Melanoma Recurrence After Mohs Micrographic Surgery with MART-1: A Systematic Review and Meta-Analysis

Cutaneous melanoma (CM) affects about 287,000 people each year, worldwide. Worse prognoses are associated with CM located on the head, neck, and trunk compared to extremities with the exception of acral melanoma. Surgical management continues to be the primary therapeutic approach for CM. Current surgical recommendations for primary CM include surgical excision with pre-determined surgical margins based on tumor characteristics and subsequent microscopic margin assessment as needed. Wide local excision (WLE) continues to be the primary recommendation of the American Academy of Dermatology (AAD), with margins predicated by tumor depth. Extra consideration must be made for anatomically constrained sites and/or when subclinical spread is suspected. Mohs micrographic surgery (MMS) is recommended to treat primary melanoma in situ and lentigo maligna located on the head and neck, acral sites, genitalia, and pretibial leg, and for locally recurrent melanoma in situ lentigo maligna in any location. The 2019 AAD workgroup has encouraged further study of MMS on these anatomically constrained areas.

Cutaneous melanoma (CM) affects about 287,000 people each year, worldwide. 1 Worse prognoses are associated with CM located on the head, neck, and trunk compared to extremities -with the exception of acral melanoma. 2 Surgical management continues to be the primary therapeutic approach for CM. 3 Current surgical recommendations for primary CM include surgical excision with pre-determined surgical margins based on tumor characteristics and subsequent microscopic margin assessment -as needed. Wide local excision (WLE) continues to be the primary recommendation of the American Academy of Dermatology (AAD), with margins predicated by tumor depth. 3 Extra consideration must be made for anatomically constrained sites and/or when subclinical spread is suspected. Mohs micrographic surgery (MMS) is recommended to treat primary melanoma in situ and lentigo maligna located on the head and neck, acral sites, genitalia, and pretibial leg, and for locally recurrent melanoma in situ lentigo maligna in any location. 4 The 2019 AAD workgroup has encouraged further study of MMS on these anatomically constrained areas.
Mohs micrographic surgery offers particular advantages for tumors located in areas that require tissue conservation and when subclinical spread is suspected. MMS is ideal for the head and neck due to the high INTRODUCTION recurrence rates of CM due to particular risk features, such as, subclinical spread, larger preoperative size, more likely to have had previous treatment, and requiring tissuerearranging reconstruction when compared to keratinocytic tumors. 5 Immediate margin assessment provides the confidence of complete tumor removal. Unfortunately, hematoxylin and eosin frozen sections are notoriously difficult to interpret for CM, resulting in poor sensitivity and specificity. [6][7][8][9][10][11][12] Over the years, immediate margin assessment has evolved to a more specific analysis. The utilization of immunostains, such as melanoma-associated antigen recognized by T-cells (MART-1), has improved the confidence in and histologic accuracy of MMS in some subtypes of CM. [13][14][15] This is most evident in melanocytic lesions demonstrating extensive subclinical spread. 14-21 Furthermore, MART-1 staining of frozen sections has become more rapid and as reliable as permanent sections. 13,22 A recent survey of Mohs surgeons and database analysis suggests wide-spread use of MART-1 in MMS for CM. 23,24 With techniques in MMS for treating melanocytic lesions evolving, there is an impetus to continue to evaluate emerging research for further clinical guideline development. To date, no systematic reviews have specifically examined the utilization of MART-1 staining in frozen sections for CMparticularly its effect on local recurrence. Sharma and colleagues 25 performed a broad systematic review to compare the local recurrence of MMS with that of traditional surgical techniques for lentigo maligna/lentigo maligna melanoma (LM/LMM) but only included 2 studies and failed to compare with MART-1.
In this study, the authors aim to systematically review and analyze the rate of local recurrence after utilization of Mohs micrographic surgery with MART-1 immunohistochemistry in frozen sections in patients being treated for primary and recurrent cutaneous melanoma. The increasing utilization of MART-1 staining in MMS supports the further evaluation of the practice.
The authors followed the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions to execute the study. The authors followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) statement and checklist to publish and report this metaanalysis. The protocol was determined before the search and data extraction took place. [PROSPERO ID: CRD42020221826]

Search Strategy
The authors conducted a systematic search with Pubmed, Medline, and Cochrane databases for articles published from their inception to October 30, 2020, to identify relevant studies published in English. Electronic searches were performed with the following search terms: ["Mohs" AND "Melanoma"].
The ClinicalTrials.gov database was searched for current trials. The authors manually checked the bibliographies of the identified articles, including any relevant reviews or metaanalyses to further identify eligible studies.

Data Extraction and Outcomes
All relevant information regarding study characteristics -design, level of evidence, the methodological quality of evidence, population, outcome measures, and followup time points -were collected by two independent reviewers using a predetermined datasheet. Two independent reviewers oversaw all review process steps, and data were independently obtained and their inclusion criteria agreed upon. Any disagreement over data relevancy were settled by a third reviewer. The primary extracted outcome was local recurrence rates.

Quality Assessment
Patient characteristics and control protocols were reviewed to evaluate for possible confounding variables. The Risk-of-bias (RoB) plot was created using The Cochrane Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool. RoB judgments were made following the supplementary guidance of Sterne et. al. 26

Data Analysis
The number of local recurrences was homogenized by determining the percent affected. A χ² test of independence was performed to examine the relationship between recurrence location and rate of recurrence. A comparison of proportions was utilized to analyze demographic data. A P-value <0.05 was considered statistically significant, and 95% confidence intervals (CIs) were reported.

Literature Search
The literature search identified 617 total manuscripts. Once articles were screened for inclusion and exclusion criteria, 17 fulltext articles were assessed for eligibility (Figure 1).
Four manuscripts with a total of 2,896 lesions were included in the review. 17,18,27,28 Nine manuscripts were excluded due to the lack of recurrence rates. 4,[13][14][15][16]20,[29][30][31] Single manuscripts were excluded due to the lesion type 32 , modified procedures 19 , digital melanoma 33 , or inclusion of the same data set. 21 Clinicaltrials.gov did not provide any additional information concerning on-going research on this subject.

Study Characteristics
All four included studies, summarized in Table 1, were retrospective. There were 492 invasive melanomas (IM) and 2,404 melanoma in-situ (MIS), of which 2,556 were primary lesions. Of the included lesions, 1,724 were located on the head/neck and 1,172 were located on the trunk/extremities. The outcomes utilized in the included studies were local recurrence, rate of metastasis, overall survival, disease-free survival rates, upstaging of the lesion, and/or surgical margins required for complete removal. Mean follow-up periods ranged from 1,026 to 1,361 days, with a reported minimum and maximum of 3 and 3167 days, respectively.

Demographics
Age was reported as a mean with a range in all of the included studies. There was no report of standard deviation. The ages of the RESULTS patients in this review ranged from 12 to 99. The average age of participants was 66 in three of the four included studies, and not reported in the last.
Sex demographics were reported in all of the included studies. A (chi-squared test) significant majority of the participants were male, with a total of 61.5% [P < 0.0001].
The location of the lesions was reported in all of the included studies. While some studies were more specific with particular locations, all of them included head, neck, trunk, and limbs. In total, 1,724 melanomas were on the head and neck; 1,172 were elsewhere ( Table 1).
Tumor specifics included whether or not the lesion was primary or recurrent and invasive or in-situ. The average Breslow depth of invasive melanomas was also included. There were 2,556 (88%) primary melanomas. Of these melanomas, 2,404 (83%) were MIS. The average Breslow depth (of IM) was estimated to be 0.768mm with a range of 0.1-5.8mm. Two studies did not report the range of depth, and one did not provide an average depth.

Methodology
All of the included studies were retrospective cohort studies that gathered data from their respective institutions only. The MMS technique described varied slightly. Two of the studies report utilizing a central debulking excision for bread-loafing in permanent paraffin-embedded H&E for the staging of the melanoma. 17,28 The remaining studies either did not report such a technique or were vague about their technique, merely referring to previous reports. Two studies described an initial debulking margin of 2-3mm with an additional 2-3mm Mohs margin that was subsequently utilized for frozen sections. 17,28 Degesys et al. reported an initial 5mm margin with 3mm subsequent Mohs margins. 27 Valentin et al. did not report the specific margins utilized. 18 Positive margin determination was relatively homogenous across the studies. They all utilized, at a minimum, the following: pagetoid spread, nests of at least 3 atypical melanocytes, and confluent melanocyte hyperplasia along the basement membrane. Follow-up data was collected similarly across the included studies. With the exception of one, all of the included studies utilized over-the-phone screening of the participants.

Local Recurrence
Among the included studies, there were 15 (0.52%) local recurrences of melanoma in situ and invasive melanoma after MMS with MART-1. The descriptions of each recurrence can be found in Table 2. MIS did not convert to IM in any of the reported recurrences. In primary vs recurrent melanomas (IM and MIS), X 2 (1, N = 2,894) = 3.26, p = .071, the chance of recurrence was not significantly different. By contrast, the chance of recurrence was significantly higher in invasive melanoma when compared to melanoma in situ: X 2 (1, N = 2,896) = 19.76, p < .001. There was no significant difference in CM recurrence between the two general locations of "head/neck" and "other location": X 2 (1, N = 2,896) = 2.65, p = 0.10. Notably, all "other location" recurrences were on the feet; no recurrences were observed on the proximal extremities. Of the head and neck melanomas, most local recurrence (1.99%) occurred on the scalp. All were initially invasive melanomas.

Risk of Bias
The review authors judged that three studies had a moderate RoB and the final study had a low RoB (Figure 2). All of the included studies were able to measure important known confounding domains, such as, patient age, patient sex, tumor location, Breslow depth, primary and recurrent lesions, and lesion type. The cohort research design without the use of control groups suggests that the effect of confounding variables within the treatment groups is unlikely. Bias in the selection of participants into the study was moderate for Degesys et. al. 27   MMS with MART-1 demonstrates lower recurrence rates than conventional excision or MMS without MART-1. 34 Of the included studies, the rate of recurrence of MIS after MMS with MART-1 was just 0.25% (6/2404). According to the work of Nosrati et. al., if the MART-1 stain is not used with MMS, the recurrence rate increases to 1.8% (5/277).
The recurrence rate of MIS after WLE has been reported to be as high as 5.7% (22/395). 35 The utilization of MART-1 in MMS for CM demonstrates advantages over classical MMS in recurrence 27,36 and in surgeon preference.
The transition to MART-1 from H&E alone and/or HMB-45 took place in many important retrospective studies. 14,20,21,33,36,37 This transition appeared to occur around 2002, when further investigation suggested MART-1 was the superior immunohistochemistry stain in frozen sections due to reliable epidermal staining, ease of interpretation, and highest histopathologic concordance. 14,15,38-41 Subsequently, studies began to demonstrate improved recurrence rates after transitioning to exclusively using MART-1. 21,33,36 In a recent study by Burnett et al., 37 a retrospective analysis of IM on the trunk and proximal extremities treated by MMS observed only 2 (0.14%) recurrences. The authors established that the sample included procedures that utilized H&E alone, then DISCUSSION HMB, then MART-1 exclusively. Recurrences occurred only prior to the incorporation of MART-1 immunostaining in MMS. While the sample included in this study was much larger than the invasive melanoma collective sample in the present study, it corroborates an extremely low recurrence rate when MART-1 is utilized for IM for the trunk and proximal extremities.
From 2001 to 2016, the chances of receiving MMS for melanoma increased 304%. Twenty-seven percent of the MMS cases reported use of Immunohistochemistry stains (IHC), with an increase in utilization over recent years with further acceptance of its benefit. 24 In addition to ease of use/interpretability and outcome improvement, it is important to investigate financial feasibility. Wilson et al. demonstrated that among comprehensive margin assessment techniques for MIS/IM, staged excision using en-face margin assessment (SEEM) and MMS with MART-1 were similar in cost for the healthcare system. The authors noted that SEEM incurred lower costs due to decreased excision code reimbursement and lower number of stages, while MMS with MART-1 required less complex reconstruction and the multiple procedure reduction rule. MMS "classic" with H&E generally required more stages, larger margins, and a flap or graft. 31 Recent epidemiological research suggests that there are no significant differences in local recurrence or overall survival when comparing MMS and WLE for head, neck, trunk, and extremities. [42][43][44] Conversely, Valentin et al. provides 5-year Kaplan-Meier local recurrence rates for primary CM treated with MMS with MART-1 that are significantly lower than historical controls of conventional wide-local excision. 18 We feel that these discordances may have occurred due to lack of considering specific MM  demonstrated strong evidence to support its routine use for invasive melanoma at "special sites" (head, neck, acral sites, genitalia, and pretibial leg) and on the trunk/proximal extremities.
The limitations of this review and the included studies are inherent and technical. The primary limitation of the included studies is that they were all observational studies that were retrospective in nature. A limitation of the review was the exclusion of alternative surgical techniques utilizing MART-1, such as modified staged-excision. Furthermore, there may be a serious confounding variability in surgical technique, frozen section technique, and MART-1 interpretation skill. Lastly, this review did not differentiate the lentigo maligna and melanoma in situ as separate entities and were all treated as MIS.
This review demonstrates that the technique of MMS with MART-1 IHC in frozen sections produces satisfactory recurrence rates for melanoma in situ and invasive melanoma. There appears to be discordance between current guidelines and research supporting broader use of MMS with MART-1. Further prospective research should be done to evaluate clinical outcomes with the use of MMS for invasive and noninvasive CM in various body locations.