution of different skin-cancer types suggests that short-duration UV-light exposure is sufficient to increase CMM risk substantially, but has little influence on SCC risk. With the increase of exposure, however, SCC rises more steeply than BCC. Age-related behaviour (i.e., another indirect indicator of duration of exposure to UV light) is consistent with the anatomical distribution of skin cancer. o 1996 Wiley-Liss, Inc. Exposure to ultra-violet (UV) radiation plays a major role in the aetiology of various skin-cancer types, i.e., cutaneous malignant melanoma (CMM), basal-cell carcinoma (BCC) and squamous-cell carcinoma (SCC). The incidence of all major types of skin cancer has appreciably increased in most white populations over the last few decades, in parallel with the increase of sun exposure, chiefly for tanning purposes (Parkin et al., 1992; Levi et al., 1995). The doses and patterns of sun exposure most strongly related to the risk of various skincancer types seem, however, to vary substantially. While SCC is believed to rise steadily with continuous intense sun exposure (such as occupational exposure) up to high cumulative dose, the strongest evidence for CMM has been for recreational or intermittent exposure, and history of early sunburn (Armstrong, 1988; IARC? 1992). BCC has been less often studied, but some data indicate that the risk of BCC may not increase beyond a certain level of sun exposure (Kaldor et aL, 1993; Kricker et al., 1995a) and may be maximal, as for CMM, for intermittent exposure (Kricker et al., 19956). Since the investigation, at an individual level, of various aspects of lifetime sun exposure, however, remains difficult, comparison of the site distribution and age pattern of different types of skin cancer can be an important source of aetiological clues, especially if based on a large and unselected populationbased series (Levi et al., 1995). MATERIAL AND METHODS The data included in the present analysis were derived from the Vaud Cancer Registry datafile, which includes information concerning incident cases of malignant neoplasms in the canton (whose population, according to the 1990 census, was about 580,000 inhabitants; Levi et al., 1992). Population estimates were based on decennial censuses (1970, 1980, 1990), and on estimates by the Cantonal Office of Statistics for each calendar year and 5-year age group, based on official numbers of births, deaths, immigrations and emigrations. Notification is based on a voluntary agreement between the recording medical institutions of the canton and the Registry. All hospital, pathology laboratories (including those specializing in dermatopathology) and most practitioners are asked to report all cases of cancer. In this population, specific attention has traditionally been paid to diagnosis, pathological confirmation and registration of skin neoplasms (Levi et al., 1988,1995). Most cases are notified repeatedly and from different institutions to the Registry, thus increasing completeness and accuracy of notification. Information collected by the registry includes general demographic characteristics of the patient (age, gender, municipality of residence), site and histological type according to the standard International Classification of Diseases for Oncology (ICD-0; WHO Health Organization, 1976), and time of diagnostic confirmation. The present series comprises 11,883 incident, histologically verified skin-cancer primaries (6,046 males and 5,837 females) registered from 1976 to 1992. Multiple skin tumours (either synchronous or metachronous) are classified by the site of the first recognized tumour of the same morphological type. For the present report, cases were grouped into the following 3 morphological categories: (1) malignant melanoma (ICD-0 M: 8720-8790, excluding 8742.2, lentigo maligna, but including 8742.3, lentigo maligna melanoma, n = 1,149); (2) basal-cell (ICD-0 M: 809CL8095, n = 7,685), and (3) squamous-cell (ICD-0 M: 8070-8076, n = 3,049) carcinoma. Cancers whose histological type was other or unspecified (n = 378, 3.1%) were not considered, nor were cancers arising from skin of genital organs (e.g., labia majora or minora, vulva, penis or scrotum (ICD-0 T: 184, 187). The following 5 topographical categories were considered: (1) face (ICD-0 T: 173.0-.l, .3); (2) neck, ears and scalp (ICD-0 T: 173.2, .4); (3) trunk (ICD-0 T: 173.5); (4) upper limbs, including shoulders (ICD-0 T: 173.6, and (5) lower limbs (ICD-0 T: 173.7). Skin cancers whose anatomical sub-site was not specified (2%) were not considered. Site- and type-specific, age-standardized (world standard population as reference population) incidence rates per 100,000 population were calculated. Site-specific rates were also computed per 100,000 unit surface, where the unit is the proportion of total skin surface area for each anatomical site (Green jTo whom correspondence and reprint requests should be sent, at Registre vaudois des tumeurs, Institut universitaire de mkdecine sociale et prtventive, Centre Hospitalier Universitaire Vaudois, Falaises l, 1011 Lausanne, Switzerland. Fax: (41) 21 323-03-03. Received: January 6,1996 and in revised form March 11,1996. 25 SITE DISTRIBUTION OF SKIN CANCER TABLE I - NUMBER OF CASES AND AGE-STANDARDIZED (WORLD) INCIDENCE RATES OF DIFFERENT TYPES OF SKIN CANCER PER 100 000 POPULATION AND PER UNIT SURFACE AREA PER 100,000IN VAUD, SWITZERLAND, IN 1976-1992. BY GENDER AND ANATOMICAL SITE Males Females Site Cutaneous malignant meIanoma Face Neck, ears and scalp Upper limbs and shoulders Trunk Lower limbs Whole body* Basal-cell carcinoma Face Neck, ears and scalp Upper limbs and shoulders Trunk Lower limbs Whole body* Squamous-cell carcinoma Face Neck, ears and scalp Upper limbs and shoulders Trunk Lower limbs Whole body* 60 44 0.9 0.6 37.4 9.7 104 26 1.o 0.3 42.5 4.7 0.9 2.1 78 1.3 6.7 124 1.8 9.3 0.7 252 73 507 4.2 1.3 8.3 13.0 3.3 132 256 642 2.2 3.9 9.1 6.7 9.8 1.9 0.3 0.9 2247 517 33.3 7.4 1449.4 112.2 2520 377 28.5 3.9 1237.6 59.0 1.2 1.9 20 1 3.1 16.1 131 1.7 8.9 1.8 798 93 3856 12.0 1.3 57.1 37.4 3.3 639 162 3829 8.4 1.8 44.3 26.3 4.6 1.4 0.7 1.3 852 500 11.6 6.3 504.4 95.0 896 102 8.1 1.0 352.7 15.3 1.4 6.2 191 2.5 13.2 188 1.7 9.1 1.5 93 47 1.3 0.7 4.2 1.7 76 104 1366 0.8 1.0 12.7 2.6 2.6 1.6 0.7 1.8 1683 ~... 22.4 ~~ *Sites "other or unspecified" excluded. et al., 1993). Briefly, for a specific site, rates per unit surface were calculated by dividing the rates for each 5-year age group and each gender group by the proportion of the total surface area of the body occupied by that site (Pearl and Scott, 1986). For example, for trunk, each age-specific rate was divided by 0.29 because the trunk (comprising trunk/back, trunk/front and buttocks) is estimated as 29% of the total body surface area. In addition, age-standardized incidence rates per 100,000 units were also calculated, separately for each gender, topographical category and morphologic group, relative to incidence rate for the body as a whole (set to unity) (Doll and Smith, 1982; Green et al., 1993). RESULTS Table I gives the number of cases, and the corresponding age-standardized (world population) incidence rate per unit surface for each histotype, site and gender. Obviously, for whole body, rates per los and per unit surface per los are identical. CMM rates per 100,000 were highest in the trunk for males (4.2/100,000) and in the lower limbs for females (3.9/ 100,000). However, when expressed per unit surface, the highest rates were in the face for both genders (37.4/100,000 males, 42.5/100,000 females). In females, rates per unit surface in lower and upper limbs were similar (9.8 and 9.3 per unit surface per 100,000 respectively), whereas in males the rates TABLE 11 - IIAI'IOS 01, AGt:-STASDARDIZED (WORLD) IIVC'IDESC'E IIAI'ES OF l)lI.'I:FRFKTTYPF.S O F SKIK ('AKC'EK I'EK UNIT SUKk'ACL: AIIEA AND 100.000 POPUl ATION IN VAI'D. SWIl'ZEKLAND. I N l Y 7 h - l Y Y ? . BY GENDER AND ANATOMICAL SITE Site Face Neck, ears and scalp Upper limbs and shoulders Trunk Lower limbs Whole bodv Males Females CMMISCC BCCiSCC CMMISCC BCCISCC 0.07 0.10 0.51 2.9 1.2 1.2 0.12 0.31 1.0 3.5 3.9 1.0 3.1 1.9 0.37 8.9 1.9 2.6 2.6 3.8 0.73 10.1 1.8 3.5 were higher in the upper limb. With reference to BCC, by far the highest incidence was observed in the face for both genders (1449/100,000 and 1238/100,000 per unit surface per 100,000 in males and females respectively), followed by neck, ears and scalp (112.2 and 59.0 per unit surface per 100,OOO in males and females respectively) and the trunk. Also for SCC the highest incidence site was the face in both genders (504.4 and 352.7, respectively, per unit surface per lOO,OOO), followed in both genders by the neck, ears and scalp (95.0 and 15.3 per unit surface per 100,000 in males and females), and by the upper limbs and shoulders. Site-specific male-to-female ratios were rather consistent in different skin-cancer types except for cancers of the 26 FRANCESCHI ETAL. CMM . MALES w . FEMALES T v Q CMM 5 4.57 4.40 Lower limb Upper limb Neck Trunk Neck F2Ce Trunk Upper Limb BCC FEMALES BCC - MALES Trunk Neck SCC MALES ~ Lower Limb FW3 - 27.56 r T v U p p r Limb Face ~ 22.20 Lower Limb Lower Limb Upper Limb Trunk Neck SCC .FEMALES FUa 27.55 22.20 4.19 5 4 3 2 1 0.07 0 Lower Limb Trunk Upper Limb Neck FKa BODY SITE Lower Limb Trunk Upper Limb Neck FaCe BODY S I E FIGURE 1 - Incidence rates of CMM, BCC and SCC by body site relative to those for the whole body, by gender. Vaud, Switzerland, 1976-1992. Site-specific rates are age-adjusted (world population); those for the whole body are set to 1 for each gender. neck, ears and scalp, where the male-to-female ratio was approximately 2 for CMM and BCC, but over 6 for SCC (Table I). The ratios of age-standardized incidence rates per unit surface between CMM and SCC and between BCC and SCC in each sex are given in Table 11. For the whole body the ratio CMM/SCC was 0.37 for males but 0.73 for females. At specific sites, it ranged between 0.07 in the face and 3.1 in the trunk for males, and between 0.12 in the face and 3.8 in the lower limbs in females. The overall BCC/SCC ratio was 2.6 for males and 3.5 for females. In specific sites, it ranged for males between SITE DISTRIBUTION OF SKIN CANCER 1.2 in the neck, ears and scalp and the upper limbs and shoulders and 8.9 in the trunk, and, for females, between 1.0 in the upper limbs and shoulder and 10.1 in the trunk. Figure 1gives the age-standardized incidence rates per unit surface per 100,000 in specific body sites relative to the whole body for each gender, site and histotype. Relative incidence of BCC and SCC shows very similar patterns, with more than 20-fold excess in the face compared with total body. Relative incidence rates per unit surface below 1 were found for all other sites except neck and ears and scalp. With respect to these latter sites, however, a 4-fold higher incidence of SCC in the male neck was observed. Rates per unit showed a much smaller variation for CMM, with rates in the face about 4-fold higher than in the whole body in both genders. Relative CMM incidence below 1 could be found for lower and upper limbs in males, and neck, ears and scalp and trunk in females, whereas trunk in males showed a 53% excess. Figure 2 illustrates the age- and gender-specific incidence curves for each skin-cancer histological type and gender. CMM increased much less steeply than BCC and SCC with age. Absence of SCC, as compared with BCC, emerged in young and middle-aged adults of both genders. At older age, SCC incidence in males approached that of BCC, and, in contrast to BCC, showed no tendency to plateau. Site-specific age curves were also examined. They were consistent with those of the whole-body curves and are, therefore, not shown. DISCUSSION The present incidence rates are consistent with previous reports (i.e., highest frequency of CMM in the trunk in males and lower limbs in females, and on the face for BCC and SCC in both genders) (Magnus, 1991). Incidence rates per unit surface have rarely been assessed (Pearl and Scott, 1986; 0sterlind et al., 1988; Green et al., 1993), but disclose a somewhat different pattern, with face being the highest incidence area also for CMM, although to a lesser extent than for BCC and SCC. This cannot be attributed to the high frequency on the face of lentigo-maligna lesions (which were excluded) or lentigo-maligna melanomas (which represented only 8% of CMM), and indicates some role of cumulative sun exposure for all 3 major skin-cancer types. In terms of relative incidence per unit surface, as compared with total body incidence (for which, obviously, incidence per lo5 and incidence per unit surface coincide), not only do the 3 major types of skin cancer differ substantially with respect to the face, where BCC and SCC relative excess is about 5-fold BCC - M l0O0T SCC F ~ 100 = CMM - F 10 1 l- 27 greater than that of CMM, but also to the neck, ears and scalp, sites more frequently sun-exposed in males than in females. In males, the relative excess compared with other body sites is similar for CMM and BCC, but 2-to-3 fold greater for SCC. Although relative incidence per unit surface is not a completely new measure, it has been rarely used. Site-specific rates per 100,000 individuals have been generally preferred, leading, however, for instance, to some under-estimation of the incidence of CMM on the face (the surface of which is relatively small) as compared with the lower limbs. Comparison of skin types at specific sites (with SCC as a reference, on account of the better understood UV-lightrelated risk pattern; Council for Scientific Affairs, 1989) showed that CMM/SCC ratios ranged from around 0.1 in the face to around 4 in the limbs in females and 3 in the trunk in males. The comparison between BCC and SCC showed ratios between about 1 (neck, ears, and scalp in males and upper limbs and shoulders in both genders) and 9-10 in the trunk. Intermediate ratios of BCC/SCC were found for the face (Moan and Dahlback, 1992; Kaldor et al., 1993). Anatomic sites can be considered as indirect indicators of different patterns of UV-light exposure, with respect to lifetime duration, intensity and intermittency of exposure. Age is another indicator of duration of exposure. The relatively high incidence rates of CMM in rarely sun-exposed areas and in young individuals suggest that short-duration UV-light exposure is sufficient to increase cancer risk substantially. SCC risk rises more steeply with age than BCC, also in the absence of a major cohort effect (Marks et al., 1993; Levi et al., 1995). Very low SCC incidence rates for short exposures ( e g , individuals below age 30 and sites other than the face and, in males, scalp, ears and neck) also suggest lack of tangible risk increase below a certain threshold and/or latent period. The age-pattern of SCC is consistent with that of most common non-cutaneous cancers. Indeed, SCC may arise in any epithelium (Stoll and Schwartz, 1987). Conversely, BCC virtually never arises from stratified squamous epithelium other than the skin. It derives from the immature basal cells of the epidermis and always contains epidermal cellular and dermal stromal components (Carter, 1987). Very often, BCC starts in the epithelium of pilosebaceous follicles, thus partly explaining the anatomic distribution of BCC (eg., relatively high frequency in the trunk). Unfortunately, the present descriptive data do not allow a distinction to be made between short-duration exposure and intermittent exposure, i.e., the type of exposure putatively more dangerous for CMM (IARC, 1992) and BCC (Kricker et al., 19956). Furthermore, some under-reporting of skin cancer has probably occurred, although this cannot be quantified, and may have been somewhat different for various histological types, sites and age groups. Nevertheless, the present site- and age-specific incidence rates are derived from a population under particularly careful surveillance (Levi et al., 1988, 1995), as confirmed by the comparatively high incidence of all types of skin cancer in this population (Levi et aL, 1992, 1993). Apart from important similarities in risk factors for all skin-cancer types (ix.,skin phenotype, IARC, 1992; immunological deficiencies, Adami et al., 1995; exposure to chemicals, Stoll and Schwartz, 1987; and, possibly, human papillomavirus, Shamanin and de Villiers, 1995), different sun-related risk patterns are suggested by the site distribution, as well as type of increase with age, of CMM, BCC and SCC. -1'7-,*' ,.=..,,,' , ; : ; ; : : : : : : 20- 25- 30- 35- 4 0 - 4 5 - 5 0 - 5 5 - 6 0 - 6 5 - 7 0 - 75- 8 0 AGE GROUP FIGURE 2 - Age-specific incidence curves for CMM, BCC and SCC by gender. Vaud, Switzerland, 1976-1992. ACKNOWLEDGEMENTS The contributions of the Swiss League against Cancer (Grant FOR 368.89.3), and of the Vaud Cancer Registry's staff are gratefully acknowledged. 28 FRANCESCHI E T A L REFERENCES ADAMI,J., FRISCH, M., YUEN,J.. GLIMELIUS, B. and MELBYE,M., Evidence of an association between non-Hodgkin's lymphoma and skin cancer. Brit. med. J., 310,491495 (1995). ARMSTRONG, B.K., Epidemiology of malignant melanoma: intermittent or total accumulated exposure to the sun? J. dermatol. Surg. Oncol., 14,835-849 (1988). CARTER,D.M., Basal cell carcinoma. In: J.D., Jeffers, E. Scott, J. White (eds). Dermatology in generalpractice tatbook and atlas, 3rd ed., pp. 759-766, McGraw-Hill, New York (1987). COUNCIL OF SCIENTIFIC AFFAIRS,Harmful effects of ultraviolet radiation. J. Amer. med. Ass., 262,380-384 (1989). DOLL, R. and SMITH,P.G., Comparison between registries: agestandardized rates. In: J.A.H. Waterhouse, C.S. Muir, K. Shanmugaratnam, J. Powell, D. Peacham, S. Whelan (eds.), Cancer incidence in five continents, Vol. IV, IARC Scientific Publication 42, pp. 671-675. IARC, Lyon (1982). GREEN, A., MACLENNAN, R., YOUL,P. and MARTIN, N.. Site distribution of cutaneous melanoma in Queensland. Int. J. Cancer, 53,232-236 (1993). IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. 55, Solar and ultraviolet radiation. IARC, Lyon (1992). WORLD HEALTHORGANIZATION INTERNATIONAL CLASSIFICATION OF DISEASES FOR ONCOLOGY (ICD-0), 1st ed.. WHO, Geneva (1976). KALDOR,J., SHUGG,D., YOUNG,B., DWYER,T. and WANG,Y-G., Non-melanoma skin cancer: ten-year experience of cancer-registrybased surveillance. In!. J. Cancer, 53,886-891 (1993). KRICKER: A.? ARMSTRONG, B.K., ENGLISH, D.R. and HEENAN, P.J., A dose-response curve for sun exposure and basal-cell carcinoma. Int. J. Cancer, 60,482-488 (1995~). KRICKER; A., ARMSTRONG, B.K., ENGLISH, D.R. and HEENAN,P.J., Does intermittent sun exposure cause basal-cell carcinoma? A casecontrol study in Western Australia. Int. J. Cancer, 60,489-494 (19956). S., TE, V-C., RANDIMBISON, L. and LA VECCHIA, LEVI,F., FRANCESCHI, C., Trends of skin cancer in the Canton of Vaud, 1976-92. Brit. J. Cancer, 71,1047-1053 (1995). LEVI, F., LA VECCHIA,C., LUCCHINI, F. and BOYLE,P., Cancer incidence and mortality in Europe. Soz. Praeventivmed., 38 (Suppl. 3), S155-S229 (1993). LEVI,F., LA VECCHIA, C., TE, V-C. and MEZZANOITE, G., Descriptive epidemiology of skin cancer in the Swiss Canton of Vaud. Int. J. Cancer, 42,811-816 (1988). LEVI.F., TE, V.-c., RANDIMBISON, L. and LA VECCHIA, c., Statistics from the Registry of the Canton of Vaud, Switzerland, 1983-87. In: D.M. Parkin, C.S. Muir, C.S. Whelan, Y.T. Gao, J. Ferlay and J. Powell (eds.). Cancer incidence in five continents, Vol. VI, IARC Scientific Publication 120, pp. 762-765, IARC, Lyon (1992). MAGNUS, K., The Nordic profile of skin-cancer incidence. A comparative epidemiological study of the three main types of skin cancer. Int. J. Cancer, 47,12-19 (1991). MARKS,R., STAPLES, M. and GILES,G.G., Trends in non-melanocytic skin cancer treated in Australia: the second National Survey. Int. J. Cancer, 53,585-590 (1993). MOAN,J. and DAHLBACK, A,, The relationship between skin cancers, solar radiation and ozone depletion.Brit. J. Cancer, 65,916-921 (1992). DSTERLIND, A,, HOU-JENSEN, K. and JENSEN,O.M., Incidence of cutaneous malignant melanoma in Denmark 1978-1982. Anatomic site distribution, histologic types, and comparison with non-melanoma skin cancer. Brit. J. Cancer, 58,383-391 (1988). PARKIN, D.M., MUIR,C.S., WHELAN,S.L., GAO,Y.-T., FERLAY, J. and POWELL, J., (eds.), Cancer incidence m five continents, Vol. VI, IARC Scientific Publication 120. IARC, Lyon (1992). PEARL,D. and SCOTT, E.L., The anatomical distribution of skin cancers. Int. J. Cancer, 15,502-506 (1986). SHAMANIN, V.A. and DE VILLIERS,E.M., HPV in non-melanoma skin cancers. In: Abstracts of the 14th International Papillomavirus Conference, Quebec City, Canada, July 23-28. 1995. Universite Laval, Quebec City (1995). STOLL,H.L. and SCHWARTZ, R.A., Squamous-cell carcinoma. In: J.D. Jeffers, E. Scott and J. White (eds.), Dermatology in generalpractice, textbook and atlas, 3rd ed., pp. 746-758, McGraw-Hill, New York (1987).