STUDY Dermoscopy of Pigmented Seborrheic Keratosis A Morphological Study Ralph Peter Braun, MD; Harold S. Rabinovitz, MD; Joachim Krischer, MD; Ju¨rgen Kreusch, MD; Margaret Oliviero, ARNP; Luigi Naldi, MD; Alfred W. Kopf, MD; Jean H. Saurat, MD Objectives: To describe morphological features of seb- Main Outcome Measures: Identification of new orrheic keratosis as seen by dermoscopy and to investi- morphological criteria and evaluation of frequency. gate their prevalence. Results: A total of 15 morphological dermoscopic crite- Design: Prospective cohort study using macrophotog- ria were identified. Standard criteria such as milialike cysts raphy and dermoscopy for the documentation of sebor- and comedolike openings were found in a high number of rheic keratosis. cases (135 and 144, respectively). We found network and networklike structures to be present in 94 lesions (46%). Using standard diagnostic criteria for seborrheic kerato- Settings: Seborrheic keratoses were prospectively col- sis, 30 lesions would not have been diagnosed as such. lected in 2 sites: a private practice in Plantation, Fla (site 1), and the Department of Dermatology at the Univer- Conclusions: The classic dermoscopic criteria for seb- sity Hospital Geneva in Switzerland (site 2). orrheic keratosis (milialike cysts and comedolike open- ings) have a high prevalence but the use of additional der- Patients: A total of 203 pigmented seborrheic kerato- moscopic criteria such as fissures, hairpin blood vessels, ses (from 192 patients) with complete documentation sharp demarcation, and moth-eaten borders improves the were collected (111 from site 1 and 93 from site 2). diagnostic accuracy. The proper identification of pig- ment network and networklike structures is important Interventions: Screening for new morphological fea- for the correct diagnosis. tures of seborrheic keratosis and evaluation of all le- sions for the prevalence of these criteria. Arch Dermatol. 2002;138:1556-1560 N THE PAST 2 decades, there has includes 2 different levels. In the first level, been a rising incidence of malig- one must determine if a lesion is of mela- nant melanoma.1-6 Due to a lack of nocytic or nonmelanocytic origin. For this adequate therapies for metastatic decision, an algorithm using morphologi- melanoma, the best treatment is cal key criteria has been proposed. Once the still early diagnosis and prompt surgical lesion is identified to be of melanocytic ori- I 3 From the Pigmented Skin excision of the primary cancer. Dermo- gin, it must be categorized as benign, sus- Lesion Clinic, Department scopy (also known as epiluminescence pect, or malignant (second level). To ac- of Dermatology, University microscopy, dermatoscopy, and skin- complish this, 4 different approaches have Hospital Geneva, Geneva, surface microscopy) is a simple, noninva- been proposed: ABCD rule of dermos- Switzerland (Drs Braun, 24,28 17,29,30 Krischer, Naldi, and Saurat); sive, in vivo method that has been de- copy, Menzies method, 7-point 10 19 Skin and Cancer Associates, scribed as a useful tool for the early checklist, and pattern analysis. Diagno- Plantation, Fla, and recognition of malignant melanoma.7-24 The sis of seborrheic keratosis is, in general, a Department of Dermatology, performance of dermoscopy has been in- clinical diagnosis, but in a certain percent- University of Miami School vestigated by many authors. Its use in- age of cases, differential diagnosis be- of Medicine, Miami, Fla creases diagnostic accuracy 5% to 30% over tween pigmented seborrheic keratosis and (Dr Rabinovitz and clinical visual inspection alone, depend- malignant melanoma is difficult. The most Ms Oliviero); and Ronald ing on the type of skin lesions and experi- common dermoscopic characteristics for O. Perelman Department 25,26 of Dermatology, New York ence of the physician. According to the seborrheic keratoses are comedolike open- University School of Medicine, guidelines of the Consensus NetMeeting on ings and milialike cysts (Figure 1). Mil- 27 New York (Dr Kopf). Dermoscopy (held in Rome in February ialike cysts are round, whitish or yellow- Dr Kreusch is in private 2001; see Web site at http://www ish structures that correspond to small practice in Lu¨beck, Germany. .dermoscopy.org), the diagnostic strategy intraepidermal, keratin-filled cysts.17,31-35 (REPRINTED) ARCH DERMATOL / VOL 138, DEC 2002 WWW.ARCHDERMATOL.COM 1556 ©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 Frequencies and Distribution of the Criteria Identified in Step 1 According to Elementary Lesion Type No. (%) of Lesions Papule/ Total Patch Plaque Nodule No. Face 18 (64) 8 (29) 2 (7) 28 Border Moth eaten 31 (33) 44 (47) 19 (20) 94 Sharply demarcated 32 (17) 91 (50) 60 (33) 183 Criteria Hairpin vessels 8 (6) 65 (50) 56 (43) 129 Comedolike openings 13 (9) 79 (55) 52 (36) 144 Fissures 12 (10) 69 (56) 43 (35) 124 Milialike cysts 19 (14) 68 (50) 48 (36) 135 Networklike structures 27 (29) 54 (57) 13 (14) 94 Prominent network 19 (29) 37 (56) 10 (15) 66 Thickened network 11 (19) 38 (66) 9 (16) 58 Heterogenic network 23 (38) 31 (51) 7 (11) 61 Fingerprinting 10 (100) 0 0 10 Blotch 0 12 (75) 4 (25) 16 Crust 0 13 (42) 18 (58) 31 Dots 9 (35) 14 (54) 3 (12) 26 Whitish veil 0 0 3 (100) 3 Exophytic papillary structure 0 12 (75) 4 (25) 16 Colors Light brown 38 (20) 97 (51) 56 (29) 191 Dark brown 38 (20) 100 (52) 55 (28) 193 Blue gray 12 (11) 55 (51) 41 (38) 108 Yellowish 1 (8) 4 (33) 7 (58) 12 Figure 1. A, Macroscopic photograph of a seborrheic keratosis Maroonish 0 4 (44) 5 (56) 9 (papular/nodular type). B, Dermoscopy shows a sharply demarcated lesion Black 4 (8) 23 (47) 22 (45) 49 with multiple milialike cysts (arrows) and comedolike openings (asterisk). Total 39 (19) 101 (50) 63 (31) 203 They may also be seen in some congenital nevi and in request or in difficult cases (ie, melanocytic lesion could not some papillomatous melanocytic nevi. Comedolike be ruled out with certainty). All slides (clinical and dermo- openings (pseudofollicular openings, crypts) are scopic), were reviewed by 2 physicians experienced in dermos- mainly seen in seborrheic keratosis or papillomatous copy (R.P.B. and H.S.R.). During this first evaluation, both au- thors recorded all morphological findings to define more melanocytic nevi. Keratin-filled invaginations of the precisely the entity of seborrheic keratosis. In the second step, epidermis correspond histopathologically to comedo- both authors reevaluated all lesions for presence or absence of 17,31-35 like structures. the morphological features previously identified. For the de- The purpose of this report was to review dermo- scriptive statistics, the SPSS 9.0 software package (SPSS Inc, scopic criteria of pigmented seborrheic keratosis, some Chicago, Ill) was used. of which have not been clearly defined in the literature. RESULTS METHODS A total of 203 pigmented seborrheic keratoses were col- Two sources of clinical cases were used in this study. These lected (111 lesions from site 1 and 92 lesions from site sites were Skin and Cancer Associates, Plantation, Fla (site 1), and the Pigmented Skin Lesion Clinic, Department of Der- 2). Histopathologic examination was performed in 99 le- matology, University Hospital Geneva, Geneva, Switzerland sions (89/2%) from site 1 and 85 lesions (91%) from site (site 2). 2. In the first step (examination of both macroscopic and All pigmented seborrheic keratoses were examined by ex- dermoscopic photographs of all lesions by both investi- perienced “dermoscopists” using a handheld dermoscope (Der- gators), a total of 15 morphological dermoscopic crite- matoscope [Heine AG], Episcope [Welch-Allyn], or Dermo- ria and 6 colors were identified (Table). The elemen- Genius basic [Rodenstock Pra¨zisionsoptik]). All pigmented tary lesion (patch, plaque, and papular/nodular) was seborrheic keratoses data were collected prospectively (3 years identified on the clinical image. The reticulated type of at site 1 and 6 months at site 2). A macroscopic clinical pho- seborrheic keratosis was mainly found in patch lesions tograph (Slue imaging system; Canfield Clinical Systems, Fair- and the acanthotic type was mainly found in thicker le- field, NJ) and documentation of the dermoscopic finding (Der- maphot lens, Heine AG) was performed for every lesion. Since sions (plaque and papular/nodular). the diagnosis of pigmented seborrheic keratoses is usually clini- In the second step, all lesions were evaluated for cal, we considered systematic biopsy (and histopathologic ex- the presence of the morphological criteria identified in amination) as unethical, even though this might introduce a the first step. The frequencies of the criteria are shown selection bias. Therefore, a biopsy was performed on patient in the Table. We found 28 lesions of the face while 175 (REPRINTED) ARCH DERMATOL / VOL 138, DEC 2002 WWW.ARCHDERMATOL.COM 1557 ©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 A A B B ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ Figure 2. A, Macroscopic photograph of a seborrheic keratosis Figure 3. A, Macroscopic photograph of a seborrheic keratosis (plaque type). (papular/nodular type). B, Dermoscopy shows a sharply demarcated lesion B, Dermoscopy shows a sharply demarcated lesion with multiple fissures that with a moth-eaten border (arrows), networklike structures (asterisks), give the lesion a brainlike appearance. milialike cysts, and some comedolike openings. lesions were located elsewhere. Elementary lesions (47%), and 19 papular/nodular lesions (20%). By defini- were distributed as follows: 39 patch lesions (19%), 101 tion, a “moth-eaten” border is sharply demarcated, but not plaque lesions (50%), and 63 papular/nodular lesions all sharply demarcated lesions have a moth-eaten border. (31%). Hairpin vessels were found in 129 lesions (63%). They were mainly found in thicker lesions (94% COMMENT plaque or papular/nodular lesions). Of the 203 sebor- rheic keratoses, 183 lesions were sharply demarcated There are many publications on the differential diagno- (90%).