Prevalence and Gross Morphologic Features of Perianal Melanocytic Nevi

Key Points

Question  How common are perianal melanocytic nevi?

Findings  In a prevalence study of 236 adults attending 1 dermatologist’s practice, perianal nevi of any size were evident in nearly half of patients. In whites, presence of at least 1 perianal nevus was significantly associated with history of any atypical nevi removed, atypical nevus pattern, and at least 1 extant atypical nevus anywhere on the skin.

Meaning  Perianal nevi may be associated with significant melanoma risk factors, and consideration should be given to examination of the perianal area during melanoma screening and surveillance.

Importance  The presence of numerous melanocytic nevi is a significant melanoma risk factor, but there are scant data related to prevalence and morphologic features of melanocytic nevi in the perianal area. The prognosis of perianal melanoma is often dismal because of hidden location and diagnosis delay.

Objective  To determine prevalence and morphologic features of perianal melanocytic nevi.

Design, Setting, and Participants  This study was conducted for 11 months during 2013 and 2014 at an outpatient dermatology clinic in Chicago, Illinois, with a convenience sample of 236 adults (men and women of all races, ≥18 years) presenting to 1 dermatologist for melanoma and/or skin cancer screening or surveillance. The analysis was conducted during April through July 2015.

Main Outcomes and Measures  Prevalence and morphologic features of perianal nevi according to race/ethnicity, sex, and age.

Results  Of 236 participating patients, 219 were non-Hispanic white; 4, Hispanic white; and 13, nonwhite. Patients included 138 men and 98 women, ages 23 to 84 years (median age, 55 years; mean [SD], 53 [15] years). Perianal nevi of any size, at least 2 mm in diameter, and at least 5 mm in diameter were evident in 48.9% (107 of 219), 39.7% (87 of 219), and 5.5% (12 of 219) of non-Hispanic whites, respectively; 50.0% (2 of 4), 0 (0 of 4), and 0 (0 of 4) of Hispanic whites, respectively; and 38.5% (5 of 13), 38.5% (5 of 13), and 0 (0 of 13) of nonwhites, respectively. In non-Hispanic whites, the presence of at least 1 perianal nevus was significantly associated with history of atypical nevus excision (odds ratio [OR], 2.9; 95% CI, 1.5-5.7); and extant findings of at least 1 atypical nevus (OR, 2.2; 95% CI, 1.3-3.9); atypical nevus pattern (≥20 nevi that were ≥2 mm in diameter), plus at least 5 nevi that were 5 mm or greater in diameter (OR, 1.8; 95% CI, 1.1-3.1); and at least 4 atypical nevi 5 mm or greater in diameter (OR, 1.9; 95% CI, 1.1-3.3).

Conclusions and Relevance  In this study, perianal melanocytic nevi were common and were associated with prominent and atypical nevi elsewhere. The perianal area is worthy of attention during melanoma screening and surveillance.

The perianal area—encompassing the anal margin, gluteal cleft, and perineum—occupies about 0.5% of the total cutaneous surface. Perianal melanoma is rare.^1,2 accounting for 2% to 4% of all malignant anorectal neoplasms and 0.4% to 1.6% of all melanomas.^3-5 The US incidence of perianal melanoma is increasing in both men and women,⁵ estimated to be 1.7 cases per million people per year.^5-8 Perianal melanoma is associated with a poor prognosis,^7,9,10 likely related to diagnosis delay. Five-year survival for perianal melanoma is 3% to 22%; median survival for recurrent or metastatic disease is 10 months or less.⁷

The perianal area is often ignored during cutaneous melanoma (CM) screening and surveillance, perhaps because of a dearth of information regarding prevalence of nevi in this site. Our study goal was to determine prevalence and morphologic features of melanocytic nevi in the perianal area of adults undergoing CM and/or skin cancer screening and surveillance.

The study population consisted of patients presenting to a senior dermatologist (A.R.R.) in the outpatient Dermatology Service of Rush University Medical Center (RUMC), Chicago, Illinois, between November 25, 2013, and October 16, 2014. Patients provided verbal and written informed consent to participate. Consent procedures were approved by the RUMC institutional review board on Human Studies. Patients were not compensated for their participation in this study.

Adults (≥18 years old) were included regardless of ethnicity, sex, or pregnancy status but were excluded if the medical encounter was related to a perianal lesion or if exigencies precluded examination or data collection. Race/ethnicity was determined by self-report and observation according to classifications outlined by the US Office of Management and Budget and the US Census Bureau. Non-Hispanic white was used to describe people of European heritage, Hispanic white for people with white skin and self-described as Hispanic, and nonwhite for all other races/ethnicities. Data acquisition occurred when documentation assistance was available (by A.S. or A.B.).

For each patient, the most urgent reason for medical encounter was recorded. Data collected included personal and family history of CM, number of nevi removed, history of atypical (dysplastic) nevi removed based on available report(s), skin phototype, and natural hair color at age 20 years.

The following were recorded if available from current or prior examination(s): number of melanocytic nevi of any size; number of nevi at least 2 mm in diameter and at least 5 mm in diameter; and presence of extant atypical melanocytic nevi (AMN), a largest nevus at least 8 mm in diameter, red hair, and freckling on sun-exposed sites. We defined AMN as nevi at least 5 mm in diameter, plus 1 or more of the following: pigmentation variegation, fried-egg pattern, irregular and/or ill-defined margin, and/or black pigmentation in toto. Typical nevus pattern was defined as the presence of fewer than 20 nevi at least 2 mm in diameter, fewer than 5 nevi at least 5 mm in diameter, and no extant AMN. If criteria for typical nevus pattern were not met, the pattern was designated atypical.

“Perianal” was defined as encompassing anal margin, perineum, and gluteal cleft. “Anal margin” refers to area of skin folds extending from external anal sphincter to nonfolded gluteal cleft skin. “Gluteal cleft” refers to skin surrounding the anal margin, visible when buttock cheeks are physically spread in the prone position. “Perineum” defines the area between the anal margin and posterior vulvar vestibule in women and posterior scrotum in men.

The perianal area was examined by A.R.R. plus A.S. or A.B. Well-circumscribed, flat pigmented lesions in the perianal area were presumed junctional nevomelanocytic nevi, lentigo simplex, typical junctional nevi, or atypical junctional (dysplastic) nevi. Raised, smooth-surfaced, pigmented plaques or papules were presumed compound typical or atypical (dysplastic) nevi. Well-circumscribed, raised, skin-colored, smooth-surfaced, dome-shaped papules or plaques were presumed intradermal nevi (excluding skin tags and lipofibromas). Perianal “freckling” (pale tan, ill-defined macules) was recorded as present or absent.

The following were collected for each lesion, including patient prior awareness: location; longest diameter; perpendicular greatest diameter; maximum height; degree of pigmentation variegation; lesion color(s); border irregularity and demarcation; and the presence of halo depigmentation (without Wood’s light), scaling, ulceration, or erosion. Measures were obtained with a millimeter scale or template.

Nevus topography was graded flat, slightly raised, or markedly raised to visual inspection. Pigmentation variegation was graded as none (1 shade and/or hue), 2 shades/hues, 3 shades/hues, or more than 3 shades/hues. Border irregularity was graded as none, slight, moderate, or marked; border demarcation was graded as well-defined or ill-defined (slight, moderate, or marked).

Constitutive skin color was assessed on the patient’s medial upper arm. Constitutive skin color and lesion color were assessed using a universal color scheme (Pantone Color Overlay Selector; Letraset USA). Skin color and lesion color were graded white (absence of color); very light brown (Pantone 466-A, walnut shell); light brown (Pantone 465-A superimposed on 475-A, skin of lightly toasted almond); medium brown (Pantone 464-A, Brazil nut shell); dark brown (Pantone 469-A, milk chocolate); very dark brown (Pantone 497-A, bittersweet chocolate); and black (Pantone opaque black-A, charcoal).

All data were placed on Microsoft Excel Office 2011 spreadsheets. The main outcome parameter was presence of 1 or more perianal nevi, with comparisons with demographic and clinical variables. Categorical variables were compared using Pearson χ² test when every table cell value was at least 5, and Fisher exact test when at least 1 table cell value was less than 5. Crude odds ratios (ORs) and corresponding 95% CIs were calculated. Reported P values are 2-tailed, and significance was set at P < .05.

In a convenience sample of 255 consecutive adults asked to participate, 18 declined, and 1 was excluded because of excessive perianal soiling. Of 236 participating patients, 219 were non-Hispanic white; 4, Hispanic white; and 13, nonwhite. Nonwhites included 6 non-Hispanic black/African Americans; 2, nonwhite Hispanics; 2, Asian; 2, mixed non-Hispanic black/African American and white; and 1, mixed Asian and white.

Patients included 138 men and 98 women, ages 23 to 84 years (median age, 55 years; mean [SD], 53 [15] years). The 219 non-Hispanic whites included 128 men and 91 women, ages 23 to 84 years (median age, 55 years; mean [SD], 53 [15] years). The 4 Hispanic whites were males, ages 48 to 75 years (median age, 63 years; mean [SD], 62 [12] years). The 13 nonwhites included 3 men and 10 women, ages 28 to 81 years (median age, 56 years; mean [SD], 55 [15] years). Small numbers of Hispanic whites and nonwhites precluded statistical comparisons with non-Hispanic whites.

Medical encounters included initial visits (51 of 236 [21.6%]) and follow-up visits (185 of 236 [78.4%]) for CM and/or skin cancer screening (46 of 236 [19.5%]), CM and/or skin cancer surveillance (178 of 236 [75.4%), or other (12 of 236 [5.1%]).

Prevalence of perianal nevi according to constitutive skin pigmentation for all 236 patients is shown in Table 1. Numbers were insufficient for valid statistical comparisons according to constitutive skin pigmentation subgroups.

Of 219 non-Hispanic whites, perianal nevi were detected in 48.9% (107 of 219). Examples of perianal nevi are shown in eFigures 1A, 1B, and 2 in the Supplement. Of the 107 non-Hispanic whites who had perianal nevi, 1 or more nevi were detected in the anal margin in 21.5% (47 of 219), gluteal cleft in 36.1% (79 of 219), and perineum in 3.2% (7 of 219) (Table 1). Of the 4 Hispanic whites, perianal nevi were detected in 50.0% (2 of 4), all lesions occurring in the gluteal cleft. For the 13 nonwhite patients, perianal nevi were detected in 38.5% (5 of 13), anal margin in 30.8% (4 of 13), gluteal cleft in 23.1% (3 of 13), and perineum in 7.7% (1 of 13).

Prevalence of perianal nevi in non-Hispanic white females (58.2% [53 of 91]) was significantly greater than in non-Hispanic white males (42.2% [54 of 128]) (OR, 1.91; 95% CI, 1.11-3.30; P = .02). In non-Hispanic whites, there was no significant difference in prevalence of perianal nevi in patients younger than 50 years (52.9% [46 of 87]) vs 50 years or older (46.2% [61 of 132]) (OR, 1.31; 95% CI, 0.76-2.25; P = .33). For non-Hispanic whites, Hispanic whites, and nonwhites combined, perianal freckling was noted in 3.2% (7 of 219) (data were missing for 17 patients).

Of 47 non-Hispanic whites who had anal margin nevi, 48.9% (23 of 47) had only 1, 21.3% (10 of 47) had 2, 12.8% (6 of 47) had 3, and 17.0% (8 of 47) had 4 or more. Of 79 non-Hispanic whites who had gluteal cleft nevi, 59.5% (47 of 79) had 1, 20.3% (16 of 79) had 2, 10.1% (8 of 79) had 3, and 10.1% (8 of 79) had 4 or more. Of 7 non-Hispanic whites who had nevi in the perineum, 71.4% (5 of 7) had only 1, and 1 patient each had 4 and 6 nevi, respectively.

Of the 2 Hispanic whites who had perianal nevi, both patients had 1 nevus in the gluteal cleft and no nevi in other perianal locations. Of 3 nonwhites who had gluteal cleft nevi, 2 patients had only 1 nevus and 1 patient had 2 nevi. Of 4 nonwhites who had anal margin nevi, 1 patient had 2, 1 patient had 3, and 2 patients had 5. The nonwhite patient who had any nevi in the perineum had only 1 nevus in that location.

In non-Hispanic whites, the greatest diameter of nevi ranged from 1 to 15 mm in the anal margin, 0.5 to 9 mm in the gluteal cleft, and 1 to 2 mm in the perineum (Table 2). In Hispanic whites, the greatest diameter of nevi ranged from 0.5 to 1 mm. In nonwhites, greatest diameter of nevi ranged from 2 to 3 mm in the anal margin, 1 to 3 mm in the gluteal cleft, and 2 mm in the perineum.

For all 236 patients combined, the diameter of the largest anal margin nevus was 5 mm or greater in 3.0% (7 of 236), and 10 mm or greater in 0.85% (2 of 236)—10 mm and 15 mm, respectively. The largest gluteal cleft nevus was 5 mm or greater in 2.1% of patients (5 of 236), the largest being 9 mm in 1 patient. The largest perineal nevus was 2 mm, in 2.5% of patients (6 of 236).

In non-Hispanic whites, perianal nevi of any size, 2 mm or greater in diameter, and 5 mm or greater in diameter were evident in 48.9% (107 of 219), 39.7% (87 of 219), and 5.5% (12 of 219) of patients, respectively. Of 107 non-Hispanic whites who had any perianal nevi, 11.2% (12 of 107) had at least 1 nevus 5 mm or greater in diameter. Of 3 non-Hispanic whites who had perianal nevi 8 mm or greater in diameter, the largest nevi were 9 mm, 10 mm, and 15 mm, respectively. In the 4 Hispanic whites, perianal nevi of any size were evident in 2, both nevi 1 mm diameter. In nonwhites, perianal nevi of any size and 2 mm or greater in diameter were evident in 38.5% (5 of 13) and 38.5% (5 of 13), respectively. None of the nonwhites had perianal nevi 5 mm or greater in diameter.

In non-Hispanic whites, the following features of perianal nevi were noted: raised topography 9.4% (10 of 107); maximum height 0.5 mm or greater in 9.4% (10 of 107); dark brown or darker pigmentation in 5.6% (6 of 107); red color in 0.9% (1 of 107); 2 or more shades or hues in 0.9% (1 of 107); and border irregularity in 0.9% (1 of 107). None of the perianal nevi in non-Hispanic whites displayed black color, ill-defined border demarcation, halo depigmentation, scaling, erosion, or ulceration (Table 3).

In the 2 Hispanic whites who had perianal nevi, the largest nevus was very light brown in 1 and light brown in 1. In the 5 nonwhites who had perianal nevi, the largest nevus was dark brown in all 5. None of the largest nevi in the 2 Hispanic whites or 5 nonwhites had an elevated surface, very dark brown or darker pigmentation, pigmentation variegation, border irregularity, ill-defined border, halo depigmentation, scaling, ulceration, or erosion.

In non-Hispanic whites, the presence of perianal nevi was significantly associated with history of atypical nevus removal, and presence of at least 4 atypical nevi, at least 1 extant atypical nevus, atypical nevus pattern, and at least 20 nevi that were 2 mm or greater in diameter plus at least 5 nevi that were 5 mm or greater in diameter (Table 4 and Table 5). The presence of perianal nevi in non-Hispanic whites showed a nonsignificant trend for association with personal history of CM or CM in a first-degree relative, and presence of atypical nevi and family history of CM with or without a personal history of CM, but no significant association with total number of nevi removed, skin phototype, natural hair color at age 20 years, largest nevus 8 mm or greater in diameter, red hair, or sun-induced freckling density.

Of 101 patients who had perianal nevi, 36 (35.6%) were previously aware of their nevi, 30 patients had been informed previously (by A.R.R.), and 5 claimed prior awareness without explanation (data were missing for 1 patient).

Our study determined the prevalence and morphologic features of perianal nevi in a convenience sample of 236 patients attending an outpatient practice of a dermatologist specializing in screening and surveillance for CM and keratinocytic skin cancer. We found that perianal nevi are relatively common and significantly associated with history of atypical nevus removal and presence of prominent numbers of nevi and atypical nevi. We detected more perianal nevi in women than in men, possibly related to excessive perianal hair in men obscuring visual detection or a true finding related to the greater incidence of anorectal melanoma in women.⁵ In non-Hispanic whites, the diameters of perianal nevi ranged from 0.5 to 15 mm; and 5.5% of patients (12 of 219) had at least 1 perianal nevus 5 mm or greater in diameter. Most nevi detected were flat and lightly pigmented. Darkly pigmented nevi at least 5 mm in diameter were uncommon in our sample, and very darkly pigmented perianal nevi of any size were detected in only 1 of 219 non-Hispanic whites and in none of the Hispanic whites or nonwhites. We had insufficient numbers of Hispanic whites and nonwhites for valid comparisons with non-Hispanic whites. Numbers were insufficient to compare prevalence of perianal nevi according to constitutive pigmentation.

We are unaware of published studies investigating prevalence and morphological features of perianal nevi, or association of perianal nevi with significant CM risk factors. One study¹¹ reported a positive association between personal and family history of CM and a personal history of genital or anorectal melanoma. In our study, we found a significant association between the presence of perianal nevi and a personal history of CM, and any atypical nevi excised. We found a nonsignificant trend for presence of perianal nevi and presence of atypical nevi plus a family history of CM, with or without a personal history of CM, but our sample size was not sufficiently powered to fully examine this latter association.

While perianal nevi are being followed closely in our study patients undergoing CM or keratinocytic skin cancer surveillance, some with digital images because of atypical features, we did not find sufficient cause to biopsy any lesion in the reported sample. Several studies^12-15 suggest that nevi of special anatomical sites, including perianal and genital sites, tend to be atypical histopathologically. However, specimens in “special site” studies are gleaned from dermatopathology files, resulting in a significant proportion being atypical or malignant, regardless of anatomic site. Clinicians excise perianal nevi if there is sufficient cause (ie, atypical appearance, instability, or discomfort). Prevalence, gross morphologic features, and histopathological features of unselected perianal nevi are underrepresented in the literature,^14-18 precluding generalizations.

The perianal area is difficult to self-examine, and physicians may not routinely examine this site. One study¹⁹ found that 30% of dermatologists reported performing “full-body” skin cancer screening on all of their patients, and 42% of respondents reported lack of time as an impediment to full-body screening. In a study of primary care physicians, nearly 60% reported routinely performing full-body examinations for high-risk patients, including buttocks and genitalia.²⁰ It is unclear in these studies if full-body includes the perianal area.

Skin cancer screening by physicians tends to be low. US adults reporting “ever having had a skin examination” by a physician was 20.6% in 1992, 20.9% in 1998, and 14.5% in 2000; “a recent skin examination” was reported by 10.3% of respondents in 1992, 11.0% in 1998, and 8.0% in 2000, with non-Hispanic whites reporting more often screened than other ethnic groups.²¹ A history of recent skin cancer screening was more common in whites who had a family history of CM.²¹ Women are more likely than men to undergo skin cancer screening.^22-24 We are unaware of data related to frequency of perianal screening by physicians.

Skin cancer screening is associated with thinner, more curable cases of CM.^25-27 While total mucocutaneous examination as a mode of secondary prevention of CM has been controversial because there are no randomized prospective trials assessing mortality outcomes,^28,29 a large population-based study in Germany in which 19% of the adult population in 1 state was screened showed a mortality decrease of 50% compared with other German states and surrounding countries.^30-32 There are no data assessing the usefulness of perianal area screening for CM and other tumors.

Our study has significant limitations and potential biases. Our study population comprised mostly non-Hispanic whites in only 1 physician’s outpatient dermatology clinic specializing in CM/keratinocytic skin cancer screening and surveillance. Numbers of Hispanic whites and nonwhites and numbers of patients with dark constitutive pigmentation were small. We demonstrated that the presence of perianal nevi appeared to be significantly associated with CM risk factors, likely accounting for a relatively high prevalence of perianal nevi in the population studied.

An additional limitation is lack of precision in assessment of gross morphologic features of perianal nevi. Measures were naked-eye estimates using a millimeter ruler or template placed adjacent to lesions and not directly on perianal skin. Dermoscopy was not used in morphologic assessments. Biopsy specimens were not obtained to confirm that detected lesions were in fact melanocytic nevi.

While we did not detect perianal CM or other significant perianal tumors in the small number of study patients reported, perianal nevi in some of our study patients are being followed closely for instability using digital imaging.

The perianal area is difficult to self-examine and may harbor a malignant tumor, including CM. Given the high prevalence of perianal nevi in our study patients, many of whom are known to be at high risk for developing CM, one may conclude that the perianal area should not be ignored during CM screening and surveillance. Additional studies in a more general population are warranted to determine the prevalence of perianal nevi and to clarify the relationship among perianal nevi, CM risk factors, and perianal CM.

Corresponding Author: Arthur R. Rhodes, MD, MPH, Department of Dermatology, Rush University Medical Center, 1653 W Congress Pkwy, Annex Building 220, Chicago, IL 60612 (Arthur_Rhodes@Rush.edu).

Accepted for Publication: April 29, 2016.

Published Online: June 15, 2016. doi:10.1001/jamadermatol.2016.1885

Author Contributions: Ms Socik and Dr Rhodes had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Socik, Rhodes.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Socik, Burnes, Rhodes

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Socik, Rhodes.

Administrative, technical, or material support: All authors.

Study supervision: Burnes, Rhodes.

Conflict of Interest Disclosures: Dr Rhodes has consulted with Castle Bioscience on their gene array prognostic test for melanoma; has received attorney payments related to consulting on 1 or 2 medical liability cases every year or every other year, for the past 35 years; and has received a 1-year grant from the State of Massachusetts to study the usefulness of Wood’s light examination of neonates for the earliest manifestations of tuberous sclerosis. No other disclosures are reported.

Additional Contributions: The Department of Dermatology, Rush University Medical Center, Chicago, Illinois, provided material support (paper and use of copy machine).