- Email: [email protected]
The Australian Mohs database
The Australian Mohs Database Periocular Squamous Cell Carcinoma Raman Malhotra, FRCOphth,1 Shyamala C. Huilgol, FACD,2 Nghi T. Huynh, MBBS,3 Dinesh Selva, FRANZCO1 Objective: To present the results of all patients with periocular squamous cell carcinoma (SCC) treated with Mohs’ micrographic surgery (MMS) in Australia between 1993 and 1999. Design: Prospective, noncomparative, multicenter, interventional case series. Participants: Seventy-nine patients undergoing MMS for periocular SCC over a 7-year period (1993– 1999). Main Outcome Measures Recurrence after MMS, site, size, prior recurrence, defect size, histologic subtype, and presence of perineural invasion (PNI). Results: Seventy-nine patients had 54 (68%) lower eyelid, 19 (24%) medial canthus, and 6 (7.6%) upper eyelid SCCs. The most common histologic subtypes were well differentiated (48.7%) and moderately differentiated (35.1%) SCC. Seventy-three percent were primary tumors and 27% were recurrent tumors. Three (4.3%) cases had histologically confirmed PNI. Two of the 3 cases with PNI were recurrent (P ⫽ 0.1355), and surgical excision was the most frequent prior treatment (P ⫽ 0.0192), with up to 3 prior excisions in 1 case. None of the patients with PNI received adjunctive radiotherapy, and none have recurred with follow-ups to date of 4, 5, and 6 years, respectively. A median follow-up of 73 months (mean, 77.3; range, 42–117) was available in 56 (71%) cases, of which 2 (3.64%; exact 95% confidence interval, 0.44%–12.53%) recurred. Both cases were primary, moderately differentiated SCCs. Conclusion: This is the largest reported prospective series of periocular SCC managed by MMS, with a median follow-up of over 5 years. It demonstrates that MMS has the lowest reported recurrence rate (3.64%) of any treatment modality for periocular SCC. That there were no recurrences for cases with PNI further emphasizes the importance of margin-controlled excision for periocular SCC. Ophthalmology 2004;111:617– 623 © 2004 by the American Academy of Ophthalmology.
Squamous cell carcinoma (SCC) is the second most common malignant tumor of the eyelids, accounting for 5% to 10% of periocular malignancies.1,2 It occurs most commonly in fair-skinned elderly individuals with a history of chronic sun exposure and skin damage, with some areas of Australia reporting incidences to be as high as 600/100 000 and 298/100 000 for men and women, respectively, 20 to 69 years old.1,3 In contrast, the reported incidence rates in the United States have ranged between 40/100 000 and 158/100 000 and between 13/100 000 and 56/100 000 for men and women, respectively.4 Local recurrence and metastatic rates for SCC of the Originally received: February 3, 2003. Accepted: July 25, 2003.
Manuscript no. 230066.
1
Oculoplastic and Orbital Unit, Department of Ophthalmology, Royal Adelaide Hospital, University of Adelaide, Adelaide, Australia.
2
Department of Dermatology, Royal Adelaide Hospital, University of Adelaide and Wakefield Clinic, Adelaide, Australia.
3
Skin and Cancer Foundation Australia, Sydney, Australia. Correspondence to Dr Dinesh Selva, Oculoplastic and Orbital Unit, Department of Ophthalmology, Royal Adelaide Hospital, North Terrace, Adelaide, South Australia 5000, Australia. E-mail: Awestwoo@mail. rah.sa.gov.au.
© 2004 by the American Academy of Ophthalmology Published by Elsevier Inc.
skin, ear, or lip managed by unmonitored surgical excision and destructive modalities are high, varying up to 25% and 45%, respectively.5 Mohs’ micrographic surgery (MMS) is a method of tumor excision with frozen-section margin assessment. It offers the lowest recurrence rate for SCC in comparison to other treatment modalities, together with maximal preservation of normal tissue, and has been recommended as a treatment of choice for periocular SCC.6 Mohs’ micrographic surgery has been shown to achieve the lowest 5-year recurrence rates for skin (3.1%), lip (2.3%), ear (5.3%), and recurrent (10%) SCC.5 The Australian Mohs’ surgery database was started in 1993 by the Skin and Cancer Foundation Australia, after the establishment of the Mohs subdivision of the Australasian College of Dermatologists. Its aim was to collect prospective data for the evaluation of outcomes after MMS, and all Mohs surgeons in the country agreed to collaborate. The aim of this study is to present the findings and outcomes of all patients with periocular SCC treated with MMS in Australia between 1993 and 1999, the largest reported prospective nationwide series of periocular SCC managed by MMS. ISSN 0161-6420/04/$–see front matter doi:10.1016/j.ophtha.2003.07.020
617
Ophthalmology Volume 111, Number 4, April 2004
Materials and Methods
Table 1. Reason Stated for Referral for Mohs’ Micrographic Surgery
We conducted a prospective, noncomparative, interventional, multicenter case series of periocular SCC treated with MMS in Australia between 1993 and 1999. We included in the study all patients with periocular SCC treated by MMS. Periocular was defined as medial canthus, upper or lower eyelid. Tumors outside this region were excluded from this study. The following parameters were recorded: patient identification number, age, gender, reason for referral, histologic classification of SCC and evidence of perineural invasion (PNI), duration of tumor, periocular site, number of recurrences and previous treatments before MMS, preoperative tumor size and postoperative defect size, number of levels to obtain histologic clearance, type of reconstructive surgical repair used, complications during MMS or reconstruction, and recurrence at 5-year follow-up after MMS. Tumor and defect sizes were defined into 8 groups based on the maximum diameter using a straight rule: 0 to 0.9 cm, 1 to 1.9 cm, 2 to 2.9 cm, 3 to 3.9 cm, 4 to 4.9 cm, 5 to 5.9 cm, 6 to 7.9 cm, and 8 to 10 cm. All patients were treated by an accredited Mohs surgeon who had undertaken a 1-year Mohs fellowship in a training program approved by the American College of Mohs Micrographic Surgery and Cutaneous Oncology. Standard fresh-frozen MMS techniques were used. Whenever necessary, if the tumor mass was debulked by the use of a curette, this was not counted as the first level of excision. The first level was then removed by excision beyond the visible margin of the tumor. A tissue map was constructed, and excised tissue was color coded. Frozen sections of the entire outer margin in a continuous layer (rather than vertical or breadloaf sections through the tissue) were prepared, and the tissue was stained with hemotoxylin– eosin. All frozen-section preparation was carried out by a trained technician. Residual tumors were mapped, and serial levels were excised until the surgical margins were cleared of tumor. Resection, orientation, mapping, and examination of sections were carried out by the Mohs surgeon. Data were collected by the Skin and Cancer Foundation, and 5-year follow-up data were based on a questionnaire sent out to all Mohs surgeons. Where follow-up data were not available, patients were contacted in December 2002 and a telephone interview aimed at determining recurrence was conducted. Patients were interviewed to determine the outcome of their most recent clinical inspection by their dermatologist or oculoplastic or Mohs surgeon. If on direct questioning the possibility of recurrence was suspected, patients were instructed to contact their clinician. All contactable patients had been clinically assessed within the previous year with no new recurrence since their last assessment. The primary outcome measure was recurrence.
Reason for Referral
No. of Patients (79)
Periocular site Recurrent tumor Poorly defined tumor Incomplete excision of tumor Tumor type No details available
44 (64%) 13 (19%) 4 (6%) 5 (7%) 3 (4%) 10
Statistical Analysis Associations between categorical variables were analyzed using chi-square tests, with the Mantel–Haenszel test for linear association where appropriate. The Fisher exact test was used if expected values were less than 5. Comparison of normally distributed variables among groups was performed using t tests and analysis of variance; their nonparametric equivalent was used for non–normally distributed data. Exact 95% confidence intervals (CIs) were calculated for the recurrence rates. Analyses were performed using SAS 8.2 (SAS Institute Inc., Cary, NC).
Results Between 1993 and 1999, 79 patients (60 male and 19 female) underwent MMS for periocular SCC. The mean age was 66⫾13 years (median, 67; range, 25–76). The most common reason given for referral for MMS (Table 1) was the periocular tumor site (64%); 59% (26/44) of the patients with this reason had the lower eyelid affected, and 30% (13/44) the medial canthal. The most common site was the lower eyelid (68%), followed by the medial canthus (24%), and the most common histologic subtypes were well differentiated (48.7%) and moderately differentiated (35.1%) SCCs. Moderately differentiated SCCs were associated with the greatest subclinical extension, being most likely to have a larger difference in tumor and defect size (P ⫽ 0.0021; Table 2). The majority (58%) of tumors were present for less than a year, and 4%3 were present for more than 5 years.
Perineural Invasion Three (4.3%) cases involved histologically confirmed PNI (data unavailable in 9 of 79 cases) and sizes of less than 1 cm, 1 to 1.9
Table 2. Difference in Tumor and Defect Sizes with Histological Subtype and Prior Recurrence Difference in Tumor and Defect Sizes (Size Categories)
SCC histology Acantholytic (5) Poorly differentiated (4) Moderately differentiated (26)* Well differentiated (36) Primary SCC (57) Previously recurrent SCC (19) SCC ⫽ squamous cell carcinoma. No details were available for 8 patients. *P ⫽ 0.0021.
618
0 (28)
1 (32)
2 (12)
ⱖ3 (1)
1 (20%) 1 (33%) 8 (31%) 18 (50%) 22 (39%) 7 (37%)
3 (60%) 2 (67%) 11 (42%) 16 (44%) 25 (44%) 9 (47%)
0 0 7 (27%) 2 (6%) 10 (17%) 2 (11%)
1 (20%) 0 0 0 0 1 (5%)
Malhotra et al 䡠 The Australian Mohs Database: Periocular Squamous Cell Carcinoma Table 3. Prior Treatments for Previously Recurrent Squamous Cell Carcinoma Prior Treatment
No. of Treatments
Cryotherapy (7)
Curettage and Cautery (2)
1 2 3 ⱖ4
29% 43% 14% 14%
50% 50% 0 0
Surgical Excision (10)
Mohs’ Excision (1)
90% 0 10% 0
100% 0 0 0
Complete data available in 20 cases.
cm, and 2 to 2.9 cm, respectively. All 3 were lower eyelid, moderately differentiated SCCs, with a difference of 2 size groups between the tumor and defect. Two of the 3 cases involved previous recurrence (P ⫽ 0.1355), and surgical excision was the most frequent prior treatment (P ⫽ 0.0192), with up to 3 prior excisions in 1 case. Perineural invasion was found to be extratumoral in 1 of the 3 cases; however, none of the patients had neurologic symptoms attributable to PNI before MMS. After excision, no patient received adjuvant radiotherapy, and none has recurred with follow-ups to date of 4, 5, and 6 years, respectively.
Differences between Primary and Recurrent Squamous Cell Carcinomas Data regarding previous tumor occurrence were available in 78 cases. Twenty-seven percent (21/78) were recurrent, after prior treatment. The remaining 73% (57/78) were primary SCCs. Prior treatments are listed in Table 3. Previously recurrent SCCs were larger (P ⫽ 0.0002) and had larger defects (P ⫽ 0.0006) than primary tumors (Table 4). However, there was no association between prior occurrence (P ⫽ 0.3214; Table 2) or, for that matter, any form of prior treatment and subclinical extension (difference in tumor and defect sizes). There was no association between gender (P ⫽ 0.2217), age (P ⫽ 0.2800), or periocular site (P ⫽ 0.4651) and prior recurrence (Table 5). However, although not statistically significant, poorly
differentiated SCCs were more likely to be recurrent (P ⫽ 0.1954; Table 5). In the recurrent group, there was no association between periocular site, tumor size, or defect size and any prior form of treatment. Of the 21 patients with recurrent SCCs, 1 had recurrence after prior MMS: a large (3–3.9 cm), lower eyelid, poorly differentiated SCC that had also been surgically excised before this (Table 3).
Levels Required There was no association between prior recurrence and the number of levels required to confirm complete excision (primary and recurrent SCCs: mean, 1.8 [range, 1– 4], and mean, 2.0 [range, 1– 4], respectively; P ⫽ 0.2337). Lymph node metastases was found at the same time as MMS in 1 case: a 1- to 1.9-cm, moderately differentiated, primary SCC of the upper eyelid, necessitating lymph node dissection and radiotherapy in 1999. He remains asymptomatic 3 years later.
Recurrence at Follow-up During a median follow-up of 73 months (mean, 77.3; range, 42–117), 7 of 79 patients died of non–tumor related causes, and 16 were not contactable. Follow-ups to date were therefore available in the 56 (71%) remaining cases, of which 2 (3.64%; exact 95% CI, 0.44%–12.53%) recurred. The first case recurred within the scar, 6 years after MMS (1996), for a small (⬍1 cm), moderately differentiated, primary SCC of the lower eyelid. The recurrence was surgically excised and reported to be basosquamous carcinoma. The second case recurred 30 months after MMS (1998) for a 1- to 1.9-cm, moderately differentiated, primary SCC of the medial canthus. The recurrence was treated with MMS and reported to be carcinoma in situ.
Discussion The Australian MMS database has the largest reported prospective series of periocular SCCs managed by MMS using fresh-frozen tissue sections and with a median follow-up of over 5 years. The most common site was the
Table 4. Tumor and Defect Sizes in Relation to Prior Recurrence †
Tumor* and Defect Sizes (cm)
Overall (79)
Primary SCC
Previously recurrent SCC*†
0–0.9 0–0.9 1–1.9 1–1.9 2–2.9 2–2.9 3–3.9 3–3.9 4–4.9 4–4.9
36 (46.8%) 9 (11.5%) 30 (38.0%) 33 (42.3%) 9 (11.7%) 23 (29.5%) 2 (2.5%) 10 (12.8%) 0 3 (3.9%)
33 (92%) 9 (100%) 22 (73%) 29 (88%) 3 (33%) 15 (65%) 0 3 (30%) 0 1 (33%)
3 (8%) 0 8 (27%) 4 (12%) 6 (67%) 8 (35%) 2 (100%) 7 (70%) 0 2 (67%)
SCC ⫽ squamous cell carcinoma. No details were available in 3 cases. Periocular site was not associated with tumor size (P ⫽ 0.4606) or defect size (P ⫽ 0.2236). *P ⫽ 0.0002. † P ⫽ 0.0006.
619
Ophthalmology Volume 111, Number 4, April 2004 Table 5. Periocular Site of Squamous Cell Carcinoma (SCC) and Histology Site of SCC (%)
Overall (79) SCC histology Acantholytic (5) Poorly differentiated (4) Moderately differentiated (26) Well differentiated (36)
Lower Eyelid
Upper Eyelid
Medial Canthus
Primary SCC (%)
Recurrent SCC (%)
54 (68.4)
6 (7.6)
19 (24)
55
19
4 (80) 4 (100) 16 (61) 27 (75)
0 0 2 (8) 3 (8)
1 (20) 0 8 (31) 6 (17)
4 (80) 1 (25) 21 (81) 26 (72)
1 (20) 3 (75)* 5 (19) 10 (28)
No details were available for 8 patients. Periocular site was not associated with any histological subtype (P ⫽ 0.0799). There was no association between prior recurrence and gender (P ⫽ 0.2217), age (P ⫽ 0.2800), or periocular site (P ⫽ 0.4651). *P ⫽ 0.1954.
lower eyelid, followed by the medial canthus, and the majority of tumors were well differentiated. We found a recurrence rate of 4% for periocular SCC. That the recurrence rate is so low, with a median follow-up of over 5 years, is particularly impressive in view of the high-risk nature of periocular SCC in this series (Tables 1, 4, 5). Our series included a significant proportion of recurrent (27%), large (14% were greater than 2 cm), or upper eyelid and medial canthal (32%) SCCs. Furthermore, a significant number of cases had aggressive histology (38% were either moderately or poorly differentiated), had been incompletely excised before MMS (7%), or had PNI (4.3%). These features are all known to carry a significant risk of invasive spread, recurrence, or metastasis.1,5,7,8 The recurrent tumors in our series, for example, had previously recurred over 4 times in one particular case.
Comparison to Published Data for Treatment of Squamous Cell Carcinomas A variety of surgical and destructive techniques have been described in the management of SCC; however, 5-year local recurrence rates for non-Mohs modalities are high and, depending on the site, vary between 3% and 23%. Furthermore, 5-year metastatic rates vary between 5% and 45%, with an overall rate of less than 10%.5,9 In the absence of randomized trials, or large studies comparing the results of different modalities in the treatment of periocular SCC, clear conclusions on the effectiveness of various treatments cannot be drawn, particularly where selection criteria for published studies are different. In 1992, however, Rowe et al5 concluded that non-Mohs modalities such as curettage and electrodissection, unmonitored surgical excision, cryotherapy, and radiotherapy are unacceptable treatment options for high-risk lesions. Such lesions would include those of the lip, ear, or eyelid, or those with a size of ⬎2 cm; poorly histologically differentiated lesions; deep lesions; lesions with PNI; previously recurrent lesions; scar carcinomas; or SCCs in immunocompromised patients.5,10 More recently, in the most extensive review of treatments for periocular malignancies, Cook and Bartley6 recommended MMS and excision with either a frozen-section or a paraffin-section control as 2 treatment options for periocular SCC. The interpretation of breadloaf sectioning
620
of standard frozen or paraffin specimens, however, rather than en-face sections, requires assumptions based on random sampling and is subject to skip areas large enough to allow a residual tumor to remain undetected. In contrast, MMS examines en-face sections of the entire outer surface of excised tissue. In the largest study of standard surgical excision for periocular SCC under frozen en-face section control, Donaldson et al11 recently reported a mean follow-up of 31.1 months (range, 3–108) for 36 periocular SCCs, with 1 (2.8%) recurrence at 6 years. In the case that recurred, mild dysplasia without invasive disease was noted at the histologic margin at the time of initial excision (Sullivan TJ, personal communication, 2002). Mohs’ micrographic surgery has consistently been shown to achieve the lowest 5-year recurrence rates for skin (3.1%), lip (2.3%), ear (5.3%), and recurrent (10%) SCCs.5 Although much of the data supporting this conclusion relies heavily on Mohs’ seminal work,12,13 the results of other studies using micrographic surgery, summarized in Table 6, support his findings. Our study is the largest reported prospective study of periocular SCC since Mohs’ report,13 and our overall group seems to comprise a greater percentage of large periocular tumors, SCCs with prior recurrence, and SCCs with aggressive histology (Table 6). Both studies report follow-up data of 5 years or more for a significant proportion of cases, in contrast to Dzubow et al,14 Turner et al,15 Riefkohl et al,16 and Miller et al.17 Furthermore, although the other 4 studies report the findings of 1229 SCCs managed by MMS, they specify only 8 periocular SCCs in total.14 –17 The mean age of 66 years at diagnosis and the preponderance of males (76%) in this study are consistent with published data.14,15
Periocular Site We found the most common site for SCC to be the lower eyelid (68%), followed by the medial canthus (24%). The periocular distribution of SCC remains a subject of dispute, whereby, in contrast, a recent review found that SCCs are most frequently found on the lower eyelid (44%), followed by the upper eyelid (31%), medial canthus (15%), and the lateral canthus (9%).1 However, hospital-based statistics may not be reliable for the epidemiologic study of SCC, due
Malhotra et al 䡠 The Australian Mohs Database: Periocular Squamous Cell Carcinoma Table 6. Comparative Clinical and Recurrence Data on 6 Reports of Micrographic Surgery for Squamous Cell Carcinoma (SCC) Dzubow†
Mohs* Total no. of SCCs Periocular SCC Previously recurrent SCC Tumor size ⬍1 cm 1–1.9 cm ⱖ2 cm Percentage of moderately to poorly differentiated SCC histology Follow-up
Local recurrence Lymph node metastases Total recurrences
213
#
213# 68** 15%††
57%†† 35%†† 7%†† 34% (grades 2–4)§§
ⱖ5 yrs
Primary SCC, 1.5% Recurrent SCC, 4.2% 3 Primary SCC, 1.5% Recurrent SCC, 4.2% Overall, 1.9%
Riefkohl㛳
Turner‡§
Miller¶ 711
Present Study
414
61
43
79
No data 47%
6 21% (13)
0 73%
2 13% (90)
27% (21)
19%‡‡ 69% 1–4 cm‡‡ No data No data (specified as “majority”)
23%‡‡ No data 31%‡‡ 13%
No No No No
6%‡‡ 49%‡‡ 45%‡‡ No data
47% 38% 14% 38%
Mean, 19 mos (range, 1–136)
“Less than 5 years”
1–6 yrs
3.4% (14)
Median, 4 yrs Mean, 3.4 yrs (range, 1–6) 3.3% (2)
No data
1.4% (10)
0.7% (3)
6.6% (4)
No data
No data
1
8.2% (5)
11.9%
1.4% (10)
3.64% (2 primary SCCs) (exact 95% CI, 0.44%–12.53%)
79
4.1% (17) Primary SCC, 8.3%㛳㛳 Recurrent SCC, 10%㛳㛳 Overall, 6.7%㛳㛳
data data data data
Median, 6 yrs Mean, 6.5 yrs (range, 42–117 months) 3.64% (2 primary SCCs)
CI ⫽ confidence interval. *Mohs FE. Micrographic surgery for the microscopically controlled excision of eyelid cancers. Arch Ophthalmol 1986;104:901–9. † Dzubow LM, Rigel DS, Robins P. Risk factors for local recurrence of primary cutaneous squamous cell carcinomas. Treatment by microscopically controlled excision. Arch Dermatol 1982;118:900 –2. ‡ Formalin-fixed paraffin-embedded tissue sections. § Turner RJ, Leonard N, Malcolm AJ, et al. A retrospective study of outcome of Mohs micrographic surgery for cutaneous squamous cell carcinoma using formalin fixed sections. Br J Dermatol 2000;142:752–7. 㛳 Riefkohl R, Pollack S, Georgiade GS. A rationale for the treatment of difficult basal cell and squamous cell carcinomas of the skin. Ann Plast Surg 1985;15:99 –104. ¶ Miller PK, Roenigk RK, Brodland DG, Randle HW. Cutaneous micrographic surgery: Mohs procedure. Mayo Clin Proc 1992;67:971– 80. # Fixed-tissue and frozen-tissue techniques. **Number of cases using frozen-tissue technique with strict 5-year follow-up. †† Percentage of 161 cases with strict 5-year follow-up. ‡‡ Percentage of SCCs with all sites included. §§ Textbook of Dermatology on CD-ROM. 6th ed. Oxford, United Kingdom: Blackwell Science; 1999. 㛳㛳 Five-year adjusted Kaplan-Meier mortality table recurrence rates.
to the fact that a significant proportion of SCCs in the United States, for example, may be diagnosed and treated in physicians’ offices. Despite the fact that our series reflects the referral selection bias towards MMS, our results are more in keeping with those of Cook and Bartley,2 who, in their retrospective, population-based incidence cohort study, found that although SCCs occurred most frequently on the lower eyelid (60%), the next most common site was the medial canthus (20%) rather than the upper lid.
Squamous Cell Carcinoma Malignancy Squamous cell carcinoma malignancy relates to factors including differentiation, degree of atypia, and depth of inva-
sion.1 Broders18 developed a grading system based on the proportion of differentiating (keratinizing) cells present, in which grades 2, 3, and 4 were defined as ⬍75%, 50%, and 25%, respectively. The degree of malignancy is now more commonly reported semiquantitatively by way of well, moderately, or poorly differentiated tumors.1 Degrees of cytologic atypia, extent of keratinization, and the tumor border are utilized, and therefore a direct comparison with Broders’ classification cannot be made. Furthermore, with 3 categories, it is said that there is a tendency for dermatopathologists to report a significant proportion of SCCs as moderately differentiated when in doubt. Despite this, most periocular SCCs are well differentiated and less likely to invade the
621
Ophthalmology Volume 111, Number 4, April 2004 orbit, in contrast to those of the nasopharynx or sinuses.1 This is consistent with our findings, whereby almost half of the tumors were well differentiated (48.7%) and a third were moderately differentiated (35.1%), despite a significant proportion of recurrent, large, previously incompletely excised and often poorly defined SCCs. Poorly differentiated SCCs were more likely to be previously recurrent (3 of 4, P ⫽ 0.1954). It is now becoming more apparent that, in addition to the above grading, prognosis is also related to other features that require documentation, the data for which are unavailable in our study. These other features would include the immune status of the patient and any lesion pre-existing the SCC. Immunosuppressed patients (e.g., after renal transplantation) are at high risk of SCC and more aggressive tumor behavior. Squamous cell carcinoma arising from preexisting actinic keratosis, Bowen’s disease or previous burn scar, or chronic ulcer is also noteworthy, as the latter 3 are known to carry a more aggressive prognosis.19 Tumor thickness and depth of invasion were also not recorded in our study. Several studies have shown a correlation between tumor thickness and anatomical depth of invasion with metastatic potential for SCC.20 Tumors larger than 4 mm, for example, have been shown to have a 45% rate of metastasis.5 However, many of these factors are not considered in the current staging classification developed by the American Joint Committee on Cancer21 and the International Union against Cancer.22 This staging system is based primarily on tumor size in diameter and not depth.
Excision Margins for Squamous Cell Carcinoma and Subclinical Extension Larger, high-risk locations and more aggressive histologic grade tumors are more prone to subclinical tumor extension. These were the findings of Brodland and Zitelli,23 who determined guidelines for excision margins of SCC, based on MMS studies of 141 primary tumors. The authors found that 4-mm margins were required to achieve a 95% clearance rate in low-risk SCC. However, 6-mm margins were required to clear 95% of low-risk site SCCs of ⱖ2 cm, periocular SCCs, and those with histologic grades 2 to 4. The authors also recommended that excision of SCCs should also include subcutaneous fat, as at least 30% invade to this level. Our findings are consistent with the above and further emphasize the importance of margin-controlled excision. Moderately differentiated SCCs were more likely to have significant subclinical extension with a larger difference in tumor and defect sizes (P ⫽ 0.0021; Table 2). Furthermore, of the 36 SCCs of ⬍1 cm, defect sizes were 2 to 2.9 cm in 8 (22%) cases and 3 to 3.9 cm in 1 case (3%). The majority of these cases were moderately differentiated SCCs. Although recurrent SCCs were larger (P ⫽ 0.0002), with larger defects (P ⫽ 0.0006), we found no evidence to suggest that recurrent SCCs have greater subclinical tumor extension (P ⫽ 0.3214).
622
Perineural Invasion Perineural invasion describes tumor growth in or around a nerve and is associated with large SCCs (⬎2 cm), head and neck tumors, previous recurrence, a tendency to be poorly differentiated, aggressive tumor behavior, and a poor prognosis. It is reported to occur in 2.5% to 14% of SCCs.7 In our study, 3 (4.3%) cases had histologically confirmed PNI and 2 were previously recurrent, with up to 3 prior excisions in 1 case. Conventional surgical excision of SCC with PNI is associated with recurrence and metastatic rates of 47% and 35%, respectively.24 In contrast, Cottel25 reported a local recurrence rate of zero and 1 metastasis (5.9%) among 17 cases of SCC with PNI treated by MMS, 9 of which also received adjunctive radiotherapy. That none of the patients with PNI in our study received adjunctive radiotherapy and yet have remained free of recurrence with follow-ups to date of 4 to 6 years supports the recommendation that MMS is the treatment of choice for cutaneous SCC with PNI, to achieve clear margins of an involved nerve.7 In conclusion, the Australian MMS database has the largest reported prospective series of periocular SCC managed by MMS. We found the most common site to be the lower eyelid, followed by the medial canthus. The majority of periocular tumors were well differentiated; however, a proportion demonstrated significant subclinical tumor extension. With median follow-up of over 5 years, MMS appears to have the lowest reported recurrence rate of 3.64% relative to any other treatment modality for periocular SCC. That there were no recurrences for cases with PNI further emphasizes the importance of margin-controlled excision for periocular SCC. Acknowledgments. The Skin and Cancer Foundation, Australia has collected and recorded the information for this article in the Mohs’ surgery database. The Mohs surgeons involved were Drs Phillip Artemi, Timothy Elliott, John Coates, Brian De’Ambrosis, Gregory Goodman, Irene Grigoris, Michelle Hunt, Dudley Hill, Shyamala Huilgol, David Leslie, Robert Paver, Shawn Richards, William Ryman, Robert Salmon, Margaret Stewart, Howard Studniberg, Carl Vinciullo, and Perry Wilson. The authors thank the Foundation for providing data to make this research possible and Dr Celia Chen for her help during the final data collection. The authors also acknowledge Kristyn Willson (Statistician), Department of Public Health, Adelaide University, for her help and advice in the statistical analysis of data.
References 1. Reifler DM, Hornblass A. Squamous cell carcinoma of the eyelid. Surv Ophthalmol 1986;30:349 – 65. 2. Cook BE Jr, Bartley GB. Epidemiological characteristics and clinical course of patients with malignant eyelid tumors in an incidence cohort in Olmsted County, Minnesota. Ophthalmology 1999;106:746 –50. 3. Green A, Battistutta D. Incidence and determinants of skin cancer in a high-risk Australian population. Int J Cancer 1990;46:356 – 61. 4. Scotto J, Kopf AW, Urbach F. Non-melanoma skin cancer among Caucasians in four areas of the United States. Cancer 1974;34:1333– 8.
Malhotra et al 䡠 The Australian Mohs Database: Periocular Squamous Cell Carcinoma 5. Rowe DE, Carroll RJ, Day CL Jr. Prognostic factors for local recurrence, metastasis, and survival rates in squamous cell carcinoma of the skin, ear, and lip. Implications for treatment modality selection. J Am Acad Dermatol 1992;26:976 –90. 6. Cook BE Jr, Bartley GB. Treatment options and future prospects for the management of eyelid malignancies: an evidence-based update. Ophthalmology 2001;108:2088 –98. 7. Feasel AH, Brown TJ, Bogle MA, et al. Perineural invasion of cutaneous malignancies. Dermatol Surg 2001;27:531– 42. 8. Birge HL. Cancer of the eyelids. I. Basal cell and mixed basal cell and squamous cell epithelioma. Arch Ophthalmol 1939; 19:700 – 8. 9. Petter G, Haustein UF. Histologic subtyping and malignancy assessment of cutaneous squamous cell carcinoma. Dermatol Surg 2000;26:521–30. 10. Committee on Guidelines of Care, Task Force on Cutaneous Squamous Cell Carcinoma. Guidelines of care for cutaneous squamous cell carcinoma. J Am Acad Dermatol 1993;28:628 – 31. 11. Donaldson MJ, Sullivan TJ, Whitehead KJ, Williamson RM. Squamous cell carcinoma of the eyelids. Br J Ophthalmol 2002;86:1161–5. 12. Mohs FE. Chemosurgery for skin cancer: fixed tissue and fresh tissue techniques. Arch Dermatol 1976;112:211–5. 13. Mohs FE. Micrographic surgery for the microscopically controlled excision of eyelid cancers. Arch Ophthalmol 1986;104: 901–9. 14. Dzubow LM, Rigel DS, Robins P. Risk factors for local recurrence of primary cutaneous squamous cell carcinomas. Treatment by microscopically controlled excision. Arch Dermatol 1982;118:900 –2.
15. Turner RJ, Leonard N, Malcolm AJ, et al. A retrospective study of outcome of Mohs micrographic surgery for cutaneous squamous cell carcinoma using formalin fixed sections. Br J Dermatol 2000;142:752–7. 16. Riefkohl R, Pollack S, Georgiade GS. A rationale for the treatment of difficult basal cell and squamous cell carcinomas of the skin. Ann Plast Surg 1985;15:99 –104. 17. Miller PK, Roenigk RK, Brodland DG, Randle HW. Cutaneous micrographic surgery: Mohs procedure. Mayo Clin Proc 1992;67:971– 80. 18. Broders AC. Practical points on the microscopic grading of carcinoma. State J Med 1932;32:667–71. 19. Textbook of Dermatology on CD-ROM. 6th ed. Oxford, United Kingdom: Blackwell Science; 1999. 20. Khanna M, Fortier-Riberdy G, Smoller B, Dinehart S. Reporting tumor thickness for cutaneous squamous cell carcinoma. J Cutan Pathol 2002;29:321–3. 21. American Joint Committee on Cancer. AJCC Cancer Staging Manual. 5th ed. Philadelphia: Lippincott-Raven; 1997. 22. Sobin LH, Wittekind C, International Union against Cancer, eds. TNM Classification of Malignant Tumors. 5th ed. New York: John Wiley & Sons; 1997. 23. Brodland DG, Zitelli JA. Surgical margins for excision of primary cutaneous squamous cell carcinoma. J Am Acad Dermatol 1992;27:241– 8. 24. Goepfert H, Dichtel WJ, Medina JE, et al. Perineural invasion in squamous cell skin carcinoma of the head and neck. Am J Surg 1984;148:542–7. 25. Cottel WI. Perineural invasion by squamous-cell carcinoma. J Dermatol Surg Oncol 1982;8:589 – 600.
623
Squamous cell carcinoma (SCC) is the second most common malignant tumor of the eyelids, accounting for 5% to 10% of periocular malignancies.1,2 It occurs most commonly in fair-skinned elderly individuals with a history of chronic sun exposure and skin damage, with some areas of Australia reporting incidences to be as high as 600/100 000 and 298/100 000 for men and women, respectively, 20 to 69 years old.1,3 In contrast, the reported incidence rates in the United States have ranged between 40/100 000 and 158/100 000 and between 13/100 000 and 56/100 000 for men and women, respectively.4 Local recurrence and metastatic rates for SCC of the Originally received: February 3, 2003. Accepted: July 25, 2003.
Manuscript no. 230066.
1
Oculoplastic and Orbital Unit, Department of Ophthalmology, Royal Adelaide Hospital, University of Adelaide, Adelaide, Australia.
2
Department of Dermatology, Royal Adelaide Hospital, University of Adelaide and Wakefield Clinic, Adelaide, Australia.
3
Skin and Cancer Foundation Australia, Sydney, Australia. Correspondence to Dr Dinesh Selva, Oculoplastic and Orbital Unit, Department of Ophthalmology, Royal Adelaide Hospital, North Terrace, Adelaide, South Australia 5000, Australia. E-mail: Awestwoo@mail. rah.sa.gov.au.
© 2004 by the American Academy of Ophthalmology Published by Elsevier Inc.
skin, ear, or lip managed by unmonitored surgical excision and destructive modalities are high, varying up to 25% and 45%, respectively.5 Mohs’ micrographic surgery (MMS) is a method of tumor excision with frozen-section margin assessment. It offers the lowest recurrence rate for SCC in comparison to other treatment modalities, together with maximal preservation of normal tissue, and has been recommended as a treatment of choice for periocular SCC.6 Mohs’ micrographic surgery has been shown to achieve the lowest 5-year recurrence rates for skin (3.1%), lip (2.3%), ear (5.3%), and recurrent (10%) SCC.5 The Australian Mohs’ surgery database was started in 1993 by the Skin and Cancer Foundation Australia, after the establishment of the Mohs subdivision of the Australasian College of Dermatologists. Its aim was to collect prospective data for the evaluation of outcomes after MMS, and all Mohs surgeons in the country agreed to collaborate. The aim of this study is to present the findings and outcomes of all patients with periocular SCC treated with MMS in Australia between 1993 and 1999, the largest reported prospective nationwide series of periocular SCC managed by MMS. ISSN 0161-6420/04/$–see front matter doi:10.1016/j.ophtha.2003.07.020
617
Ophthalmology Volume 111, Number 4, April 2004
Materials and Methods
Table 1. Reason Stated for Referral for Mohs’ Micrographic Surgery
We conducted a prospective, noncomparative, interventional, multicenter case series of periocular SCC treated with MMS in Australia between 1993 and 1999. We included in the study all patients with periocular SCC treated by MMS. Periocular was defined as medial canthus, upper or lower eyelid. Tumors outside this region were excluded from this study. The following parameters were recorded: patient identification number, age, gender, reason for referral, histologic classification of SCC and evidence of perineural invasion (PNI), duration of tumor, periocular site, number of recurrences and previous treatments before MMS, preoperative tumor size and postoperative defect size, number of levels to obtain histologic clearance, type of reconstructive surgical repair used, complications during MMS or reconstruction, and recurrence at 5-year follow-up after MMS. Tumor and defect sizes were defined into 8 groups based on the maximum diameter using a straight rule: 0 to 0.9 cm, 1 to 1.9 cm, 2 to 2.9 cm, 3 to 3.9 cm, 4 to 4.9 cm, 5 to 5.9 cm, 6 to 7.9 cm, and 8 to 10 cm. All patients were treated by an accredited Mohs surgeon who had undertaken a 1-year Mohs fellowship in a training program approved by the American College of Mohs Micrographic Surgery and Cutaneous Oncology. Standard fresh-frozen MMS techniques were used. Whenever necessary, if the tumor mass was debulked by the use of a curette, this was not counted as the first level of excision. The first level was then removed by excision beyond the visible margin of the tumor. A tissue map was constructed, and excised tissue was color coded. Frozen sections of the entire outer margin in a continuous layer (rather than vertical or breadloaf sections through the tissue) were prepared, and the tissue was stained with hemotoxylin– eosin. All frozen-section preparation was carried out by a trained technician. Residual tumors were mapped, and serial levels were excised until the surgical margins were cleared of tumor. Resection, orientation, mapping, and examination of sections were carried out by the Mohs surgeon. Data were collected by the Skin and Cancer Foundation, and 5-year follow-up data were based on a questionnaire sent out to all Mohs surgeons. Where follow-up data were not available, patients were contacted in December 2002 and a telephone interview aimed at determining recurrence was conducted. Patients were interviewed to determine the outcome of their most recent clinical inspection by their dermatologist or oculoplastic or Mohs surgeon. If on direct questioning the possibility of recurrence was suspected, patients were instructed to contact their clinician. All contactable patients had been clinically assessed within the previous year with no new recurrence since their last assessment. The primary outcome measure was recurrence.
Reason for Referral
No. of Patients (79)
Periocular site Recurrent tumor Poorly defined tumor Incomplete excision of tumor Tumor type No details available
44 (64%) 13 (19%) 4 (6%) 5 (7%) 3 (4%) 10
Statistical Analysis Associations between categorical variables were analyzed using chi-square tests, with the Mantel–Haenszel test for linear association where appropriate. The Fisher exact test was used if expected values were less than 5. Comparison of normally distributed variables among groups was performed using t tests and analysis of variance; their nonparametric equivalent was used for non–normally distributed data. Exact 95% confidence intervals (CIs) were calculated for the recurrence rates. Analyses were performed using SAS 8.2 (SAS Institute Inc., Cary, NC).
Results Between 1993 and 1999, 79 patients (60 male and 19 female) underwent MMS for periocular SCC. The mean age was 66⫾13 years (median, 67; range, 25–76). The most common reason given for referral for MMS (Table 1) was the periocular tumor site (64%); 59% (26/44) of the patients with this reason had the lower eyelid affected, and 30% (13/44) the medial canthal. The most common site was the lower eyelid (68%), followed by the medial canthus (24%), and the most common histologic subtypes were well differentiated (48.7%) and moderately differentiated (35.1%) SCCs. Moderately differentiated SCCs were associated with the greatest subclinical extension, being most likely to have a larger difference in tumor and defect size (P ⫽ 0.0021; Table 2). The majority (58%) of tumors were present for less than a year, and 4%3 were present for more than 5 years.
Perineural Invasion Three (4.3%) cases involved histologically confirmed PNI (data unavailable in 9 of 79 cases) and sizes of less than 1 cm, 1 to 1.9
Table 2. Difference in Tumor and Defect Sizes with Histological Subtype and Prior Recurrence Difference in Tumor and Defect Sizes (Size Categories)
SCC histology Acantholytic (5) Poorly differentiated (4) Moderately differentiated (26)* Well differentiated (36) Primary SCC (57) Previously recurrent SCC (19) SCC ⫽ squamous cell carcinoma. No details were available for 8 patients. *P ⫽ 0.0021.
618
0 (28)
1 (32)
2 (12)
ⱖ3 (1)
1 (20%) 1 (33%) 8 (31%) 18 (50%) 22 (39%) 7 (37%)
3 (60%) 2 (67%) 11 (42%) 16 (44%) 25 (44%) 9 (47%)
0 0 7 (27%) 2 (6%) 10 (17%) 2 (11%)
1 (20%) 0 0 0 0 1 (5%)
Malhotra et al 䡠 The Australian Mohs Database: Periocular Squamous Cell Carcinoma Table 3. Prior Treatments for Previously Recurrent Squamous Cell Carcinoma Prior Treatment
No. of Treatments
Cryotherapy (7)
Curettage and Cautery (2)
1 2 3 ⱖ4
29% 43% 14% 14%
50% 50% 0 0
Surgical Excision (10)
Mohs’ Excision (1)
90% 0 10% 0
100% 0 0 0
Complete data available in 20 cases.
cm, and 2 to 2.9 cm, respectively. All 3 were lower eyelid, moderately differentiated SCCs, with a difference of 2 size groups between the tumor and defect. Two of the 3 cases involved previous recurrence (P ⫽ 0.1355), and surgical excision was the most frequent prior treatment (P ⫽ 0.0192), with up to 3 prior excisions in 1 case. Perineural invasion was found to be extratumoral in 1 of the 3 cases; however, none of the patients had neurologic symptoms attributable to PNI before MMS. After excision, no patient received adjuvant radiotherapy, and none has recurred with follow-ups to date of 4, 5, and 6 years, respectively.
Differences between Primary and Recurrent Squamous Cell Carcinomas Data regarding previous tumor occurrence were available in 78 cases. Twenty-seven percent (21/78) were recurrent, after prior treatment. The remaining 73% (57/78) were primary SCCs. Prior treatments are listed in Table 3. Previously recurrent SCCs were larger (P ⫽ 0.0002) and had larger defects (P ⫽ 0.0006) than primary tumors (Table 4). However, there was no association between prior occurrence (P ⫽ 0.3214; Table 2) or, for that matter, any form of prior treatment and subclinical extension (difference in tumor and defect sizes). There was no association between gender (P ⫽ 0.2217), age (P ⫽ 0.2800), or periocular site (P ⫽ 0.4651) and prior recurrence (Table 5). However, although not statistically significant, poorly
differentiated SCCs were more likely to be recurrent (P ⫽ 0.1954; Table 5). In the recurrent group, there was no association between periocular site, tumor size, or defect size and any prior form of treatment. Of the 21 patients with recurrent SCCs, 1 had recurrence after prior MMS: a large (3–3.9 cm), lower eyelid, poorly differentiated SCC that had also been surgically excised before this (Table 3).
Levels Required There was no association between prior recurrence and the number of levels required to confirm complete excision (primary and recurrent SCCs: mean, 1.8 [range, 1– 4], and mean, 2.0 [range, 1– 4], respectively; P ⫽ 0.2337). Lymph node metastases was found at the same time as MMS in 1 case: a 1- to 1.9-cm, moderately differentiated, primary SCC of the upper eyelid, necessitating lymph node dissection and radiotherapy in 1999. He remains asymptomatic 3 years later.
Recurrence at Follow-up During a median follow-up of 73 months (mean, 77.3; range, 42–117), 7 of 79 patients died of non–tumor related causes, and 16 were not contactable. Follow-ups to date were therefore available in the 56 (71%) remaining cases, of which 2 (3.64%; exact 95% CI, 0.44%–12.53%) recurred. The first case recurred within the scar, 6 years after MMS (1996), for a small (⬍1 cm), moderately differentiated, primary SCC of the lower eyelid. The recurrence was surgically excised and reported to be basosquamous carcinoma. The second case recurred 30 months after MMS (1998) for a 1- to 1.9-cm, moderately differentiated, primary SCC of the medial canthus. The recurrence was treated with MMS and reported to be carcinoma in situ.
Discussion The Australian MMS database has the largest reported prospective series of periocular SCCs managed by MMS using fresh-frozen tissue sections and with a median follow-up of over 5 years. The most common site was the
Table 4. Tumor and Defect Sizes in Relation to Prior Recurrence †
Tumor* and Defect Sizes (cm)
Overall (79)
Primary SCC
Previously recurrent SCC*†
0–0.9 0–0.9 1–1.9 1–1.9 2–2.9 2–2.9 3–3.9 3–3.9 4–4.9 4–4.9
36 (46.8%) 9 (11.5%) 30 (38.0%) 33 (42.3%) 9 (11.7%) 23 (29.5%) 2 (2.5%) 10 (12.8%) 0 3 (3.9%)
33 (92%) 9 (100%) 22 (73%) 29 (88%) 3 (33%) 15 (65%) 0 3 (30%) 0 1 (33%)
3 (8%) 0 8 (27%) 4 (12%) 6 (67%) 8 (35%) 2 (100%) 7 (70%) 0 2 (67%)
SCC ⫽ squamous cell carcinoma. No details were available in 3 cases. Periocular site was not associated with tumor size (P ⫽ 0.4606) or defect size (P ⫽ 0.2236). *P ⫽ 0.0002. † P ⫽ 0.0006.
619
Ophthalmology Volume 111, Number 4, April 2004 Table 5. Periocular Site of Squamous Cell Carcinoma (SCC) and Histology Site of SCC (%)
Overall (79) SCC histology Acantholytic (5) Poorly differentiated (4) Moderately differentiated (26) Well differentiated (36)
Lower Eyelid
Upper Eyelid
Medial Canthus
Primary SCC (%)
Recurrent SCC (%)
54 (68.4)
6 (7.6)
19 (24)
55
19
4 (80) 4 (100) 16 (61) 27 (75)
0 0 2 (8) 3 (8)
1 (20) 0 8 (31) 6 (17)
4 (80) 1 (25) 21 (81) 26 (72)
1 (20) 3 (75)* 5 (19) 10 (28)
No details were available for 8 patients. Periocular site was not associated with any histological subtype (P ⫽ 0.0799). There was no association between prior recurrence and gender (P ⫽ 0.2217), age (P ⫽ 0.2800), or periocular site (P ⫽ 0.4651). *P ⫽ 0.1954.
lower eyelid, followed by the medial canthus, and the majority of tumors were well differentiated. We found a recurrence rate of 4% for periocular SCC. That the recurrence rate is so low, with a median follow-up of over 5 years, is particularly impressive in view of the high-risk nature of periocular SCC in this series (Tables 1, 4, 5). Our series included a significant proportion of recurrent (27%), large (14% were greater than 2 cm), or upper eyelid and medial canthal (32%) SCCs. Furthermore, a significant number of cases had aggressive histology (38% were either moderately or poorly differentiated), had been incompletely excised before MMS (7%), or had PNI (4.3%). These features are all known to carry a significant risk of invasive spread, recurrence, or metastasis.1,5,7,8 The recurrent tumors in our series, for example, had previously recurred over 4 times in one particular case.
Comparison to Published Data for Treatment of Squamous Cell Carcinomas A variety of surgical and destructive techniques have been described in the management of SCC; however, 5-year local recurrence rates for non-Mohs modalities are high and, depending on the site, vary between 3% and 23%. Furthermore, 5-year metastatic rates vary between 5% and 45%, with an overall rate of less than 10%.5,9 In the absence of randomized trials, or large studies comparing the results of different modalities in the treatment of periocular SCC, clear conclusions on the effectiveness of various treatments cannot be drawn, particularly where selection criteria for published studies are different. In 1992, however, Rowe et al5 concluded that non-Mohs modalities such as curettage and electrodissection, unmonitored surgical excision, cryotherapy, and radiotherapy are unacceptable treatment options for high-risk lesions. Such lesions would include those of the lip, ear, or eyelid, or those with a size of ⬎2 cm; poorly histologically differentiated lesions; deep lesions; lesions with PNI; previously recurrent lesions; scar carcinomas; or SCCs in immunocompromised patients.5,10 More recently, in the most extensive review of treatments for periocular malignancies, Cook and Bartley6 recommended MMS and excision with either a frozen-section or a paraffin-section control as 2 treatment options for periocular SCC. The interpretation of breadloaf sectioning
620
of standard frozen or paraffin specimens, however, rather than en-face sections, requires assumptions based on random sampling and is subject to skip areas large enough to allow a residual tumor to remain undetected. In contrast, MMS examines en-face sections of the entire outer surface of excised tissue. In the largest study of standard surgical excision for periocular SCC under frozen en-face section control, Donaldson et al11 recently reported a mean follow-up of 31.1 months (range, 3–108) for 36 periocular SCCs, with 1 (2.8%) recurrence at 6 years. In the case that recurred, mild dysplasia without invasive disease was noted at the histologic margin at the time of initial excision (Sullivan TJ, personal communication, 2002). Mohs’ micrographic surgery has consistently been shown to achieve the lowest 5-year recurrence rates for skin (3.1%), lip (2.3%), ear (5.3%), and recurrent (10%) SCCs.5 Although much of the data supporting this conclusion relies heavily on Mohs’ seminal work,12,13 the results of other studies using micrographic surgery, summarized in Table 6, support his findings. Our study is the largest reported prospective study of periocular SCC since Mohs’ report,13 and our overall group seems to comprise a greater percentage of large periocular tumors, SCCs with prior recurrence, and SCCs with aggressive histology (Table 6). Both studies report follow-up data of 5 years or more for a significant proportion of cases, in contrast to Dzubow et al,14 Turner et al,15 Riefkohl et al,16 and Miller et al.17 Furthermore, although the other 4 studies report the findings of 1229 SCCs managed by MMS, they specify only 8 periocular SCCs in total.14 –17 The mean age of 66 years at diagnosis and the preponderance of males (76%) in this study are consistent with published data.14,15
Periocular Site We found the most common site for SCC to be the lower eyelid (68%), followed by the medial canthus (24%). The periocular distribution of SCC remains a subject of dispute, whereby, in contrast, a recent review found that SCCs are most frequently found on the lower eyelid (44%), followed by the upper eyelid (31%), medial canthus (15%), and the lateral canthus (9%).1 However, hospital-based statistics may not be reliable for the epidemiologic study of SCC, due
Malhotra et al 䡠 The Australian Mohs Database: Periocular Squamous Cell Carcinoma Table 6. Comparative Clinical and Recurrence Data on 6 Reports of Micrographic Surgery for Squamous Cell Carcinoma (SCC) Dzubow†
Mohs* Total no. of SCCs Periocular SCC Previously recurrent SCC Tumor size ⬍1 cm 1–1.9 cm ⱖ2 cm Percentage of moderately to poorly differentiated SCC histology Follow-up
Local recurrence Lymph node metastases Total recurrences
213
#
213# 68** 15%††
57%†† 35%†† 7%†† 34% (grades 2–4)§§
ⱖ5 yrs
Primary SCC, 1.5% Recurrent SCC, 4.2% 3 Primary SCC, 1.5% Recurrent SCC, 4.2% Overall, 1.9%
Riefkohl㛳
Turner‡§
Miller¶ 711
Present Study
414
61
43
79
No data 47%
6 21% (13)
0 73%
2 13% (90)
27% (21)
19%‡‡ 69% 1–4 cm‡‡ No data No data (specified as “majority”)
23%‡‡ No data 31%‡‡ 13%
No No No No
6%‡‡ 49%‡‡ 45%‡‡ No data
47% 38% 14% 38%
Mean, 19 mos (range, 1–136)
“Less than 5 years”
1–6 yrs
3.4% (14)
Median, 4 yrs Mean, 3.4 yrs (range, 1–6) 3.3% (2)
No data
1.4% (10)
0.7% (3)
6.6% (4)
No data
No data
1
8.2% (5)
11.9%
1.4% (10)
3.64% (2 primary SCCs) (exact 95% CI, 0.44%–12.53%)
79
4.1% (17) Primary SCC, 8.3%㛳㛳 Recurrent SCC, 10%㛳㛳 Overall, 6.7%㛳㛳
data data data data
Median, 6 yrs Mean, 6.5 yrs (range, 42–117 months) 3.64% (2 primary SCCs)
CI ⫽ confidence interval. *Mohs FE. Micrographic surgery for the microscopically controlled excision of eyelid cancers. Arch Ophthalmol 1986;104:901–9. † Dzubow LM, Rigel DS, Robins P. Risk factors for local recurrence of primary cutaneous squamous cell carcinomas. Treatment by microscopically controlled excision. Arch Dermatol 1982;118:900 –2. ‡ Formalin-fixed paraffin-embedded tissue sections. § Turner RJ, Leonard N, Malcolm AJ, et al. A retrospective study of outcome of Mohs micrographic surgery for cutaneous squamous cell carcinoma using formalin fixed sections. Br J Dermatol 2000;142:752–7. 㛳 Riefkohl R, Pollack S, Georgiade GS. A rationale for the treatment of difficult basal cell and squamous cell carcinomas of the skin. Ann Plast Surg 1985;15:99 –104. ¶ Miller PK, Roenigk RK, Brodland DG, Randle HW. Cutaneous micrographic surgery: Mohs procedure. Mayo Clin Proc 1992;67:971– 80. # Fixed-tissue and frozen-tissue techniques. **Number of cases using frozen-tissue technique with strict 5-year follow-up. †† Percentage of 161 cases with strict 5-year follow-up. ‡‡ Percentage of SCCs with all sites included. §§ Textbook of Dermatology on CD-ROM. 6th ed. Oxford, United Kingdom: Blackwell Science; 1999. 㛳㛳 Five-year adjusted Kaplan-Meier mortality table recurrence rates.
to the fact that a significant proportion of SCCs in the United States, for example, may be diagnosed and treated in physicians’ offices. Despite the fact that our series reflects the referral selection bias towards MMS, our results are more in keeping with those of Cook and Bartley,2 who, in their retrospective, population-based incidence cohort study, found that although SCCs occurred most frequently on the lower eyelid (60%), the next most common site was the medial canthus (20%) rather than the upper lid.
Squamous Cell Carcinoma Malignancy Squamous cell carcinoma malignancy relates to factors including differentiation, degree of atypia, and depth of inva-
sion.1 Broders18 developed a grading system based on the proportion of differentiating (keratinizing) cells present, in which grades 2, 3, and 4 were defined as ⬍75%, 50%, and 25%, respectively. The degree of malignancy is now more commonly reported semiquantitatively by way of well, moderately, or poorly differentiated tumors.1 Degrees of cytologic atypia, extent of keratinization, and the tumor border are utilized, and therefore a direct comparison with Broders’ classification cannot be made. Furthermore, with 3 categories, it is said that there is a tendency for dermatopathologists to report a significant proportion of SCCs as moderately differentiated when in doubt. Despite this, most periocular SCCs are well differentiated and less likely to invade the
621
Ophthalmology Volume 111, Number 4, April 2004 orbit, in contrast to those of the nasopharynx or sinuses.1 This is consistent with our findings, whereby almost half of the tumors were well differentiated (48.7%) and a third were moderately differentiated (35.1%), despite a significant proportion of recurrent, large, previously incompletely excised and often poorly defined SCCs. Poorly differentiated SCCs were more likely to be previously recurrent (3 of 4, P ⫽ 0.1954). It is now becoming more apparent that, in addition to the above grading, prognosis is also related to other features that require documentation, the data for which are unavailable in our study. These other features would include the immune status of the patient and any lesion pre-existing the SCC. Immunosuppressed patients (e.g., after renal transplantation) are at high risk of SCC and more aggressive tumor behavior. Squamous cell carcinoma arising from preexisting actinic keratosis, Bowen’s disease or previous burn scar, or chronic ulcer is also noteworthy, as the latter 3 are known to carry a more aggressive prognosis.19 Tumor thickness and depth of invasion were also not recorded in our study. Several studies have shown a correlation between tumor thickness and anatomical depth of invasion with metastatic potential for SCC.20 Tumors larger than 4 mm, for example, have been shown to have a 45% rate of metastasis.5 However, many of these factors are not considered in the current staging classification developed by the American Joint Committee on Cancer21 and the International Union against Cancer.22 This staging system is based primarily on tumor size in diameter and not depth.
Excision Margins for Squamous Cell Carcinoma and Subclinical Extension Larger, high-risk locations and more aggressive histologic grade tumors are more prone to subclinical tumor extension. These were the findings of Brodland and Zitelli,23 who determined guidelines for excision margins of SCC, based on MMS studies of 141 primary tumors. The authors found that 4-mm margins were required to achieve a 95% clearance rate in low-risk SCC. However, 6-mm margins were required to clear 95% of low-risk site SCCs of ⱖ2 cm, periocular SCCs, and those with histologic grades 2 to 4. The authors also recommended that excision of SCCs should also include subcutaneous fat, as at least 30% invade to this level. Our findings are consistent with the above and further emphasize the importance of margin-controlled excision. Moderately differentiated SCCs were more likely to have significant subclinical extension with a larger difference in tumor and defect sizes (P ⫽ 0.0021; Table 2). Furthermore, of the 36 SCCs of ⬍1 cm, defect sizes were 2 to 2.9 cm in 8 (22%) cases and 3 to 3.9 cm in 1 case (3%). The majority of these cases were moderately differentiated SCCs. Although recurrent SCCs were larger (P ⫽ 0.0002), with larger defects (P ⫽ 0.0006), we found no evidence to suggest that recurrent SCCs have greater subclinical tumor extension (P ⫽ 0.3214).
622
Perineural Invasion Perineural invasion describes tumor growth in or around a nerve and is associated with large SCCs (⬎2 cm), head and neck tumors, previous recurrence, a tendency to be poorly differentiated, aggressive tumor behavior, and a poor prognosis. It is reported to occur in 2.5% to 14% of SCCs.7 In our study, 3 (4.3%) cases had histologically confirmed PNI and 2 were previously recurrent, with up to 3 prior excisions in 1 case. Conventional surgical excision of SCC with PNI is associated with recurrence and metastatic rates of 47% and 35%, respectively.24 In contrast, Cottel25 reported a local recurrence rate of zero and 1 metastasis (5.9%) among 17 cases of SCC with PNI treated by MMS, 9 of which also received adjunctive radiotherapy. That none of the patients with PNI in our study received adjunctive radiotherapy and yet have remained free of recurrence with follow-ups to date of 4 to 6 years supports the recommendation that MMS is the treatment of choice for cutaneous SCC with PNI, to achieve clear margins of an involved nerve.7 In conclusion, the Australian MMS database has the largest reported prospective series of periocular SCC managed by MMS. We found the most common site to be the lower eyelid, followed by the medial canthus. The majority of periocular tumors were well differentiated; however, a proportion demonstrated significant subclinical tumor extension. With median follow-up of over 5 years, MMS appears to have the lowest reported recurrence rate of 3.64% relative to any other treatment modality for periocular SCC. That there were no recurrences for cases with PNI further emphasizes the importance of margin-controlled excision for periocular SCC. Acknowledgments. The Skin and Cancer Foundation, Australia has collected and recorded the information for this article in the Mohs’ surgery database. The Mohs surgeons involved were Drs Phillip Artemi, Timothy Elliott, John Coates, Brian De’Ambrosis, Gregory Goodman, Irene Grigoris, Michelle Hunt, Dudley Hill, Shyamala Huilgol, David Leslie, Robert Paver, Shawn Richards, William Ryman, Robert Salmon, Margaret Stewart, Howard Studniberg, Carl Vinciullo, and Perry Wilson. The authors thank the Foundation for providing data to make this research possible and Dr Celia Chen for her help during the final data collection. The authors also acknowledge Kristyn Willson (Statistician), Department of Public Health, Adelaide University, for her help and advice in the statistical analysis of data.
References 1. Reifler DM, Hornblass A. Squamous cell carcinoma of the eyelid. Surv Ophthalmol 1986;30:349 – 65. 2. Cook BE Jr, Bartley GB. Epidemiological characteristics and clinical course of patients with malignant eyelid tumors in an incidence cohort in Olmsted County, Minnesota. Ophthalmology 1999;106:746 –50. 3. Green A, Battistutta D. Incidence and determinants of skin cancer in a high-risk Australian population. Int J Cancer 1990;46:356 – 61. 4. Scotto J, Kopf AW, Urbach F. Non-melanoma skin cancer among Caucasians in four areas of the United States. Cancer 1974;34:1333– 8.
Malhotra et al 䡠 The Australian Mohs Database: Periocular Squamous Cell Carcinoma 5. Rowe DE, Carroll RJ, Day CL Jr. Prognostic factors for local recurrence, metastasis, and survival rates in squamous cell carcinoma of the skin, ear, and lip. Implications for treatment modality selection. J Am Acad Dermatol 1992;26:976 –90. 6. Cook BE Jr, Bartley GB. Treatment options and future prospects for the management of eyelid malignancies: an evidence-based update. Ophthalmology 2001;108:2088 –98. 7. Feasel AH, Brown TJ, Bogle MA, et al. Perineural invasion of cutaneous malignancies. Dermatol Surg 2001;27:531– 42. 8. Birge HL. Cancer of the eyelids. I. Basal cell and mixed basal cell and squamous cell epithelioma. Arch Ophthalmol 1939; 19:700 – 8. 9. Petter G, Haustein UF. Histologic subtyping and malignancy assessment of cutaneous squamous cell carcinoma. Dermatol Surg 2000;26:521–30. 10. Committee on Guidelines of Care, Task Force on Cutaneous Squamous Cell Carcinoma. Guidelines of care for cutaneous squamous cell carcinoma. J Am Acad Dermatol 1993;28:628 – 31. 11. Donaldson MJ, Sullivan TJ, Whitehead KJ, Williamson RM. Squamous cell carcinoma of the eyelids. Br J Ophthalmol 2002;86:1161–5. 12. Mohs FE. Chemosurgery for skin cancer: fixed tissue and fresh tissue techniques. Arch Dermatol 1976;112:211–5. 13. Mohs FE. Micrographic surgery for the microscopically controlled excision of eyelid cancers. Arch Ophthalmol 1986;104: 901–9. 14. Dzubow LM, Rigel DS, Robins P. Risk factors for local recurrence of primary cutaneous squamous cell carcinomas. Treatment by microscopically controlled excision. Arch Dermatol 1982;118:900 –2.
15. Turner RJ, Leonard N, Malcolm AJ, et al. A retrospective study of outcome of Mohs micrographic surgery for cutaneous squamous cell carcinoma using formalin fixed sections. Br J Dermatol 2000;142:752–7. 16. Riefkohl R, Pollack S, Georgiade GS. A rationale for the treatment of difficult basal cell and squamous cell carcinomas of the skin. Ann Plast Surg 1985;15:99 –104. 17. Miller PK, Roenigk RK, Brodland DG, Randle HW. Cutaneous micrographic surgery: Mohs procedure. Mayo Clin Proc 1992;67:971– 80. 18. Broders AC. Practical points on the microscopic grading of carcinoma. State J Med 1932;32:667–71. 19. Textbook of Dermatology on CD-ROM. 6th ed. Oxford, United Kingdom: Blackwell Science; 1999. 20. Khanna M, Fortier-Riberdy G, Smoller B, Dinehart S. Reporting tumor thickness for cutaneous squamous cell carcinoma. J Cutan Pathol 2002;29:321–3. 21. American Joint Committee on Cancer. AJCC Cancer Staging Manual. 5th ed. Philadelphia: Lippincott-Raven; 1997. 22. Sobin LH, Wittekind C, International Union against Cancer, eds. TNM Classification of Malignant Tumors. 5th ed. New York: John Wiley & Sons; 1997. 23. Brodland DG, Zitelli JA. Surgical margins for excision of primary cutaneous squamous cell carcinoma. J Am Acad Dermatol 1992;27:241– 8. 24. Goepfert H, Dichtel WJ, Medina JE, et al. Perineural invasion in squamous cell skin carcinoma of the head and neck. Am J Surg 1984;148:542–7. 25. Cottel WI. Perineural invasion by squamous-cell carcinoma. J Dermatol Surg Oncol 1982;8:589 – 600.
623