Article Text
PDF
Abstract
Aims To determine the number of excisions needed to achieve clear margins and the prognostic value of the 7th edition of American Joint Committee on Cancer (AJCC) classification for eyelid melanoma.
Methods Retrospective chart review of consecutive patients treated for eyelid melanoma from January 2006 through May 2013 by the senior author at a tertiary care cancer centre.
Results Of the 64 patients (25 men and 39 women), clear surgical margins were achieved with a single excision in 38 patients (62%), 2 excisions in 21 patients (34%), and 3 excisions in 2 patients (3%). Need for repeat excision was not correlated with the size of the surgical margin (p=0.14) or AJCC TNM classification (p=0.15). Nodal disease at presentation was significantly associated with T category greater than T2b (p=0.0026) and shorter time to disease progression (p=0.007). Patients followed for a minimum of 1 year with T category greater than T2b had a significantly higher risk of nodal or distant metastasis (p=0.0061).
Conclusions More than a third of patients with eyelid melanoma required more than 1 excision to achieve clear margins, supporting delayed reconstruction for eyelid melanoma. Nodal metastasis at presentation was significantly correlated with AJCC T category and time to progression.
- melanoma
- AJCC
Statistics from Altmetric.com
Introduction
Melanoma accounts for less than 1% of eyelid malignancies, and the age-adjusted incidence of eyelid melanoma is 0.08 cases per 100 000 person-years.1 Histopathological features such as Breslow thickness of 1.5 mm or greater and Clark level of IV or greater have been correlated with increased mortality in eyelid melanoma.2
The standard of care in the management of cutaneous melanoma is wide local excision. In the periocular region, a 5-mm margin of excision for thin (≤1 mm) eyelid melanomas has been recommended, given the desire to preserve vital tissues and ocular function.3 Several studies have shown that frozen section technique is unreliable for evaluating surgical margins for melanoma.4 ,5 In up to 33% of cases, margin status has been falsely read as negative on frozen sections when it is positive on permanent sections.4 Some have, therefore, advocated for delayed reconstruction (as opposed to immediate) after examination of paraffin-embedded sections.
To provide additional evidence about whether reconstruction should be delayed until after examination of permanent sections in patients with eyelid melanoma, we aimed to determine the frequency of a need for repeat excision to achieve clear margins in such patients. In addition, we aimed to validate the seventh-edition American Joint Committee on Cancer (AJCC) classification for eyelid melanomas.
Materials and methods
After obtaining institutional review board approval, a retrospective review of records of consecutive patients treated for eyelid or periocular melanoma during the period from January 2006 to May 2013 by the senior author (BE) at a single centre were reviewed. A total of 74 patients were identified with a diagnosis of eyelid or periocular melanoma. Patients were excluded if they had predominantly conjunctival melanoma, declined treatment, or presented for follow-up after excision elsewhere without need for further surgery.
In addition to patient demographics, the following data were collected: surgical details (number of excisions required to achieve clear margins and surgical margin width), baseline tumour characteristics and staging, local recurrence, regional or distant metastasis, and adjuvant therapy received. Tumour histological characteristics recorded were in situ versus invasive disease; Breslow thickness; Clark level; number of mitoses; and presence of ulceration, regression, lymphovascular invasion, perineural invasion, satellite lesions, associated naevus, and radial and vertical growth patterns. Disease was staged according to the seventh edition of the AJCC staging system (box 1). Pathological stage was available for 63 patients; 1 patient had only clinical stage available.
American Joint Committee on Cancer (AJCC) staging system for melanoma of the skin, seventh edition, based on histopathological criteria15
T Category
TX—Primary tumour cannot be assessed
T0—No evidence of primary tumour
Tis—Melanoma in situ
T1—Melanoma ≤ 1.0 mm thick
T1a—without ulceration and mitosis <1/mm2
T1b—with ulceration or mitoses ≥1/mm2
T2—Melanoma 1.01–2.0 mm thick
T2a—without ulceration
T2b—with ulceration
T3—Melanoma 2.01–4.0 mm thick
T3a—without ulceration
T3b—with ulceration
T4—Melanoma >4.00 mm thick
T4a—without ulceration
T4b—with ulceration
N Category
NX—regional lymph nodes cannot be assessed
N0—no regional lymph node metastasis
N1—1 metastatic node
N1a—micrometastasis*
N1 b—macrometastasis**
N2—2–3 metastatic nodes
N2a—micrometastasis*
N2b—macrometastasis**
N2c—in-transit metastasis or metastases/satellite(s) without metastatic nodes
N3—4 or more metastatic nodes or matted nodes, or in-transit metastasis or metastases/satellite(s) with metastatic node(s)
M Category
M0—no metastasis
M1a—metastasis to skin, subcutaneous tissues, or distant lymph nodes
M1b—metastasis to lung
M1c—metastasis to all other visceral sites or distant metastasis to any site combined with an elevated serum lactate dehydrogenase level
*Micrometastasis refers to microscopic metastasis found on sentinel lymph node biopsy. **Macrometastasis refers to clinically palpable or radiographically identifiable nodal metastasis.
At baseline, all patients had a complete history and physical examination and staging with ultrasonography of the parotid and neck and radiography of the chest. The need for further imaging was determined on the basis of the clinical and histological characteristics of each melanoma. Surgical resection margins were planned by the surgeon (BE) on the basis of tumour thickness and anatomical location. All surgical margins were evaluated by review of permanent sections with confirmatory immunochemical staining when needed.
Fisher's exact test and the Wilcoxon rank-sum test were used to examine associations with the need for re-excision. Progression-free survival and metastasis-free survival were analysed using the Kaplan-Meier method and Cox proportional hazards regression technique. A progression event was defined as a recurrence or metastasis after eyelid surgery. All p values were two-tailed and considered statistically significant at α <0.05. Statistical analysis was conducted using SAS (release 9.2, SAS Institute, Cary, North Carolina, USA).
Results
Sixty-four patients with cutaneous eyelid melanoma were included in the study, 39 women (61%) and 25 men (39%). The median age at diagnosis was 65.9 years (range, 18–96 years). Sixty-one patients (95%) were non-Hispanic white and three (5%) were Hispanic. Patients were followed from day of eyelid surgery for a median of 1.8 years (range, 63 days to 6.9 years).
Baseline tumour characteristics
Forty-three patients (67%) presented with primary melanoma, while 21 (33%) presented with recurrent disease.
The median largest tumour diameter was 15 mm (range, 1–79 mm). Thirty-one patients (49%) presented with melanoma in situ. Among the 33 patients with invasive melanoma, the median Breslow thickness was 1.98 mm (range, 0.14–20.86 mm) in 32 patients, ulceration was seen in 8 of 31 patients (26%), and perineural invasion was identified in 4 of 32 patients (13%) for whom information was available.
The AJCC TNM stages and classifications for each of the patients in the cohort are listed in table 1.
American Joint Committee on Cancer stage and TNM classification at presentation for patients with eyelid melanoma
Treatment
The median time from diagnosis to surgery was 1.9 months (range, 0.7–83.6 months). Three patients had distant metastases at the time of diagnosis and were excluded from this analysis. Two of these patients were treated with palliative resection followed by chemoradiation; the other patient was treated with palliative resection followed by a trial of resiquimod.
Among the 61 patients treated with intent to cure, clear margins were obtained after one excision in 38 patients (62%), two excisions in 21 patients (34%) and three excisions in 2 patients (3%). Eight patients had in situ melanoma present at one margin after final excision. Five of these eight patients had recurrent in situ melanoma (rTisN0M0) at presentation to our institution; the remaining three patients presented with rT2aN0M0, rT4bN0M0 and T2aN0M0 melanomas. Three patients were treated with adjuvant topical imiquimod for positive in situ margin; the remaining five patients were observed. The median planned clinical margin width at first excision was 5.0 mm (range, 1–20 mm).
Thirteen patients had sentinal lymph node (SLN) biopsy; four of these patients had positive SLNs, and biopsy results were indeterminate for one patient.
Predictors of need for two or more excisions
Need for two or more excisions was not associated with surgical margin width, largest tumour diameter (p=0.11), Breslow thickness (p=0.54), AJCC TNM classification or disease status at presentation (primary vs recurrent disease) (table 2).
Association between re-excision requirement and various tumour and treatment features in 61 patients with eyelid melanoma
The median Breslow thickness was 1.95 mm (range, 0.2–4.25 mm) for melanomas that required at least two excisions to achieve negative margins and 2 mm (range, 0.14–21 mm) for melanomas requiring a single excision. The median largest tumour diameter was 15 mm (range, 1.5–60 mm) for melanomas that required at least two excisions and 13 mm (range, 1–70 mm) for melanomas with clear margins achieved after requiring a single excision.
Local recurrence
Among the 61 patients who had surgery with curative intent, local recurrence was observed in only 1 patient, 21 months after surgery. This patient had presented to our institution with an incompletely excised melanoma of the upper eyelid and was treated with complete excision. At her follow-up 21 months later, she was found to have subcutaneous nodules that showed dermal melanoma on excision. Her margins of resection had been deemed to be free of melanoma after one excision.
Prognostic importance of T category
At presentation, 6 of the 64 patients in the study (10%) had nodal metastases. Five of the 14 patients who had SLN biopsy had a positive SLN detected, and one additional patient was diagnosed with nodal metastasis at presentation on the basis of positive findings on ultrasound-guided fine needle aspiration biopsy. All five patients with positive SLNs had T category greater than T2b, Breslow thickness greater than 1.5 mm and more than one mitotic figure per mm2. Six of the 17 patients (35%) presenting with T category greater than T2b had nodal metastasis at presentation, compared with none of the 46 patients presenting with T category less than or equal to T2b (p=0.0026) (figure 1). Nodal metastasis at presentation was also correlated with shorter time to disease progression (p=0.007).
(A) Proportion of patients with positive sentinel lymph nodes at presentation according to American Joint Committee on Cancer T category at presentation shows a statistically significant association between positive nodal metastasis and T category more advanced than T2b at presentation on Fisher's exact test (p=0.0026). (B) Proportion of patients with regional or distant metastasis either at presentation or during follow-up as a function of T category (T2b or less advanced vs more advanced than T2b) using Fisher's exact test (p=0.0061). AJCC, American Joint Committee on Cancer.
One patient developed nodal metastases in the parotid and neck 4 months after surgery at our institution. This patient had had normal findings on imaging studies at presentation to our institution, including sonography of nodal basins, CT of the orbit, head and neck, and positron emission tomography. This patient had undergone surgical excision and SLN biopsy at another institution 11 months before presentation to our institution. The SLN biopsy had revealed a positive SLN, but no additional treatment had been delivered to address this positive SLN. The patient was referred to our centre for management of a local subcutaneous recurrence in the area of the previous surgery (lower eyelid).
Another patient developed metastasis to the lung 9 months after surgery at our institution. This patient had had normal findings on whole body CT at baseline.
Survival
During follow-up, three patients died of metastatic disease. The overall survival rate at 6 months was 96.5% (95% CI 86.5% to 99.1%). At 1 year after surgical resection, the metastasis-free survival rate was 96.05% (95% CI 85.0% to 99.0%), and the progression-free survival rate was 95.9% (95% CI 84.6% to 99.0%) (figure 2). Of patients followed for at least 1 year, 4 of 18 patients (22.2%) with T category greater than T2b at presentation had regional or distant metastasis at presentation or during the follow-up, compared with 0 of 32 patients with T category T2b or less (p=0.0061) (figure 1B).
Kaplan–Meier progression-free survival curve for the entire cohort of patients with eyelid melanoma.
Discussion
Our study suggests that more than a third of patients with eyelid melanoma will need at least two excisions to achieve clear margins. There was a higher proportion of recurrent melanomas than primary melanomas needing more than one excision to achieve negative margins (53% vs 31%), but this difference was not statistically significant (p=0.15).
We also found that AJCC T category greater than T2b was associated with a significantly higher risk of lymph node metastasis at presentation or during the follow-up period. Furthermore, nodal status at presentation (N category of AJCC) was significantly associated with a shorter time to progression.
The AJCC T category is based on Breslow thickness, presence of ulceration and presence of mitotic figures (box 1). Thirty-five per cent of patients who had eyelid melanoma >T2b had lymph node metastasis at presentation, compared with 0% of patients with tumours ≤T2b. Furthermore, for patients followed for at least 1 year, T category >T2b was correlated with a significantly higher risk of nodal and distant metastasis at presentation or during the follow-up; 100% of patients with nodal and distant metastasis had T category >T2b. These findings corroborate with some of the findings of the Collaborative Eyelid Skin Melanoma Group (CESMG),3 which analysed a cohort of 44 patients treated by 23 different surgeons at 23 centres and found that each millimetre increase in Breslow thickness was associated with an increase in HR of 1.78 for local recurrence and 2.16 for nodal metastasis. The CESMG study did not find increased Breslow thickness to be associated with significant increase in HR for distant metastasis.
The significant correlation between the presence of nodal metastasis at presentation and a shorter time to progression (local recurrence or distant metastasis) in our study underscores the importance of evaluating the regional lymph nodes at presentation in patients with eyelid melanoma. We advocate baseline ultrasonography of parotid and submandibular nodal basins and consideration of SLN biopsy for patients with eyelid melanomas with T category >T2b.
No deaths were observed among the 61 patients in our cohort who were treated with intent to cure, a better prognosis than previously reported for thin cutaneous melanoma (<1 mm) elsewhere in the body.6 We observed local recurrence in only 1 out of 61 patients (2%) who were surgically treated with the intent to cure; this is a lower local recurrence rate than previously reported.3 We believe our method of relying on paraffin evaluation of margins and delayed reconstruction only after negative margin status is confirmed may be partly responsible for this excellent local control rate.7 Examination of surgical margins on permanent sections is considered the gold standard in margin control for melanoma.4 Recommendations regarding the width of the excision margins for cutaneous melanomas are somewhat variable. A systematic review of randomised controlled trials in cutaneous melanoma recommended excision margin of 2–5 mm for in situ melanoma, margins of 2 cm for melanoma 2–4 mm thick, and margins of 2–3 cm for melanoma >4 mm thick.8 In contrast, the WHO recommends a 1 cm margin of resection for melanoma up to 2 mm thick and a 2 cm margin of excision for thicker melanoma. For melanomas in the periocular region where complicated reconstruction of the upper and lower eyelids is often necessary, we recommend that the repair of the defect be delayed until margin status is confirmed. We acknowledge, however, that there may be some disadvantages to this approach. First, the management of an open wound for up to more than 10 days9 after the initial surgery may be psychologically and logistically unacceptable to patients. All patients in our report had 24 h ‘rush’ permanent section evaluation of the margins, and the defect was closed the next day. Thus, prolonged delays in reconstruction were not an issue in our cohort.
Second, in patients with comorbidities such as diabetes or immune compromise, an open wound may be associated with increased risk for infection compared with immediate closure of the wound. This, however, was not found to be the case in a series of 280 patients with defects from skin cancer treated with delayed reconstruction.10 Demirci et al11 adopted the ‘square procedure’, excision of a 2 mm strip of margins around the clinical tumour (without excision of the tumour) until margins are cleared, as a way to avoid an open wound yet still evaluate margins on paraffin-embedded sections.11 The ‘square procedure’, however, did not provide control for deep margins and did not obviate the need for repeated surgical excisions: a median of two excisions (range one to five) were needed to achieve negative margins.
Immunochemical staining in conjunction with frozen section examination has been advocated by the dermatological community as an alternative to permanent sections for evaluating melanoma margins. In one report (N=7), 100% correlation was found between results on paraffin-embedded and MART-1-stained frozen sections, with a turnaround time of 90 min.12 However, the cost of 90 min in the operating room will equate to an additional $5580 per section margin evaluated according to a 2005 study.13 The shortened 20 min MART-1 staining14 approach has not yet been validated against paraffin section technique.
Our study is limited by its retrospective nature and the relatively short median follow-up time of 2 years. Our lower than expected rates of local, regional and distant recurrence may be secondary to the short follow-up time; however, this is unlikely to be a compounding factor since CESMG found that regional and distant metastases plateau after 2 years.3 To our knowledge our study is the largest single-centre study of eyelid melanomas that looks at the issue of surgical margins. It is also the first cohort to validate AJCC seventh-edition T category as a prognostic correlate for nodal metastasis and disease progression. In summary, we recommend that delayed periocular reconstruction be carried out only after margins have been cleared on paraffin-embedded sections, particularly with tumours of the upper and lower eyelids, for which complex eyelid-sharing reconstruction may be needed. Undoing and redoing a flap in these cases may not be advisable and can lead to significant ocular morbidity. Our study also highlights the importance of histological characteristics of eyelid melanoma. Breslow thickness, presence of ulceration, and number of mitotic figures per mm2 are the basis of AJCC T category designation and should be routinely reported by pathologists who review eyelid melanoma specimens. The management algorithm for eyelid melanoma should include evaluation of lymph nodes at presentation and consideration of SLN biopsy, particularly for eyelid melanomas with T category greater than T2b.
References
Footnotes
-
Contributors All listed coauthors have contributed to the content of this manuscript and fit the definition of a contributing author.
-
Funding This research was supported in part by the National Institutes of Health through MD Anderson’s Cancer Center Support Grant, CA016672.
-
Competing interests None.
-
Ethics approval MD Anderson Cancer Center IRB.
-
Provenance and peer review Not commissioned; externally peer reviewed.