Non-Melanoma Skin Cancer

Non-Melanoma Skin Cancer

Trends in Sun Exposure Knowledge, Attitudes, and Behaviors: 1986 to 1996

Introduction.—The national Melanoma/Skin Cancer Detection and Prevention program of the American Academy of Dermatology was implemented in response to the rising incidence of invasive melanoma in the U.S. Though the long-term goal is to reduce the incidence and mortality of skin cancer, the short-term goal is to change knowledge, attitudes, and behaviors affecting skin cancer risk. A follow-up survey was performed to assess these changes in knowledge, attitudes, and behaviors about sun exposure, sun protection, and skin cancer/melanoma.

Methods.—The first national survey concerning these issues was per-formed in 1986. A follow-up survey performed in 1996 used the same questions to evaluate change over time, and classified questions to gain greater insight into attitudes and behaviors.

Non-Melanoma Skin CancerResults.—Knowledge of the dangers of sun exposure increased significantly between the 2 surveys, while the ability to recognize skin cancer as a danger declined. The percentage of respondents who believed a tan looked healthy decreased from 66% to 56%; men were especially likely to believe that a tan enhanced appearance. The incidence of sunburn, sunscreen use, and tanning bed use all increased. Responses suggested that subjects knew sun exposure is harmful, but did not recognize the potential for harm to themselves. Sunburn was associated with male sex, younger age, and more education. Weekend exposure was mainly recreational in nature among indoor workers living in areas of low daily sunshine; weekday exposure was mainly occupational in nature. Intentional tanning was associated with female sex, younger age, and living in areas with fewer sunny days. Younger women and whites were more likely to use a tanning booth or sunlamp. Forty-two percent of respondents used sunscreen, 39% sought out shady areas, and 32% wore hats. Some sun-protective measures were significantly related to a perceived elevation in skin cancer risk. Older age, white race, increased education, and female sex were related to awareness of skin cancer prevention. Factors related to increased sunscreen use were a personal or family history of skin cancer, self-reported sun sensitivity, self-reported skin type, and history of sunburn.

Conclusions.—The findings are consistent with early changes in cognitive and emotional variables related to sun exposure. The study identifies some high-risk populations who can be targeted for future interventions. Though knowledge about prevention has improved, this may not be enough for effective behavioral change. If young people can be convinced that sun exposure knowledge is relevant to them personally, then further behavioral changes leading to more effective sun protection may take place.

► Since 1985, the American Academy of Dermatology, through the media and through free skin cancer screenings, has attempted to promote the early detection of skin cancer, to curb the rising incidence of skin cancer, and to educate the public about the harmful effects of excessive sun exposure. In this study the authors used a survey to assess the impact of this effort. Although people were more aware of the harmful effects of excessive sun exposure, there appeared to be less focus on the development of skin cancer as a result of it. Moreover, respondents did not seem to be able to relate these harmful effects to themselves. There did not appear to be an impact on those purposely seeking a tan, even though sunscreen use increased. The majority of respondents also still felt a tan “looked good.” Respondents continued to experience sunburns and to receive intense weekend sun exposure. Regular use of suntanning devices increased during this period. Women, younger people, people residing in areas with fewer sunny days, and whites were more likely to tan intentionally. Men living in the south were more likely to sunburn. Although this survey suggests that the efforts of the American Academy of Dermatology have had an impact on knowledge about the harmful effects of ultraviolet light and have resulted in increased sunscreen use, overall there has been minimal behavior modification as it relates to sun exposure. However, this data should allow for a better targeting of high-risk groups and for a better chance of behavior modification.

Sunbathing Habits and Sunscreen Use Among White Adults: Results of a National Survey

Objective.—One of the ways of preventing melanoma is to wear sunscreen. Data on use of sun protection measures is lacking. The sunbathing habits and prevalence of sunscreen use in a representative national sample is presented.

Methods.—A telephone survey of 3,042 households was conducted in July through September 1991 to assess recreational aquatic activities during the past 122 months. To control for skin type, the analysis was restricted to 2,459 whites. Respondents were asked about the number of days they had sunbathed, use of sunscreen, and solar protection factor (SPF) of the sunscreen. Demographic information was collected.

Results.—A total of 59% (n = 1,440) of respondents reported sunbathing at least once during the year. Of the two thirds (n = 1,614) of respondents reporting activities on or near the water, 33% (n = 530) said they used sunscreen. Of those 530,248 (47%) said they used sunscreen regularly, with women being regular users more often than men. Younger users (ages 16-25 years) and those with the lowest level of education used sunscreen least often. Income did not affect sunscreen use. Routine users were more likely to use a sunscreen with an SPF of at least 15, although 45% of routine users did not use a sunscreen with an SPF of at least 15. Sixty percent of sporadic users did not use a sunscreen with SPF of at least 15.

Although 59% of white Americans sunbathe and 25% do so regularly, fewer than 50% use sunscreen. Only half of routine users use sunscreens with an SPF 15 or greater. Younger sunbathers were least likely to follow recommendations, creating a substantial lifetime risk for melanoma.

Conclusion.—Only 47% of respondents reported regular sunscreen use, and only 55% of those used a sunscreen with an SPF 15 or greater.

► In this telephone survey, which confined its analysis to whites, the investigators found that a quarter of the responders were frequent sunbathers and that the most frequent sunbathers were women and those in the youngest age group. They also found that only 25% of sunbathers used a sunscreen with an SPF of at least 15. Men, those in the youngest age group, and those with a lower educational background were less apt to use a sunscreen. Skin type and the use of other protective measures when not sunbathing were not addressed. Regular use of a sunscreen with an SPF of at least 15 has been shown to decrease the risk of skin cancer, and the majority of one’s sun damage occurs in the first 18 years of life. Thus, to decrease the increasing incidence of skin cancer, these high-risk groups need to be educated about the harmful effects of sunbathing and the importance of protecting the skin through a variety of measures, including regular use of a sunscreen with SPF 15 or greater.

Activating Patients to Practice Skin Cancer Prevention: Response to Mailed Materials From Physicians Versus HMOs

Introduction.—The benefits of activating patients to perform preventive health behaviors are recognized by HMOs. The source and emphasis of mailed messages regarding skin cancer were assessed in 981 randomly selected patients to determine if this approach could differentially activate them to initiate skin cancer prevention by calling a toll-free number.

Methods.—Research subjects were mailed a questionnaire evaluating their concern about and risk for skin cancer and an educational booklet. These were accompanied by a letter inviting them to call a toll-free number. Letters were sent from 1 of 3 sources: their physician, their HMO, or a fictitious junk mail organization. Mailings had 1 of 3 different messages emphasizing the effects of ultraviolet rays on either the risk of skin cancer, aging and wrinkling of the skin, or aging and wrinkling accompanied by a book further emphasizing these harmful effects of the sun.

Results.—Forty-eight packets were returned undelivered. The overall activation rate was only 7%. The source of the preventive message significantly affected whether research subjects called the toll-free number. Messages from physicians and HMOs were more activating than those from the fictitious junk mail organization. There were no differences in activation in messages mailed from physicians and HMOs. The call-in rate was not affected by the emphasis of the message. Activation was positively associated with risk for skin cancer. It was not associated with attitudes or beliefs about skin cancer prevention.

Conclusion.—These findings indicate that physicians and HMOs can effectively use preventive health outreach as a part of their multicomponent prevention efforts. Physicians and HMOs can activate patients at high risk for skin cancer by emphasizing both the risks of skin cancer and aging and wrinkling when they deliver a skin cancer prevention message.

► Although the activation rate of less than 10% was disappointing, this percentage simply measured those patients who called a toll-free number to report their skin cancer risk scores and to request free sunscreen. The mailing may have affected patient behavior in other ways, such as by increasing skin cancer awareness and sunscreen use.

Beachfront Screening for Skin Cancer in Texas Gulf Coast Surfers

Introduction.—Free skin cancer screening programs can improve early detection. However, because of self-selection, such programs may miss the patients at highest risk. As mortality from nonmelanoma skin cancer continues to rise, this seems to be the case with skin cancer screening programs. In an attempt to target high-risk populations, a beachfront skin cancer screening program for surfers was studied.

Methods.—The program took place at a 1992 surfing competition. The surfers were asked about skin cancer risk factors and knowledge. They were then screened for skin cancer by a dermatologist, with special attention to the identification of skin lesions such as actinic keratoses, atypical moles, basal cell carcinoma, benign-appearing nevi, freckles, melanoma, and squamous cell carcinoma. The findings were compared with those of patients attending a free skin cancer program with subject self-selection held the same year.

Results.—Fifty-five percent of 110 surfers registered for the event completed a questionnaire; more than 90% of these respondents were male. Forty-nine adult surfers (mean age, 30 years) were screened. Thirty-six of this group had atypical moles, actinic keratoses, or basal cell carcinoma. Two thirds of patients with abnormal screening findings did not use sunscreens regularly. Overall, 41% of subjects had lesions consistent with actinic keratoses. Sixteen percent of subjects had basal cell carcinomas; mean age in this group was 38 years. No cases of squamous cell carcinoma or melanoma were noted. Subjects in the comparison sample were significantly older and more likely to be female. The number of basal cell carcinomas detected in this sample was significantly lower than among the surfers.

Conclusions.—Targeting a high-risk population, in this case surfers, appears to be a helpful approach to the diagnosis of premalignant and malignant skin conditions. The study is limited by the age and sex composition of the 2 groups; however the findings suggest that screening and education programs for high-risk groups should significantly reduce morbidity and mortality from all forms of skin cancer. Measures are needed to overcome outdoor athletes’ resistance to the use of sunscreens.

► Although the free mass skin cancer screenings that have been sponsored by the American Academy of Dermatology (AAD) are commendable, unfortunately there has been no guarantee that patients at highest risk for developing melanoma or nonmelanoma skin cancer will participate. To more effectively screen and educate high-risk patients, it is probably necessary to identify these patients and seek them out rather than depend on them to avail themselves of free screenings. To test this hypothesis, the authors offered a free skin cancer screening to participants in a surfing contest and compared these findings with those accumulated at an AAD-sponsored screening during the same month and in the same region. When matched for age, a greater number of basal cell carcinomas were detected in the surfers. Although the two groups screened were markedly different with regard to age and sex distribution, it is still clear that many high-risk patients do not avail themselves of free screenings and must be sought out.

Screening for Skin Cancer in Primary Care Settings

Introduction.—As the incidence and mortality of skin cancers increase, there are variable recommendations for screening. The primary care office seems to be an appropriate site for skin cancer screening, but it is unknown whether such screening is taking place. The frequency of documented skin cancer screening by primary care physicians was assessed.

Methods.—The retrospective cohort study included a random sample of 200 patients, age 50 or older, receiving outpatient care at 2 Veterans Affairs general medical clinics. Ninety-seven percent of the patients were men. The medical records were reviewed to determine the frequency of documented skin examination, compared with other selected components of the physical examination.

Results.—Skin examination was performed and documented in 28% of patients. By comparison, fecal occult blood testing was done in 60%, rectal examination in 64%, and sigmoidoscopy in 47%; prostate examination was done in 59% of the men. Among patients without skin-related complaints, the frequency of skin examination fell to 18%.

Conclusion.—Skin examination is rarely documented in the primary care setting. Other screening studies are performed with much greater frequency. The results may have important implications for assessment of the quality of dermatologic care.

► It is no secret among dermatologists that examination of the skin is often neglected in the primary care setting. Although many patients visiting a primary care physician have a diagnosable skin disease, often the skin is examined only if the patient brings a specific problem to the physician’s attention. A complete skin cancer screening is the exception rather than the rule, even among patients with risk factors for cutaneous malignancy.

To Burn or Not to Burn: Use of Computer-enhanced Stimuli to Encourage Application of Sunscreens

Purpose.—It has proven difficult to educate the public about the dangers of sun exposure to the skin. Increasing awareness does not translate into lasting improvement in protective measures. One contributing factor appears to be the “optimistic bias” that leads people to underestimate their future likelihood of damage. This study used computer simulations of the sun’s damaging effects in an attempt to motivate the use of sunscreen.

Methods.—The study included 30 white female high school students working at a summer camp. One group of girls was shown a computer simulation of their own faces with 25 years’ aging to resemble the effect caused by long-standing skin exposure. In part of this group, the faces were further disfigured by the addition of unsightly skin cancers. Another group was shown no pictures. The students were provided with sunscreen and asked to log their sunscreen use and time spent outdoors over 6 weeks. The effects of viewing the computer-altered pictures on sunscreen use were analyzed.

Results.—In the girls who were shown the computer simulations, sunscreen use increased significantly. The increase was greater for girls who viewed both the aged and disfigured faces. For girls who did not view simulations, sunscreen use actually decreased. The effects of viewing the simulations decreased after a week or 2.

Conclusions.—Viewing computer simulations of the effects of sun on the facial skin can increase the short-term use of sunscreen by adolescent girls. The effect is greatest when the girls view simulations of not only aging but also disfiguring cancers. The study is limited by concerns over reliability and the size and nature of the subject population.

► Although intense education of the public about the dangers of sun exposure and its consequences has been advocated, studies have shown that awareness does not necessarily correlate with the institution of protective measures, especially in adolescents who feel they are immune to the harmful rays of ultraviolet light. Since many people correlate a tan with a healthy appearance, the author wondered whether behavior modification might result if adolescents were shown computer-generated photographs of themselves that reflected the ravages of chronic sun exposure. He found that adolescents shown altered pictures of themselves, which reflected photoaging and the development of disfiguring skin lesions, were more likely to use sunscreens. Unfortunately, by week 6 there was a significant decline in sunscreen use.

Clothing as Protection From Ultraviolet Radiation: Which Fabric Is Most Effective?

Introduction.—The incidence of skin cancers is increasing rapidly in Caucasian populations, a development attributed to depletion of the ozone layer and the resulting exposure to elevated levels of ultraviolet radiation (UVR). Sunscreen agents and protective clothing are recommended for individuals who must be outdoors every day. Twenty-eight fabrics were examined to determine which types are most effective in screening UVR.

Methods.—Transmission of UVR through the fabrics was measured using spectrophotometric techniques. Fabric characteristics evaluated included cover (the degree of closed vs. open space in a fabric), mass, fiber type, fabric count, structure, and color. All tested fabrics were white, bleached or unbleached; types tested were cotton, wool, polyester, nylon, linen, acetate, acrylic, rayon, a polyester/wool blend, and a polyester/ cotton blend. An average sun protection factor (SPF) was calculated for the 5 tested specimens of each test fabric.

Results.—The cellulose-based cotton, linen, and rayon fabrics are comfortable to wear in summer, but most had low (3-5) SPF values. Two heavy weight cellulose-based fabrics had higher SPF values (9 and 13). Polyester fabrics provided higher levels of protection, and initial findings suggest that a polyester/cotton blend would provide significantly better protection than cotton alone. Fabrics of the highest mass or weight were more protective than lighter weight fabrics. When blue dye was applied to cotton fabrics, including denim, the SPF was increased by more than a factor of 3.

Conclusions.—Clothing can provide an effective barrier to UVR, but 19 of the 28 fabrics tested offered less protection than a sunscreen with SPF 15. Cotton was less effective than polyester, white fabrics were less effective than dyed fabrics, and fabric with greater mass or weight usually offered more protection. The best protection against UVR would appear to come from a dark fabric, made of polyester or a polyester blend, and having a heavy weight and close weave.

► Patients need to be educated that thin, lightweight clothing is no substitute for a good sunscreen. In this study, of 28 white fabrics tested, 19 offered less protection than an SPF 15 sunscreen.

Death From Skin Cancer Among the Elderly: Epidemiological Patterns

Introduction.—The leading cause of death from skin disorders is melanoma. The mortality rate from melanoma is increasing. Both melanoma and nonmelanoma cancer types affect adults of all ages, but the impact of the malignancies varies with age. The patterns of mortality from skin cancer throughout life were documented.

Methods.—A review was conducted of persons who died from skin cancer in the United States during a total of 8 years within a 13-year period to determine the distribution of deaths from skin cancer by cause of death, age, gender, and race.

Results.—During the 8 years studied, skin cancer was responsible for 61,458 deaths, and 72% of these were attributed to melanoma. At the end of each 5-year period studied, the number of deaths from skin cancer steadily increased. About 90% of deaths from skin cancer among whites younger than 50 years were attributed to melanoma, but only a minority of deaths among blacks and whites older than 85 years was caused by melanoma.

Conclusion.—Melanoma and nonmelanoma skin cancer should be the focus of efforts at early detection among the elderly white population. Although nonmelanoma skin cancer is 25 times more common than melanoma, it is less likely to lead to death. Unlike melanoma, mortality rates for nonmelanoma skin cancer have been declining for at least 20 years. Among the elderly, the pattern of mortality from skin cancer differs from the pattern at younger ages. Melanoma carries an important risk among young adults, but the mortality from nonmelanoma skin cancer is greater in older persons.

► Because melanoma is more likely to cause death than is nonmelanoma skin cancer, much of the emphasis on early detection has been focused on it and not on nonmelanoma skin cancer. However, as emphasized by Wein- stock, in older whites, nonmelanoma skin cancer accounts for more deaths than melanoma. Thus, from a public health standpoint, it is just as important to focus on the early detection of nonmelanoma skin cancer as it is to focus on melanoma.

Analysis of Risk Factors for Cutaneous Warts in Renal Transplant Recipients

Background.—Cutaneous warts are one of the most common dermatological side effects of immunosuppressive therapy, occurring in 20% to 70% of patients who have renal transplantation. Such patients have a 3% to 10% incidence of skin neoplasms and a risk of skin cancer 20-40 times higher than in individuals who are immunocompetent. Risk factors for cutaneous warts were examined in patients treated with 3 immunosuppressive regimens. The effect of several risk factors for skin cancer on the onset of cutaneous warts was also examined.

Methods.—The onset of cutaneous warts was analyzed in 199 consecutive patients after renal transplantation. The mean patient age was about 44, and the mean age at transplantation was about 38. Patients were divided into 3 groups: 38 patients were given prednisolone/azathioprine, 82 patients were given prednisolone/azathioprine/cyclosporine, and 79 patients were given prednisolone/cyclosporine. The effects of types of immunosuppressive drugs used, age at transplantation, and HLA-A11 and HLA-B mismatch on the onset of cutaneous warts were also analyzed.

Results.—Kaplan and Meier analysis showed that the onset of cutaneous warts was significantly earlier in patients given prednisolone/azathioprine/cyclosporine and prednisolone/cyclosporine. This may have been related to the enhanced immunosuppressive action of the regimens that included cyclosporine. There was a significant difference between the groups given prednisolone/azathioprine and prednisolone/azathioprine/cyclosporine, the groups given prednisolone/azathioprine and prednisolone/ cyclosporine, and the groups given prednisolone/cyclosporine and prednisolone/azathioprine/cyclosporine. Cox analysis showed a possible relationship between the maintenance protocol of immunosuppressive drugs and warts, with the risk significantly increasing with higher doses of azathioprine and cyclosporine, but decreasing with higher doses of prednisolone. Of 7 patients with skin cancer, 4 also had multiple warts at the cancer site; histological examination showed the neoplastic transformation of a wart in 2 patients.

Discussion.—Maintenance treatment with azathioprine, cyclosporine, and prednisolone may be a major risk factor for cutaneous warts. Patients taking effective immunosuppressive regimens may have earlier onset of warts. It appeared that HLA-B mismatch and HLA-A11 had no effect. Accurate dermatological screening for early diagnosis and treatment of warts may be beneficial to patients undergoing renal transplantation.

► In our institution, we urge our transplant surgery colleagues to allow us to perform a complete skin examination on candidates for such surgery. This allows us to treat any obvious viral or precancerous lesions before the patient is subjected to immunosuppressive therapy. Clearly, cyclosporine and azathioprine, two drugs commonly used for these patients, are associated with a significant risk for the development of viral warts and cutaneous carcinogenesis.

Prevailing Papillomavirus Types in Non-Melanoma Carcinomas of the Skin in Renal Allograft Recipients

Introduction.—Human papillomavirus (HPV) is a causative factor in genital tract carcinomas, both in situ and invasive. Epidermodysplasia verruciformis (EV) is a rare disease characterized by extensive, unusual warts and cutaneous squamous cell cancers on light-exposed skin. The carcinogenic process of EV is thought to involve a specific group of HPVs. The lack of any sensitive test for a wide range of cutaneous HPV types has made it difficult to assess the role of these viruses in nonmelanoma skin cancers. Such cancers are common in immunosuppressed organ transplant recipients. Two sets of polymerase chain reaction (PCR) primers were used to clarify the types of HPV involved in benign and malignant skin lesions of renal allograft recipients.

Methods.—The study included 68 benign and malignant tumors taken from 25 renal allograft recipients over a 12-year period. Two sets of degenerate primers for HPV were used in the PCR studies, the second set being designed to detect EV-specific types of HPV. The PCR products were cloned and sequenced to differentiate between HPV DNA and other cellular sequences.

Results.—All viral warts tested expressed HPV DNA, as did 65% of keratoses, 91% of intraepidermal cancers, and 91% of invasive squamous cell cancers. The HPVs detected included both cutaneous and EV types, 18 of which were newly identified types. The EV types prevailed among 4 patients with multiple cancers: HPV 20, 23, and 38, and the novel types DL40 and DL267, which are related to HPV 10 and 38, respectively. Together, these 5 HPV types were found in nearly three fourths of the malignant lesions tested.

Conclusion.—The technique employed in this study, using 2 sets of PCR primers, represents a reliable approach to detecting HPV DNA in nonmelanoma skin cancer specimens. Cancers from patients with renal allografts show predominantly EV HPVs. However, many different HPV types can be found, with some lesions showing double infections. Thus, the carcinogenic process may involve not only virus-host interactions but also virus-virus interactions.

► In this study, a number of different HPV types were detected in different tumors, either singly or in combination. This raises the possibility of virus-virus interaction, resulting in complementation of otherwise defective types. Further studies will be necessary to elucidate the relevant virus-virus and virus-host interactions. Equally interesting, from a clinical point of view, would be the results of a similar study to characterize the HPV types present, if any, in presumably nonimmunosuppressed patients with severely photodamaged skin and multiple skin cancers. It is tempting to speculate that these individuals have sufficient UV-induced immunosuppression to allow for the development of viral carcinogenesis.

Induction of Cancer, Actinic Keratosis, and Specific p53 Mutations by UVB Light in Human Skin Maintained in Severe Combined Immunodeficient Mice

Introduction.—Depletion of stratospheric ozone and increases in environmental levels of genotoxic UVB in sunlight may be responsible for a rise in skin cancer frequency. Human skin transplanted to severe combined immunodeficient mice was exposed to daily doses of UVB for about 2 years to evaluate the mechanism and risk of human skin cancer from solar light.

Methods.—Human skins resected from 5,4, and 2 patients, respectively, undergoing mammectomy for breast cancer, phimectomy for phimosis, and excision of actinic keratosis and skin grafting were used for hetero- transplantation to SCID mice. Two weeks later, about half of the transplanted human skins from each donor underwent daily exposure to UVB from fluorescent sunlamps. The other half were irradiated. Daily doses of UVB were 8,000 J/m2 on Monday, Wednesday, and Saturday and 2,000 J/m2 on remaining days. Unirradiated and UVB-irradiated skin pieces were examined for mutations of p53 and K-ras genes.

Results.—A significantly high incidence of actinic keratosis was induced in the UVB-irradiated human skin. There was no incidence of actinic keratoses in unirradiated human skins. No histologic changes were observed in mouse skin around the transplanted site because of hair, but skin erosion and necrosis were induced in the ears of all UVB-irradiated mice. Of 18 normal skins exposed to doses of 7.3 X 105 to 1.8 X 106 J/m2, 14 skins (77.8%) subsequently had actinic keratoses and 3 (16.7%) had squamous cell carcinomas. In none of the skins not exposed to UVB did actinic keratoses or cancer develop. Donor skin susceptibility varied. Skins sensitive for actinic keratosis were also susceptible for cancer induction. Mutations of the K-ras oncogene were not observed in any of the UVB- induced skin cancers and actinic keratoses. An assortment of mutations were observed in 4 exons of the p53 gene. The p53 mutation at codon 242 (C TGC —» C CGC: Cys —» Arg) was specifically seen in skin cancers and actinic keratoses. Double or triple mutations were created in all UVB- induced skin cancers and 3 of 8 actinic keratoses. Most mutations were observed at dipyrimidine sites.

Conclusion.—A UVB-induced mutation at codon 242 may be crucial and initial to cancer and keratoses susceptibility. Combinations with other mutations in the p53 gene and others might be needed for malignant transformation.

► The authors succeeded in inducing actinic keratoses and squamous cell carcinomas in human skin exposed to daily doses of UVB after transplantation to severe combined immunodeficient mice. Although skin from different donors showed variation in susceptibility to dysplastic changes, skin in which squamous cell carcinoma developed showed early development of actinic keratoses. Thus, in addition to elucidating some of the molecular mechanisms underlying UV-induced carcinogenesis, this study also provides evidence of the importance of treating actinic keratoses before they evolve into invasive tumors.

High Rate of Malignant Transformation in Hyperkeratotic Actinic Keratoses

Introduction.—The most common epithelial precancerous lesions are actinic keratoses, which are divided into 5 types: hypertrophic, atrophic, bowenoid, acantholytic, and pigmented. The histologic features of these lesions cannot be predicted by clinical examination because there are no established criteria. It is also unknown which features could indicate which actinic keratoses are more strongly associated with malignancy. Clinically diagnosed hyperkeratotic actinic keratoses were examined and classified histologically.

Methods.—In 43 patients, 50 papular hyperkeratotic actinic keratoses on the dorsum of the hand, wrist, and arm, less than 1 cm in diameter were identified. Histologic evaluation was conducted by 2 observers in a blinded fashion.

Results.—Invasive squamous cell carcinoma was found in 38 lesions (36%), and squamous cell carcinoma in situ was found in another 7 lesions (14%). The next most common lesions identified were proliferative actinic keratoses in (26%), followed by hypertrophic actinic keratoses (10%), and lichenoid actinic keratoses (4%). The remainder of the biopsy specimens consisted of basal cell carcinomas (4%) and other benign lesions (6%).

Conclusions.—The actinic keratoses that have a greater tendency toward malignant change can be predicted clinically with greater certainty than previously thought. A malignancy rate of 50% was found in clinical hyperkeratotic actinic keratoses of less than 1 cm in diameter on the dorsum of the hand, wrist, or forearms of white patients. Lesions with this clinical description deserve special attention.

► In today’s environment of cost containment and managed care, insurance carriers sometimes will use data from the literature to either no longer provide coverage for a given service or to dictate how a given condition is managed. Such is the case of the actinic keratosis (AK). Recent literature suggests that many of these lesions involute spontaneously, and only a small percentage evolve into squamous cell carcinomas (SCC). This obviously raises the question of whether these lesions require treatment. From an epidemiologic standpoint, it would appear that AK treatment is an ideal example of preventive medicine, but data to support this contention are needed. Suchniak et al. selected for study hypertrophic AKs of less than 1 cm in diameter that were located on the dorsal hands, wrists, and forearms of patients with a history of significant actinic damage. On biopsy, 50% of the lesions were found to contain foci of either invasive or in situ SCC. The authors conclude, and rightfully so, that this subset of AKs deserves aggressive management.

Solar Keratosis Can Be Distinguished From Superficial Basal Cell Carcinoma by Expression of bcl-2

Introduction.—Solar keratosis (SK), or actinic keratosis, is an epidermal dysplasia that may be regarded as a squamous cell carcinoma in situ. The histologic features of SK are well defined and are distinct from those of superficial basal cell carcinoma (SBCC). The distinctions between SK and SBCC are sometimes difficult, particularly when examining small-punch or superficial shave biopsy specimens in which the overall global architectural pattern of the lesion is not represented. The presence or absence of the bcl-2 gene may provide a way for SK and SBCC to be separated histologically.

Methods.—Hematoxylin and eosin-stained slides from all recent patients with SK and SBCC were reviewed and 10 examples of each were selected for comparison. All lesions were “gold standard” examples. Immunoperoxidase staining was performed and Dako monoclonal mouse anti-human bcl-2 oncoprotein was used as the primary antibody.

Results.—All SBCC slides showed uniform strong positive epithelial bcl-2 immunoreactivity in the characteristic budding basaloid nests. Negative epithelial immunoreactivity was observed in all SK slides. Subepithelial lymphocytes demonstrated variable scattered positive cells in both groups of slides.

Conclusion.—Earlier reports have shown that normal keratinocytes in the basal layer are bcl-2 negative. Basal cell carcinoma (BCC) is derived from basaloid stem cells in the basal layer of follicular or interfollicular epidermis. BCC may result from bcl-2 expression. It appears that the BCC tumor is “frozen” at the stem-cell level. The solar keratosis-squamous cell carcinoma sequence appears to reflect clonal expansion of keratinocytes arrested at a more mature level of development. Differential expression of bcl-2 may be used to distinguish SK from SBCC.

► Dermatopathologists are often required to distinguish between SBCC and SK in biopsy specimens that are less than optimal, e.g., curettings or small punch biopsy specimens. Both conditions may show buds of atypical basaloid keratinocytes projecting from the undersurface of the epidermis. This study by Mills demonstrates a 100% expression of bcl-2 in SBCC and lack of expression in all SKs. Because many dermatologists would choose to treat SBCC by excision with margin control, but would treat SK by a destructive modality such as liquid nitrogen cryotherapy, it is important to distinguish accurately between these neoplasms. This simple immunoperoxidase stain could prove helpful in situations that are difficult to resolve with conventional light microscopy.

Differences in Age and Body Site Distribution of the Histological Sub-types of Basal Cell Carcinoma: A Possible Indicator of Differing Causes

Background.—No previous studies have examined the subtypes of basal cell carcinoma (BCC) in terms of anatomical site, patient age, and patient sex. Variations in these characteristics among the different histologic sub- types of BCC were assessed.

Methods.—The findings of 3,885 patients with BCCs were analyzed. The lesions were analyzed by a major Australian pathology laboratory over a 6-month period.

Results.—The BCCs came from 1,688 men, aged 15-99 years, and 1,254 women, aged 17—101 years. Fifty-eight percent of lesions were found on the head and neck, 25% on the trunk, 6% on the upper limb, and 9% on the lower limb. Women had a higher percentage of BCCs on the lower limb than men, whereas men had more lesions on the trunk. Patients with superficial BCCs had a significantly younger mean age (57 years) than patients with other tumor subtypes (63-66 years). Seventy-three percent of superficial BCCs occurred on the trunk and limbs, whereas all other subtypes occurred most frequently on the head and neck. Two percent of lesions had features of both superficial and nodular BCC.

Conclusions.—Superficial BCCs differ from other histologic subtypes in terms of age and body distribution. The findings may result from differences in sun exposure needed to cause BCC. Superficial BCC does not appear to reflect an evolutionary phase of other BCC subtypes.

► This study confirms what clinicians have observed for years, i.e., superficial multicentric basal cell carcinomas (SMBCCs) are most commonly found on the trunk and extremities. Not surprisingly these lesions are more likely to be found on the legs of women than men. The authors suggest that intermittent intense sun exposure may explain the difference in the anatomical distribution of the SMBCC vs. other variants of BCC. Certain patients seem to have a proclivity to develop multiple SMBCCs. Host factors, other possible carcinogens or cocarcinogens, and sun exposure habits may all play a role in determining the anatomical distribution of SMBCC.

Characteristics of Incompletely Excised Basal Cell Carcinomas of the Skin

Introduction.—The most common form of skin cancer is basal cell carcinoma. Raised papular or nodular lesions or flat plaques are the clinical characteristics of basal cell carcinomas. The lesions can grow widely and deeply if left untreated, and the most common form of treatment is surgical excision. The proportion of, and characteristics of, in completely excised basal cell carcinomas in excision biopsy specimens were determined.

Methods.—A total of 268 patients with 353 histologically confirmed basal cell carcinomas were studied. Records were reviewed for the age and sex of patients, sites of basal cell carcinoma, discipline of referring doctor, and histologic growth pattern, macroscopic features, and completeness of excision of the basal cell carcinoma.

Results.—Of the 353 basal cell carcinomas studied, 58 (16%) extended to the margin of surgical excision. Most of these basal cell carcinomas (74%) were located on the head or neck. The majority were flat (81%). An infiltrative growth pattern was present in 36% of incompletely excised basal cell carcinomas. There was a greater tendency for the 8% of recurrent basal cell carcinomas, which were identified from the history or because of histologically identified surgical scarring, to be flat and to show a microscopic infiltrative growth pattern. Incomplete excision was found in 7 (25%) of the recurrent basal cell carcinomas, all of which were on the head, and 5 had an infiltrative growth pattern.

Conclusions.—Incompletely excised basal cell carcinomas are those with the greatest likelihood of recurring. Pathologists should report the microscopic growth pattern of basal cell carcinomas and the completeness of excision. Where possible, wide margins should be used to excise head and neck basal cell carcinomas. Immediate reexcision should be considered in patients with tumors extending to the margin of excision and those on the head and neck, because those are the most likely to recur.

► Although this study presents no new data, the significance of a message sometimes is not appreciated until it is heard over and over again. The study is flawed in that the “surgeons” submitting specimens were a heterogeneous group (plastic surgeons, general practitioners, and “surgical specialists”) and so may have been the patient population; moreover, no comment was made regarding the size of the lesions nor the margins of excision. Nonetheless, as has been reported previously, basal cell carcinomas (BCCs) with an infiltrative growth pattern (clinically, these lesions are flat), BCCs located on the head and neck, and BCCs that were recurrent were more likely to have marginal involvement. Although most of the authors’ conclusions appear reasonable, their recommendation that a questionable lesion be excised with a wide margin seems rather drastic and untenable. Why not perform a punch or shave biopsy before definitive therapy is administered?

Keratins (K8 and K19) as Potential Markers of Recurrent Basal Cell Carcinoma

Introduction.—Trials evaluating keratin expression in basal cell carcinomas (BCCs) have reported conflicting results. A panel of specific monoclonal keratin antibodies was applied to BCCs to determine potential differences in biological behavior between primary untreated and recurrent BCCs.

Methods.—Ten white patients with recurrent BCCs underwent Mohs micrographic surgery. Controls were 10 white patients with primary untreated nodular BCCs. Biopsy specimens of patients with primary and recurrent BCCs were assayed immunohistochemically using a panel of monospecific monoclonal antibodies to keratins.

Results.—All 20 tumors showed homogeneous staining for the basal cell keratins K5, K14, and K17. Nine recurrent BCCs and 2 primary BCCs, expressed K8 or K19. None of the 20 tumors expressed K10, a marker for keratinization, or K18, 1 of the simple epithelial markers.

Conclusion.—The expression of simple epithelial keratins may be biologically important in BCCs and may be related to tumor invasion or changes in epithelial-mesenchymal interactions. This information may be useful diagnostically and prognostically.

► Most dermatologists have noted that the histologic subtype of BCC may vary from the time of biopsy to the final excision or from the time of excision to recurrence. This, of course, could be related to a sampling error or the interpretative abilities of the dermatopathologist. Offering an alternative explanation, Krekels et al. suggest that the biological behavior of the tumor may reflect the expression of simple epithelial keratins, which could affect tumor aggressiveness and tumor invasion.

Regression of Basal Cell Carcinoma by Intralesional Interferon-alpha Treatment Is Mediated by CD95 (Apo-1/Fas)-CD95 Ligand-induced Suicide

Objective.—Basal cell carcinoma (BCC), the most common form of human skin cancer, rarely metastasizes. However, tumor invasion and destruction of surrounding tissues can occur. Regression of BCCs can be induced by intralesional injection with interferon (IFN)-α, although the mechanism of this effect is unknown. Various cell types express the cell surface molecule CD95 receptor (CD95), a member of the tumor necrosis factor receptor superfamily, whereas CD95 ligand (CD95L) is expressed mainly on activated T cells. The role of the CD95 receptor and ligand in the destruction of BCC cells after intralesional IFN-α injection was studied.

Methods.—The study included 15 patients with histologically confirmed BCC, 9 of whom received IFN-α treatment. The dosages were 1.5 x 106 IU 3 times weekly for 3 weeks in 6 patients, 3.0 X 106 IU per injection 3 times weekly in 2 patients, and 3 injections of 1.5 X 106 IU over 1 week in 1 patient. Immunohistochemical studies were performed in skin biopsy specimens from patients with and without IFN-α treatment.

Results.—Tumor cells from untreated patients expressed CD95L, but not the CD95 receptor. This suggests that tumor expansion could occur by preventing attack by activated CD95 receptor-positive lymphoid effector cells. CD95-positive cells incubated on BCC cryosections underwent apoptosis and lysis, which demonstrated the functionality of the CD95L of BCC cells. Expression of both the CD95L and receptor was noted in BCC cells from patients treated with IFN-α. In contrast, the infiltrate around the tumor consisted mainly of CD4+ T cells, which contained CD95 receptor with few cells positive for CD95L.

Conclusions.—In BCCs treated with intralesional IFN-α, apoptosis appears to be the major mechanism of cell death. The findings suggest that BCCs may form because of the capability of tumor cells to lyse CD95- expressing effector T cells via their CD95L. After treatment with IFN-α, BCC cells not only express CD95L but become positive for CD95 receptor. The study’s original hypothesis—that CD95L-positive cytotoxic T lymphocytes have a major lytic effect on BCCs—seems unlikely. Apoptotic cell death in BCCs appears to result from specific interactions between the CD95 receptor and ligand on the tumor cells. In effect the tumor cells commit suicide.

► Carcinogenesis clearly is associated with an increased longevity of cells, which in epithelial cell lines, to some extent, involves an escape from normal apoptotic mechanisms. Buechner et al. suggest that, by mediating apoptosis, intralesional injections of INF-α can induce BCCs to regress. The authors have demonstrated that BCCs do not express the CD95 receptor (Apo-1/Fas), although they do express the ligand. After IFN-α injections, these basal cell tumors are induced to express the receptor, and, when CD95L and receptor are co-expressed, the cells undergo apoptosis, which eventuates in tumor regression.

Successful Surgery of Multiple Recurrent Basal Cell Carcinomas Guided by Photodynamic Diagnosis

Purpose.—Surgical excision is generally the best treatment for epithelial skin tumors. Recurrence rates are higher with the other tested approaches, including cryosurgery, laser therapy, and photodynamic therapy. It can be difficult to determine the extent of basal cell carcinomas (BCCs), particularly in anatomically difficult areas such as the face. Preoperative delineation of tumor tissue may make surgery more effective. The successful use of photodynamic diagnosis as an adjunct to surgery in a patient with facial BCC is reported.

Case.—Man, 56, had a 20-year history of facial BCC. About 200 BCC lesions, most often found on the face and scalp, had been excised over the years. The excisions led to extensive tissue defects that required several large skin grafts for closure. Even with the excision margins tumor free, new tumors appeared close to the former skin grafts. Preoperative photodynamic diagnosis was performed by applying 20% &-aminolevulinic acid (ALA) to the presumed area and surrounding skin. After 4-6 hours of occlusion, the skin was illuminated with a Wood’s lamp. Nineteen widely spread areas of deep red fluorescence were marked and excised for histopathologic examination. Most of the fluorescing lesions appeared near the skin grafts. Even in areas that looked clinically normal, the fluorescence was sharply distinguished. Guided by the fluorescence extension, the tumors were completely removed. Histologic examination showed that all were BCCs.

Discussion.—The use of photodynamic diagnosis as an adjunct to excision in a patient with recurrent facial BCCs is reported. Areas of detectable fluorescence corresponded to lateral histologic tumor extension. This approach may help to conserve tissue in patients with BCCs in anatomic areas of cosmetic and functional importance. The authors have subsequently used photodynamic diagnosis to guide surgery in additional patients with skin tumors.

► In recent years more and more studies have appeared on the use of photodynamic therapy to manage precancerous and cancerous lesions of the skin.’ In this article, the authors employed topical ALA under occlusion in conjunction with Wood’s light irradiation to delineate subclinical tumor deposits of basal cell carcinoma. The fluorescent areas were then excised and submitted for pathologic examination. In these patients with a history of multiple BCCs of the head and neck, the authors’ goal was to preserve healthy tissue. This article brings to mind the question often asked by patients: “Can’t you inject a dye that would show where the cancer is and just cut out that amount of tissue?” This case report has important clinical implications, but many questions are left unanswered, for example, the limitations of the technique for deeply penetrating tumors and how far beyond the fluorescent outline the tumor extends. If tissue conservation is critical, why not combine the technique with Mohs surgery? Long-term follow-up data regarding recurrence are obviously needed.

Photodynamic Therapy of Superficial Basal Cell Carcinoma by Instillation of Aminolevulinic Acid and Irradiation With Visible Light

Introduction.—An effective nonsurgical treatment modality for basal cell carcinoma may be photodynamic therapy with aminolevulinic acid (ALA). Applied epicutaneously, ALA, a precursor of porphyrins, penetrates through abnormal epidermis when in aqueous solution and is metabolized by tumor cells into photosensitizing concentrations of porphyrins. Tumor cells are selectively destroyed by subsequent exposure to photoactivating light. Because of its capacity for specific photosensitization of epithelial cells, ALA is unique compared with other photosensitizers. High recurrences have been found with this therapy, however. This therapy was used differently in a selected patient, in whom 10% ALA in aqueous solution was injected intralesionally and the basal cell carcinoma was subsequently exposed to visible light.

Case Report.—Man, 63, who had a large, 4-cm superficial basal cell carcinoma on the right side of his lower back refused surgical intervention. Intralesion ALA photodynamic therapy was given to the patient. A syringe was used to inject 2 mL of 10% ALA in 0.9% sodium chloride solution intralesionally. After light-shielding occlusion for 4 hours, the typical red fluorescence of porphyrin was noted to be limited to the tumor under Wood’s light examination in a darkened room. The tumor was then exposed to 180-J/cm2 full-spectrum visible light.

Results.—After intralesional instillation of ALA, there was no sign of systemic photosensitivity. The treated lesion developed a thin crust within 3-5 days, and within 4 weeks it completely healed. Only mild hypopigmentation remained after therapy. There was no sign of recurrence after a follow-up period of 54 months.

Conclusions.—Further research should be conducted using other oil-in-water emulsions as vehicles for ALA, or other modes of application, including intralesional instillation of the photosensitizer. Photodynamic therapy using intralesional instillation of ALA may be an effective treatment for basal cell carcinoma.

► Large superficial basal cell carcinomas (BCCs) can be problematic from a management standpoint. Topical 5-fluorouracil (5FU) can be used, but the treatment is prolonged and unpleasant. Intralesional 5FU and interferon are limited by the amount of drug that can be injected. Surgical excision necessitates a wound that may take months to heal or requires grafting. Electro-desiccation and curettage and cryosurgery are not always feasible or effective. Recently, photodynamic therapy (PDT), using either topical or intralesional &-ALA in conjunction with a visible light source (e.g., a slide projector) has been offered as an alternative. Fink-Puches et al. relate their experience with PDT in a patient with a large BCC who refused surgery. The tissue reaction was minimal and healing was complete by 4 weeks. Follow-up after 54 months revealed no recurrence. In their discussion, the authors state that epicutaneous application of ALA is a less reliable means of delivery and inadequate penetration may be observed with thicker lesions, thus resulting in treatment failure.

Photodynamic Therapy of Actinic Keratosis With Topical 5-Aminolevulinic Acid: A Pilot Dose-Ranging Study

Introduction.—Photodynamic therapy (PDT) with topical 5-aminolevulinic acid (ALA) is an effective treatment for basal cell carcinoma, squamous cell carcinoma in situ, and other superficial epithelial cutaneous malignancies. Studies of ALA photodynamic therapy for actinic keratosis (AK) have been reported as well. However, they have included few patients, used a fixed ALA concentration of 20%, and used only visible light. This dose-ranging study assessed the use of topical ALA at various concentrations and laser-delivered visible red light in PDT for AKs.

Methods.—The study included 40 patients, each with 6 clinically typical, previously untreated AKs. The patients were assigned to treatment with ALA concentrations of 0%, 10%, 20%, or 30%. After 3 hours of occlusion, the lesions were checked for fluorescence. They were then treated with an argon pumped dye laser at a wavelength of 630 nm and fluences of 10 to 150 J/cm2. The treatment was assessed for safety and efficacy at intervals of up to 16 weeks. By 8 weeks, there was a total of 218 evaluable lesions.

Results.—After 3 hours of occlusion, the lesions showed weak to moderate red fluorescence, regardless of the ALA concentration used. Clinical improvement was greatest in grade 1 and 2 lesions treated with 30% ALA, with a complete response rate of 61% and partial response rate of 26%; however, all ALA concentrations were effective. The treatment response was better for lesions on the face and scalp (total clearance achieved in 91% of lesions) than on the trunk and extremities (total clearance in 45% of lesions). A maximal response was achieved at a fluence of 10 J/cm2; higher fluences did not yield a higher response. Greater fluorescence was associated with an improved clinical response and greater photoxic effects on the AK and surrounding skin. Hypertrophic grade 3 AKs did not respond well to any ALA concentration. Mild phototoxic effects were noted 3 days to 1 week after PDT, but resolved by 4 weeks. The patients reported mild to moderate burning and stinging during PDT.

Conclusions.—A single session of PDT using topical ALA is an effective treatment for typical AKs. At an ALA concentration of 10% to 30%, good lesion clearance with good patient tolerance can be obtained. This treatment is more effective for lesions on the face and scalp than for lesions on the trunk and extremities; it is not effective for hypertrophic AKs.

► Extensive AKs may be managed by a number of methods, including topical 5-FU and medium-depth chemical peeling. In this study the authors offer yet another modality, photodynamic therapy. Topical ALA under occlusion was used in combination with an argon pumped dye laser to treat patients with extensive AKs. As with other modalities used to treat AKs, lesions of the face and scalp responded much better (91 % clearance rate) than those of the trunk and extremities (45% clearance rate), probably because lesions in the latter locations tend to be thicker, thus limiting penetration of the ALA. Future studies need to address several issues, including relapse rate, cost-effectiveness, and pretreatment with keratolytic agents to increase the response of thick lesions.

Chemotherapy for Disseminated Actinic Keratoses With 5-Fluorouracil and Isotretinoin

Background.—Patients with actinic keratoses (AKs) in photodamaged skin are commonly treated with topical 5-fluorouracil (5-FU). To achieve a high cure rate, 5-6 weeks of treatment may be necessary; the results are not so good when the AKs are located on the hands and forearms. Combination therapies have been tried to improve the efficacy of treatment for AKs. An experience with low-dose oral cis-retinoic acid and topical 5-FU for the treatment of disseminated AKs is reported.

Methods.—Twenty-seven patients at 2 sites were treated. All had severe photodamage with numerous, sometimes confluent AKs. All were treated with 5% 5-FU applied to the photodamaged area twice a day plus oral isotretinoin, 20 mg/day. The treated areas included the dorsal hands, forearms, face, and scalp. Treatment continued for a median of 21 days.

Results.—Twenty-two patients had complete regression of palpable AKs, and the other 5 had nearly complete regression. By the end of the second week of treatment, the patients had redness and tenderness of the keratoses, followed by exudative crusting and painful erosions. Pruritus, burning, and tenderness were intense, but resolved within a few weeks after the end of treatment. There were no serious problems with mucocutaneous signs typically accompanying higher doses of oral isotretinoin, and serum laboratory levels were unchanged. At a median follow-up of 12 months, 3 patients had undergone podophyllin therapy for a few AKs.

Discussion.—The effective use of combined 5-FU and isotretinoin for patients with disseminated AKs on photodamaged skin is reported. This treatment causes severe local side effects, but in most patients dense, thick AKs are rapidly eliminated. The combination approach, compared with either agent alone, greatly reduces the duration of treatment. Further study is needed.

► Thick AKs of the hands and arms may be quite resistant to topical 5-FU, even when prolonged treatment is prescribed. Combining topical 5-FU with occlusion and topical tretinoin increases its efficacy, but is not always successful. In this small study (27 patients), the authors used oral isotretinoin, at a fixed daily dose of 20 mg, in conjunction with 5% 5-FU cream, applied twice daily. The median period of treatment was 3 weeks. Most patients had complete regression of their lesions. The only significant side effect was severe local irritation. Median follow-up time was 12 months. In only 44.4% of the patients were the lesions located on the hands and arms. No mention was made whether these patients had failed conventional topical 5-FU treatment. Although the combination of topical 5-FU and oral isotretinoin has potential therapeutic usefulness, a number of issues need to be addressed, such as the added cost of oral isotretinoin and perhaps additional laboratory studies. Should this therapy be reserved for patients failing conventional topical 5-FU treatment, and should it be used only for lesions in certain locations (hands, arms)? When treating the face, could one use the 1 % concentration of 5-FU? Was the follow-up period long enough to determine the true relapse rate?

An Open Study to Assess the Efficacy and Safety of Topical 3% Diclofenac in a 2.5% Hyaluronic Acid Gel for the Treatment of Actinic Keratoses

Background.—If left untreated, actinic keratoses can lead to invasive squamous cell carcinoma. However the current topical treatments can be painful and cause skin irritation. The current study evaluated the utility and adverse effect profile of topical 3% diclofenac in a 2.5% hyaluronic acid gel in treating actinic keratoses.

Methods.—The subjects were 29 white patients between the ages of 18 and 82 who had at least 1 actinic lesion. The following were exclusion items: (1) current usage of systemic corticosteroids, systemic or topical retinoids, antineoplastic drugs, cyclosporine, or systemic nonsteroidal anti-inflammatory drugs (NSAIDs); (2) past or present usage of etretinate, isotretinoin, or similar drugs; (3) hypersensitivity to NSAIDs or diclofenac; (4) coexistent skin conditions (such as basal cell carcinoma); (5) pregnancy; (6) tanning; and (7) the use of cosmetics that could interfere with test results. Target lesions were chosen, and subjects were instructed to apply 1.0 g (in a premeasured applicator) of the study gel twice daily for up to 180 days. Once lesions cleared, treatment stopped. Lesions were assessed at baseline and after 60,120, and 180 days of treatment, or at 30 days after lesions cleared. Physicians assessed the treatment response on a 7-point scale, and efficacy and tolerance were assessed on a 4-point scale.

Findings.—Treatment lasted a median of 62 days (range 33-176 days), and 27 of the patients were reassessed 30 days after lesions cleared and drug treatment ceased. AH but 1 subject (26 of 27, 96%) had at least marked clinical improvement, and 22 of 27 subjects (81 %) had a complete response. The mean time to achieve a complete response in these 22 subjects was 94 days (range 56-194 days). Skin irritation was reported by 21 of 29 subjects (72%). Furthermore, 7 of the subjects had an initial accentuation of their lesions, but by follow up all lesions in these subjects had complete responses. Finally, contact dermatitis at the treatment site required discontinuation of treatment in 7 subjects. The reaction resolved after 1-2 weeks, and at follow-up 5 of these patients had a complete response and 2 had marked improvement.

Conclusions.—The efficacy of topical 3% diclofenac in 2.5% hyaluronic acid gel was similar to that of current topical treatments with masoprocol or fluorouracil. However, side effects with the study gel were generally mild, and treatment was successful despite their occurrence. The mechanism of action of diclofenac with hyaluronic acid may be a reduction in angiogenesis in the lesions. Further research is needed to confirm the utility of this combination in a double-blind randomized study.

► A double-blind randomized trial is warranted to clearly establish the efficacy of this formulation and the durability of the treatment response. Histologic evaluation of the treated lesions was conspicuously absent.

Analysis of Tumor-infiltrating Lymphocytes in Cutaneous Squamous Cell Carcinoma

Background.—Squamous cell carcinoma (SCC) of the skin is a common cancer with well-recognized environmental causes. However, relatively little is known about the possible host factors; evidence suggests that the immune system has important effects on the development and course of cutaneous SCC. It has been suggested that exposure to ultraviolet (UV) rays causes changes in the host immune system that prevent it from eliminating neoantigens on UV-irradiated cells. The tumor-infiltrating lymphocytes (TILs) of SCC lesions were studied in detail.

Methods.—The study included specimens from 21 patients with confirmed squamoproliferative disorders: 11 invasive SCCs, 4 SCCs in situ, and 6 actinic keratoses. All were studied using immunohistologic techniques, molecular biologic analysis of T cell clonality, and in vitro functional assessment of cytolytic potential.

Results.—A variably dense inflammatory infiltrate was found in all specimens, usually in the stroma adjacent to and at the interface with tumor cell nests, with additional inflammatory cells within the tumor cell foci. The main components of the infiltrate were T cells and macrophages, with few or no natural killer cells, B cells, granulocytes, and T cell receptor-&+ cells. Each specimen showed Langerhans cells and indeterminate cells; the concentration of Langerhans cells was unrelated to tumor size or invasion. The ratio of CD4 to CD8 T cells ranged from 2 to 4, and 20% to 40% of CD3+ TILs were CD8+. Most of the CD8+ cells had a phenotype consistent with major histocompatibility complex-restricted cytotoxic T lymphocytes. They were also HLA-DR positive, suggesting in vivo activation. In all cases, the TILs appeared to be polyclonal. On functional chromium 51-release assays, TILs derived from SCC lesions appeared to have cytotoxic effects on autologous tumor cell targets.

Conclusions.—This study demonstrates the presence of a polyclonal inflammatory reaction containing cytotoxic T lymphocytes in and around SCCs. Though no specific antitumoral activity is demonstrated, functional studies suggest that the TILs have specific antiautologous tumor cell cytotoxicity. These findings may help to explain the relatively slow growth of SCCs in healthy subjects vs. the more aggressive growth in patients with immune system impairment.

► Based on the observation that there is an increased incidence of SCC of the skin in immunosuppressed patients and in patients receiving oral psoralen and UVA irradiation (which induces a state of immunosuppression and is, in itself, carcinogenic), there is little question that the immune system is important in the pathogenesis of SCC. It has long been recognized that biopsy specimens from SCCs commonly contain an inflammatory infiltrate in which T lymphocytes predominate. These investigators carry this observation a step further by looking at the subsets of inflammatory cells and trying to determine if these cells are activated and capable of destroying neoplastic cells. Using a variety of assays, they found that both T-helper and T-suppressor cells are present, the former predominating and the latter being cytotoxic. They also found that the T cells were activated and capable of killing autologous tumor cells. This study adds further support to the concept that the immune system plays an important role in the pathogenesis of SCC and perhaps could be manipulated to prevent the development of tumors in high-risk patients.

Expression of PACE4 in Chemically induced Carcinomas Is Associated With Spindle Cell Tumor Conversion and Increased Invasive Ability

Introduction.—An important feature of malignancy is the gradual acquisition of a more aggressive phenotype, also described as tumor progression. This phenomenon in squamous cell carcinoma (SCC) is commonly associated with partial or total loss of squamous differentiation and the appearance of a more anaplastic or spindle cell morphology, along with a more invasive and metastatic behavior. Differential display was used to identify genes in addition to those known to be involved in the conversion of a SCC to a more aggressive, poorly differentiated spindle cell tumor. A differentiated proprotein convertase, PACE4, was identified.

Findings.—A murine model of human SCC progression was used. About 50% of mouse spindle cell carcinomas displayed moderate to high levels of PACE4 expression; only one fourth of well-differentiated mouse SCCs were moderately to weakly stained for PACE4 expression, using immunohistochemical staining. The exogenous overexpression of PACE4 cDNA in mouse SCC cells of low invasive ability caused enhanced tumor cell invasiveness that was not present in parental or mock-transfected SCC cells. The PACE4-transfected cells became able to process prostromelysin 3 into its active enzyme form.

Conclusion.—Together, these findings demonstrate that up-regulation of PACE4 expression is related to SCC conversion to spindle cell carcinoma and indicate that activation of essential PACE4 substrates, such as the metalloproteinase stromelysin 3, is needed for tumor cell invasion.

► In an animal model of SCC progression, Hubbard et al. were able to alter the invasive characteristics of these tumors by manipulating the expression of PACE4. PACE4 is one of a number of calcium-dependent serine proteases (or proprotein convertases) that activate prostromelysin 3. In this study, the expression of PACE4 directly correlated with invasive ability of SCCs. Activation of this metalloproteinase may thus enable the expressing cells to penetrate into the surrounding stroma, creating a more aggressive phenotype.

p21 WAF1/CIP1 Expression in Non-Melanoma Skin Tumors

Introduction.—The p53 tumor suppressor gene is a major subject of research on cancer, and mutations of the gene are seen as early events in non-melanoma skin cancer. A universal inhibitor of Gl cyclin-dependent kinase, p21WAF1/CIP1 protein is induced by p53-dependent and -independent pathways. Immunohistochemical analyses were performed to examine the role of the p21 protein in benign, premalignant, and malignant keratinocytic tumors.

Methods.—The study material consisted of formalin-fixed, paraffin-embedded sections from 10 cases of seborrheic keratosis, 5 of keratoacanthoma, 10 of actinic keratosis, 10 of Bowen’s disease (BD), 30 of squamous cell carcinoma (SCC), and 8 of basal cell carcinoma (BCC). Sections were incubated with the primary antibody, a mouse monoclonal antibody that recognizes human p21. Expression of p53 protein was also examined in selected cases with positive cells.

Results.—Positive staining was observed in normal skin, in maturing cells in the centers of sebaceous glands suspected to be in apoptosis. Eight of the 10 cases of seborrheic keratosis and 4 of 5 cases of keratoacanthoma stained positively for p21 protein and exhibited a similar pattern of staining. Nine of 10 cases of BD were positive, with 8 showing a differentiation pattern and 1 a mixed pattern of staining. Seven of 10 cases of actinic keratosis were positive for p21, and their staining pattern was similar to that of BD (a differentiation pattern in 6 and a mixed pattern in 1). No positive cells were seen in the 8 cases of BCC, but 25 of 30 cases of SCC were positive. More than 50% of cells stained ( +++) in 8 SCC cases, a finding rarely noted in other tumor types. Twenty-one of the p21-positive tumors were also stained with anti-p53 antibody. The location and grades of p21 and p23 expression in these tumors were generally inconsistent, indicating that the pathway for p21 expression in skin tumors is independent of p53.

Conclusion.—Although p21 protein was expressed in benign skin tumors, premalignant lesions, and SCC, positive staining was not detected in BCC. Identification of the exact mechanisms of p21 expression and its biological role in skin tumors may require analysis of p21 and p53 genes in identical tumors.

► This survey of non-melanoma cutaneous tumors describes the expression of p21. As a downstream effector of p53, p21 negatively regulates cyclin-dependent kinases, resulting in G1 arrest. Ahmed, et al. demonstrate that normal skin typically expresses p21 only within sebaceous glands. In contrast, seborrheic keratoses variably express p21 within the basal layer and lower epidermis (proliferation pattern) and occasionally as high as the stratum granulosum (differentiation pattern). Keratoacanthomas, Bowen’s disease, and actinic keratoses predominantly demonstrated a differentiation staining pattern, while squamous cell carcinomas demonstrated a proliferation pattern. Basal cell carcinomas failed to exhibit any significant degree of p21 staining.

Squamous Cell Carcinoma of the External Ear: A Clinico-Pathological Study of 8 Japanese Patients

Background.—Most squamous cell carcinomas (SCCs) of the external ear are believed to be caused by exposure to sunlight. Sun-induced SCCs of the external ear are more likely to metastasize than those at other sites. The clinical and histologic findings in a series of SCCs of the external ear are reported.

Methods.—The study included 8 Japanese patients with primary cutaneous SCCs of the external ear. After confirmatory incisional biopsies, all patients underwent full-thickness resections of the lesions with lateral margins of at least 5 mm. Depth of tumor invasion was assessed according to Clark’s classification; Clark’s level V was used to denote invasion beyond the full thickness of the dermis. Maximal tumor thickness was also measured with an ocular micrometer. The incidence of DNA aneuploidy was used as a marker of malignant tumor cells.

Findings.—Four tumors were on the uppermost part of the helix and 4 on the middle part of the antihelix. Mean maximum tumor width was 17 mm. The patients were followed for a mean of 75 months. Though none had nodal metastases at presentation, cervical lymph node involvement later developed in 1 patient. Five lesions had invaded to the reticular dermis, 3 to the perichondrium, and 1 to the posterior skin of the auricle. As measured by ocular micrometer, the vertical tumor thickness averaged 4.2 mm, with 4 lesions being less than 4 mm thick and none being thicker than 10 mm. One tumor invaded the perichondrium even though it was only 1.1 mm thick. The resection margins were deemed sufficient in every case. On flow cytometry, one fourth of the cancers showed DNA-aneuploidy; the rest had a diploid pattern only.

Conclusions.—Squamous cell carcinomas of the external ear tend to occur on the uppermost part of the helix or the middle part of the antihelix. Even thin tumors may invade to deep dermis near the underlying auricular cartilage. To prevent auricular SCCs, the implicated areas of the auricle should be included in skin protection. Even for thin lesions, surgery should include resection of the underlying cartilage. The authors report excellent outcomes with total-thickness resection of the pinna, including a margin of 5-10 mm.

► Squamous cell carcinomas of the external ear, when compared to other anatomical locations, excluding the lip, may be at higher risk for metastasizing. In this article the authors review their experience with 8 patients, only 1 of whom developed regional metastases. Although the number of patients is small, several important observations can be made. Aneuploidy, which often correlates with a greater proclivity for metastasis, did not appear to be a factor in these patients. Although Breslow thickness and Clark’s level have been shown to be important prognostic variables, they are not infallible. There were several patients with SCCs thicker than that of the 1 patient developing metastases. Finally, the authors’ suggestion that cartilage routinely be excised, especially for lesions of the anthelix, appears rather radical. Using the Mohs technique, many of these tumors could be removed with preservation of the perichondrium, with the wound being allowed to heal by secondary intention with an excellent cosmetic result.

Merkel Cell Carcinoma: Analysis of Clinical, Histologic, and Immunohistologic Features of 132 Cases With Relation to Survival

Objective.—Merkel cell carcinomas (MCCs) are potentially aggressive, commonly recurrent, and metastatic skin cancers. No large histologic and immunochemical studies have been conducted. Results of a retrospective study of 132 MCCs diagnosed between 1970 and 1989 are presented.

Methods.—Tumor size, cell size, mitotic rate, vascular proliferation, vascular invasion, fibrosis, stromal mucin, solar elastosis, epidermal involvement, squamous differentiation, bowenoid change, nuclear molding, growth pattern, lymphocytic infiltration, crush artifact, necrosis, ulceration, apoptosis, involvement of reticular dermis or subcutaneous adipose tissue, and spindle cell change were recorded for hematoxylin-and-eosin-stained and immunohistochemically stained MCC tumor sections. Cell size was determined from appearance of hematoxylin-and-eosin-stained sections. The small cell variety, with diameters equal to or less than 14 pm, contained scant cytoplasm and little organoid growth. Intermediate cell types, with diameters greater than 14 pm, had more gray cytoplasm and more organoid growth sections. Tumors were divided into 3 groups based on mitotic rate per high-power field (HPF). Group 1 had 5 or fewer mitoses, group 2 had 6-10 mitoses, and group 3 had more than 10 mitoses. Mitotic growth patterns included trabecular, organoid, ribbonoid, and diffuse or nodular. Cells were immunochemically stained with cytokeratin, neuron-specific enolase, BER-EP4, epithelial membrane antigen, chromogranin, Leu-7, and carcinogenic antigen.

Results.—The 132 patients (64 females), age 15-94, had 57 tumors on the head or neck, 33 on the upper extremity, 13 on the lower extremity, 19 on the buttock or thigh area, 9 on the truck, and 1 in the axilla. Patients were followed for 1-10 years. There were 57 alive without tumor, 28 died because of tumor, 36 died of other or unknown causes, and 10 were lost to followup. Tumor was recurrent in 37, and metastatic, mainly to the local lymph nodes, in 43 with 18 metastases occurring after local recurrence. Histologically the most significant features were cell size, mitotic rate, and tumor size. All tumors stained for neuron-specific enolase and cytokeratin, 48 of 50 stained for BER-EP4, 31 of 41 stained for chromogranin, and 17 of 17 stained for epithelial membrane antigen. Sensitivity, specificity, false-positive rate, false-negative rate, and predictive value for tumor size were 40.7%, 88.0%, 35.3%, 26.7%, and 71.4%; for cell size were 74.1%, 94.0%, 13.0%, 13.0%, and 87.0%; and for mitoses were 81.5%, 92.0%, 15.4%, 9.8%, and 88.3%. Whereas multivariate analysis showed no significant risk factors associated with the small cell variant, mitosis was a significant risk factor for the intermediate cell size. MCC, thought to be of epithelial origin, carries a mortality rate of at least 25%.

Conclusion.—Mitotic rate, cell size, and tumor size have a significant-predictive value for survival in patients with MCC.

► Merkel cell carcinoma is an uncommon tumor thought to be of epithelial origin. Its course and prognosis are variable. Very few large studies have addressed the factors that affect prognosis. In this retrospective study of 132 cases, the authors analyzed, in detail, the histologic and clinical features that impacted survival. Cell size, tumor size, and mitotic rate were the three variables that most influenced prognosis. Of these, cell size seemed to have the greatest impact. Unfortunately, this study did not address if or how these variables might influence the clinician’s decision regarding treatment, and if so, whether the prognosis could be altered.

Merkel Cell Carcinoma: Comparison of Mohs Micrographic Surgery and Wide Excision in Eighty-Six Patients

Introduction.—An aggressive malignancy with a poor prognosis, Merkel cell carcinoma has local persistence in up to 44% of surgically treated patients, requiring multiple reexcisions. Anastomosing cords of undifferentiated small cells are the histologic characteristics. Other characteristics include abundant mitoses, foci of tumor necrosis, and vascular and lymphatic invasion. Treatment recommendations are based on individual cases because of the rarity of the tumor, but the current standard of care is wide excision with 2.5-3 cm margins and resection of any palpable regional lymph nodes. The use of Mohs micrographic surgery was evaluated for this aggressive neoplasm.

Methods.—A total of 85 patients with Merkel cell carcinoma who were studied retrospectively to establish rates of local persistence and the development of regional metastasis after standard surgical excision. On a subgroup of 13 patients treated with Mohs surgery, detailed follow-up was available. The mean follow-up time was 60 months for the entire group and 36 months for the group treated with Mohs surgery.

Results.—High rates of local persistence (13 of 41 [31.7%]) and regional metastasis (20 of 41 [48.8%]) were associated with standard surgical excision. One Mohs-treated patient had very extensive disease and died shortly after surgery. Local persistence of disease was found in 1 of the remaining 12 (8.3%) Mohs-treated patients with histologically confirmed clearance. A second Mohs excision was performed on this patient, and he has remained disease free for 84 months. In 4 of these 12 patients (33.3%) regional metastasis occurred. These included 4 of the 8 patients treated without postoperative radiotherapy. None of the 4 patients treated with radiotherapy after Mohs surgery had regional metastasis.

Conclusions.—When compared with standard surgical excisions, Mohs surgery compares favorably. In transit metastases and nodal disease may be further reduced with radiotherapy after Mohs surgery.

► Because Merkel cell carcinoma (MCC) frequently gives rise to satellite lesions and regional and systemic metastases, there is controversy over the role of Mohs surgery (MS) in its management. Many of these tumors arise on the head and neck, and it is not always possible to excise them with wide margins. In these patients, MS has a role. However, the use of surgery alone to treat these tumors is doomed to failure in a significant number of patients. Adjunctive radiation therapy delivered to the wound bed, intervening lymphatics, and regional nodes is essential; however, even with radiation therapy recurrence is possible. The modifications used for MS when treating MCC are similar to those suggested for melanoma, i.e., start with a margin (3 mm) of clinically normal skin when debulking, and take a wider margin (at least 3 mm) than usual when obtaining the Mohs layer. I also prefer the use of permanent sections. Although I agree with the concept that wider margins are necessary, it is unclear why one author removed an additional 1-cm margin of tissue once a cancer-free wound had been achieved.

Mohs Micrographic Surgery for the Treatment of Dermatofibrosarcoma Protuberans: Results of a Multiinstitutional Series With an Analysis of the Extent of Microscopic Spread

Introduction.—An uncommon locally invasive soft-tissue tumor, dermatofibrosarcoma protuberans accounts for less than 0.1% of all malignancies, but it has extensive penetration of cutaneous and subcutaneous tissues with patients requiring multiple resections. The treatment of choice is considered to be wide excision down to the fascia. However, the primal width of excision around the primary tumor has not been defined prospectively. Mohs micrographic surgery has been generating interest as a treatment because of the high rates of recurrence with standard excisions. A review of patients with primary and recurrent dermatofibrosarcoma protuberans treated with Mohs micrographic surgery was conducted.

Methods.—A review was conducted of 58 patients with primary and recurrent dermatofibrosarcoma protuberans who were treated with Mohs micrographic surgery. Recordings were taken of the macroscopic and microscopic extent of the tumors. The efficacy of various excision widths was estimated.

Results.—There was an overall local recurrence rate of 2%, a zero recurrence rate for primary tumors, and a 4.8% recurrence rate for recurrent tumors. For example, 1 patient with a twice-recurrent dermatofibrosarcoma protuberans had another recurrence after Mohs micrographic surgery. No distant or regional metastases were seen. Tumors in only 12 (20.7%) patients were completely excised in a single procedure, and most patients required several staged excisions for complete tumor clearance, with the average number required being 2.4. Microscopic postoperative tumor size ranged from 1.8 X 1.0 cm to 35 X 40 cm, and macroscopic tumor size ranged from 0.3 X 0.6 cm to 30 X 20 cm. The entire microscopic extent of the tumor would be cleared with a given calculated width of excision around the macroscopic tumor. The calculated widths are as follows: 1 cm around the primary tumor would leave microscopic residual tumor in 70.7%, 2 cm would leave 39.7%, 3 cm would leave 15.5%, and 5 cm would leave 5.2%. The microscopic extent of some tumors would not be cleared even with an excision width of 10 cm.

Conclusion.—There is a low recurrence of primary and recurrent dermatofibrosarcoma protuberans after treatment by Mohs micrographic surgery because of the technique’s ability to detect and excise microscopic tumor elements in even the most asymmetric tumors. In this series, the extent of the tumors was extensive and eccentric, and a standard wide excision would not have eradicated all the disease in many of these patients. It is necessary to achieve tumor-free margins in all directions and assess the entire perimeter of the tumor for microscopic extension, whatever type of surgery is chosen to treat dermatofibrosarcoma protuberans.

► This is a multicenter series of 58 patients with primary and recurrent DFSP treated with Mohs micrographic surgery. The recurrence rates of DFSP presented in this series are significantly lower than those in published series using standard surgical excision. The predicted rate of inadequate incision using traditional surgical margins ranges from 71% with a 1 cm margin to 15.5% with a 3 cm margin. Even a 10 cm margin would not have cleared all the tumors. Because of the eccentric growth of DFSP, the Mohs micrographic technique is a logical choice for the treatment of DFSP. Immunohistochemical staining with CD34 may improve the effectiveness of surgical margin evaluation, whether Mohs micrographic surgery or traditional surgical excision is used.

Paget’s Disease of the Groin: Report of Seven Cases

Introduction.—The first described cases of Paget’s disease (1874) involved the nipple in women with concomitant carcinoma of the breast. Fifteen years later, a patient with scrotal and penile involvement represented the first reported case of extramammary Paget’s disease. The patients described here had Paget’s disease of the groin, a very infrequent diagnosis.

Methods.—The 7 patients were seen at Mayo Clinic over a 25-year period (1970 to 1995). All were followed as outpatients at the clinic or through individual contact. In a retrospective study, the cases were evaluated for patient gender and age, associated location of extramammary Paget’s disease, type of surgery, histology, associated tumors, complications, and survival. All patients underwent wide local excision with curative intent. Re-excision was performed when needed to achieve free margins. Histologic examination confirmed the diagnoses.

Results.—The patients, all men, had a mean age of 73 years; duration of symptoms ranged from 2 months to 10 years. Most patients complained of rash and pruritus, and 4 had associated scrotal or penile involvement. After surgery, 3 patients with extensive lesions had recurrent disease. Two of these patients had previously had an associated malignancy (prostatic or renal cell carcinoma); the third died of Paget’s disease with multiple pulmonary metastases.

Discussion.—A 1985 review reported only 27 published cases of Paget’s disease of the groin. The disease usually affects elderly men, many of whom have involvement of the scrotum and penis. Patients typically have a nonhealing eczematous lesion that can resemble other types of skin pathologies. A wide local excision with free margins is effective treatment in most cases. Because metastatic disease will develop if complete control is not achieved, patients require long-term follow-up.

Paget’s Disease of the Vulva: Report of Five Cases

Introduction.—Paget’s disease of the vulva is a relatively rare disease that usually affects postmenopausal Caucasian women. Findings, treatment, and outcome were reported for 5 cases.

Case Reports.—The patients were women aged 47, 61, 66, 80, and 84 years. All reported a history of vulvar pruritus ranging in duration from 8 months to 5 years. Examination revealed either vulvular lesions or masses in each case. Treatment consisted of surgery in the 3 younger women. The 2 older women were in poor medical condition; one received intratumoral injections of bleomycin, then carbon dioxide laser treatment, and the other received topical 5-fluorouracil ointment. The 84-year-old patient (the one who received bleomycin) improved for 7 months, then was lost to follow-up. Two patients died of disease and 2 were doing well at short-term follow-up.

Discussion.—Paget’s disease of the vulva involved the labium majus in 4 of these patients, the labium minus in 2, the clitoris in 1, and the pubic and perineal area in 1. Invasive disease is associated with a poor prognosis. In 25% of patients, Paget’s disease is related synchronously or asynchronously with breast, gynecologic, urologic, or gastrointestinal malignancies. None of the patients in this series, however had another primary malignant neoplasm. Surgery is the treatment of choice. Local recurrence (usually in situ lesions) remains a problem. Patients should, thus, be monitored indefinitely for invasive disease or recurrence.

► Both of these articles on extramammary Paget’s disease (Abstracts 18-30 and 18-31) stress the importance of careful follow-up, given the known propensity of the disease for local recurrence. An underlying adenocarcinoma is uncommon but must be searched for.

Cutaneous Angiosarcomas With a Starry-sky Pattern

Background.—Angiosarcoma typically develops in the skin and superficial soft tissue of the head and neck of elderly persons. They also develop on the extremity in patients with acquired or congenital chronic lymphedema. A subgroup of epithelioid angiosarcomas with a prominent starry-sky histologic pattern has been identified on the extremities of patients with chronic edema and on the scalp of elderly persons.

Methods and Findings.—The 4 affected patients were 3 men and 1 woman aged 52-90 years. All died within 5 years of the diagnosis of angiosarcoma, 3 within 30 months. Histopathologic analysis and immunohistochemical stains were performed on all 4 neoplasms. The immunohistochemical staining patterns were similar in all tumors. All were positive for vimentin, factor VIH-related antigen, and CD34. The other immunohistochemical stains, including cytokeratin (which may be positive in other epithelioid angiosarcomas), were negative.

Conclusions.—A distinctive starry-sky histologic pattern has been identified in a group of epithelioid angiosarcomas. Awareness of this pattern may be important in the early diagnosis of this rapidly fatal tumor.

► A starry-sky pattern with numerous phagocytic macrophages is seen most often in lymphomas, especially Burkitt’s and non-Burkitt’s-type lymphoma, mantle cell lymphoma, and lymphoblastic lymphoma. This pattern is nonspecific and may occur whenever there is a high rate of cellular proliferation along with a high rate of cellular apoptosis and/or necrosis.