Frequency of Facial Basal Cell Carcinoma Does Not Correlate With Site-Specific UV Exposure

BASAL CELL carcinoma (BCC) is the most common skin cancer in whites, with a steadily increasing incidence.¹ Ultraviolet radiation is considered as the single most important risk factor for BCC, because the tumor occurs most frequently on sun-exposed areas of the body, with approximately 4 of 5 BCCs occurring on the face.^{1 - 2} Furthermore, BCC has a higher incidence in patients who have a fair complexion, reflecting an inadequate biological ability for UV protection by tanning,^{3 - 5} and patients exposed to excessive UV radiation in the course of phototherapy have a higher risk of developing BCC.⁶

However, a simple correlation such as "the more UV radiation, the more skin cancer" does not explain adequately the pattern of occurrence of BCC: Why is the tumor so rarely found in certain extrafacial localizations, eg, the dorsal aspect of the hand despite considerable UV exposure,⁷ and why is the tumor conspicuously frequent in certain facial regions, such as the orbit, which are less exposed to UV radiation than the buccal or zygomatic regions? Unfortunately, most authors reporting larger series of BCC^{2 ,8 - 10} are referring to rather crude anatomical divisions (eg, forehead, nose, and ears), without acknowledging that UV exposure is critically dependent on anatomical details of the elaborate facial contour.¹¹

Another enigma of BCC is its way of spreading, which correlates in part to particular histologic features¹² : morpheaform BCC has been reported to show more widespread and irregular tumor formation than solid or adenoid-cystic BCC.¹³ Superficial BCC, which occurs more frequently on the trunk than on the face, is the least agressive type. Therefore, we decided to investigate whether anatomical prominences with peak UV exposure have correlating peak frequencies and/or distinct histologic features of BCC.

This report summarizes data derived from 3065 facial BCCs in 2785 patients, giving a comprehensive account of localization with respect to particular facial areas, histopathologic differentiation, and correlation to UV exposure.

The study material included clinical reports of patients who were admitted to the Department of Dermatology, Ludwig Maximilians University, Munich, Germany, for treatment of BCC between the years 1979 and 1994 and who fulfilled the following criteria: the BCC was located on the face and the clinical diagnosis was confirmed by the findings of microscopic examination of a skin biopsy specimen. Based on the predominant histopathologic appearance, BCCs were categorized as either solid, morpheaform, or adenoid-cystic. Pigmented BCCs were categorized according to their predominant histologic features.

The exact anatomical localization and extension of BCCs were documented on a uniform schematic drawing in a 1:1 scale. The periocular area was further subdivided into 4 equal segments (ie, the lateral, medial, upper, and lower quadrants [Figure 1A]). Likewise, the nose was divided into 4 segments by taking a lateral view with 2 segments on each side: the basal segment of the nose bordering to the nasolabial fold extending midway up the side of the nose, and the upper segment of the nose extending from midway to the back of the nose, including the apex (Figure 1A).

Data collection and processing were carried out using commercially available computer software (Excel; Microsoft Worldwide, and SPSS; SPSS Inc, Chicago, Ill). The χ² test was used to calculate the level of significance for differences of BCC frequencies within comparable segments.

A total of 3065 BCCs were studied in 2785 patients, 297 (11%) of whom had more than 1 facial BCC at the time of diagnosis. Men were slightly more prevalent (53%) than women. The youngest patient was 20 years old; the oldest, 101. The mean ± SD age of the patients at the time of diagnosis was 63.2 ± 12.9 years. The incidence of BCC increased steadily up to the eighth decade of life (Figure 2). Only 16 patients (0.6%) were younger than 30 years, and 496 (17.8 %) were younger than 50 years. There was no difference between men and women regarding age at the time of diagnosis of BCC.

The localizations of facial BCCs are shown in Figure 1B. The nose, as the most prominent facial anatomical structure, was also the most frequent site of BCC (n = 1373, or 44.8%). Of the 1373 nasal BCCs, however, 851 (62.0%) occurred on the base of the nose, while only 292 (21.3%) occurred on the apex and 230 (16.8%) occurred on the dorsum of the nose (Figure 1A). Assuming that the surface areas of the upper (apex and dorsum) and lower (alae and base) segments of the nose were comparable, the difference was significant at P<.05.

The orbital region harbored 12.6% (386/3065) of all facial BCCs, which were unevenly distributed, with 58.3% (225/386) in the medial quadrant and only 6.2% (24/386) in the lateral quadrant (P<.001) (Figure 3A). Of 269 auricular BCCs, 105 (39.0%) were located at the preauricular crest, 99 (36.8%) in the retroauricular fold, and 65 (24.2%) on the helix of the ear (Figure 3B). When the putative effects of hairstyles that cover the ears were taken into account, there was no difference in frequencies of auricular and retroauricular BCCs in men and women. The least affected facial prominence was the chin (47/3065, or 1.5%). For each region specified in Figure 1B, there was no significant difference (P>.05) when patients aged 63 years or older (n = 1804) were compared with younger patients (n = 1261).

The predominant histopathologic subtypes in descending order of frequency were solid (1514/3065, or 49.4%), morpheaform (1149/3065, 37.5%), and adenoid-cystic (402/3065, or13.1%). Of the 1149 morpheaform tumors, 928 displayed exclusively morpheaform features, while the remaining 221 were composed of solid tumor islands mixed with morpheaform features. The distribution of these histopathologic types in the aforementioned regions is shown in Figure 4. In descending order, the buccal, temporal, perioral, and frontal regions displayed a relatively high percentage of morpheaform BCCs, exceeding the percentage of solid BCCs. Comparison of histopathologic types in both sexes revealed that solid BCCs occurred 1.2 times more frequently in men than in women, while morpheaform BCCs were 1.3 times more frequent in women (P<.001).

Chronic UV exposure has been postulated to be a causative factor for the development of skin cancer. The capacity of DNA repair after UV irradiation is lower in patients with BCC than in controls.¹⁴ A number of UV-induced phenomena have been identified as putative mechanisms of BCC development. Clonal expansion of mutant p53 is frequently overexpressed in and around BCCs on sun-exposed skin.^{15 - 16} Mutations of the oncogene H-ras occur with increased frequencies in BCCs,¹⁷ while the expression of the oncogene c-fos is reduced.¹⁸ Levels of interleukin 4 and interleukin 10 are elevated by UV irradiation, reducing immunologic tumor surveillance.¹⁹ Interstitial collagenase is UV inducible,^{20 - 21} fostering local tumor invasion.²² If these mechanisms, however, which are operating in a UV dose–dependent fashion, are the driving force of tumor development, one would expect the highest frequency of BCCs in those sites of the face that are receiving the most exposure to UV radiation.

The doses of UV radiation received in various parts of the face have been measured with UV-sensitive halogen-free radicals on the surface of mannequin heads with respect to topographic anatomy.¹¹ Accordingly, the apex of the nose and the helical part of the ears were the most exposed areas, receiving twice as much UV radiation as the buccal and zygomatic regions and up to 10 times more than the orbital region. We found that the orbital region was more frequently affected than the buccal and zygomatic regions, indicating a mismatch of BCC frequencies with UV exposure. These areas may differ in size and are therefore difficult to compare numerically. However, comparing the equally sized lateral and medial quadrants of the orbit, we found an even more conspicuous misbalance of BCC frequency, which is unexplainable by UV exposure. The hypothesis that UV radiation may be intensified unproportionately by glasses worn to correct impaired vision in the elderly could not be confirmed by the respective UV measurements using such glasses.¹¹ Another sign of a lack of correlation of UV exposure and BCC frequency is the inverse distribution of BCCs on less exposed basal segments of the nose and the retroauricular fold. Finally, BCC is a tumor of advanced age. Our findings of peak incidence within the eighth decade of life confirm the observations of a recent study.⁸ Consequently, the differences of cumulative UV exposure in various sites should be more pronounced over time. In our study, however, the age of the patients had no influence on the frequency pattern of BCCs.

Descriptively, those areas with high BCC frequency despite low UV exposure are characterized by a concave shape, reduced skin tension, and the presence of marked skin folds. These features may reflect reduced dermal thickness and altered connective tissue arrangements. This may influence the expression of metalloproteases such as collagenases, which are responsible for tissue breakdown in the course of tumor expansion²³ and which are overexposed in and around BCCs.^{22 ,24} Furthermore, metalloproteases in BCCs and adjacent stromal cells are dependent on integrins.²⁵ Several integrins that operate as cell matrix receptors were found to be reduced²⁶ or even completely absent²⁷ in BCCs.

In conclusion, the development of facial BCCs appears to be poorly correlated to the cumulative effects of UV radiation alone. Specific textural changes, such as reduced tension and dermal thickness, may putatively explain this mismatch and point toward disturbed cell matrix interactions as a cofactor for the development of BCCs.

Accepted for publication May 16, 2002.

This research was supported by Ludwig-Maximilians University, Munich, Germany, and by grant GSF 07 UVB 59/2 from the Bundesministerium für Bildung, Forschung, Wissenschaft und Technologie (BMBF), Berlin, Germany.

Corresponding author: Marc Heckmann, MD, Department of Dermatology, Ludwig-Maximilians University, Frauenlobstrasse 9-11, 80337 München, Germany (e-mail: heckmann@derma.de).