A, Junctional melanocytic proliferation with focal bridging of rete ridges by atypical melanocytes with mild-to-moderate cytologic atypia (originally read as junctional melanocytic nevus with mild dysplasia). B, High-power view of areas of mild-to-moderate cytologic atypia and bridging of melanocytic nests.
A, Irregular nesting and extensive lentiginous proliferation with extension beyond the biopsy scar (read as atypical intraepidermal melanocytic proliferation, favor early melanoma in situ, lentigo maligna type). B, Melan-A stain highlights the confluent junctional atypical melanocytes with irregular nesting and pagetoid upward scatter that extends to the biopsy margin. C, Hyperchromatic, enlarged nuclei of the atypical melanocytes are observed with extensive pagetoid scatter. D, Melan-A stain aids in the identification of the atypical melanocytes present in nests and lentiginous spread within the epidermis.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Fleming NH, Egbert BM, Kim J, Swetter SM. Reexamining the Threshold for Reexcision of Histologically Transected Dysplastic Nevi. JAMA Dermatol. 2016;152(12):1327–1334. doi:https://doi.org/10.1001/jamadermatol.2016.2869
What proportion of incompletely excised, mildly to moderately dysplastic nevi progress to melanoma with long-term patient follow-up?
In this cohort study of 590 dysplastic nevi, favorable patient outcomes were observed over 20 years in most dysplastic nevi with positive histologic margins that were not further excised. Only 1 case of melanoma in situ arose 5 years later from an excisionally biopsied moderately dysplastic nevus with positive histologic margins, and 5 other melanomas developed at partial biopsy sites that likely represented sampling error.
Mildly to moderately dysplastic nevi biopsied with excisional intent but with histologically involved margins can generally be clinically monitored without reexcision.
Controversy persists regarding the appropriate management of incompletely excised, biopsy-proven, mild and moderate dysplastic nevi (DN).
To determine long-term risk of associated melanoma in biopsied mild or moderate DN with positive histologic margins that were clinically observed vs reexcised with negative margins.
Design, Setting, and Participants
Retrospective cohort study of mixed referral and community patients from an academic pigmented lesion clinic and dermatology clinics of the affiliated Veteran Affairs medical center with biopsy-confirmed DN with positive histologic margins diagnosed from May 15, 1991, to July 8, 2015, and followed up through May 30, 2016. A consecutive sample of 1473 histologically confirmed DN was identified using surgical pathology databases at the study sites; 590 cases in 498 patients met eligibility criteria.
Main Outcomes and Measures
The primary outcome was the proportion of biopsied DN that progressed to histologically confirmed invasive or in situ melanoma. Secondary outcomes included local nevus recurrence and development of primary melanoma at other anatomic sites.
The 498 patients had a mean (range) age of 57.6 (14-93) years and 90% were male. Among 590 positive-margin DN, 191 were reexcised and 399 clinically observed without further surgery; 170 reexcised and 304 observed DN had available follow-up data, with mean (SD) follow-up of 5.5 (4.6) years. Cases in the observation group were more likely to demonstrate nevus recurrence than those that were reexcised (3.3% vs 0%; P = .02). Six of 304 (2.0%) observed DN subsequently developed melanoma at the same site, compared with 1 of 170 (0.6%) that were reexcised (P = .43). Five of 6 observed patients who developed melanoma initially underwent partial biopsy with grossly positive margins; 1 melanoma in situ evolved from an excisionally biopsied moderately dysplastic nevus 5 years later. Only 1 case of thin invasive melanoma (≤1 mm) was observed, and no deaths from melanoma arising from biopsy-proven DN occurred through the latest dermatology follow-up. New primary melanoma developed at other sites in 9.9% of excised and 9.4% of resected DN.
Conclusions and Relevance
In cases of mild and moderate DN with microscopically positive margins and no concerning clinical residual lesion, observation, rather than reexcision, was a reasonable management option. Partial biopsies of pigmented lesions suspicious for melanoma may lead to delayed melanoma diagnosis and should be discouraged.
Atypical nevi (AN) are melanocytic neoplasms with clinical features that may simulate melanoma (eg, diameter >6 mm, topographical asymmetry, color variegation, border irregularity) but are largely benign. The histologic correlate of the clinical atypical nevus is commonly termed a “dysplastic nevus,” which is believed by some to lie on the spectrum between a common (or “typical”) nevus and melanoma.1,2 Others have promoted the concept that dysplastic nevi (DN) are a distinct but entirely benign histomorphologic nevus variant (eg, the so-called Clark nevus).3-5(pp273-284) The clinical significance of AN and DN is complicated by the fact that whereas a small proportion of these lesions have the potential to transform into primary melanoma, the majority remain clinically and biologically stable and pose no malignant risk—or at least no greater risk than do common or typical nevi.6,7 The presence of clinical AN and/or histologic DN is an established risk factor for melanoma,8-13 but rates of nevus-associated melanoma appear to be low, with recent estimates ranging from 18% to 26%.7,14,15
The optimal clinical management of histologic DN remains controversial and has been compounded by a lack of evidence regarding subsequent melanoma development in cases in which incompletely excised lesions were observed rather than surgically removed. For histologically transected severe dysplasia, there is general agreement that reexcision is appropriate, typically with 2- to 5-mm margins, as proposed by the National Institutes of Health Consensus Development Conference in 1991.16-18 However, there is substantial variability in opinion and clinical practice regarding the need for further surgery in cases of incompletely excised mild or moderate DN.17 Concerns for “missing” melanoma in histologically transected DN have resulted in the common practice of reexcision, regardless of the extent of cytomorphologic atypia on initial biopsy.19,20 This may be the most prudent course of action in the setting of small, partial biopsy of a clinical AN, in which incomplete sampling fails to identify an existing melanoma.
In contrast, recent retrospective studies examining outcomes of mild to moderate DN that were clinically observed have not identified any cases with positive histological margins that progressed to melanoma, suggesting that clinical observation may be appropriate, particularly when excisional biopsy was intended at the outset.21-23 This view has been expressed in recent literature reviews and is the consensus opinion of the Pigmented Lesion Subcommittee of the Melanoma Prevention Working Group.24,25 However, a total of only 210 DN were examined in these studies, limiting the ability to draw conclusions regarding appropriateness of reexcision. The Pigmented Lesion Subcommittee is embarking on a multicenter study of up to 3000 retrospectively collected histologic DN to identify risk of malignant transformation in excisionally biopsied lesions with positive histologic margins.
In the meantime, additional institutional high-quality, long-term analyses of patient outcome, including melanoma development in incompletely excised DN, are critical to provide evidence to guide consensus recommendations for the most appropriate management. The aim of this study was to ascertain the risk of melanoma resulting from incompletely excised mild or moderate DN managed by observation, using a unique data set with a limited number of academic dermatopathologists and dermatopathologists who diagnosed DN and observed the patients over a 20-year period in an academic pigmented lesion clinic and the Veterans Affairs medical center affiliate. Secondary aims included rates of nevus recurrence and development of melanoma at other sites.
The study population consisted of patients who received a diagnosis of histologic DN at the Veterans Affairs Palo Alto Health Care System (VAPAHCS) and the Pigmented Lesion and Melanoma Clinic (PLMC) at Stanford University Medical Center (SUMC) and Cancer Institute, the latter directed since January 1995 by a single clinician (S.M.S.), who also guided the treatment of most patients with DN at VAPAHCS during the study period. A single dermatopathologist (B.M.E.) has assisted in histopathologic diagnoses of DN at VAPAHCS since 1980, and similar diagnostic criteria were used to grade DN by the Stanford dermatopathologists over the entire study period.2 These criteria consisted of a combination of architectural disorder (asymmetry, superficial dermal concentric fibroplasia, and rete ridge bridging between adjacent rete ridges) and random cytologic atypia of melanocytes. In compound lesions, extension of the junctional component beyond the dermal component (shouldering) was characteristic, and a lymphocytic infiltrate host response, as well as superficial dermal vessels, was frequently encountered.26
Search terms were used to extract eligible cases from surgical pathology databases at VAPAHCS (May 15, 1991, through May 27, 2015) and at SUMC (June 26, 1997, through July 8, 2015); 2417 unique pathology records were identified (2228 VAPAHCS, 189 SUMC) and subsequently screened. Fewer cases were eligible at SUMC because of the smaller patient base of the specialized PLMC, a higher clinical threshold for biopsy of AN, and a greater percentage of histologically confirmed excisional biopsies performed at the outset. Cases were included only if they represented the initial biopsy of the site in question, and the final pathologic diagnosis consisted of dysplastic nevus (accounting also for variations in terminology such as “melanocytic nevus with dysplasia or cytologic atypia”). Reasons for exclusion included absence of dysplastic features, presence of suspected or definitive melanoma in situ, histologic features suggestive of early or evolving melanoma in situ (MIS), lentigo maligna (LM) type, that is, atypical intraepidermal melanocytic proliferation (AIMP), and biopsies or reexcisions of previously biopsied DN. Patients and pathology records containing multiple unique diagnoses of DN occurring on distinct sites were included as separate cases for the purpose of the study. Of 1473 eligible histologically confirmed DN (1287 VAPAHCS, 186 SUMC), 590 cases (399 VAPAHCS, 191 SUMC) from 498 patients (447 VAPAHCS, 51 SUMC) had positive biopsy margins and known excision status and were further analyzed.
For each eligible case, clinical, histopathologic, and dermatology follow-up data were extracted from the corresponding pathology reports and medical record. Clinical information included patient age, sex, race/ethnicity (if known), personal and family history of melanoma, estimated nevus count, presence of atypical mole phenotype, and prior UV radiation exposure (if documented). Pathologic data included initial and subsequent biopsy and/or excision dates, anatomic site, and method (partial or excisional in intent, if reported); prebiopsy nevus dimensions; degree of histopathologic atypia/dysplasia (identified as mild, moderate, or severe at SUMC and as mild, mild to moderate, moderate, moderate to severe, and severe at VAPAHCS); and presence, degree, and location (peripheral vs deep) of histologic margin involvement. All histologic grades of dysplasia were included to reduce the potential for interobserver variability in DN diagnosis among the study pathologists. Patient follow-up (by dermatology) occurred through May 30, 2016, and included dermatologic management (observation vs reexcision) of the dysplastic nevus, nevus recurrence at the biopsy site (defined as clinical repigmentation that was subsequently biopsied and histologically confirmed), diagnosis of melanoma at the biopsy site, diagnosis of melanoma at another anatomic site, and patient outcome in cases of melanoma (eg, death from disease or not). For the purpose of this study, reexcision was defined as an excisional procedure resulting in negative histologic margins that was performed at the first follow-up visit or shortly thereafter.
For all cases of melanoma that developed at the initial biopsy site, slides and blocks from all procedures (biopsies and excisions) were obtained and reviewed by a study dermatopathologist (B.M.E.) and pigmented lesion specialist (S.M.S.) in a blinded manner. Additional Melan-A immunostaining was performed when appropriate to confirm pathologic diagnoses.
For statistical analysis, patients were divided into 2 groups according to treatment type (observation vs reexcision). Fisher exact test, the χ2 test, and 2-sample t test were used for comparisons between the 2 groups. Patients who incurred an event at the site of the original DN (defined as an additional biopsy or excisional procedure) were right-censored at the time of the additional procedure. Statistical analyses were performed in Stata MP, version 14.0. The study was approved by the institutional review boards at both VAPAHCS and SUMC. Informed consent was waived due to the retrospective nature of the study.
Baseline characteristics of the patients studied are presented in Table 1. As anticipated, more moderate to severe and severe DN were in the excised than clinically observed group (P < .001). There were no significant differences between the 2 groups in sex, age, personal history of melanoma, biopsy type, or location of positive margin involvement. In the clinically observed group, 105 of 399 initial biopsies (26%) were performed with partial intent (ie, partial sampling of the clinically apparent lesion).
Only cases with available follow-up data were included in the analysis (n = 474: 416 VAPAHCS, 58 SUMC). The mean (SD) follow-up time was 5.5 (4.6) years; 345 DN were followed for at least 2 years and 298 for least 3 years (range, 1 month to 20.7 years at VAPAHCS and 1 month to 13.8 years at SUMC).
Cases of local nevus recurrence are summarized in Table 2. Among 304 observed DN (276 VAPAHCS, 28 SUMC), 10 (3.3%) recurred, ranging histologically from common nevi to moderate DN; no recurrences were observed in the 170 reexcised DN (P = .02).
Among 159 clinically observed DN in which a documented excisional biopsy was performed, only 1 case of MIS, LM type, developed at the biopsy site 5 years later (Figure 1 and Figure 2). In contrast, 5 cases of MIS or invasive melanoma were diagnosed 1 to 10 years later in clinically observed cases in which an initial partial biopsy was performed. In addition, among the 170 reexcised DN (140 VAPAHCS, 30 SUMC), 1 case of MIS developed at the biopsy site.
Clinical data and pathologic analysis from these 7 melanoma cases are summarized in Table 3. In the observed group, 5 of the 6 melanomas were classified as MIS, LM type, or AIMP favoring early/evolving LM. One case of invasive melanoma (T1a, stage IA) was identified on biopsy 5 years later, with Breslow thickness of 0.8 mm. No patients in either group died of melanoma arising from DN biopsy sites. Of note, blinded retrospective review of the initial biopsies using Melan-A immunostains demonstrated focal confluence of junctional atypical melanocytes suggestive of early MIS, LM type, in 4 clinically observed cases.
New primary melanoma at an unrelated site was diagnosed during the follow-up period in 42 (9.7%) of cases overall, with similar incidence between the observed and reexcised groups (30 [9.9%] and 16 [9.4%], respectively).
Our long-term analysis of DN outcome augments the current evidence21,22 in support of clinical observation for biopsy-proven mild or moderate DN with positive histologic margins. Although we observed a higher incidence (2.0%) of melanoma arising from observed DN, 5 of 6 (83%) arose in patients in whom a partial biopsy was performed at the outset (typically comprising <50% of the clinical AN). This suggests that transformation of microscopic residual nevus was not a major factor in our population, but rather incomplete initial sampling of a de novo or nevus-associated early melanoma (most often MIS). Only 1 subsequent melanoma was invasive, and no metastases or deaths from melanoma were observed as a result of observation of the incompletely excised DN. Consequently, there is little evidence to support routine reexcision of mild to moderate DN with microscopically positive margins, and the increased morbidity and medical costs associated with additional surgery should be taken into account.
There are several potential causes of a melanoma diagnosis in a site of previously biopsied DN.25 These include initial underdiagnosis of the original biopsy, sampling error within an incomplete initial biopsy, later misdiagnosis of melanoma due to overlapping histologic features of recurrent nevi,27,28 and, finally, true malignant transformation of residual cells following biopsy.
Our blinded pathology rereview resulted in a change in diagnostic classification with underdiagnosis of the initial biopsy in 4 of 6 observed cases. With the benefit of Melan-A immunostains on blinded reexamination, we observed focal confluence of junctional atypical melanocytes consistent with AIMP or early MIS, LM type, and, in retrospect, would have recommended reexcision of these lesions following initial biopsy. We have previously reported dysplastic nevus–like features in LM, including bridging of melanocytic nests and underlying lamellar fibrosis, with 58% of 65 analyzed LM cases showing mixed or predominant dysplastic nevus–like features histologically, rather than characteristic epidermal atrophy and flattening of rete ridges.29 Thus, careful attention to confluent atypical junctional melanocytes or any focal AIMP within a dysplastic nevus is critical to avoid misdiagnosis of LM, particularly when located on sun-damaged skin in older patients.
Our findings support the practice of performing an excisional biopsy of clinically and dermoscopically concerning nevi to ensure that any adjacent areas of severe dysplasia or existing melanoma are not missed on partial biopsy.30 We advocate the saucerization/deep shave biopsy technique as the most practical to remove all of the clinically apparent nevus, providing adequate sampling of the deeper reticular dermis to remove any compound portion, with histologic transection most commonly noted of the peripheral junctional component.31 Our melanoma cases typically arose from an initial partial punch biopsy, which was a more common practice in the first decade of the study. With broader acceptance of the saucerization technique, fewer partial punch biopsies were performed in the second decade (48% prior to 2006 vs 24% thereafter).
Residual or recurrent melanocytic nevi may demonstrate postbiopsy changes that mimic severe dysplasia or melanoma, particularly in the setting of recurrent junctional hyperplasia.28 However, this finding can generally be histologically distinguished by experienced dermatopathologists based on cytologic and architectural atypia restricted to areas overlying the dermal scar. Our blinded pathology review of all subsequent melanoma specimens excluded confounding by recurrent junctional hyperplasia.
We observed only 1 case of early MIS (LM type) arising from a moderately dysplastic nevus that was excisionally biopsied but had positive histologic margins. Interestingly, we also observed 1 case of MIS arising from a moderately to severely dysplastic nevus that was initially reexcised, suggesting that residual atypical cells may sometimes extend beyond the borders of the excision margin. Finally, the rate of local nevus recurrence (common or dysplastic) at the same site was 3.3%, which is consistent with a study by Goodson et al,21 although this percentage only reflects clinical repigmentation that was rebiopsied and so may not represent all instances of local nevus recurrence.
Importantly, our data also revealed a relatively high rate (9.7%) of “other site” primary melanomas in patients with histologic DN, developing from 2 months to 14 years after DN biopsy. Our study cohort represented a high-risk group, composed largely of older men with substantial sun exposure or patients with an atypical mole phenotype, in whom 25% had a prior melanoma. This is consistent with studies collectively showing a 3- to 20-fold increased melanoma risk in patients with AN or DN,10 supporting the need for long-term dermatologic surveillance, especially for those with additional risk factors.
Our study is one of the largest to examine rates and patterns of recurrence of histologic DN with positive margins that were clinically observed. Strengths include a large sample size from a single integrated health care system (VAPAHCS) and academic PLMC that allowed for patient follow-up for up to 20 years, full access to all relevant pathology specimens for review, and low rates of interobserver histopathologic diagnoses due to the stability of practitioners and consistent diagnostic criteria for DN over the duration of the study period.
An important limitation is the lack of uniform pathologic criteria for reproducible assessment of the degree of atypia in DN across institutions and among pathologists.32,33 There remains considerable disagreement and a high level of interobserver variability regarding the histopathologic classification of DN, which impedes the development of consensus recommendations for clinical management. We recommend conservative reexcision of severe (or moderate to severe) DN, which are more reproducibly diagnosed among pathologists.32 Some histopathologic variability between different “low-grade” variants of DN (eg, mildly, mildly moderately, and moderately dysplastic) appears to be less important because our results and other studies have shown that these lesions may safely be followed without reexcision, particularly when the majority of the lesion is removed on initial biopsy with excisional intent.
Further limitations of our study are the notable high-risk features of our study population, as well as the retrospective design. If anything, we would have expected increased rates of melanoma development in histologically transected DN in both our predominantly older, male population and those with clinical AN who resided in a high–UV radiation region. The proportion of DN or other nevus subtypes that progress to melanoma remains unknown34 and presents an additional limitation in determining the statistical power needed to analyze malignant transformation rates, as well as melanoma mortality in this context. As is planned, a multicenter study with larger numbers of cases will further clarify the risks and most appropriate management options for incompletely excised DN in both higher risk groups and the general population.
Our study addresses a contemporary question frequently encountered in practice regarding whether clinical observation is a reasonable option for histologically transected mild-to-moderate DN. Our findings support this conclusion when excisional biopsy is intended at the outset, and no residual, clinically apparent pigmented lesion is evident. Among the 6 of 304 melanomas that arose from incompletely excised DN, 5 were melanomas in situ and only 1 arose from an intended excisional biopsy. Partial biopsies of concerning pigmented lesions may result in sampling error and should be avoided when possible.
Our study used both clinicians and dermatopathologists with expertise with pigmented lesion clinical-pathologic correlation, resulting in low clinical and histopathologic variability over the 20-year study period; management of AN and DN may differ in practice settings without similar levels of communication and experience. However, our findings add to recent publications suggesting that reexcision of mild to moderate DN is generally not warranted in the absence of severe dysplasia or a nonrepresentative biopsy. It is time to dispel long-held notions regarding the “premalignant” status of the majority of AN and DN, which more accurately serve as risk markers for melanoma development as demonstrated by the high rate of unrelated primary melanomas in our study cohort.
Accepted for Publication: June 21, 2016.
Corresponding Author: Susan M. Swetter, MD, Dermatology/Cutaneous Oncology, Stanford University Medical Center and Cancer Institute, 900 Blake Wilbur Dr, W3045, Stanford, CA 94305 (firstname.lastname@example.org).
Correction: This article was corrected on February 1, 2017, to fix an error in a percentage in the abstract Results.
Published Online: August 17, 2016. doi:10.1001/jamadermatol.2016.2869
Author Contributions: Mr Fleming and Dr Swetter 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.
Concept and design: Fleming, Swetter.
Acquisition, analysis, or interpretation of data: All Authors.
Drafting of the manuscript: Fleming, Swetter.
Critical revision of the manuscript for important intellectual content: All Authors.
Obtaining funding: Fleming, Swetter.
Administrative, technical, or material support: Egbert, Kim, Swetter.
Study supervision: Swetter.
Conflict of Interest Disclosures: None reported.
Funding/Support: Mr Fleming’s funding was provided by the Stanford Medical Scholars Fellowship Program.
Role of the Funder/Sponsor: The Stanford Medical Scholars Fellowship Program had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Robert V. Rouse, MD, Pathology Service, Veterans Affairs Palo Alto Health Care System, provided assistance with pathology case finding. Dr Rouse received no compensation.
Create a personal account or sign in to: