Dermoscopy Patterns of Halo Nevi

OBSERVATION Dermoscopy Patterns of Halo Nevi

Isabel Kolm, MD; Alessandro Di Stefani, MD; Rainer Hofmann-Wellenhof, MD; Regina Fink-Puches, MD; Ingrid H. Wolf, MD; Erika Richtig, MD; Josef Smolle, MD; Helmut Kerl, MD; H. Peter Soyer, MD; Iris Zalaudek, MD

Background: Halo nevi (HN) are benign melanocytic globular and/or homogeneous patterns in more than 80% nevi surrounded by a depigmented area (halo). This study of HN. Follow-up of 33 HN revealed considerable size aims to evaluate the dermoscopic features of HN and their reduction of the nevus component, but this was not as- changes during digital dermoscopic follow-up and to in- sociated with significant structural changes. Of a total of vestigate the frequency of the halo phenomenon in a se- 475 melanomas, only 2 revealed an encircling halo, but ries of melanomas. both displayed clear-cut melanoma-specific patterns according to dermoscopy. Observations: In a retrospective study, digital dermoscopic images of HN from patients who attended the Pig- Conclusions: Halo nevi exhibit the characteristic der- mented Skin Lesions Clinic of the Department of Der- moscopic features of benign melanocytic nevi, repre- matology, Medical University of Graz, between October sented by globular and/or homogeneous patterns that are 1, 1997, and March 31, 2004, were reviewed and classi- typically observed in children and young adults. Halo nevi fied by dermoscopic morphologic criteria. For HN that reveal considerable changes of area over time during digi- were followed up with digital dermoscopy, the percent- tal dermoscopic follow-up, albeit their structural pat- ages of changes in the size of the nevus and halo com- terns remain unchanged. For this reason and because ponents were calculated. In addition, digital dermo- melanoma with halolike depigmentation, despite being scopic images of histopathologically confirmed melanomas rare, additionally exhibits melanoma-specific dermo- obtained from the same database were reviewed for the scopic criteria, the role of digital dermoscopic fol- presence of an encircling halolike depigmentation. We low-up in the diagnosis of HN is insignificant. classified 138 HN in 87 patients (mean age, 22.4 years). The most common dermoscopic structures were the Arch Dermatol. 2006;142:1627-1632

ALO NEVI (HN), ALSO yet fully understood, but T lymphocytes termed Sutton nevi or leu- are considered to play a key role in the pro- koderma acquisitum cen- gressive destruction of nevus cells.7-11 Al- trifugum, are defined as be- though the halo phenomenon is most com- nign melanocytic nevi that mon in benign melanocytic nevi, reports are surrounded by a rim of depigmenta- of HN in individuals with a family and/or H 1-3 tion, resembling a halo. Halo nevi are personal history of melanoma and mela- common in children and young adults, with nomas with halo prompted concerns about a mean age at onset of 15 years.4 Affected the diagnosis and management of mela- individuals frequently have multiple HN, nocytic skin lesions that exhibit an encir- which are usually localized on the back and cling halo.12-15 4 Author Affiliations: may be clustered. The incidence of HN in Dermoscopy gained popularity be- Department of Dermatology, the population is estimated to be approxi- cause it permits a more accurate diagnosis Medical University of Graz, mately 1%.5 There is no predilection for sex, and management of melanocytic skin le- Graz, Austria (Drs Kolm, and all races are susceptible to the devel- sions compared with examination with the Hofmann-Wellenhof, opment of these lesions. A familial ten- naked eye.16-19 Although the dermoscopic Fink-Puches, Wolf, Richtig, dency for HN has been reported,5 and HN criteria of many benign melanocytic skin Smolle, Kerl, Soyer, and may be associated with atopic dermatitis or lesions and melanoma have been studied Zalaudek); and Department of with autoimmune disorders such as viti- extensively, the dermoscopic structures that Dermatology, University of ligo and Hashimoto thyroiditis.4-6 characterize HN have not yet been de- Rome Tor Vergata, Rome, Italy (Dr Di Stefani). Dr Zalaudek is For a period of months or even years, scribed in detail. We analyzed the dermo- currently with the Department HN tend to undergo 4 clinical stages char- scopic structures in HN and correlated the of Dermatology, Second acterized by a progressive involution with dermoscopic findings with changes in the University of Naples, Naples, subsequent total regression of the central size of the nevus and halo components of Italy. nevus. The underlying pathogenesis is not HN over time. Moreover, we investigated

(REPRINTED) ARCH DERMATOL/ VOL 142, DEC 2006 WWW.ARCHDERMATOL.COM 1627

©2006 American Medical Association. All rights reserved. Downloaded From: http://archderm.jamanetwork.com/ by a UQ Library User on 11/12/2015 100 Table 1. Demographics of Patients With Halo Nevi 90 Revealing a History of Concomitant Disease

80 Patient No./ Sex/Age, y Concomitant Disease 60 1/F/45 Personal history of melanoma 2/F/58 Personal history of melanoma 40 3/M/14 Family history of melanoma (grandfather)

No. of Halo Nevi 4/F/24 Family history of melanoma (mother) 5/F/35 Dysplastic nevus syndrome 20 16 17 9 6/M/33 Dysplastic nevus syndrome 2 4 7/F/28 History of autoimmune Hashimoto thyroiditis 0 Legs Head and Arms Abdomen Chest Back 8/F/10 History of familial halo nevi (sister and mother) Neck 9/F/8 Vitiligo Body Site 10/M/28 Vitiligo 11/F/20 Atopic dermatitis Figure 1. Geometric body site–related distribution of halo nevi (N=138). 12/F/9 Hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber disease)

the frequency of halo phenomenon in melanomas to evaluate the occurrence of this peculiar type of regression. Forty-six patients (52.9%) were women and 41 (47.1%) were men, with a mean age for all patients of 22.4 years (range, 5-69 years). METHODS From data included in the patient registration forms, we recorded concomitant diseases (Table 1). In 6 pa- For this retrospective study, dermoscopic images of HN were tients, generally accepted risk factors for melanoma were selected from a database that contained 29 383 digital images of pigmented skin lesions from 6079 patients who attended the recorded. These patients were older (mean age, 34.8 years; Pigmented Skin Lesions Clinic of the Department of Derma- range, 14-58 years) with respect to the mean age of 23.2 tology, Medical University of Graz, between October 1, 1997, years in our study population. However, in none of these and March 31, 2004. The digital dermoscopic images were re- patients was the time at onset of HN related to mela- corded with a digital epiluminescence microscopic system (Mole noma development and/or melanoma progression. Max II; Derma Medical Systems, Vienna, Austria) and stored In 58 nevi (42.0%), we observed a combination of ho- at 30-fold magnification in JPEG format, with a resolution of mogeneous and globular patterns in different areas of the 640ϫ480 pixels at 24-bit color depth. The diagnosis of HN was lesion. The homogeneous-globular pattern was fol- made by experienced dermatologists (I.K., A.D.S., and R.H.- lowed in frequency by the homogeneous pattern (32 nevi; W.) and was based on clinical and dermoscopic examination. 23.2%) and the globular pattern (24 nevi; 17.4%) The HN were classified according to the primary dermoscopic structural patterns (reticular, globular, or homogeneous).18 (Figure 2 and Figure 3). Variations on the reticular When 2 different dermoscopic structural patterns were pre- patterns were recorded in only 9 nevi, and no single ne- sent, the HN were classified as reticular-globular, homogeneous- vus in the study showed a combination of all 3 patterns globular, or homogeneous-reticular. The HN that showed al- (reticular, globular, and homogeneous). Three nevi most total disappearance of the nevus were classified as regressed. showed complete regression of the nevus component, If a nevus could not be classified because of poor quality of the characterized by a light brown to pink central area that digital dermoscopic image, it was labeled as “not classifiable.” exhibited dotted vessels as the only dermoscopic fea- When available, follow-up digital images of a given nevus ture (Figure 4). were measured and compared with the earlier digital images. From a total of 138 HN, we were able to follow changes Because of the retrospective study design, we were unable to in 33 HN (23.9%) from 16 patients. The mean fol- clarify the exact time of onset of HN in an individual patient. However, given that the approximate waiting time for consul- low-up period for these nevi was 26 months (range, 4-61 tation at our specialized clinic is 2 months, we assumed that months). We observed a reduction in the nevus area in the baseline visit for treatment of HN was sought within the all HN, with a mean reduction of 2.2% per month. For first 3 months of onset. Measurement was performed on- 12 HN a follow-up image was available within 6 months screen (semiautomatically) with the Mole Max “Expertizer” pro- after the initial baseline visit, and these nevi showed a gram (Derma Medical Systems), and the percentage of change mean reduction in the nevus area of 5.2% per month dur- in the area per month was calculated for both the nevus and ing the first 6 months (Table 2). halo components. In addition, we reviewed the dermoscopic Halo size decreased over time in 17 (51.5%) of the 33 images of histopathologically confirmed melanomas obtained cases for which follow-up images were available from the same database for the presence of an encircling halo. (Figure 5), with a mean reduction in the halo area of 0.85% per month. Nine (27.3%) of the 33 HN showed a RESULTS considerable enlargement of the halo component, with a mean increase in size of 1.3% per month. Poor quality We analyzed 138 digital dermoscopic images of HN from of the follow-up images or large size of the HN, extend- 87 patients. Of the 138 HN, 123 (89.1%) were located ing beyond the maximum field of view of the image frame, on the trunk, 12 (8.7%) on upper or lower extremities, precluded our measuring change in halo area for the re- and 3 (2.2%) on the head and neck region (Figure 1). maining 7 HN (21.2%).

(REPRINTED) ARCH DERMATOL/ VOL 142, DEC 2006 WWW.ARCHDERMATOL.COM 1628

60 58

No. of Halo Nevi 30

9 10 8

3 3 1 0 Not Regressed Reticular Reticular- Homogeneous- Globular Homogeneous Homogeneous- Classifiable Globular Reticular Globular Pattern

Figure 2. Geometric distribution of dermoscopic morphologic patterns in 138 halo nevi.

Figure 3. Dermoscopically, this nevus exhibits a globular pattern and an Figure 4. This completely regressed nevus reveals a central light brown to encircling rim of a white structureless area corresponding to the surrounding pink area with some dotted vessels as the only dermoscopic structure. In halo (hematoxylin-eosin, original magnification ϫ10). addition, telangiectases can be seen within the whitish halo and the normal skin (hematoxylin-eosin, original magnification ϫ10).

Remarkably, of 475 melanomas, only 2 (0.4%) showed cal pigment network, blue-white veil, irregular dots and an encircling halo. The 2 melanomas were located on the globules, streaks, and regression structures) and colors backs of a 33-year-old man (patient 1) and a 44-year old that varied from white to brown to blue to black man (patient 2). In the first case, histologic analysis re- (Figure 6). vealed a superficial spreading melanoma with focal regression (Clark level III; tumor thickness, 1.25 mm with COMMENT focal regression), whereas in the second case, an exact histopathologic measurement of the level of invasion and In our study, we found the globular and/or homoge- tumor thickness was not possible because of the pres- neous patterns to be the typical dermoscopic features in ence of significant extensive regression. Dermoscopy of HN. These patterns are generally associated with be- both melanomas revealed a multicomponent pattern, char- nign melanocytic nevi and represent, together with the acterized by a marked asymmetry of structures (ie, atypi- reticular patterns, the main criteria in the dermoscopic

(REPRINTED) ARCH DERMATOL/ VOL 142, DEC 2006 WWW.ARCHDERMATOL.COM 1629

©2006 American Medical Association. All rights reserved. Downloaded From: http://archderm.jamanetwork.com/ by a UQ Library User on 11/12/2015 Table 2. Nevus Size at Baseline and During Follow-up of 33 Halo Nevi

Change in Nevus Size, % Halo Nevus Nevus Size No. Location at Baseline, mm2 First 6 mo of Follow-up Total Follow-up Follow-up Period, mo 1 Back 18.77 −5.47 −0.69 52 2 Neck 19.31 ND −0.47 47 3 Back 28.30 −2.10 −1.10 12 4 Back 6.31 ND −4.17 24 5 Back 8.08 ND −6.67 15 6 Arm 11.49 ND −6.67 15 7 Back 34.03 −4.85 −4.85 6 8 Back 5.58 ND −1.64 61 9 Back 9.77 ND −1.64 61 10 Hip 5.70 ND −1.64 61 11 Back 54.19 ND −0.88 22 12 Back 9.99 ND −2.14 36 13 Arm 10.22 ND −0.11 34 14 Breast 20.24 ND −0.54 40 15 Back 17.12 ND −0.90 41 16 Back 34.34 −1.04 −0.30 29 17 Back 14.97 ND −0.15 25 18 Back 46.14 ND −0.31 14 19 Back 9.36 ND −0.14 7 20 Back 9.87 ND −0.43 20 21 Neck 7.13 ND −0.62 28 22 Breast 12.50 −1.46 −1.46 4 23 Breast 11.96 −3.50 −3.50 4 24 Abdomen 10.13 −1.53 −1.53 32 25 Abdomen 0.97 ND −1.68 16 26 Back 4.76 ND −4.55 22 27 Breast 8.21 −1.13 −2.56 10 28 Back 13.95 −5.20 −4.80 10 29 Back 3.02 −2.81 −2.81 25 30 Hip 3.61 −8.50 −3.30 25 31 Leg 4.84 −25.00 −4.00 25 32 Arm 5.92 ND −0.32 16 33 Back 1.18 ND −6.25 16

Abbreviation: ND, not done.

A B

0.5 cm

Figure 5. Halo nevus on the back of a 20-year-old man. A, Dermoscopic Figure 6. Dermoscopy of a halo melanoma located on the back of a image of the halo nevus shows the globular pattern (inset: clinical image). 33-year-old man (patient 1). A multicomponent pattern is represented by a The area of the nevus is 8.1 mm2. B, Dermoscopic image of the same nevus blue-white veil, irregular dots and globules, irregular streaks, atypical 10 months later. The nevus still exhibits a globular pattern but the area is network, and regression structures showing linear-irregular vessels and blue decreased to 6.1 mm2 (original magnification ϫ10 [A and B]). to gray pepperlike granules (hematoxylin-eosin, original magnification ϫ10).

classification of melanocytic nevi.17,18 Notably, only a small terns in our series of HN is an intriguing observation that proportion of HN in our series displayed reticular and/or requires further examination. unspecific patterns. The high frequency of globular and One explanation might be that halo phenomenon com- homogeneous patterns compared with reticular pat- monly occurs in melanocytic nevi, such as compound

(REPRINTED) ARCH DERMATOL/ VOL 142, DEC 2006 WWW.ARCHDERMATOL.COM 1630

©2006 American Medical Association. All rights reserved. Downloaded From: http://archderm.jamanetwork.com/ by a UQ Library User on 11/12/2015 nevi, congenital melanocytic nevi, and less frequently, curred in adults and exhibited a multicomponent pat- compound Spitz nevi and blue nevi.3-6,9,10 All these nevi tern, including an atypical pigmented network, irregu- are common in children and young adults and often ex- lar dots and globules, irregular streaks, blotches, blue- hibit globular and/or homogeneous patterns according white veil, and atypical vascular structures. These features to dermoscopy, including the so-called cobblestone pat- are strongly associated with melanoma and therefore in- terns as a variant of the globular patterns.17,18 We re- dicate the necessity of subsequent histopathologic diag- cently demonstrated that nevi exhibit considerable age- nosis.17,26 Despite this finding, it must be noted that digi- related differences in their dermoscopic patterns, with tal dermoscopic follow-up carries the certain risk that a variations on the globular pattern as the characteristic lesion that displays extensive regression at baseline could feature of nevi in children and of the reticular patterns eventually completely regress during follow-up and sub- in adults.20 Given that a mean age of 22.4 years in the sequently would no longer be detectable at the next visit. present study reflects a relatively young study popula- We therefore conclude that digital dermoscopic fol- tion, it could therefore be assumed that the predomi- low-up of HN does not provide additional diagnostic in- nance of the globular and/or homogeneous patterns in formation compared with a good clinical dermoscopic our series of HN was influenced by the age of our study correlation at baseline. population. Our study has some limitations. The small number of Only 6 patients had recorded risk factors for mela- patients with concomitant diseases associated with in- noma. These patients tended to be older at onset of HN creased risk of melanoma is not enough to draw definite compared with those without risk factors, although no conclusions about an eventual association of HN with mela- significant differences in dermoscopic pattern of HN be- noma development and/or the prognostic significance for tween these 2 patients groups were observed. In no pa- melanoma progression. However, all patients with pig- tient was the onset of HN directly time related to mela- mented skin lesions, regardless of whether they display HN, noma development and/or progression. Accordingly, our should undergo a routine full-body examination.12-15 data do not support previous reports that have sug- Furthermore, the retrospective study design did not gested an association between HN and melanoma. How- allow us to exactly define the date of onset in patients ever, following the general rule that the examination of with HN who were followed up. Given that the waiting all patients with pigmented skin tumors should include time for consultation at our clinic is approximately 2 an inspection of the entire body, we consider the same months, we assumed that onset of HN occurred no longer procedure also valid for patients with HN, regardless of than 3 months before baseline examination was per- associated risk factors. On the other hand, the decision formed. Accordingly, our observation of more rapid to perform a biopsy on a halo lesion that reveals clini- change in size during the first 6 months of follow-up must cally and/or dermoscopically atypical features must not be seen in light of a speculative assumption of the time be based on the patient’s age.12-15 of HN development in a given individual. Finally, it must In addition, we analyzed follow-up images of 33 HN be underlined that we observed only 2 melanomas that and compared these images with those obtained at the displayed an encircling halo, but in both cases clear-cut initial visit. Although we were unable to clarify the ex- melanoma-specific features were observed. act onset of HN development because of the retrospec- In summary, our study indicates that HN are com- tive study design, we observed in 12 HN a marked re- mon, although not exclusive, in children and young adults duction in size of the nevus component during the first and display the typical dermoscopic features of benign 6 months of the follow-up period, whereas changes in melanocytic nevi, represented by variations on the globu- size were less evident after the first 6 months. lar pattern. The role and diagnostic impact of digital der- In this context it seems critical to underline that the moscopic follow-up of HN in these patients, however, primary purpose of short-term follow-up is to monitor seem questionable. By contrast, halo melanoma is rare eventual changes over time, which aids the more accu- but reveals evident melanoma-specific patterns accord- rate differentiation of melanoma from nevi. It has been ing to dermoscopy. shown that significant changes in size, structure, and/or morphologic findings that occur in the short time of 3 Accepted for Publication: April 22, 2006. to 6 months are often associated with melanoma, whereas Correspondence: Rainer Hofmann-Wellenhof, MD, De- such changes are unusual in benign nevi.21-25 partment of Dermatology, Medical University of Graz, Because we found remarkable changes in the size of Auenbruggerplatz 8, 8036 Graz, Austria (rainer.hofmann HN during follow-up, albeit the dermoscopic patterns re- @meduni-graz.at). mained basically unchanged, our observation questions Author Contributions: Study concept and design: Hof- in part the value of short-term digital dermoscopic fol- mann-Wellenhof, Kerl, and Zalaudek. Acquisition of data: low-up for the diagnosis of these particular nevi and the Kolm, Di Stefani, Hofmann-Wellenhof, Fink-Puches, impact of follow-up on their differentiation from mela- Wolf, and Richtig. Analysis and interpretation of data: noma. This is also because we found that halo mela- Kolm, Di Stefani, Hofmann-Wellenhof, Smolle, and Soyer. noma not only was rare (2 of 478 melanomas) but also Drafting of the manuscript: Kolm, Hofmann-Wellenhof, revealed evident melanoma-specific patterns according Fink-Puches, Wolf, Richtig, and Smolle. Critical revi- to dermoscopy that allowed its diagnosis with high con- sion of the manuscript for important intellectual content:Di fidence and without the need for additional digital der- Stefani, Hofmann-Wellenhof, Kerl, Soyer, and Za- moscopic follow-up. Thus, the halo was more irregular laudek. Administrative, technical, and material support: than that seen in HN. Furthermore, both melanomas oc- Kolm, Hofmann-Wellenhof, Wolf, and Zalaudek. Study

(REPRINTED) ARCH DERMATOL/ VOL 142, DEC 2006 WWW.ARCHDERMATOL.COM 1631

©2006 American Medical Association. All rights reserved. Downloaded From: http://archderm.jamanetwork.com/ by a UQ Library User on 11/12/2015 supervision: Di Stefani, Hofmann-Wellenhof, Fink- 14. Whimster IW. The halo naevus and cutaneous malignant melanoma. Br J Dermatol. Puches, and Soyer. 1974;90:111-113. 15. Inamadar AC, Palit A, Athanikar SB, et al. Unusual course of a halo nevus. Pe- Financial Disclosure: None reported. diatr Dermatol. 2003;20:542-543. 16. Kittler H, Pehamberger H, Wolff K, Binder M. Diagnostic accuracy of dermoscopy. REFERENCES Lancet Oncol. 2002;3:159-165. 17. Argenziano G, Soyer HP, Chimenti S, et al. Dermoscopy of pigmented skin lesions: results of a consensus meeting via the Internet. J Am Acad Dermatol. 2003; 1. Sutton RL. An unusual variety of vitiligo (leucoderma acquisitum centrifugum). 48:679-693. J Cutan Dis. 1916;34:797-800. 18. Hofmann-Wellenhof R, Blum A, Wolf IH, et al. Dermoscopic classification of atypi- 2. Kopf AW, Morril SD, Silberberg I. Broad spectrum of leukoderma acquisitum cal melanocytic nevi (Clark nevi). Arch Dermatol. 2001;137:1575-1580. centrifugum. Arch Dermatol. 1965;92:14-35. 19. Teban L, Pehamberger H, Wolff K, et al. Klinische Wertigkeit einer dermatosko- 3. MacKie RM. Disorders of the cutaneous melanocyte: halo naevus. In: Burns T, pischen Klassifikation von Clark-Na¨vi. J Dtsch Dermatol Ges. 2003;1:292-296. Breathnach S, Cox N, Griffith C, eds. Rook’s Textbook of Dermatology. Vol 2. 20. Zalaudek I, Grinschgl S, Argenziano G, et al. Age-related prevalence of dermos- 7th ed. Oxford, England: Blackwell Scientific Publications; 2004:1-39. 4. Fritsch P. Dermatologie und Venerologie. Grundlagen, Klinik und Atlas. 2nd ed. copy patterns in acquired melanocytic nevi. Br J Dermatol. 2006;154:299-304. Berlin, Germany: Springer; 2004:627. 21. Menzies SW, Gutenev A, Avramidis M, Batrac A, McCarthy WH. Short-term digi- 5. Herd RM, Hunter JA. Familial halo naevi. Clin Exp Dermatol. 1998;23:68-69. tal surface microscopic monitoring of atypical or changing melanocytic lesions. 6. Kerr OA, Schlofield O. Halo congenital nevi. Pediatr Dermatol. 2003;20:541-542. Arch Dermatol. 2001;137:1583-1589. 7. Mooney MA, Barr RJ, Buxton MG. Halo nevus or halo phenomenon? a study of 22. Kittler H, Seltenheim M, Dawid M, Pehamberger H, Wolff K, Binder M. Fre- 142 cases. J Cutan Pathol. 1995;22:342-348. quency and characteristics of enlarging common melanocytic nevi. Arch Dermatol. 8. Akasu R, From L, Kahn HJ. Characterization of the mononuclear infiltrate in- 2000;136:316-320. volved in regression of halo nevi. J Cutan Pathol. 1994;21:302-311. 23. Kittler H, Pehamberger H, Wolff K, Binder M. Follow-up of melanocytic skin le- 9. Harvell JD, Meehan SA, LeBoit PE. Spitz’s nevi with halo reaction: a histopatho- sions with digital epiluminescence microscopy: patterns of modifications ob- logic study of 17 cases. J Cutan Pathol. 1997;24:611-619. served in early melanoma, atypical nevi, and common nevi. J Am Acad Dermatol. 10. Zeff RA, Freitag A, Grin CM, Grant-Kels JM. The immune response in halo nevi. 2000;43:467-476. J Am Acad Dermatol. 1997;37:620-624. 24. Kittler H, Binder M. Follow-up of melanocytic skin lesions with digital dermos- 11. Koh HK, Sober AJ, Nakagawa H, et al. Malignant melanoma and vitiligo-like leuko- copy: risks and benefits. Arch Dermatol. 2002;138:1379. derma: an electron microscopic study. J Am Acad Dermatol. 1983;9:696-708. 25. Skvara H, Teban L, Fiebiger M, Binder M, Kittler H. Limitations of dermoscopy in 12. Fishman HC. Malignant melanoma arising with two halo nevi. Arch Dermatol. the recognition of melanoma. Arch Dermatol. 2005;141:155-160. 1976;112:407-408. 26. Zalaudek I, Argenziano G, Ferrara G, et al. Clinically equivocal melanocytic skin 13. Pantoja E, Wendth AJ, Beecher TS. Perilesional vitiligo in melanoma. Cutis. 1977; lesions with features of regression: a dermoscopic-pathological study. Br J 19:51-53. Dermatol. 2004;150:64-71.

(REPRINTED) ARCH DERMATOL/ VOL 142, DEC 2006 WWW.ARCHDERMATOL.COM 1632