STUDY Prevalence of Clinically Resembling Seborrheic Keratosis Analysis of 9204 Cases

Leonid Izikson, BS; Arthur J. Sober, MD; Martin C. Mihm, Jr, MD, FRCP; Artur Zembowicz, MD, PhD

Objective: To estimate the prevalence of melanoma clini- Main Outcome Measure: Histological diagnosis, which cally mimicking seborrheic keratosis. was correlated with the preoperative clinical diagnosis.

Design: Retrospective review of cases submitted for his- Results: Melanoma was identified in 61 cases (0.66%) tological examination with a clinical diagnosis of sebor- submitted for histological examination with a clinical rheic keratosis or with a differential diagnosis that in- diagnosis that included seborrheic keratosis. Melanoma cluded seborrheic keratosis. was in the clinical differential diagnosis of 31 cases (51%). The remaining lesions had a differential diagno- Setting: A tertiary medical care center–based der- sis of seborrheic keratosis vs melanocytic (17 matopathology laboratory serving academic der- cases, 28%), basal cell carcinoma (7 cases, 12%), or a squa- matology clinics that have a busy pigmented lesion mous proliferation (3 cases, 5%). In 3 cases (5%), seb- clinic. orrheic keratosis was the only clinical diagnosis. All histological types of melanoma were represented. Materials and Methods: A total of 9204 consecutive pathology reports containing a diagnosis of seborrheic Conclusions: Our results confirm that melanoma can keratosis in the clinical information field were identi- mimic seborrheic keratosis. These data strongly support fied between the years 1992 and 2001 through a com- the current policy of submitting for histological examina- puter database search. Reports with a final histological tion all specimens that have been removed from patients. diagnosis of melanoma were selected for further review and clinicopathological analysis. Arch Dermatol. 2002;138:1562-1566

EBORRHEIC KERATOSIS is one of dures and treatments may be subopti- the most common benign neo- mal. For these reasons, clinicians must bal- plasms in adults. Seborrheic ance a high index of suspicion for keratoses begin to appear af- malignancy in clinically diagnosed sebor- ter the age of 30 years in ge- rheic keratoses against the impracticality Snetically susceptible individuals and con- and costs of histological evaluation of all tinue to develop throughout their life span. lesions. Thus, appropriate clinical man- These lesions account for a large number agement of seborrheic keratosis requires of physician visits and have an associated an understanding of the frequency with significant health care cost. Most clinically which melanoma clinically mimics this be- diagnosed seborrheic keratoses are un- nign lesion. treated or removed by cryotherapy with- We have encountered several cases of out histological confirmation of the diag- melanoma clinically resembling sebor- nosis despite limited accuracy of the clinical rheic keratosis in our dermatopathology diagnosis of seborrheic keratosis.1-5 practice at the Massachusetts General Hos- From the Dermatopathology Seborrheic keratoses are most often pital Dermatopathology Unit and in our Unit, Departments of confused clinically with actinic keratoses clinical practice at the Massachusetts Gen- Dermatology (Drs Sober and or benign conditions such as solar len- eral Hospital Pigmented Lesions Clinic, Mihm) and Pathology tigo or viral warts, with relatively little Boston. Two small retrospective series ad- (Drs Mihm and Zembowicz), consequence for prognosis or therapy. In dressing this issue identified 2 cases of Massachusetts General Hospital 6 and Harvard Medical School, contrast, misdiagnosis of , es- melanoma among 577 specimens and 1 4 Boston, Mass. Mr Izikson is a pecially melanoma, as seborrheic kerato- case among 328 specimens submitted for third-year medical study at sis might have untoward implications for histological examination with a clinical di- Harvard Medical School. the patient, as both diagnostic proce- agnosis that included seborrheic kerato-

(REPRINTED) ARCH DERMATOL / VOL 138, DEC 2002 WWW.ARCHDERMATOL.COM 1562

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 sis. Seborrheic keratosis was one of the most common nonmelanoma diagnoses in retrospective studies that ex- Table 1. Demographic and Clinical Characteristics amined the accuracy of clinical diagnosis of histologi- of Patients With Melanoma Clinically Resembling Seborrheic Keratosis cally confirmed .7-16 In the present article, we assess the prevalence of Mean melanoma resembling seborrheic keratosis in our ter- Clinical Diagnostic Group (n) Age, y M/F Ratio Distribution* tiary medical care center–based dermatopathology prac- All cases (61) 60.5 37/24 (1.54) 20 H/N, 23 T, 18 E tice serving a large academic dermatology practice and a Seborrheic keratosis (3) 61.3 0/3 1 H/N, 2 E busy pigmented lesions clinic and provide a clinicopatho- Seborrheic keratosis vs 75.3 1/2 (0.5) 1 H/N, 2 E logical analysis of identified cases. squamous proliferation (3) Seborrheic keratosis vs 61.8 3/4 (0.75) 1 H/N, 4 T, 2 E basal cell carcinoma (7) METHODS Seborrheic keratosis vs 55.8 9/8 (1.12) 4 H/N, 7 T, 6 E (17) Pathology reports generated at the Massachusetts General Hos- Seborrheic keratosis vs 61.3 24/7 (3.4) 13 H/N, 12 T, 6 E pital Dermatopathology Unit from 1992 to 2001 were searched melanoma (31) to identify those cases in which the clinical diagnosis of seb- orrheic keratosis was entered as the primary diagnosis or was *H/N indicates head/neck; T, trunk; and E, extremities. included in the differential diagnosis for the submitted speci- men. This search yielded 9204 reports, from which 61 cases with the histopathological diagnosis of melanoma were iden- ential diagnosis was between seborrheic keratosis and a tified for further analysis. The vast majority of our laboratory’s melanocytic nevus. material is submitted by dermatologists from Massachusetts Gen- Demographic and clinical information about the iden- eral Hospital dermatology clinics, including the Massachu- tified cases is summarized in Table 1. We found a slight setts General Hospital Pigmented Lesion Clinic, which are staffed male predominance among the patients, and 54 patients by full- and part-time clinical staff supervising residents and (80%) were 50 years of age or older. The lesions were dis- fellows. Community dermatologists and Massachusetts Gen- tributed almost equally on the head and neck region, ex- eral Hospital surgical and medical services submit a small per- tremities, and trunk. As expected, the trunk was the most centage of cases. In our practice, residents and fellows do not common site in men (16 of the 37 cases); the extremities make decisions regarding biopsy specimens without the input of staff members. were the most common site in women (11 of the 24 cases). Clinical analysis involved a review of patients’ charts for Of note, but without obvious explanation, all 3 cases in any clinical and historical information relevant to the diagno- which seborrheic keratosis was the only clinical consid- sis of melanoma, including site, clinical appearance, clinical dif- eration occurred in women. The choice of diagnostic pro- ferential diagnosis, and personal or family history of mela- cedure depended on the clinical differential diagnosis. Eight noma. Data were collected in accordance with the institutional lesions (85%) with the clinical differential diagnosis of squa- and National Institutes of Health guidelines for clinical stud- mous proliferation or basal cell carcinoma were sampled ies, and the experimental protocols were approved by the in- as shave biopsy specimens. Shave biopsy specimens were stitutional review board. obtained in 18 cases (37%) in which a melanocytic lesion Pathological analysis included a review of available diag- was included in the clinical differential diagnosis. The rest nostic material from biopsy or excision specimens, which con- sisted of formalin-fixed, paraffin-embedded sections stained with of the cases were submitted as punch biopsy (n=19 [39%]) hematoxylin-eosin. Histological information included maxi- or excision (n=9 [19%]) specimens. One case clinically mal tumor thickness, Clark anatomical level of invasion, pres- diagnosed as seborrheic keratosis was submitted as a cu- ence of intralesional inflammatory response (absent or pres- rettage specimen. The remaining 2 cases were shave bi- ent, brisk or nonbrisk), number of mitoses, evidence of opsy specimens. ulceration, evidence of regression, presence of microsatellites, A clinicopathological analysis of the cases is sum- evidence of precursor lesion, and lymphovascular invasion. marized in Table 2. The diagnosis of melanoma was con- sidered in the clinical differential diagnosis of 31 speci- mens (51%). Seborrheic keratosis was the only clinical RESULTS diagnosis in 3 cases (5%). One lesion was described as an inflamed stucco keratosis. All 3 lesions were invasive In our series, of the 9204 consecutive lesions clinically melanomas, with a mean thickness of invasion of 0.84 diagnosed as seborrheic keratosis or with seborrheic kera- mm. One case was a superficial spreading melanoma, one tosis in the differential diagnosis, 61 (0.66%) revealed was maligna melanoma, and the third was an in- melanoma on histological examination. Of note, mela- vasive melanoma not further classified. noma was included in the clinical differential diagnosis In 3 cases (5%), a differential diagnosis of sebor- in only 50% of these cases. Based on the clinical descrip- rheic keratosis vs a squamous lesion was considered. Sus- tion and clinical differential diagnosis, the remaining cases pected squamous lesions included squamous cell carci- can be divided into 4 broad diagnostic categories: (1) cases noma (2 cases) and clear cell acanthoma vs large cell in which seborrheic keratosis was the only diagnostic con- acanthoma (1 case). Both cases with the differential di- sideration; (2) cases in which the differential diagnosis agnosis including squamous cell carcinoma turned out was between seborrheic keratosis and a squamous le- to be superficial spreading melanomas invasive to 2.3 and sion, malignant or benign; (3) cases in which the differ- 2.8 mm, respectively. ential diagnosis was between seborrheic keratosis and In 7 cases (12%), biopsy specimens were obtained basal cell carcinoma; and (4) cases in which the differ- to differentiate between seborrheic keratosis and basal

(REPRINTED) ARCH DERMATOL / VOL 138, DEC 2002 WWW.ARCHDERMATOL.COM 1563

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 Table 2. Clinicopathological Analysis of Melanomas Clinically Resembling Seborrheic Keratosis

No. of Mean Depth No. (%) Melanomas, of Invasive Histological Subtype Clinical Diagnostic Group of Cases Alternative Clinical Diagnoses (n) In Situ/Invasive Tumors, mm of Melanoma* Seborrheic keratosis 3 (5) None 0/3 0.84 1 LMM, 1 SSM, 1 NOS Seborrheic keratosis vs 3 (5) Squamous cell carcinoma (2), 1/2 2.55 2 NOS squamous proliferation clear cell acanthoma vs inflamed stucco keratosis (1) Seborrheic keratosis vs 7 (12) Pigmented basal cell carcinoma (5), 3/4 1.20 1 LMM, 1 NM, 2 NOS basal cell carcinoma basal cell carcinoma (2) Seborrheic keratosis vs 17 (28) Lentigo (4), nevus (3), 7/10 0.98 3 LMM, 2 SSM, 1 nevoid, melanocytic nevus dysplastic/atypical nevus (10) 1 desmoplastic, 1 acral, 2 NOS Seborrheic keratosis vs 31 (51) (12), lentigo 8/23 0.95 8 LMM, 7 SSM, 2 NMM, melanoma maligna melanoma (6), 1 desmoplastic, melanoma not otherwise 1 nevoid, 1 acral, specified (16) 3 NOS

*LMM indicates ; SSM, superficial spreading melanoma; NM, ; and NOS, not otherwise specified.

cell carcinoma. In 5 cases, the clinical impression was lished smaller retrospective series,4,6 which established pigmented basal cell carcinoma. Histologically, 3 le- a prevalence of 0.3% for melanoma in lesions submitted sions were melanomas in situ. Four lesions were inva- for histological examination in which a clinical diagno- sive melanomas, with a mean thickness of 1.2 mm. His- sis of seborrheic keratosis had been made. tological types of invasive melanoma included lentigo Our results are in agreement with those of a num- maligna (1 case) and nodular (1 case) and 2 cases in which ber of previous studies that retrospectively investigated subtyping could not be made. the clinical diagnoses of histologically confirmed mela- Finally, a differential diagnosis of seborrheic kera- nomas.7-17 Those studies clearly illustrated that the clini- tosis vs was considered in 17 cal diagnostic accuracy of melanoma ranges from 48% cases (28%). Clinical diagnoses included dysplastic/ to 67%. They also confirmed that seborrheic keratosis is atypical nevus (10 cases), lentigo (4 cases), and benign one of the lesions for which melanoma is commonly mis- nevus (5 cases). In this group, histological examination diagnosed. This error occurred in 7.7% to 31.0% of cases, revealed that 7 melanomas were in situ. Ten melanomas depending on the study. were invasive, with a mean maximal depth of invasion Seborrheic keratoses come to clinical attention for of 0.98 mm. Histological types of melanoma in this group cosmetic reasons; for symptoms associated with the le- included superficial spreading (2 cases), lentigo ma- sions, such as itching or inflammation; or for an atypi- ligna (3 cases), nevoid (1 case), desmoplastic (1 case), cal clinical appearance that necessitates the exclusion of and acral lentiginous (1 case). Two melanomas were not a malignant lesion. Most seborrheic keratoses in the first further classified. None of the cases represented mela- category and many in the second are correctly diag- noma arising in a seborrheic keratosis or collision be- nosed clinically and followed up or are treated using cryo- tween melanoma and a seborrheic keratosis. therapy, without histological confirmation of the diag- Reflecting the pattern of referrals in our laborato- nosis. Lesions removed for histological examination are ry’s practice, in 54 cases the diagnosis was made by a der- more likely to be atypical. Consequently, our results re- matologist. A plastic surgeon and internist submitted 1 flect the prevalence of melanoma in lesions resembling case each. In 5 cases submitted from outside institu- seborrheic keratoses submitted for histopathological ex- tions as second-opinion consultations, the subspecial- amination rather than in all seborrheic keratoses. Con- ity of the physician performing the biopsy was not es- sistent with this assumption, only 3 (5%) of the 61 mela- tablished. Two of the cases submitted as seborrheic nomas had seborrheic keratosis as the only clinical keratosis without the differential diagnosis were diag- diagnosis. Detailed review of the charts revealed that 2 nosed by dermatologists, and 1 case was diagnosed by a of these 3 lesions closely resembled seborrheic kerato- plastic surgeon. sis, and 1 was believed to be an inflamed seborrheic kera- tosis. This is an important observation, as it illustrates COMMENT that melanoma can appear clinically identical to sebor- rheic keratosis. Because we studied only seborrheic kera- In our series of 9204 consecutive cases submitted for his- tosis–like lesions submitted for histological examina- tological examination in which a clinical diagnosis of seb- tion, the true prevalence of melanoma in all clinically orrheic keratosis, or in which a differential diagnosis that diagnosed and normal-appearing seborrheic keratoses can- included seborrheic keratosis, was made, the preva- not be determined. However, common sense would sug- lence of melanoma was 0.66%. Only half of the cases rep- gest that it is much lower than 0.66%. resented true diagnostic mistakes, as in 51% cases the clini- Based on our results, melanomas mimicking sebor- cal differential diagnosis included melanoma. This rheic keratoses can be divided into 5 principal clinical prevalence is similar to that reported in previously pub- categories, each associated with a different set of clini-

(REPRINTED) ARCH DERMATOL / VOL 138, DEC 2002 WWW.ARCHDERMATOL.COM 1564

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 diagnosis of seborrheic keratosis contained 2 (66%) deeply A invasive melanomas with a significant risk of metastasis and mortality. The third group comprises lesions in which sebor- rheic keratosis was clinically considered along with an atypical squamous lesion, such as invasive or in situ squa- mous cell carcinoma. Clinical management of these le- sions is likely to be appropriate, as suspected squamous lesions are usually adequately sampled. Melanomas with a clinical differential diagnosis of seborrheic keratosis and basal cell carcinoma are not a surprising occurrence, as some basal cell carcinomas, as do most melanomas, present as pigmented lesions. Mis- B diagnosis of melanoma as seborrheic keratosis or pig- mented basal cell carcinoma may have adverse manage- ment implications, as both lesions are likely to be sampled by superficial shave biopsies, which may compromise de- finitive pathological diagnosis and prognostic informa- tion needed for further management. Finally, lesions resembling benign pigmented le- sions, such as lentigo and common or dysplastic/ atypical nevi, may represent melanoma. Clinical man- agement of these lesions will most likely include sampling of representative tissue via punch biopsy, deep shave bi- opsy, or excision. Clinicopathological analysis of melanomas in the present study revealed that all histological types of Histological sections of a curettage specimen from a lesion clinically diagnosed as seborrheic keratosis. A, At scanning magnification, it has a melanoma can mimic a seborrheic keratosis. Because verrucous architecture, hyperkeratosis, and acanthosis reminiscent of the number of invasive melanomas in each diagnostic seborrheic keratosis (hematoxylin-eosin, original magnification ϫ40). group was small, it is difficult to make definitive state- B, Examination at higher magnification reveals diagnostic features of an underlying melanoma (hematoxylin-eosin, original magnification ϫ200). ments regarding the association between a particular histological type of melanoma and a particular clinical appearance of the lesion. It is likely that melanomas cal, management, and pathological issues. We have in- clinically resembling seborrheic keratosis may share cluded in our analysis the lesions in which melanoma was certain histological features of seborrheic keratosis, considered in the differential diagnosis, as this category such as acanthosis, hyperkeratosis, horn pseudocysts, represented the most frequent subset of melanomas clini- epidermal hyperpigmentation, and verrucoid or papillo- cally resembling seborrheic keratosis, even though they matous architecture. The case illustrated in the Figure do not represent diagnostic mistakes and did not result supports this notion. However, comprehensive study of in management problems. this issue would require rigorous comparison of histo- Most problematic is the group of lesions in which logical features of melanomas clinically mimicking seb- the diagnosis of melanoma or other malignancy is not orrheic keratosis with those of clinically typical melano- suspected clinically, since the presumed seborrheic kera- mas, a topic beyond the scope of the present study. Of toses were removed for cosmetic reasons or because of interest is the presence of 2 cases of nevoid melanoma, inflammation. Such lesions may not be treated at all or a verrucous hyperkeratotic variant of which is espe- may be treated by locally destructive methods, without cially likely to resemble seborrheic keratosis clinically.18 histological confirmation. As in this series, even when Finally, although several cases of melanoma arising submitted for histological examination, lesions in this cat- within seborrheic keratosis have been reported,19 none egory are most likely to be sampled by superficial shave of the lesions in our study had the features of a collision biopsies or submitted as curettage specimens. Because between melanoma and seborrheic keratosis or evi- of scant tissue, the dermatopathological interpretation of dence of melanoma arising in an existing seborrheic such specimens can be very difficult. This point is well keratosis. illustrated in the Figure, which shows histological fea- In conclusion, our results estimate the prevalence tures of a curetted lesion with the presumptive clinical of melanoma in lesions clinically resembling seborrheic diagnosis of seborrheic keratosis. The verrucous archi- keratosis and submitted for histological examination at tecture, hyperkeratosis, acanthosis, and inflammatory in- 0.66%. This observation supports the current policy of filtrate of this lesion are all very reminiscent of sebor- submitting all specimens removed from patients for his- rheic keratosis. Moreover, superficial shave biopsy or tological examination. curettage specimens do not provide optimal prognostic indicators in melanoma, eg, maximal tumor thickness, Accepted for publication July 25, 2002. which are critical for clinical management. It is impor- This study was supported in part by clinical practice tant to emphasize that this group of lesions with a sole funds.

(REPRINTED) ARCH DERMATOL / VOL 138, DEC 2002 WWW.ARCHDERMATOL.COM 1565

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 Corresponding author and reprints: Artur Zembowicz, 9. Becker SW. Pitfalls in the diagnosis and treatment of melanoma. Arch Dermatol. MD, PhD, Dermatopathology Unit, Department of Pathol- 1954;69:11-30. 10. McMullan FH, Hubener LF. Malignant melanoma: a statistical review of clinical ogy, Massachusetts General Hospital, Warren 820, 15 Fruit and histological diagnosis. Arch Dermatol. 1956;74:618-619. St, Boston, MA 02114 (e-mail: [email protected]). 11. Freeman RG, Knox JM. Clinical accuracy of clinical diagnosis of skin tumors. Geriatrics. 1963;18:546-551. REFERENCES 12. Sober AJ, Fitzpatrick TB, Mihm MC, et al. Early recognition of cutaneous mela- noma. JAMA. 1979;242:2795-2799. 1. Stern RS, Boudreaux C, Arndt KA. Diagnostic accuracy and appropriateness of 13. Grin CM, Kopf AW, Welkovich B, Bart RS, Levenstein MJ. Accuracy in the clini- care for seborrheic keratoses: a pilot study of an approach to quality assurance cal diagnosis of malignant melanoma [published correction appears in Arch Der- for cutaneous surgery. JAMA. 1991;265:74-77. matol. 1990;126:1461]. Arch Dermatol. 1990;126:763-766. 2. Basarab T, Munn SE, Jones RR. Diagnostic accuracy and appropriateness of gen- 14. Lindelof B, Hedblad MA. Accuracy in the clinical diagnosis and pattern eral practitioner referrals to a dermatology out-patient clinic. Br J Dermatol. 1996; of malignant melanoma at a dermatological clinic. J Dermatol. 1994;21:461- 135:70-73. 464. 3. Bedlow AJ, Cliff S, Melia J, Moss SM, Seyan R, Harland CC. Impact of skin can- 15. Witheiler DD, Cockerell CJ. Sensitivity of diagnosis of malignant melanoma: a cer education on general practitioners’ diagnostic skills. Clin Exp Dermatol. 2000; clinicopathologic study with a critical assessment of biopsy techniques. Exp Der- 25:115-118. matol. 1992;1:170-175. 4. Lewis FM, Start RD, Gawkrodger DJ. Diagnostic accuracy for basal cell papillo- 16. Witheiler DD, Cockerell CJ. Histologic features and sensitivity of diagnosis of clini- mas. Br J Dermatol. 1993;129:94-95. cally unsuspected cutaneous melanoma. Am J Dermatopathol. 1991;13:551- 5. Chen SC, Bravata DM, Weil E, Olkin I. A comparison of dermatologists’ and pri- 556. mary care physicians’ accuracy in diagnosing melanoma: a systematic review. 17. Rampen FH, Rumke P. Referral pattern and accuracy of clinical diagnosis of cu- Arch Dermatol. 2001;137:1627-1634. taneous melanoma. Acta Derm Venereol. 1988;68:61-64. 6. Eads TJ, Hood AF, Chuang TY, Faust HB, Farmer ER. The diagnostic yield of his- 18. Zembowicz A, McCusker M, Chiarelli C, et al. Morphological analysis of nevoid tologic examination of seborrheic keratoses. Arch Dermatol. 1997;133:1417-1420. melanoma: a study of 20 cases with a review of the literature. Am J Dermato- 7. Grant-Kels JM, Bason ET, Grin CM. The misdiagnosis of malignant melanoma. pathol. 2001;23:167-175. J Am Acad Dermatol. 1999;40:539-548. 19. Cascajo CD, Reichel M, Sanchez JL. Malignant neoplasms associated with seb- 8. Ewing MR, Powel T. Some observations on the clinically pigmented skin tu- orrheic keratoses: an analysis of 54 cases. Am J Dermatopathol. 1996;18:278- mors. Br J Surg. 1951;38:442-454. 282.

ARCHIVES Features

1. Free color publication if color illustrations enhance the didactic value of the article. 2. Rapid publication. The time from acceptance to publication in our March 2001 issue was 3.2 months. 3. Expedited review on request from authors. 4. Archives Express. Manuscripts with high priority can be reviewed, ac- cepted, and published within 6 to 8 weeks. 5. Research Trainee Award of $1000 for the resident or fellow who is the first author of the best research paper of the past year. 6. Dermatologic Surgery Trainee Award of $1000 for the resident or fellow who is the first author of the best article on surgical dermatology published in the past year.

(REPRINTED) ARCH DERMATOL / VOL 138, DEC 2002 WWW.ARCHDERMATOL.COM 1566

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021