1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References

BACKGROUND Staged excision with rush-processed paraffin-embedded tissue sections (Slow-Mohs) is an effective treatment for periocular melanoma. Although there is no consensus on initial margins of excision, narrower margins in the eyelids have the functionally and cosmetically important consequence of smaller postoperative wounds.

OBJECTIVES To report early cure rates for periocular melanoma using Slow-Mohs surgery with en-face margin sectioning.

METHODS Retrospective, multicenter, noncomparative case series. Slow-Mohs surgery in 14 patients with periocular melanoma from 2000 to 2006.

RESULTS Fourteen patients underwent 14 Slow-Mohs procedures for eight lentigo maligna, one nodular, and one superficial spreading melanoma, and four lentigo maligna, 12 primary, and two recurrent tumors. The most common site was the lower eyelid (8/14, 57.1%). Breslow thickness ranged from 0.27 to 1.70 mm, with four cases less than 0.76 mm and one case greater than 1.5 mm. Five cases were a Clark level II or greater. Complete excision was achieved with one level (6 cases) or two or three levels (8 cases), with 2- to 3-mm margins at each level in all but one case. With median follow-up of 36 months, there were two local recurrences (2/14, 14.3%).

CONCLUSION Slow-Mohs with en-face sections achieves similar early cure rates to previously published margin-controlled excision techniques. Narrow margins of excision can optimize tissue preservation without compromising outcome.

Malignant melanoma (MM) of the eyelid is rare but has the highest mortality of all primary skin tumors.1 The commonest subtype of periocular melanoma, lentigo maligna melanoma (LMM), and its precursor lentigo maligna (LM) are often large at presentation and have a propensity for subclinical extension beyond the clinically apparent margin.2,3

The difficulties in accurately interpreting the morphology of melanoma in frozen tissue section in standard Mohs micrographic surgery (MMS) are well documented,4–6 and have been overcome with a variety of modified Mohs surgery techniques using rush-processed paraffin-embedded tissue sections (Slow-Mohs).4,7–10 These techniques differ primarily in the way in which the margins of permanent sections are cut and examined—as en-face horizontal sections 11–14 or radial vertical sections (mapped serial excision (MSE)).4,7

Although there is no consensus on the initial margins of excision that should be used during staged Slow-Mohs excisions for periocular MM, several reports suggest that the majority of periocular MM are thin (≤1 mm thick) at presentation and that initial margins of excision of 5 mm or greater result in acceptably low recurrence and metastatic rates.4,7,9,11 More recently, the use of narrower margins of 2 to 3 mm at each level of excision has been reported, resulting in smaller mean postoperative defects and comparable recurrence rates with series using wider margins.15

We report early cure rates in a series of 14 patients with in situ or invasive periocular MM treated with Slow-Mohs with en-face sections in which complete excision with low recurrence rates was achieved in the majority of patients with initial margins of 2 to 3 mm and a comparable number of levels of excision to previous series.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References

This was a multicenter, retrospective, interventional case series of 14 patients with periocular MM undergoing MMS rush processing of paraffin-embedded en-face sections (Slow-Mohs) between January 2000 and March 2007. The authors invited all oculoplastic surgeons affiliated with a Mohs surgical unit in the United Kingdom to participate and, in addition, collaborated with a Mohs unit in Adelaide, Australia. Patients were subsequently identified from the database of oculoplastic and dermatology consultants at five oculoplastic surgery referral units in and around London and in Adelaide, Australia.

Patients with in situ melanoma (LM), invasive lentigo melanoma (LMM), superficial spreading melanoma (SSMM) or nodular melanoma (NM) were included in the study. Periocular was defined as predominantly involving the upper or lower eyelid or the medial or lateral canthus.

Surgical Technique: Slow-Mohs Staged Excision with Rush En-Face Permanent Sections

The histologic diagnosis was made before treatment in all patients using a punch or incisional biopsy of the most heavily pigmented or thickest areas or both.

The patients were examined preoperatively to demarcate the clinically apparent borders of the lesion; a Wood's lamp was additionally used to demarcate the borders if available. After infiltration with 1% lidocaine with adrenaline, the demarcated area was excised and sent in formalin for routine vertical section processing. A 2-mm margin of clinically normal surrounding skin was then cut, notched, and transferred so that its anatomical orientation was maintained. The tissue layer was divided, color coded, and then mapped. Paraffin sections were prepared on all specimens, stained with hematoxylin and eosin, and examined by the Mohs surgeon and a pathologist. In relation to LM, nests and contiguous melanocytes and large numbers of single melanocytes with markedly abnormal features (including nuclear pleomorphism or hyperchromatism, nucleolar prominence, and abundant granular cytoplasm) were interpreted as evidence of disease. An attempt was thus made to distinguish these from the mildly atypical melanocytes seen in high numbers in sun-damaged skin. Further layers were removed as indicated until an apparently tumor-free margin was obtained.

The minimum and maximum diameters of the clinically apparent lesion and the final surgical defect were measured, and an oculoplastic surgeon repaired the defect after histologic confirmation of clear margins.

Data Collection

Data collected included patient details, site and histologic features of the lesion (type of melanoma, Clark level of invasion, and Breslow thickness), previous tumor occurrence and treatment, minimum and maximum size of the clinically apparent lesion and final defect, margins of excision at each level and the number of levels required for complete excision, and method of surgical reconstruction.

The involved surgeon or dermatologist measured outcome at the last clinical follow-up visit.

Outcome Measures

Outcome measures were local recurrence, regional or distant metastases, and tumor-related mortality. Local recurrence was defined as biopsy-confirmed tumor within 2 cm of the surgical scar.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References


Fourteen patients, nine female (64.3%) and five male (36.7%), underwent 14 Slow-Mohs procedures. The mean age was 69.3 (range 49–90).

The majority of lesions (85.7%, 12/14) were LM or LMM (4 LM, 8 LMM). There was one case of NM involving the lower eyelid and one case of amelanotic SSMM involving the lower eyelid that had been previously misdiagnosed as sebaceous carcinoma on biopsy. No data were available as to whether invasive lesions arose de novo or from preexisting pigmented lesions.

There were 12 primary lesions and two recurrent lesions that had been previously excised. The most common site of involvement by primary MM was the lower eyelid (8 patients) followed by the upper eyelid (2 patients) and lateral canthus (2 patients). The two recurrent MM had more extensive involvement. One case of recurrent LMM involving the lower and upper eyelids had undergone previous complete excision, but no data on type of excision or time to recurrence were available. One case of recurrent LM involving the upper eyelid/lateral canthus had undergone six previous excisions of LM in the upper eyelid and temporal region, with 3 years between the last excision and most recent treatment.

No patients had loco-regional or distant metastases at the time of diagnosis, and no patients underwent sentinel lymph node biopsy or elective lymph node dissection.

Tumor Characteristics

In the 10 cases of invasive MM (8 LMM, 1 NM, 1 SSMM), Breslow thickness and Clark level of invasion had been recorded in eight and seven cases, respectively. The majority of invasive tumors were 1 mm or less in thickness (87.5%, 7/8) and Clark's level II (57.1%, 4/7) (Table 1).

Table 1. Type, Histopathologic Details, and Surgical Parameters of 14 Cases of Periocular Melanoma
CaseTypeBreslow Thickness (mm)Clark Level (1–5)Size of Lesion (mm)Margin of Excision at Each Level (mm)Levels Required (n)Size of Final Defect (mm)
  1. * Recurrent lesions that had been previously excised. Margins of lesion in Case 6 were outlined using a Wood's light.Unknown results recorded with dash (—). NA, not applicable; LM, lentigo maligna; LMM, lentigo maligna melanoma; SSMM, superficial spreading melanoma; NM, nodular melanoma.


Number of Levels Required for Complete Excision

For all lesions, the average number of levels required for complete excision was 1.64 (range 1–3 levels). Six (6/14, 42.9%) lesions required only one level of excision, with the margins at each level ranging from 2 to 5 mm. The majority of lesions (8/14, 57.1%) were completely excised with two or three levels, with 2-mm margins at each level in all but one case.

For all lesions, the total margin of excision ranged from 2 to 10 mm from the clinically apparent tumor. All cases of LM (4/4, 100%) were completely excised with total margins of 4 mm or less, whereas the majority of invasive lesions (8/10, 80%) were completely excised with margins of 5 mm or less. The majority of tumors with a recorded Breslow thickness of 1 mm or less (5/8, 62.5%) were excised with one level, as was the single tumor with a thickness greater than 1.5 mm.

Both cases of recurrent MM were completely excised with one level, with the LM requiring a 3-mm total margin and the LMM requiring a 7-mm total margin.

One of the cases of LMM (Case 11), in which a local recurrence later occurred, also required the largest total margin of excision (10 mm).

Size of Lesions and Defects

Lesions ranged in size from 4 × 10 mm to 30 × 42 mm before treatment (recorded as minimum × maximum diameter). The final postsurgical defect after complete histologic clearance ranged from 6 × 16 mm to 50 × 60 mm.

The size and location of the final defect determined method of surgical reconstruction, which included direct closure, full-thickness skin graft, and flap repair. No patients required exenteration. No patients received any adjuvant therapies.


After a mean follow-up of 35.8 months (median 36 months, range 2–72 months), there were two cases of local recurrence in patients with primary LMM (2/14, 14.3%). In the first case (Case 6), recurrent LM presented at the margin of excision at 8 months. This was completely excised with a 3-mm margin. In the second case (Case 11), time to recurrence was 18 months and presented as LM associated with primary acquired melanosis. The patient declined further surgical treatment and is being managed conservatively.

All patients (14/14, 100%) were free of regional or distant metastases at the time of last clinical review. Two patients (Cases 3 and 8) subsequently died of causes unrelated to tumor.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References

Reports on outcomes after margin-controlled excision of periocular MM are few because of the rarity of this tumor. This large series adds to the body of evidence that staged excision with rush-processed paraffin-embedded permanent tissue sections (Slow-Mohs) and conservative margins is an effective treatment for this aggressive tumor, ensuring complete excision with low recurrence rates. Our results also suggest that narrower margins of excision can be safely employed with comparable number of levels of excision with other series, achieving smaller postoperative defects without compromising early cure rates or cost or convenience to the patient.

Standard excision margins for melanoma may not be necessary in the context of periocular disease. The potential disadvantages of wide excision are unnecessary loss of periocular tissue and incomplete excision. The majority of periocular melanomas are LM or LMM.2,3 They are often thin (<1 mm) on presentation and have a tendency to unpredictable subclinical extension beyond the visibly pigmented and Wood's light margins.16 Furthermore, a number of studies have demonstrated that standard margins may be inadequate for the complete excision of more than 30% of LM and LMM7,17,18 and are associated with recurrence rates of 9% to 20%.2,19 Zitelli reported that an average of 1.8 cm more of normal skin is preserved with MMS than with standard excisional surgery. This is critical in the eyelid, whose total length is approximately 3 cm.20

Standard Mohs micrographic surgery uses en-face horizontally cut frozen tissue sections that allow examination of 100% of the lateral and deep margins, but the morphology of melanoma at these margins is difficult to interpret on frozen tissue sections. Problems with frozen sections include freeze artefact, fat tears, and relatively thick sections. These features can make it difficult to determine the peripheral margins of melanomas, particularly in the presence of scar tissue, a dense inflammatory infiltrate, or on chronically sun-damaged skin (which contains prominent single melanocytes with large, hyperchromatic nuclei).5

It may be particularly challenging to identify the periphery of LM, which may be represented only by single melanocytes, albeit with marked atypia.5,19

The use of melanocyte-specific immunohistochemical stains, with monoclonal antibody to HMB-45, S-100, Melan-A, Mel-5, and MART-1 proteins, may improve the diagnostic accuracy and facilitate margin evaluation in MMS frozen sections.12–14 HMB-45 does not distinguish benign from malignant melanocytes but has been reported as useful in demonstrating single atypical melanocyte proliferation in MMS for melanoma in situ.9 MART-1 has been reported to be more sensitive than HMB-45 in MMS frozen sections of invasive and in situ melanoma,14 although as has been previously pointed out,5,19 immunohistochemical methods can yield false negatives due to alteration or loss of the marker antigen or technical error, which means that lack of staining does not always exclude atypical melanocytic proliferations.

In common with the frozen section procedure, Slow-Mohs surgery involves the excision and orientation of the tumor with marker dyes and color-coded diagrams. Techniques differ in the manner in which the margins of specimens are sectioned for histologic examination. En-face sectioning, as used in our series, is the most commonly described and used.11,8–10 More recently, a technique known as MSE has been described for LM and LMM.4,7,15 This involves radial and vertical sectioning of the entire specimen, usually in 1-mm slices, before paraffin embedding. The advantages of radial and vertical sections over en-face sections include the ability to follow the evolution of tumor architecture from the center to the periphery. This in turn can increase the accuracy of identification of single or small nests of melanoma cells at the true periphery of LM. It may also assist in detecting invasive disease and allow measurement of tumor thickness throughout the specimen. A disadvantage of MSE is that is much more labor intensive, requiring experienced staff to process and interpret multiple sections.

Local recurrence of MM is strongly correlated with tumor thickness, widely regarded as the most important single predictor of survival in early (Stage 1) tumors.2,3 Local recurrence is uncommon in tumors of 1 to 2 mm thick or less, which account for the majority of periocular MM in most series.2,7,12 In a retrospective review of 44 cases of periocular MM (mean follow-up of 34 months), Esmaeli and colleagues reported a 50% recurrence rate in tumors 2 mm thick or thicker, compared with 6.3% recurrence in tumors 1 mm thick or less.21 The majority (7/9, 77.8%) of invasive lesions with a recorded Breslow thickness in our series were also thin (≤1 mm). The two cases of recurrence occurred in primary LMM with Breslow thicknesses of 1.70 and 0.30 mm, respectively.

Local recurrence is an important indicator of adequacy of surgical margins. Although there is no consensus on the margins that should be used for periocular MM, many margin-controlled techniques use initial excision margins of 5 mm or more and report acceptably low rates of local recurrence and regional or distant metastases.2,4,6,7,9–11

The study described above by Zalla and colleagues12 retrospectively reported the practice of surgeons managing 44 cases of eyelid MM throughout Europe, the United Kingdom, and the United States. Although no mention was made of the methods of excision used, the majority of lesions were 1 mm thick or less (43.1%) and had been excised with 5-mm margins or less (67%). As a consequence of their higher local recurrence rates, they recommended a margin of excision of at least 5 mm for tumors 1 mm or less and wider margins for thicker tumors.

Chan and colleagues2 reviewed 29 cases of periocular MM (the majority of invasive MM being less than 1.5 mm thick) managed with standard excision or margin-controlled excision with standard or modified MMS. The majority (79.3%) were excised with total margins of 5 to 10 mm. Subsequent median follow-up was 3 years, with local recurrence rates reported in 17.2% and distant metastases in 13.8%.

Malhotra and colleagues7 examined 27 cases of periocular LM and LMM (the majority of LMM being <1 mm thick) treated with MSE using 5-mm excision margins at each level. Most (55.6%) were completely excised with one level (5-mm margin), whereas the remainder required 2 to 3 levels (10- to 15-mm margins). Local recurrence rates were 7.4%, with a mean follow-up of 32 months.

Although there is clearly evidence to support the safety of 5-mm margins during staged excision of periocular MM, there is no evidence that refutes the use of narrower margins. In the eyelids, every millimeter of tissue loss is significant, and narrower margins have the functionally and cosmetically important consequence of smaller postoperative wounds.

In a retrospective review of 59 patients with cutaneous LM and LMM, the majority of which were in the head and neck (including 3 periocular cases), Bub and colleagues15 reported the long-term outcomes after treatment with MSE using 2- to 3-mm margins of initial excision. Despite narrower margins, complete excision was achieved with an average of 1.7 levels of excision. This is comparable with previous series using 5-mm margins or greater, where number of levels of excision ranged from 1.9 to 2.3,10,11,18 for lesions of similar size. In addition, the overall postoperative defect sizes were smaller than those previously reported. This was attributed to the narrower margins excised at each level, and they maintained low recurrence rates of 5% with a mean follow-up of 57 months.

Our data support the use of narrow margins of excision (<5 mm) through the use of Slow-Mohs surgery for periocular MM. The majority of lesions (10/14, 71.4%) were completely excised using 2- to 4-mm margins at each level, with an average of 1.64 levels for histologic clearance. A significant proportion of lesions (6/14, 42.9%) required only one level of excision. The majority of lesions (8/14, 57.1%) were cleared with two or three levels, using 2-mm margins at each level in all but one case. The significance of this was that the total margin of excision was 4 to 6 mm, rather than the 10 to 15 mm that would have resulted if 5-mm margins had been used. The only thick tumor (>1 mm) in this series was excised in one stage with a 3-mm margin. Our mean postoperative defect size (2.8 × 3.8 cm) is comparable with that of Bub and colleagues15 (2.5 × 3.3 cm) and therefore also smaller than those previously reported of 4.2 × 4.5 cm.18 Our recurrence rate of 14.3% was comparable with those of other series.

In conclusion, we report a large series of periocular MM treated with Mohs surgery using rush-processing of paraffin-embedded en-face sections (Slow-Mohs). Although the retrospective nature and relatively short follow-up of this series are significant limitations, our results emphasize that staged margin-controlled surgery should be the criterion standard of treatment for periocular MM. They also suggest that, where accurate histologic control is available, narrow margins of excision can be employed at each level of excision to optimize preservation of tissue without compromising outcome.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  • 1
    Cook BE Jr., Bartley GB. Treatment options and future prospects for the management of eyelid malignancies; an evidence-based update. Ophthalmology 2001;108:208898.
  • 2
    Chan F, O'Donnell B, Whitehead K, et al. Treatment and outcomes of malignant melanoma of the eyelid: a review of 29 cases in Australia. Ophthalmology 2007;114:18792.
  • 3
    Boulos P, Rubin P. Cutaneous melanoma of the eyelid. Semin Ophthalmol 2006;21:195206.
  • 4
    Hill D, Gramp A. Surgical treatment of lentigo maligna and lentigo maligna melanoma. Aust J Derm 1999;40:2530.
  • 5
    Barlow R, White C, Swanson N. Mohs micrographic surgery using frozen sections alone may be unsuitable for detecting single atypical melanocytes at the margins of melanoma in situ. Br J Dermatol 2002;146:2904.
  • 6
    Cohen LM. Lentigo maligna and lentigo maligna melanoma. J Am Acad Dermatol 1995;33:92336.
  • 7
    Malhotra R, Chen C, Huilgol S, et al. Mapped serial excision for periocular lentigo maligna and lentigo maligna melanoma. Ophthalmology 2003;110:20118.
  • 8
    Dhawan S, Wolf D, Rabinovitz H, Poulos E. Lentigo maligna: the use of rush permanent sections in therapy. Arch Dermatol 1990;126:92830.
  • 9
    Johnson T, Headington J, Baker S, Lowe L. Usefulness of the staged excision for lentigo maligna and lentigo maligna melanoma: the “square” procedure. J Am Acad Dermatol 1997;37:75864.
  • 10
    Zitelli J, Moy R, Abell E. The reliability of frozen sections in the evaluation of surgical margins for melanoma. J Am Acad Dermatol 1991;24:1026.
  • 11
    Agarwal-Antal N, Bowen GM, Gerwels JW. Histologic evaluation of lentigo maligna with permanent sections: implications regarding current guidelines. J Am Acad Dermatol 2002;47:7438.
  • 12
    Zalla MJ, Lim KK, Dicuado DJ, Gagnot MM. Mohs micrographic excision of melanoma using immunostains. Dermatol Surg 2000;26:77184.
  • 13
    Menaker GM, Chiang JK, Tabila M, Moy RL. Rapid HMB-45 staining in Mohs micrographic surgery for melanoma in-situ and invasive melanoma. J Am Acad Dermatol 2001;44:8336.
  • 14
    Albertini JG, Elston DM, Libow LF, et al. Mohs micrographic surgery for melanoma: a case series, a comparative study of immunostains, an informative case report, and a unique mapping technique. Dermatol Surg 2002;28:65665.
  • 15
    Bub J, Berg D, Slee A, Odland P. Management of lentigo maligna and lentigo maligna melanoma with stages excision: a 5 year follow-up. Arch Dermatol 2004;140:5528.
  • 16
    Robinson J. Margin control for lentigo maligna. J Am Acad Dermatol 1994;31:7985.
  • 17
    Zitelli J, Brown C, Hanusa B. Surgical margins for excision of primary cutaneous melanoma. J Am Acad Dermatol 1997;37:23645.
  • 18
    Cohen L, McCall M, Zax R. Mohs micrographic surgery for lentigo maligna and lentigo maligna melanoma: a follow-up study. Dermatol Surg 1998;24:6737.
  • 19
    Stonecipher M, Leshin B, Patrick J, White W. Management of lentigo maligna and lentigo maligna melanoma with paraffin-embedded tangential sections: utility of immunoperoxidase staining and supplemental vertical sections. J Am Acad Dermatol 1993;29:58994.
  • 20
    Zitelli J, Mohs F, Larson P, Snow S. Mohs micrographic surgery for melanoma. Dermatol Clin 1989;7:83343.
  • 21
    Esmaeli B, Youssef A, Naderi A, et al. Margins of excision for cutaneous melanoma of the eyelid skin: The Collaborative Eyelid Skin Melanoma Group Report. Ophth Plast Reconst Surg 2003;19:96101.