Low rates of clinical recurrence after biopsy of benignto moderately dysplastic melanocytic nevi

Agnessa Gadeliya Goodson, MD,a Scott R. Florell, MD,a Kenneth M. Boucher, PhD,b,c and

Douglas Grossman, MD, PhDa,b,c

Salt Lake City, Utah

From

ce

H

Supp

C

G

Conf

Acce

Repr

Corre

C

C

Publ

0190

ª 20

doi:1

Background: Little is known about the recurrence/persistence rates of dysplastic nevi (DN) after biopsy,and whether incompletely removed DN should be re-excised to prevent recurrence.

Objective: Our purpose was to determine the recurrence rates of previously biopsied DN, and to assesswhether biopsy method, margin involvement, congenital features, epidermal location, and degree ofdysplasia are associated with recurrence.

Methods: Patients having a history of a ‘‘nevus biopsy’’ at least 2 years earlier were assessed for clinicalrecurrence. Slides of original lesions were re-reviewed by a dermatopathologist.

Results: A total of 271 nevus biopsy sites were assessed in 115 patients. Of 195 DN with greater than 2 yearsof follow-up, 7 (3.6%) demonstrated recurrence on clinical examination. In all, 98 DN had a follow-upperiod of at least 4 years with no clinical recurrence. Of 61 benign nevus biopsy sites examined, clinicalrecurrence was observed in two (3.3%). For all nevi, recurrence was significantly associated with shave biopsytechnique but not with nevus dysplasia or subtype, or the presence of positive margin or congenital features.

Limitations: Most biopsies were performed in a pigmented lesion clinic at a single tertiary referral center.Determinations of nevus recurrence were made on clinical rather than histologic grounds, and follow-uptimes were limited in some cases.

Conclusion: In this cohort, rates of clinical recurrence after biopsy of DN and benign nevi were extremelylow. Re-excision of nevi, including mildly to moderately DN with a positive margin, may not be necessary.( J Am Acad Dermatol 2010;62:591-6.)

Key words: biopsy; dysplastic; nevi; recurrence.

anagement of dysplastic nevi (DN) after agree that DN demonstrating severe atypia should

M biopsy remains a controversial issue.Although most dermatologists would

the Departments of Dermatologya and Oncological Scien-

s,b and the Huntsman Cancer Institute,c University of Utah

ealth Sciences Center.

orted by the Department of Dermatology, the Huntsman

ancer Foundation, and the National Institutes of Health (Dr

rossman)

licts of interest: None declared.

pted for publication June 25, 2009.

ints not available from the authors.

spondence to: Douglas Grossman, MD, PhD, Huntsman

ancer Institute, Suite 5262, 2000 Circle of Hope, Salt Lake

ity, UT 84112. E-mail: doug.grossman@hci.utah.edu.

ished online December 17, 2009.

-9622/$36.00

09 by the American Academy of Dermatology, Inc.

0.1016/j.jaad.2009.06.080

be re-excised because they may represent earlymelanoma or a lesion evolving into melanoma, thereare no clear guidelines regarding whether an incom-pletely removed nevus with a mild or moderatedegree of dysplasia should be re-excised.Discordance among dermatopathologists as to iden-tifying dysplasia and differing degrees of atypiafurther complicate decision-making.1 Although a1992 National Institutes of Health ConsensusConference established margin guidelines for re-excision of DN (0.2-0.5 cm), it did not specify indi-cations for re-excision.2 In addition to potential risk ofmelanoma development, nevus recurrence may beassociated with histologic ‘‘pseudomelanoma,’’ a be-nign process simulating melanoma that poses a diag-nostic dilemma for the dermatopathologist.3

Most studies investigating recurrence of biop-sied nevi examined benign nevi (BN) removed for

591

J AM ACAD DERMATOL

APRIL 2010

592 Goodson et al

cosmetic purposes, and reported recurrence ratesranging from 6% to 41%.4-6 Gambichler et al7

prospectively evaluated the effectiveness of deepshave excision in removing macular melanocyticnevi and the cosmetic outcome. For 77 nevibiopsied in 45 patients, histologic evaluationrevealed that 88% of lesions had clear margins,

CAPSULE SUMMARY

d It is unclear whether incompletelyremoved dysplastic nevi should bere-excised to prevent recurrence.

d For 175 dysplastic nevi and 61 benignnevi without earlier re-excision and withgreater than 2 years of follow-up,recurrence on clinical examination wasfound in 4.0% and 3.3% of lesions,respectively.

d Nevus recurrence was significantlyassociated with shave biopsy technique.

d The low rates of recurrence after biopsysuggest that re-excision of nevi,including dysplastic nevi with a positivemargin, may not be necessary.

and 60% had atypical ordysplastic features.7 After 6months 56 of 77 biopsy siteswere reassessed, and 7 of56 (13%) were found tohave clinical recurrence.7

Given the limited data onrecurrence rates of biopsiedDN, and lack of clear guide-lines for re-excision of DN,it is not surprising that thereis significant variability inphysician management. Asurvey of 145 fellows ofthe American Academy ofDermatology found thatamong 45% of responderswho provided a reason forre-excising incompletely re-moved DN, the most com-mon was a finding of

moderate or severe cytologic atypia.8 In addition,53% stated that they re-excise incompletely re-moved DN in the majority of cases.8 In this study,we sought to determine recurrence rates of previ-ously biopsied DN, with the hope of providingsome guidance as to whether incompletely re-moved DN should be re-excised.

METHODSPatients and nevus biopsies

This study was approved by our universityinstitutional review board. Patients were recruitedfrom our pigmented lesion clinic, in which ap-proximately 7 new patients are seen each weekand more than 1000 patients with history ofnumerous or atypical nevi and/or personal orfamily history of melanoma are monitored. Thecharts of patients scheduled for examination werereviewed before their visit to determine whetherany previously biopsied melanocytic nevi wereappropriate for the study. Earlier biopsies hadbeen performed between 1998 and 2007, predom-inantly by physicians at our institution and somefrom the surrounding area. Of the 271 biopsiesstudied (Table I), shave technique was used in 163(60%), punch technique was used in 74 (27%), andelliptical excision was used in 34 (13%). Biopsy

specimens from most anatomic areas were in-cluded, with the majority (55%) from the trunk.Lesions that were nonmelanocytic, had demon-strated recurrent nevus on initial biopsy specimen,or were biopsied less than 2 years before the visit;those without a pathology report; and those forwhich a biopsy scar could not be identified on the

patient or the original slidecould not be obtained forreview were excluded.During the visit, sites ofpreviously biopsied melano-cytic nevi that met inclusioncriteria were assessed forclinical recurrence. No le-sions that had been re-excised (20 DN and 5 othernevi) (Table I) after initialbiopsy recurred, and thesewere not further analyzedfor histologic features.

Histologic examinationHematoxylin and eosine

stained slides were obtained,and re-reviewed by a derma-topathologist (S. R. F.) to de-termine whether nevi were

benign or dysplastic, nevus subtype (junctional,compound, intradermal), degree of dysplasia, margininvolvement, and presence of congenital features.Diagnosis of common (BN) and DN was based onwell-established architectural and cytologic criteria.The DN exhibited architectural disorder manifestedby irregular placement of variably sized melanocytictheques at the tips and sides of elongated and some-times bridged rete. Patterned fibroplasia of the pap-illary dermis was present. In some specimens, ‘‘mild’’cytologic atypia of melanocytes was characterized bynuclear enlargement similar to the size of a keratino-cyte nucleus with finely granular pigmented cyto-plasms. Dermal melanocytes were arrayed in thequesthat showednuclear and cytoplasmicmaturationwithprogressive descent. ‘‘Moderately’’ atypical DN dem-onstrated prominent fibroplasia of the dermis withentrapment of dermal melanocytic theques and ahostresponse of lymphocytes. Significant pagetoid scatterof melanocytes was not observed in any of the nevi.Nevi with congenital features demonstrated dermalmelanocytes arrayed in theques, sheets, or cords thatsplayed reticular dermal collagen bundles with closeassociation of nevus cells with adnexal structures.Positive margins were defined as melanocytic the-ques, either in the epidermis or dermis, that wereidentified in inked specimen margins. Negative

Table I. Nevi assessed for recurrence in this study

Nevus type Total No. Subtype Biopsy method Re-excised

DNMild atypia 179 56/179 Junctional

123/179 Compound96/179 Shave57/179 Punch26/179 Excision

17/179

Moderate atypia 16 3/16 Junctional13/16 Compound

8/16 Shave3/16 Punch5/16 Excision

3/16

Severe atypia 0 e e eBN 61 35/61 Compound

26/61 Intradermal50/61 Shave

8/61 Punch3/61 Excision

0/61

Other nevi (Spitz, blue) 15 8/15 Compound7/15 Intradermal

9/15 Shave6/15 Punch

5/15

BN, Benign nevi; DN, dysplastic nevi.

Fig 1. Nevus biopsy sites demonstrating clinical lack (A) or presence (B) of nevus recurrence.

J AM ACAD DERMATOL

VOLUME 62, NUMBER 4

Goodson et al 593

margins were defined as lack of melanocytic thequesin inked margins. Close extension (\0.2 mm) ofeither the junctional or dermal component to aninked margin was considered to be a positive margin.

Statistical analysisTwo-sided Fisher exact tests were used for all

comparisons. Statistical analysis was performed us-ing R 2.8.0 (The R Foundation for StatisticalComputing, Vienna, Austria). P values of .05 or lesswere considered statistically significant.

RESULTSRates of nevus recurrence

During an 8-month period, 271 sites of previ-ously biopsied melanocytic nevi were assessed forclinical recurrence in 115 patients (Table I). In mostcases, nevi at the follow-up visit presented as awell-healed hypopigmented scar (Fig 1, A). In somecases, on the other hand, pigmentation was seenwithin the scar (Fig 1, B), and interpreted as nevusrecurrence. After histologic re-evaluation of original

biopsy specimens, 195 lesions were classified asDN, 61 were classified as BN, and 15 were classifiedas ‘‘other’’ (which included 6 blue nevi, 8 Spitz nevior spindle cell nevi, and one combined Spitz/bluenevus). The majority (179/195) of DN in this studywere mildly dysplastic, whereas 16 of 195 lesionswere moderately dysplastic (Table I). There wereno DN with severe dysplasia. None of the 25 lesions(9.2% of total) that had been re-excised after biopsywere found to have recurred. After excluding theselesions that had been previously re-excised, only 7of 175 (4.0%) DN showed clinical recurrence at least2 years after biopsy (Table II). Of the 175 DN, 93lesions showed no evidence of recurrence 4 yearsor more after biopsy. Of the 61 BN, only two (3.3%)had clinical evidence of recurrence greater than 2years after initial biopsy (Table II). Thus, ourobserved rates of recurrence for both types ofnevi were very low, and without statistical differ-ence (P = 1.00) between the two groups. None ofthe blue nevi or Spitz/spindle nevi demonstratedrecurrence.

Table II. Features of clinically recurrent and nonrecurrent nevi

Nevus type No. Subtype Biopsy method Positive margin

Congenital

features

DN, mild 162Recurrent 7/162 1/7 Junctional

6/7 Compound7/7 Shave 5/7 3/7

Nonrecurrent 155/162 51/155 Junctional104/155 Compound

76/155 Shave54/155 Punch25/155 Excision

60/155 38/155

DN, moderate 13Recurrent 0/13 e e e eNonrecurrent 13/13 2/13 Junctional

11/13 Compound5/13 Shave3/13 Punch5/13 Excision

4/13 4/13

BN 61Recurrent 2/61 2/2 Compound 2/2 Shave 2/2 2/2Nonrecurrent 59/61 33/59 Compound

26/59 Intradermal48/59 Shave

8/59 Punch3/59 Excision

53/59 58/59

Other nevi 10Recurrent 0/10 e e e eNonrecurrent 10/10 6/10 Compound

4/10 Intradermal5/10 Shave5/10 Punch

4/10 1/10

BN, Benign nevi; DN, dysplastic nevi.

Nevi that were re-excised after original biopsy (Table I) are excluded here.

J AM ACAD DERMATOL

APRIL 2010

594 Goodson et al

Role of degree of dysplasia, lesion subtype,biopsy method, margin involvement, andcongenital features

We next analyzed whether particular nevus fea-tures and biopsy method were associated withrecurrent versus nonrecurrent nevi, excluding thelesions that had been re-excised after biopsy. Thebreakdown by lesion type, subtype, biopsy method,margin involvement, and presence of congenitalfeatures is detailed in Table II. Most DN that recurred(6/7) were of compound type and one was junc-tional, and all (7/7) had mild dysplasia and had beenbiopsied by shave technique. None of 13 DN withmoderate dysplasia recurred. The different recur-rence rates between DN with mild (7/162) versusmoderate (0/13) dysplasia were not statistically sig-nificant (P = 1.00). Of the 7 recurrent DN, 5 had apositive margin and 3 had congenital features. Of 168nonrecurrent DN, 53 (32%) were junctional and 115(68%) were compound. These nonrecurrent DNrepresented lesions biopsied by shave (48%), punch(34%), and excisional (18%) technique, and 64 of 168(38%) had a positive margin whereas 42 of 168 (25%)had congenital features. For all DN, there was not astatistically significant association of recurrence withnevus subtype (junctional or compound, P = .44),presence of positive margin (P = .11), or congenitalfeatures (P = .38). On the other hand, for all DN,shave biopsy technique was significantly (P = .033)associated with recurrence.

Of the two recurrent BN, both were compoundtype, were biopsied by shave technique, and dem-onstrated a positive margin and congenital features.Of the other 59 BN that did not recur, 33 (56%) werecompound and 26 (44%) were intradermal. Thesenonrecurrent BN represented lesions biopsied byshave (81%), punch (14%), and excisional (5%)technique, and 53 of 59 (90%) had a positive marginwhereas 58 of 59 (98%) had congenital features. Forall BN, there was not a statistically significant asso-ciation of recurrence with nevus subtype (com-pound or intradermal, P = .50), presence of positivemargin (P = 1.00), or congenital features (P = 1.00).In contrast to the DN, shave biopsy technique wasnot significantly (P = 1.00) associated with recur-rence of BN. Considering all the nevi (DN and BNcombined), shave biopsy technique was significantly(P = .045) associated with recurrence; for lesionsbiopsied only by shave or punch technique, theassociation between shave and recurrence was ofeven greater significance (P = .032).

DISCUSSIONIt is clear from the literature that patients with DN

are at increased risk for developing melanoma.9-11

Although 20% to 50% of melanomas appear to arisefrom a pre-existing nevus,12-14 the annual risk ofindividual nevi transforming into a melanoma isextremely loweestimated to be only 1 in 200,000.15

The annual risk is higher for DN (estimated 1 in

J AM ACAD DERMATOL

VOLUME 62, NUMBER 4

Goodson et al 595

10,000),16 raising the question of whether incom-pletely removed DN should be re-excised with clearmargins to prevent potential evolution into mela-noma. Although most dermatologists would agreethat DN demonstrating severe dysplasia should bere-excised given the risk of early or evolving mela-noma, management of incompletely excised DNdemonstrating mild or moderate dysplasia remainsan open question. A key factor to consider is thelikelihood of recurrence, which must be balancedagainst the cost of re-excision and risk associatedwith a surgical procedure, including a larger scar. Toour knowledge, the current study of DN recurrenceafter biopsy involves the largest number of lesionsand longest follow-up period reported in theliterature.

We found very low (3%-4%) recurrence rates forboth BN and DN after biopsy, regardless of margininvolvement, nevus subtype (junctional, compound,intradermal), or the presence of congenital features.We might have expected the recurrence rate ofincompletely excised DN to be significantly higherthan that of BN, given that DN are associated withincreased proliferation and decreased senes-cence.17,18 In addition, we might have expected therecurrence rate to be higher for compound thanintradermal lesions because nevus cells are morelikely to be proliferative and less differentiated in acompound nevus. The recurrence rates we observedwere much lower than those seen in previous stud-ies.4-7 Our lower recurrence rates are unlikely a resultof differences in follow-up times, because our fol-low-up period was greater than that in most of theother studies. A more likely explanation is that wemay perform deeper and broader shave biopsies ofclinically atypical nevi in our patients at high risk inan attempt to remove nevi completely, whereasearlier studies primarily examined recurrence of BNremoved for cosmetic purposes, where a moresuperficial shave biopsy may have been done tominimize scarring.

We also might have expected the presence ofcongenital features to be associated with nevusrecurrence, given that nevus cells in congenitalnevi tend to extend deeper into the skin throughtheir involvement with vascular and follicular struc-tures. However, because we did not rebiopsy thenevus sites in this study, it is possible that some caseslacking clinical evidence of recurrence might dem-onstrate residual nevus cells beneath the scar. Even ifwe did find nevus persistence in some sites onhistologic examination, however, the lack of visiblyapparent change over time would suggest limitedclinical significance. It is also possible that clinicallyapparent nevus in the biopsy scar may represent

pigmentation in keratinocytes or melanophagesrather than persistent or recurrent nevus cells.

Although there was an association between pos-itive margin and recurrence, it was not statisticallysignificant (P = .17 for all DN and BN combined,P = .11 for DN), perhaps because of the low numbersof recurrences. Lack of greater association of recur-rence with margin involvement was somewhat sur-prising, given that positive margin is the justificationoften used for re-excision of DN.8 Failure of neviwith positive margins to recur suggests that in mostcases, residual nevus cells in the biopsy wound arenot of sufficient number (or do not have the capac-ity) for regeneration and pigment production.

The only statistically significant association foundwith nevus recurrence was biopsy method, withshave technique being significantly associated withrecurrence. One potential explanation for a higherrecurrence rate with shave biopsies compared withpunch biopsies is that nevus recurrence may be morelikely to originate from a deep rather than lateralmarginewhich indeed we observed in most recur-rent nevi in this study (Fig 1, B). Because shavebiopsies are generally more superficial than punchbiopsies, they would be more likely to have apositive deep margin. Another explanation for theassociation between shave technique and recurrenceis that lesions that are shaved tend to be much largerthan those selected for punch biopsy, where thediameter of the punch tool (usually # 6 mm) neces-sarily limits application of punch biopsies to smalllesions. Larger lesions may contain more prolifera-tive cells or be more likely to recur from residual cellsfor other reasons that are presently unclear.

One of the factors complicating management ofDN is potential variability in histologic interpretation.In one study addressing interobserver variation inhistologic diagnosis of atypical nevi, lesions inter-preted as DN by an expert panel were diagnosed asmelanoma by other pathologists in 21% of cases.19

Conversely, lesions originally diagnosed as thin or insitu melanomas were re-read as DN in 12% of thecases.19 In our study, all of the histologic slides werere-reviewed by the same dermatopathologist to limitinterobserver variation. In addition, routine histo-logic evaluation of punch or shave biopsy specimenstypically involves examination of only a fraction ofthe lesion to determine degree of atypia and margininvolvement, leading some to propose that all DNshould undergo excisional biopsy to obtain the mostaccurate diagnosis.20 A recent study from our insti-tution, however, found that in a majority of biopsiednevi the histologic findings were homogeneous suchthat the diagnostic information in one section couldbe extrapolated to the remainder of the specimen.21

J AM ACAD DERMATOL

APRIL 2010

596 Goodson et al

We recognize the possibility that some lesions mayhave had unrecognized margin involvement, andthat the degree of dysplasia may have been under-estimated. However, this information would nothave affected the low rate of clinical recurrencethat was observed.

Our results are consistent with those of Kmetzet al,22 who found that no melanomas developedduring a 5-year period after biopsy of 55 atypical nevi(26 lesions with at least one positive margin and 29with clear margins) that were not re-excised. Basedon these findings, the authors recommended obser-vation as a safe alternative to re-excision for incom-pletely removed atypical nevi.22 To fully answer thequestion of whether incompletely removed DN withmild or moderate dysplasia should be re-excised,longer follow-up of more lesions is required. We stillbelieve that nevi with severe dysplasia should bere-excised to ensure complete removal, given thatsuch lesions may represent evolving melanoma orcould later be reinterpreted as melanoma given thepossibility of interobserver variation among pathol-ogists19 as noted above. However, our data suggestthat lesions that demonstrate only mild or moderatedysplasia may not need to be re-excised given theirlow likelihood of recurrence, and can be followedclinically for evidence of recurrence or developmentof any concerning features.

We thank Glen Bowen for helpful discussions in theplanning phase of this study.

REFERENCES

1. Shapiro M, Chren MM, Levy RM, Elder DE, LeBoit PE, Mihm MC

Jr, et al. Variability in nomenclature used for nevi with

architectural disorder and cytologic atypia (microscopically

dysplastic nevi) by dermatologists and dermatopathologists.

J Cutan Pathol 2004;31:523-30.

2. National Institutes of Health Consensus Conference. Diagnosis

and treatment of early melanoma. JAMA 1992;268:1314-9.

3. Kornberg R, Ackerman AB. Pseudomelanoma: recurrent mel-

anocytic nevus following partial surgical removal. Arch Der-

matol 1975;111:1588-90.

4. Breuninger H, Garbe C, Rassner G. Shave excision of melano-

cytic nevi of the skin: indications, technique, results. Hautarzt

2000;51:575-80.

5. Bong JL, Perkins W. Shave excision of benign facial melano-

cytic nevi: a patient’s satisfaction survey. Dermatol Surg 2003;

29:227-9.

6. Ferrandiz L, Moreno-Ramirez D, Camacho FM. Shave excision of

common acquired melanocytic nevi: cosmetic outcome, recur-

rences, and complications. Dermatol Surg 2005;31:1112-5.

7. Gambichler T, Senger E, Rapp S, Alamouti D, Altmeyer P,

Hoffmann K. Deep shave excision of macular melanocytic nevi

with the razor blade biopsy technique. Dermatol Surg 2000;26:

662-6.

8. Fung MA. Terminology and management of dysplastic nevi:

responses from 145 dermatologists. Arch Dermatol 2003;139:

1374-5.

9. Holly EA, Kelly JW, Shpall SN, Chiu SH. Number of melanocytic

nevi as a major risk factor for malignant melanoma. J Am Acad

Dermatol 1987;17:459-68.

10. Bataille V, Bishop JA, Sasieni P, Swerdlow AJ, Pinney E, Griffiths

K, et al. Risk of cutaneous melanoma in relation to the

numbers, types and sites of nevi: a case-control study. Br J

Cancer 1996;73:1605-11.

11. Tucker MA, Halpern A, Holly EA, Hartge P, Elder DE, Sagebiel

RW, et al. Clinically recognized dysplastic nevi: a central risk

factor for cutaneous melanoma. JAMA 1997;277:1439-44.

12. Marks R, Dorevitch AP, Mason G. Do all melanomas come from

‘‘moles’’? A study of the histological association between

melanocytic nevi and melanoma. Australas J Dermatol 1990;

31:77-80.

13. Kelly JW, Yeatman JM, Regalia C, Mason G, Henham AP. A high

incidence of melanoma found in patients with multiple

dysplastic nevi by photographic surveillance. Med J Aust

1997;167:191-4.

14. Bevona C, Goggins W, Quinn T, Fullerton J, Tsao H. Cutaneous

melanomas associated with nevi. Arch Dermatol 2003;139:

1620-4.

15. Tsao H, Bevona C, Goggins W, Quinn T. The transformation

rate of moles (melanocytic nevi) into cutaneous melanoma: a

population-based estimate. Arch Dermatol 2003;139:282-8.

16. Naeyaert JM, Brochez L. Clinical practice: dysplastic nevi.

N Engl J Med 2003;349:2233-40.

17. Tuthill RJ, Reed RJ. Failure of senescence in the dysplasia-

melanoma sequence: demonstration using a tissue microarray

and a revised paradigm for melanoma. Semin Oncol 2007;34:

467-75.

18. Adams PD, Enders GH. Wnt signaling and senescence: a tug of

war in early neoplasia? Cancer Biol Ther 2008;7:1706-11.

19. Brochez L, Verhaeghe E, Grosshans E, Haneke E, Pierard G,

Ruiter D, et al. Inter-observer variation in the histopathological

diagnosis of clinically suspicious pigmented skin lesions.

J Pathol 2002;196:459-66.

20. Culpepper KS, Granter SR, McKee PH. My approach to atypical

melanocytic lesions. J Clin Pathol 2004;57:1121-31.

21. Florell SR, Smoller BR, Boucher KM, Grossman D, Harris RM,

Bowen GM, et al. Sampling of melanocytic nevi for research

purposes: a prospective, pilot study to determine effect on

diagnosis. J Am Acad Dermatol 2008;59:814-21.

22. Kmetz EC, Sanders H, Fisher G, Lang PG, Maize JC Sr. The role

of observation in the management of atypical nevi. South Med

J 2009;102:45-8.