- Email: [email protected]
Cutaneous squamous cell carcinoma treated with Mohs micrographic surgery in Australia II. Perineural invasion
Cutaneous squamous cell carcinoma treated with Mohs micrographic surgery in Australia II. Perineural invasion Igal Leibovitch, MD,a Shyamala C. Huilgol, FACD,b,c Dinesh Selva, FRANZCO,a,d Dudley Hill, FACD,c Shawn Richards, FACD,e and Robert Paver, FACDe Adelaide and Sydney, South Australia Background: Perineural invasion (PNI) is an important histologic factor that plays a significant role in cutaneous tumors’ aggressiveness. Objectives: We sought to evaluate the incidence, features, and outcomes of cutaneous squamous cell carcinoma with PNI in patients treated with Mohs micrographic surgery (MMS). Method: This prospective, multicenter, case series included all patients in Australia treated with MMS for squamous cell carcinoma with PNI who were monitored by the Skin and Cancer Foundation Australia between 1993 and 2002. The parameters recorded were patient demographics, duration of tumor, site, preoperative tumor size, recurrence before MMS, histologic subtypes, postoperative defect size, and recurrence at 5 years after MMS. Results: Seventy patients were given a diagnosis of PNI. PNI was more common in men (77.1%) and in previously recurrent tumors (P = .04). The moderately and poorly differentiated histologic subtypes were more likely to be associated with PNI (P \.0001). Tumor sizes before excision, postoperative defect sizes, subclinical extension, and mean number of MMS levels were significantly larger in cases with PNI compared with cases without PNI (P \ .0001, P \ .0001, P = .002, and P \ .0001, respectively). Most patients with PNI (52.9%) were treated with adjunctive radiotherapy. In all, 25 patients completed a 5-year follow-up post-MMS, and two of them (8.0%) were given a diagnosis of recurrence. Conclusion: Although PNI is an uncommon feature of cutaneous squamous cell carcinoma, when present, it is associated with larger, more aggressive tumors, and the risk of recurrence is higher. This emphasizes the importance of tumor excision with margin control and long-term patient monitoring. ( J Am Acad Dermatol 2005;53:261-6.)
P
erineural invasion (PNI) was first described in a mammary carcinoma 170 years ago1 and describes tumor spread by growth in and around the nerve. It is regarded as an important pathway of minimal resistance for tumor extension2-6
From the Oculoplastic and Orbital Division, Department of Ophthalmology and Visual Sciences, Royal Adelaide Hospital,a Department of Dermatology, Royal Adelaide Hospital,b and Department of Surgery,d University of Adelaide; Wakefield Clinicc; and Skin and Cancer Foundation Australia, Sydney.e Funding sources: None. Conflicts of interest: None identified Reprint requests: Shyamala C. Huilgol, FACD, Wakefield Clinic, 270 Wakefield St, Adelaide, South Australia 5000. E-mail: [email protected]. 0190-9622/$30.00 ª 2005 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2005.03.048
and is now known to involve a wide range of noncutaneous and cutaneous tumors. The most frequently involved cutaneous tumors are squamous cell carcinoma (SCC), keratoacanthoma, microcystic adnexal carcinoma, primary cutaneous adenoid cystic carcinoma, malignant melanoma, eccrine epithelioma, and basal cell carcinoma (BCC).2-5 Tumors exhibiting PNI are generally regarded as more aggressive and have a higher rate of recurrence, morbidity, and mortality, particularly in the head and neck region.2,5-8 It is estimated that PNI is present in less than 5% of patients with skin carcinomas.3,5 The incidence in SCC has been reported to be between 2.5% and 14%, and is linked to an increased risk of regional and distant metastases.3,5,9 The Australian Mohs micrographic surgery (MMS) database was initiated in 1993 by the Skin and Cancer 261
262 Leibovitch et al
J AM ACAD DERMATOL AUGUST 2005
Foundation Australia (SCFA) with the aim of collecting prospective data, and involved all MMS surgeons in the country. This second part reports all patients with SCC and PNI who were treated with MMS and monitored by the SCFA between 1993 and 2002.
METHODS We conducted a prospective, noncomparative, multicenter, interventional case series of patients with cutaneous SCC treated with MMS in Australia and monitored by the SCFA between 1993 and 2002. All patients were treated by fellowship-trained MMS surgeons, using standard fresh-frozen MMS techniques, and the data were collected by the SCFA. The criteria for selection were all cases of histologically diagnosed SCC with PNI treated with MMS. The following parameters were recorded: patient identification number, age, sex, reason for referral, duration of tumor, site, recurrences before MMS, histologic classification of malignancy, evidence of PNI, preoperative tumor size, postoperative defect size, and outcome at 5-year follow-up in relation to PNI. Tumor and postoperative defect size were defined into 8 size groups based on maximum diameter using a straight ruler: 0 to 0.9, 1 to 1.9, 2 to 2.9, 3 to 3.9, 4 to 4.9, 5 to 5.9, 6 to 7.9, and 8 to 10 cm. Subclinical tumor extension was calculated as the difference in number of size groups between the defect and tumor size. The degree of anaplasia in the tumor nests was used to grade SCC, based on a subjective assessment of differentiation. The categories were well differentiated (the relative proportion of anaplastic cells was low, and the architecture resembled normal epidermis), moderately differentiated (half of the cells or more were anaplastic), and poorly differentiated (all cells were anaplastic, and diagnosis was based on keratin markers). All surgeons used a standard fresh-frozen MMS technique with a tissue map and color coding of the excised tissue. Frozen sections of the entire outer margin in a continuous layer were prepared, and the tissue was stained with hematoxylin and eosin. A trained MMS technician performed all frozen-section preparation. Residual tumor was mapped, and targeted serial excision performed until the surgical margins were clear. All excision, mapping, and tissue examination was carried out by the MMS surgeons. The decision of whether to perform initial curettage for tumor debulking was made by the individual surgeons, and this was not counted as a first level of excision. Excision margins for initial and subsequent layers varied according to tumor and site and between different surgeons.
The final histologic diagnosis was done on a debulking specimen before MMS and/or on the frozen sections seen on the routine MMS sections, according to the surgeon’s preferences. In several centers, the final histologic specimens were also reviewed by an experienced dermopathologist. The diagnosis of PNI was based on the presence of tumor cells either immediately peripheral to the perineurium of nerves, or involving the epineurium or endoneurium.2 All cases of PNI were confirmed on the debulking specimens or the final MMS sections. The diagnosis of recurrence at 5-year follow-up was based on a clinical examination of the patients by the MMS surgeons. Every patient with suggested recurrence underwent a biopsy. Statistical analysis Associations between categoric variables were analyzed using chi-square tests, with the MantelHaenszel test for linear association where appropriate. The Fisher’s 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 nonnormally distributed data. Exact 95% confidence intervals were calculated for the 5-year recurrence rates. Analyses were performed using software (SAS, Version 9.1, SAS Institute Inc, Cary, NC).
RESULTS Data on PNI were available for 1177 patients who underwent MMS for cutaneous SCC (93.2%). PNI was diagnosed in 70 patients (5.95%); 54 were men (77.1%) and 16 were women (22.9%), with a mean age of 64 6 15 years (median, 66 years; range, 35-96 years). There were 36 cases of PNI of the 772 cases of primary SCC (4.7%), and 34 cases of PNI of 491 cases with previously recurrent SCC (6.9%) (P = .04), comprising 51.4% and 48.6% of all PNI cases, respectively. Most tumors with PNI (91.4%) were present for less than 5 years (48.6%, \1 year; 42.9%, 1-5 years). The main reasons for patients’ referral to MMS were tumor recurrence (31.4%) and a poorly defined tumor (15.7%). SCC with PNI was most common in the auricular area (18 cases, 25.7%), cheek and maxilla (15 cases, 21.4%), and forehead (13 cases, 18.6%), with significant association between forehead location and PNI (P = .05). The most common histologic subtypes for patients with PNI were the moderately differentiated (38 cases, 54.3%) and poorly differentiated (20 cases, 28.6%), whereas for patients with no PNI these were
Leibovitch et al 263
J AM ACAD DERMATOL VOLUME 53, NUMBER 2
the well differentiated infiltrating (43.4%) and moderately differentiated (42.5%). There was a significant association between a moderately or poorly differentiated histology and PNI (P \.0001). (Table I). Previous treatments for recurrent SCC are summarized in Table II. PNI was associated with an increasing number of previous surgical excisions or cryotherapy (P = .04 and P = .07, respectively), but not with an increasing number of curettage and cautery or radiotherapy treatments (P = 1.0). Tumor sizes before excision and postoperative defect sizes were significantly larger in cases with PNI when compared with those with none (P\.0001 for both parameters) (Tables III and IV). Analysis of tumor sizes showed that 55.7% of tumors with PNI were 2 cm or larger compared with only 30.0% of those with no PNI. In addition, defect sizes were 3 cm or larger in 65.7% of cases with PNI compared with only 31.3% of cases without PNI. Significant subclinical extension, defined as tumor defect exceeding tumor size by at least two size groups, was seen in 47.7% of PNI cases, and only in 17.6% of cases without PNI (P = .002) (Table V). The mean number of MMS levels required for complete tumor excision in cases of PNI was 2.64 6 1.0 (median, 2; range 1-6), whereas in non-PNI cases it was 1.70 6 0.8 (median, 2; range 1-7) (P \.0001). There were 37 patients with PNI (52.9%) who were treated with adjunctive local radiotherapy post-MMS. Five-year recurrence data There were 336 patients of 1177 with SCC and data on PNI who completed a 5-year follow-up period after MMS (220 had not yet reached the 5-year follow-up, and 621 were lost to follow-up). Of those 336 patients, 25 patients had PNI (35.7% of all patients initially given a diagnosis of PNI) and 311 patients had no PNI (26.0% of all patients with no PNI). Of the other 45 patients with PNI, 16 had not yet reached the 5-year follow-up and 29 were lost to follow-up. Recurrence at 5 years post-MMS was diagnosed in two patients with PNI (8.0%; exact 95% confidence interval, 5.2%-13.5%) and in 12 of the patients with no PNI (3.7%; exact 95% confidence interval, 2.3%8.2%) (P = .02). The two cases of PNI with recurrence at 5 years are presented in Table VI. The fact that there were only two cases of PNI and recurrence at 5 years precluded statistical comparison of clinical and tumor parameters with the other 23 patients with PNI and no recurrence at 5 years. There was no statistically significant difference, for all the major demographic and clinical parameters (age, sex, previous recurrence, histology,
Table I. Histologic subtypes of squamous cell carcinoma in relation to perineural invasion Histologic subtype
Well differentiated* Moderately differentiated* Poorly differentiated* Acantholytic Data not available
No PNI (N = 1107)
480 470 64 89 4
(43.4%) (42.5%) (5.8%) (8.0%) (0.4%)
PNI cases (N = 70)
5 38 20 5 2
(7.1%) (54.3%) (28.6%) (7.1%) (2.9%)
PNI, Perineural invasion. *P \ .0001
tumor and defect size, subclinical extension, and number of MMS levels), between the 25 patients with PNI and 5-year follow-up, and the 45 patients who were initially given a diagnosis of PNI, but did not complete the 5-year follow-up period.
DISCUSSION The perineural space is a cleavage plane between the nerve and the nerve sheath.10,11 This plane extends from the small peripheral nerves, following their course to the subarachnoid space, and provides a pathway for cutaneous tumor spread to the central nervous system.12,13 This mode of spread is a significant factor that contributes to the morbidity and mortality from cutaneous SCC. In one of the first reports on PNI in skin tumors, Mohs and Lathrop5 reported 807 cases of SCC treated with MMS, with a PNI incidence of 2.5% (17 cases). They also found a PNI incidence of 2% in 508 cases of lip SCC. Cottel14 reported a PNI incidence of 4.9% in 347 cases of SCC treated with MMS. In another report, Goepfert et al9 reviewed 520 patients with SCC and found a higher PNI incidence of 14%. In our series, we identified 70 cases of PNI of 1177 patients with data available (5.95%). This number of cutaneous SCC cases with PNI makes our series one of the largest reports in the literature. The relatively high incidence of PNI in our series is partially explained by the initial referral of high-risk tumors (large, recurrent, midfacial tumors, with aggressive histology and significant subclinical extension) for MMS. In addition, by using en face sections, MMS may allow for better detection of tumors involving small nerve fibers than standard vertical pathology sections, thereby giving a higher and more accurate estimation of PNI incidence. The majority of skin tumors with PNI, previously reported in the literature, were located in the head and neck region. In general, PNI may affect most cranial nerves, but most often it involves the maxillary and mandibular divisions of the trigeminal
264 Leibovitch et al
J AM ACAD DERMATOL AUGUST 2005
Table II. Prior treatments for recurrent squamous cell carcinoma in relation to perineural invasion Prior treatments (No. of cases) Cryotherapy* (PNI/no PNI)
No. of Tx
1 2 3 $4 Total No.
5 2 1 6
(35.7%)/110 (41.8%) (14.3%)/84 (31.9%) (7.1%)/23 (8.7%) (4.3%)/46 (17.5%) 14/263
Curettage and cauteryy (PNI/no PNI)
Surgical excisionz (PNI/no PNI)
RT § (PNI/no PNI)
1 (100%)/70 (85.4%) 0/12 (14.6%) 1/82
23 (74.2%)/176 (83%) 4 (12.9%)/33 (15.6%) 4 (12.9%)/2 (0.9%) 0/1 (0.5%) 31/212
1 (100%)/11 (91.7%) 0/1 (8.3%) 1/12
PNI, Perineural invasion; RT, radiotherapy; Tx, treatments. *P = .07. y P = 1.0. z P = .04. § P = 1.0.
Table III. Preoperative tumor sizes in relation to perineural invasion for patients with squamous cell carcinoma treated with Mohs micrographic surgery Tumor sizes,* cm
\1 1-1.9 2-2.9 3-3.9 4-4.9 5-5.9 6-7.9 8-10 Data not available
No PNI (N = 1107)
251 505 210 68 30 16 6 4 17
(22.7%) (45.6%) (18.9%) (6.1%) (2.7%) (1.4%) (0.5%) (0.4%) (1.5%)
PNI cases (N = 70)
6 23 17 13 8 1
(8.6%) (32.9%) (24.3%) (18.6%) (11.4%) (1.4%) 2 (2.9%)
PNI, Perineural invasion. *P \ .0001.
Table IV. Postoperative defect sizes in relation to perineural invasion for patients with squamous cell carcinoma treated with Mohs micrographic surgery Defect size,* cm
\1 1-1.9 2-2.9 3-3.9 4-4.9 5-5.9 6-7.9 8-10 [10 Data not available
No PNI (N = 1107)
41 362 349 178 86 40 30 10 2 9
(3.7%) (32.7%) (31.5%) (16.1%) (7.8%) (3.6%) (2.7%) (0.9%) (0.2%) (0.8%)
PNI cases (N = 70)
6 14 18 14 7 3 3 1 4
(8.6%) (20.0%) (25.7%) (20.0%) (10.0%) (4.3%) (4.3%) (1.4%) (5.7%)
PNI, Perineural invasion. *P \ .0001.
nerve, and the facial nerve branches.8,9,11 Most of the tumors in our series were located in the head and neck area, and so it is not surprising that so were most cases of PNI. We found that the most commonly
Table V. Subclinical extension (difference between tumor and defect size) in relation to perineural invasion by squamous cell carcinoma Difference between tumor and defect size (No. of size categories) 0
1
2
31
No PNI 354 531 136 53 (N = 1074) (33.0%) (49.4%) (12.7%) (4.9%) PNI cases 14 19 19 11 (N = 63) (22.2%) (30.2%) (30.2%) (17.5%) PNI, Perineural invasion. Data were not available for 40 patients (3.4%).
involved sites were the auricular area (25.7%), cheek and maxilla (21.4%), and forehead (18.6%), with significant association between forehead location and PNI (P = .05). The probable explanations for the high incidence of PNI in these areas include the extensive nerve supply, the high prevalence of skin carcinomas there,2,8 and possibly a more accessible cleavage plane between the nerve and its sheath.5 One of the main risk factors for an aggressive SCC phenotype is the histologic subtype and the degree of differentiation. Tumors exhibiting lower degrees of differentiation are more commonly associated with PNI.2 The most common histologic subtypes in our patients with PNI were the moderately differentiated (54.3%) and poorly differentiated (28.6%), whereas in patients with no PNI these were the well differentiated (43.4%) and moderately differentiated (42.5%). We also found that there was a significant association between a moderately or poorly differentiated histology and PNI (P \ .0001). Our findings support previous reports by Mohs and Lathrop5 along with Cottel14 and Lawrence and Cottel15 indicating a higher risk of nerve involvement in the more aggressive and less differentiated histologic subtypes of SCC.
Leibovitch et al 265
J AM ACAD DERMATOL VOLUME 53, NUMBER 2
Table VI. Clinical characteristics of patients with perineural invasion and recurrence at 5 years Age, y/sex
Site
66/F 56/M 66/F 56/M
Forehead Auricular Forehead Auricular
Previous SCC Tumor recurrence differentiation size, cm
Yes Yes Yes Yes
Moderately Poorly Moderately Poorly
4-4.9 1-1.9 4-4.9 1-1.9
Time from Defect No. of MMS to size, excision Postexcision recurrence, y levels RT cm
6-7.9 2-2.9 6-7.9 2-2.9
2 2 2 2
Yes Yes Yes Yes
2 1 2 1
Recurrence presentation
Skin Skin Skin Skin
induration induration induration induration
Postrecurrence treatment
MMS MMS MMS MMS
1 1 1 1
RT RT RT RT
F, Female; M, male; MMS, Mohs micrographic surgery; RT, radiotherapy; SCC, squamous cell carcinoma.
Different studies show that tumor recurrence is a very important factor contributing to a more aggressive behavior in SCC, and that PNI is more common in these recurrent tumors. In the series by Mohs and Lathrop,5 one third of the 807 SCC cases were previously treated surgically or with radiotherapy, but the proportion of recurrent tumors among the PNI cases was much higher (62%). Similar findings were reported by Ballantyne et al3 (61%), Cottel14 (56%), Barrett et al13 (57%), and Mendenhall et al16 (68%). Although in our study previously recurrent tumors comprised only 48.6% of PNI cases, it was diagnosed in 6.9% of recurrent SCC cases, compared with only 4.7% of primary tumors (P = .04). These findings support the notion that PNI is more common in recurrent tumors. That previous tumor treatment or incomplete excision contributes to a more aggressive phenotype of skin carcinomas may in part be explained by the presence of scar tissue, which obscures monitoring and delays clinical detection of recurrence.17 In addition, the fibrosis may entrap malignant cells and favor deep extension by preventing upward migration. This postulation is perhaps supported by our findings that PNI was associated with an increasing number of previous surgical excisions or cryotherapy treatments (P = .04 and P = .07, respectively) (Table II). It is also possible that tumors with aggressive histology are more likely to be recurrent after any form of treatment, and it is not that recurrence is a cause of PNI, but rather that tumors that are likely to be recurrent are also likely to have PNI. Nevertheless, it is also possible that recurrence is merely a result of undetected PNI, even when not associated with scar tissue from previous treatments or aggressive histology. As discussed in part I of this report, large tumors with significant subclinical extension are strongly linked to more aggressive behavior. We found that tumors exhibiting PNI were larger than those without PNI, and the postoperative defect sizes were significantly bigger (P \ .0001) (Tables III and IV). In addition, significant subclinical extension was seen in 47.7% of PNI cases, and only in 17.6% of those without PNI (P = .002) (Table V). The mean number
of MMS levels required for complete excision in cases of PNI was also significantly larger than in cases without PNI. These are important factors that demonstrate the more invasive nature of tumors with PNI and emphasize the need for margin-controlled resection. Treatment of tumors with PNI can be a significant challenge as a result of their aggressive nature, potential route of spread, and the difficulties in determining accurate histologic margins.2 Several series were published recently on radiotherapy alone, or in combination with surgical excision, for skin tumors with PNI. These series included a mixed population of SCC and BCC, and so preclude definite conclusions on the cure rate for SCC.7,18,19 McCord et al.7,18 found that local control rate for patients with histologic evidence of PNI (3 BCC and 31 SCC) was 78% for radiotherapy alone or combined with surgical excision, whereas in another series of patients with clinical PNI (16 BCC and 46 SCC), the recurrence rate was even higher (50%). Mendenhall et al16 analyzed 25 patients with cutaneous SCC of the head and neck with clinical PNI who were treated with radiotherapy alone or in combination with operation. They found a 5-year local control rate of 20% for radiotherapy alone, and 38% for the combined treatment. The high sensitivity of MMS in detecting PNI makes it the preferred mode of treatment for SCC with PNI.2,20,21 Rowe et al21 reviewed all studies since 1940 on the prognosis of SCC. They found that the local recurrence rate for 72 cases of SCC and PNI treated with surgical excision was 47.2%, whereas no cases of recurrence were noted in 17 cases treated with MMS. Cottel14 reported one cases of metastasis (5.9%) and no cases of local recurrence in 17 patients with SCC and PNI who were treated with MMS and followed up for a period of 1 to 42 months. There were 9 patients in their series (53%) who were treated with adjunctive radiotherapy. In another report by Lawrence and Cottel,15 of the 44 cases of SCC with PNI treated with MMS, 3 (6.8%) had recurrences. More than 50% of the cases of PNI in our series were treated with adjunctive radiotherapy.
266 Leibovitch et al
J AM ACAD DERMATOL AUGUST 2005
This was decided by the MMS surgeon, based on the clinical and histologic findings. The use of adjunctive radiotherapy has been advocated in cases with perivascular or regional lymphatic spread by the tumor, involvement of multiple nerves, large nerves, or when a perineural tumor extended through a craniofacial foramen.22 The 5-year recurrence rate of 8% for patients with PNI in our series is lower than with other non-MMS modalities and is comparable with the outcome in the above-mentioned reports on MMS. Nevertheless, this 5-year recurrence rate was significantly higher than for patients without PNI (3.7%). A possible explanation for this relatively higher recurrence rate is ‘‘apparent skip lesions.’’ These are areas in the nerve that are microscopically free of tumor while there is still tumor in the more proximal or distal parts of the nerve.12 These areas are thought to represent artifacts introduced by twists or turns of the specimen during tissue manipulation and cutting.8 These apparent skip lesions preclude identification of tumor spread and may result in false-negative margins. Some authors believe that to overcome this possible false negative in MMS, an additional level should be taken beyond the point where sections are free of tumor for all cases of PNI.12 Only 35.7% of the patients with PNI in our series completed the 5-year follow-up period. Although these missing data are certainly important, we found no significant differences between the clinical and histologic parameters of patients who completed the follow-up and those who did not, and we believe that the information obtained after 5 years is valuable. In conclusion, although PNI is an uncommon feature of cutaneous SCC, it was found to be an important risk factor that is associated with more aggressive tumor behavior. Tumors with PNI were larger, had significant subclinical extension, and had larger postoperative defects compared with tumors without PNI. Although MMS offers the highest cure rate, it still results in a significant recurrence rate, emphasizing the importance of excision with margin control and long-term monitoring of these patients. The authors wish to thank the SCFA and the participating MMS surgeons for their generosity in providing the data for this research. The MMS surgeons involved were Drs Phillip Artemi, John Coates, Brian De’Ambrosis, Timothy Elliott, Gregory Goodman, Irene Grigoris, Dudley Hill, Shyamala Huilgol, Michelle Hunt, David Leslie, Robert Paver, Shawn Richards, William Ryman, Robert Salmon, Margaret Stewart, Howard Studniberg, Carl Vinciullo, and Perry Wilson. The authors also wish to thank Emmae Ramsay (statistician), Department of Public Health University of Adelaide, for her help and advice in the statistical analysis of data.
REFERENCES 1. Cruveilhier J. Maladies des nerfs. In: Cruveilhier J. Anatomie pathologique du corps humain. 2nd ed, part 35. Paris: JB Bailliere; 1835. p. 3. 2. Feasel AM, Brown TJ, Bogle MA, Tschen JA, Nelson BR. Perineural invasion of cutaneous malignancies. Dermatol Surg 2001;27:531-42. 3. Ballantyne AJ, McCarten AB, Ibanez ML. The extension of cancer of the head and neck through peripheral nerves. Am J Surg 1963;106:651-67. 4. Mendenhall WM, Amdur RJ, Williams LS, Mancuso AA, Stringer SP, Mendenhall NP. Carcinoma of the skin of the head and neck with perineural invasion. Head Neck 2002;24:78-83. 5. Mohs FE, Lathrop TG. Modes of spread of cancer of skin. Arch Dermatol Syph 1952;66:427-39. 6. Chen-Tsai CP, Colome-Grimmer M, Wagner RF Jr. Correlations among neural cell adhesion molecule, nerve growth factor, and its receptors, TrkA, TrkB, TrkC, and p75, in perineural invasion by basal cell and cutaneous squamous cell carcinomas. Dermatol Surg 2004;30:1009-16. 7. McCord MW, Mendenhall WM, Parsons JT, Amdur RJ, Stringer SP, Cassisi NJ, et al. Skin cancer of the head and neck with clinical perineural invasion. Int J Radiat Oncol Biol Phys 2000;47:89-93. 8. Matorin PA, Wagner RF Jr. Mohs micrographic surgery: technical difficulties posed by perineural invasion. Int J Dermatol 2002;31:83-6. 9. Goepfert H, Dichtel WJ, Medina JE, Lindberg RD, Luna MD. Perineural invasion in squamous cell carcinoma of the head and neck. Am J Surg 1984;148:542-7. 10. Larson DL, Rodin AE, Roberts DK, O’Steen WK, Rapperport AS, Lewis SR. Perineural lymphatics: myth or fact. Am J Surg 1966;112:488-92. 11. Dodd GD, Dolan PA, Ballantyne AJ, Ibanez ML, Chau P. The dissection of tumors of the head and neck via the cranial nerves. Radiol Clin North Am 1970;8:445-61. 12. Hanke CW, Wolf RL, Hockman SA, O’Brian JJ. Chemosurgical reports: perineural spread of basal cell carcinoma. J Dermatol Surg Oncol 1983;9:742-7. 13. Barrett TL, Greenway HT Jr, Massullo V, Carlson C. Treatment of basal cell carcinoma and squamous cell carcinoma with perineural invasion. Adv Dermatol 1993;8:277-304. 14. Cottel WI. Perineural invasion of squamous cell carcinoma. J Dermatol Surg Oncol 1982;8:589-600. 15. Lawrence N, Cottel WI. Squamous cell carcinoma of the skin with perineural invasion. J Am Acad Dermatol 1994;31:30-3. 16. Mendenhall WM, Parsons JT, Mendenhall NP, Brant TA, Stringer SP, Cassisi NJ, et al. Carcinoma of the skin of the head and neck with perineural invasion. Head Neck 1989;11:301-8. 17. Richmond J, Devie RM. The significance of incomplete excision in patients with basal cell carcinoma. Br J Plast Surg 1987;40:63-7. 18. McCord MW, Mendenhall WM, Parsons JT, Flowers FP. Skin cancers of the head and neck with incidental perineural invasion. Int J Radiat Oncol Biol Phys 1999;43:591-5. 19. Garcia-Serra A, Hinerman RW, Mendenhall WM, Amdur RJ, Morris CG, Williams LS, et al. Carcinoma of the skin with perineural invasion. Head Neck 2003;25:1027-33. 20. Shriner DL, McCoy DK, Goldberg DJ, Richard F, Wagner RF Jr. Mohs micrographic surgery. J Am Acad Dermatol 1998;39:79-97. 21. Rowe DE, Carroll RJ, Day CL. Prognostic factors for local recurrence, metastasis and survival rates in squamous cell carcinoma of the skin, ear and lip. J Am Acad Dermatol 1992; 26:976-90. 22. Osguthorpe JD, Abel CG, Lang P, Hochman M. Neurotropic cutaneous tumors of the head and neck. Arch Otolaryngol Head Neck Surg 1997;123:871-6.
P
erineural invasion (PNI) was first described in a mammary carcinoma 170 years ago1 and describes tumor spread by growth in and around the nerve. It is regarded as an important pathway of minimal resistance for tumor extension2-6
From the Oculoplastic and Orbital Division, Department of Ophthalmology and Visual Sciences, Royal Adelaide Hospital,a Department of Dermatology, Royal Adelaide Hospital,b and Department of Surgery,d University of Adelaide; Wakefield Clinicc; and Skin and Cancer Foundation Australia, Sydney.e Funding sources: None. Conflicts of interest: None identified Reprint requests: Shyamala C. Huilgol, FACD, Wakefield Clinic, 270 Wakefield St, Adelaide, South Australia 5000. E-mail: [email protected]. 0190-9622/$30.00 ª 2005 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2005.03.048
and is now known to involve a wide range of noncutaneous and cutaneous tumors. The most frequently involved cutaneous tumors are squamous cell carcinoma (SCC), keratoacanthoma, microcystic adnexal carcinoma, primary cutaneous adenoid cystic carcinoma, malignant melanoma, eccrine epithelioma, and basal cell carcinoma (BCC).2-5 Tumors exhibiting PNI are generally regarded as more aggressive and have a higher rate of recurrence, morbidity, and mortality, particularly in the head and neck region.2,5-8 It is estimated that PNI is present in less than 5% of patients with skin carcinomas.3,5 The incidence in SCC has been reported to be between 2.5% and 14%, and is linked to an increased risk of regional and distant metastases.3,5,9 The Australian Mohs micrographic surgery (MMS) database was initiated in 1993 by the Skin and Cancer 261
262 Leibovitch et al
J AM ACAD DERMATOL AUGUST 2005
Foundation Australia (SCFA) with the aim of collecting prospective data, and involved all MMS surgeons in the country. This second part reports all patients with SCC and PNI who were treated with MMS and monitored by the SCFA between 1993 and 2002.
METHODS We conducted a prospective, noncomparative, multicenter, interventional case series of patients with cutaneous SCC treated with MMS in Australia and monitored by the SCFA between 1993 and 2002. All patients were treated by fellowship-trained MMS surgeons, using standard fresh-frozen MMS techniques, and the data were collected by the SCFA. The criteria for selection were all cases of histologically diagnosed SCC with PNI treated with MMS. The following parameters were recorded: patient identification number, age, sex, reason for referral, duration of tumor, site, recurrences before MMS, histologic classification of malignancy, evidence of PNI, preoperative tumor size, postoperative defect size, and outcome at 5-year follow-up in relation to PNI. Tumor and postoperative defect size were defined into 8 size groups based on maximum diameter using a straight ruler: 0 to 0.9, 1 to 1.9, 2 to 2.9, 3 to 3.9, 4 to 4.9, 5 to 5.9, 6 to 7.9, and 8 to 10 cm. Subclinical tumor extension was calculated as the difference in number of size groups between the defect and tumor size. The degree of anaplasia in the tumor nests was used to grade SCC, based on a subjective assessment of differentiation. The categories were well differentiated (the relative proportion of anaplastic cells was low, and the architecture resembled normal epidermis), moderately differentiated (half of the cells or more were anaplastic), and poorly differentiated (all cells were anaplastic, and diagnosis was based on keratin markers). All surgeons used a standard fresh-frozen MMS technique with a tissue map and color coding of the excised tissue. Frozen sections of the entire outer margin in a continuous layer were prepared, and the tissue was stained with hematoxylin and eosin. A trained MMS technician performed all frozen-section preparation. Residual tumor was mapped, and targeted serial excision performed until the surgical margins were clear. All excision, mapping, and tissue examination was carried out by the MMS surgeons. The decision of whether to perform initial curettage for tumor debulking was made by the individual surgeons, and this was not counted as a first level of excision. Excision margins for initial and subsequent layers varied according to tumor and site and between different surgeons.
The final histologic diagnosis was done on a debulking specimen before MMS and/or on the frozen sections seen on the routine MMS sections, according to the surgeon’s preferences. In several centers, the final histologic specimens were also reviewed by an experienced dermopathologist. The diagnosis of PNI was based on the presence of tumor cells either immediately peripheral to the perineurium of nerves, or involving the epineurium or endoneurium.2 All cases of PNI were confirmed on the debulking specimens or the final MMS sections. The diagnosis of recurrence at 5-year follow-up was based on a clinical examination of the patients by the MMS surgeons. Every patient with suggested recurrence underwent a biopsy. Statistical analysis Associations between categoric variables were analyzed using chi-square tests, with the MantelHaenszel test for linear association where appropriate. The Fisher’s 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 nonnormally distributed data. Exact 95% confidence intervals were calculated for the 5-year recurrence rates. Analyses were performed using software (SAS, Version 9.1, SAS Institute Inc, Cary, NC).
RESULTS Data on PNI were available for 1177 patients who underwent MMS for cutaneous SCC (93.2%). PNI was diagnosed in 70 patients (5.95%); 54 were men (77.1%) and 16 were women (22.9%), with a mean age of 64 6 15 years (median, 66 years; range, 35-96 years). There were 36 cases of PNI of the 772 cases of primary SCC (4.7%), and 34 cases of PNI of 491 cases with previously recurrent SCC (6.9%) (P = .04), comprising 51.4% and 48.6% of all PNI cases, respectively. Most tumors with PNI (91.4%) were present for less than 5 years (48.6%, \1 year; 42.9%, 1-5 years). The main reasons for patients’ referral to MMS were tumor recurrence (31.4%) and a poorly defined tumor (15.7%). SCC with PNI was most common in the auricular area (18 cases, 25.7%), cheek and maxilla (15 cases, 21.4%), and forehead (13 cases, 18.6%), with significant association between forehead location and PNI (P = .05). The most common histologic subtypes for patients with PNI were the moderately differentiated (38 cases, 54.3%) and poorly differentiated (20 cases, 28.6%), whereas for patients with no PNI these were
Leibovitch et al 263
J AM ACAD DERMATOL VOLUME 53, NUMBER 2
the well differentiated infiltrating (43.4%) and moderately differentiated (42.5%). There was a significant association between a moderately or poorly differentiated histology and PNI (P \.0001). (Table I). Previous treatments for recurrent SCC are summarized in Table II. PNI was associated with an increasing number of previous surgical excisions or cryotherapy (P = .04 and P = .07, respectively), but not with an increasing number of curettage and cautery or radiotherapy treatments (P = 1.0). Tumor sizes before excision and postoperative defect sizes were significantly larger in cases with PNI when compared with those with none (P\.0001 for both parameters) (Tables III and IV). Analysis of tumor sizes showed that 55.7% of tumors with PNI were 2 cm or larger compared with only 30.0% of those with no PNI. In addition, defect sizes were 3 cm or larger in 65.7% of cases with PNI compared with only 31.3% of cases without PNI. Significant subclinical extension, defined as tumor defect exceeding tumor size by at least two size groups, was seen in 47.7% of PNI cases, and only in 17.6% of cases without PNI (P = .002) (Table V). The mean number of MMS levels required for complete tumor excision in cases of PNI was 2.64 6 1.0 (median, 2; range 1-6), whereas in non-PNI cases it was 1.70 6 0.8 (median, 2; range 1-7) (P \.0001). There were 37 patients with PNI (52.9%) who were treated with adjunctive local radiotherapy post-MMS. Five-year recurrence data There were 336 patients of 1177 with SCC and data on PNI who completed a 5-year follow-up period after MMS (220 had not yet reached the 5-year follow-up, and 621 were lost to follow-up). Of those 336 patients, 25 patients had PNI (35.7% of all patients initially given a diagnosis of PNI) and 311 patients had no PNI (26.0% of all patients with no PNI). Of the other 45 patients with PNI, 16 had not yet reached the 5-year follow-up and 29 were lost to follow-up. Recurrence at 5 years post-MMS was diagnosed in two patients with PNI (8.0%; exact 95% confidence interval, 5.2%-13.5%) and in 12 of the patients with no PNI (3.7%; exact 95% confidence interval, 2.3%8.2%) (P = .02). The two cases of PNI with recurrence at 5 years are presented in Table VI. The fact that there were only two cases of PNI and recurrence at 5 years precluded statistical comparison of clinical and tumor parameters with the other 23 patients with PNI and no recurrence at 5 years. There was no statistically significant difference, for all the major demographic and clinical parameters (age, sex, previous recurrence, histology,
Table I. Histologic subtypes of squamous cell carcinoma in relation to perineural invasion Histologic subtype
Well differentiated* Moderately differentiated* Poorly differentiated* Acantholytic Data not available
No PNI (N = 1107)
480 470 64 89 4
(43.4%) (42.5%) (5.8%) (8.0%) (0.4%)
PNI cases (N = 70)
5 38 20 5 2
(7.1%) (54.3%) (28.6%) (7.1%) (2.9%)
PNI, Perineural invasion. *P \ .0001
tumor and defect size, subclinical extension, and number of MMS levels), between the 25 patients with PNI and 5-year follow-up, and the 45 patients who were initially given a diagnosis of PNI, but did not complete the 5-year follow-up period.
DISCUSSION The perineural space is a cleavage plane between the nerve and the nerve sheath.10,11 This plane extends from the small peripheral nerves, following their course to the subarachnoid space, and provides a pathway for cutaneous tumor spread to the central nervous system.12,13 This mode of spread is a significant factor that contributes to the morbidity and mortality from cutaneous SCC. In one of the first reports on PNI in skin tumors, Mohs and Lathrop5 reported 807 cases of SCC treated with MMS, with a PNI incidence of 2.5% (17 cases). They also found a PNI incidence of 2% in 508 cases of lip SCC. Cottel14 reported a PNI incidence of 4.9% in 347 cases of SCC treated with MMS. In another report, Goepfert et al9 reviewed 520 patients with SCC and found a higher PNI incidence of 14%. In our series, we identified 70 cases of PNI of 1177 patients with data available (5.95%). This number of cutaneous SCC cases with PNI makes our series one of the largest reports in the literature. The relatively high incidence of PNI in our series is partially explained by the initial referral of high-risk tumors (large, recurrent, midfacial tumors, with aggressive histology and significant subclinical extension) for MMS. In addition, by using en face sections, MMS may allow for better detection of tumors involving small nerve fibers than standard vertical pathology sections, thereby giving a higher and more accurate estimation of PNI incidence. The majority of skin tumors with PNI, previously reported in the literature, were located in the head and neck region. In general, PNI may affect most cranial nerves, but most often it involves the maxillary and mandibular divisions of the trigeminal
264 Leibovitch et al
J AM ACAD DERMATOL AUGUST 2005
Table II. Prior treatments for recurrent squamous cell carcinoma in relation to perineural invasion Prior treatments (No. of cases) Cryotherapy* (PNI/no PNI)
No. of Tx
1 2 3 $4 Total No.
5 2 1 6
(35.7%)/110 (41.8%) (14.3%)/84 (31.9%) (7.1%)/23 (8.7%) (4.3%)/46 (17.5%) 14/263
Curettage and cauteryy (PNI/no PNI)
Surgical excisionz (PNI/no PNI)
RT § (PNI/no PNI)
1 (100%)/70 (85.4%) 0/12 (14.6%) 1/82
23 (74.2%)/176 (83%) 4 (12.9%)/33 (15.6%) 4 (12.9%)/2 (0.9%) 0/1 (0.5%) 31/212
1 (100%)/11 (91.7%) 0/1 (8.3%) 1/12
PNI, Perineural invasion; RT, radiotherapy; Tx, treatments. *P = .07. y P = 1.0. z P = .04. § P = 1.0.
Table III. Preoperative tumor sizes in relation to perineural invasion for patients with squamous cell carcinoma treated with Mohs micrographic surgery Tumor sizes,* cm
\1 1-1.9 2-2.9 3-3.9 4-4.9 5-5.9 6-7.9 8-10 Data not available
No PNI (N = 1107)
251 505 210 68 30 16 6 4 17
(22.7%) (45.6%) (18.9%) (6.1%) (2.7%) (1.4%) (0.5%) (0.4%) (1.5%)
PNI cases (N = 70)
6 23 17 13 8 1
(8.6%) (32.9%) (24.3%) (18.6%) (11.4%) (1.4%) 2 (2.9%)
PNI, Perineural invasion. *P \ .0001.
Table IV. Postoperative defect sizes in relation to perineural invasion for patients with squamous cell carcinoma treated with Mohs micrographic surgery Defect size,* cm
\1 1-1.9 2-2.9 3-3.9 4-4.9 5-5.9 6-7.9 8-10 [10 Data not available
No PNI (N = 1107)
41 362 349 178 86 40 30 10 2 9
(3.7%) (32.7%) (31.5%) (16.1%) (7.8%) (3.6%) (2.7%) (0.9%) (0.2%) (0.8%)
PNI cases (N = 70)
6 14 18 14 7 3 3 1 4
(8.6%) (20.0%) (25.7%) (20.0%) (10.0%) (4.3%) (4.3%) (1.4%) (5.7%)
PNI, Perineural invasion. *P \ .0001.
nerve, and the facial nerve branches.8,9,11 Most of the tumors in our series were located in the head and neck area, and so it is not surprising that so were most cases of PNI. We found that the most commonly
Table V. Subclinical extension (difference between tumor and defect size) in relation to perineural invasion by squamous cell carcinoma Difference between tumor and defect size (No. of size categories) 0
1
2
31
No PNI 354 531 136 53 (N = 1074) (33.0%) (49.4%) (12.7%) (4.9%) PNI cases 14 19 19 11 (N = 63) (22.2%) (30.2%) (30.2%) (17.5%) PNI, Perineural invasion. Data were not available for 40 patients (3.4%).
involved sites were the auricular area (25.7%), cheek and maxilla (21.4%), and forehead (18.6%), with significant association between forehead location and PNI (P = .05). The probable explanations for the high incidence of PNI in these areas include the extensive nerve supply, the high prevalence of skin carcinomas there,2,8 and possibly a more accessible cleavage plane between the nerve and its sheath.5 One of the main risk factors for an aggressive SCC phenotype is the histologic subtype and the degree of differentiation. Tumors exhibiting lower degrees of differentiation are more commonly associated with PNI.2 The most common histologic subtypes in our patients with PNI were the moderately differentiated (54.3%) and poorly differentiated (28.6%), whereas in patients with no PNI these were the well differentiated (43.4%) and moderately differentiated (42.5%). We also found that there was a significant association between a moderately or poorly differentiated histology and PNI (P \ .0001). Our findings support previous reports by Mohs and Lathrop5 along with Cottel14 and Lawrence and Cottel15 indicating a higher risk of nerve involvement in the more aggressive and less differentiated histologic subtypes of SCC.
Leibovitch et al 265
J AM ACAD DERMATOL VOLUME 53, NUMBER 2
Table VI. Clinical characteristics of patients with perineural invasion and recurrence at 5 years Age, y/sex
Site
66/F 56/M 66/F 56/M
Forehead Auricular Forehead Auricular
Previous SCC Tumor recurrence differentiation size, cm
Yes Yes Yes Yes
Moderately Poorly Moderately Poorly
4-4.9 1-1.9 4-4.9 1-1.9
Time from Defect No. of MMS to size, excision Postexcision recurrence, y levels RT cm
6-7.9 2-2.9 6-7.9 2-2.9
2 2 2 2
Yes Yes Yes Yes
2 1 2 1
Recurrence presentation
Skin Skin Skin Skin
induration induration induration induration
Postrecurrence treatment
MMS MMS MMS MMS
1 1 1 1
RT RT RT RT
F, Female; M, male; MMS, Mohs micrographic surgery; RT, radiotherapy; SCC, squamous cell carcinoma.
Different studies show that tumor recurrence is a very important factor contributing to a more aggressive behavior in SCC, and that PNI is more common in these recurrent tumors. In the series by Mohs and Lathrop,5 one third of the 807 SCC cases were previously treated surgically or with radiotherapy, but the proportion of recurrent tumors among the PNI cases was much higher (62%). Similar findings were reported by Ballantyne et al3 (61%), Cottel14 (56%), Barrett et al13 (57%), and Mendenhall et al16 (68%). Although in our study previously recurrent tumors comprised only 48.6% of PNI cases, it was diagnosed in 6.9% of recurrent SCC cases, compared with only 4.7% of primary tumors (P = .04). These findings support the notion that PNI is more common in recurrent tumors. That previous tumor treatment or incomplete excision contributes to a more aggressive phenotype of skin carcinomas may in part be explained by the presence of scar tissue, which obscures monitoring and delays clinical detection of recurrence.17 In addition, the fibrosis may entrap malignant cells and favor deep extension by preventing upward migration. This postulation is perhaps supported by our findings that PNI was associated with an increasing number of previous surgical excisions or cryotherapy treatments (P = .04 and P = .07, respectively) (Table II). It is also possible that tumors with aggressive histology are more likely to be recurrent after any form of treatment, and it is not that recurrence is a cause of PNI, but rather that tumors that are likely to be recurrent are also likely to have PNI. Nevertheless, it is also possible that recurrence is merely a result of undetected PNI, even when not associated with scar tissue from previous treatments or aggressive histology. As discussed in part I of this report, large tumors with significant subclinical extension are strongly linked to more aggressive behavior. We found that tumors exhibiting PNI were larger than those without PNI, and the postoperative defect sizes were significantly bigger (P \ .0001) (Tables III and IV). In addition, significant subclinical extension was seen in 47.7% of PNI cases, and only in 17.6% of those without PNI (P = .002) (Table V). The mean number
of MMS levels required for complete excision in cases of PNI was also significantly larger than in cases without PNI. These are important factors that demonstrate the more invasive nature of tumors with PNI and emphasize the need for margin-controlled resection. Treatment of tumors with PNI can be a significant challenge as a result of their aggressive nature, potential route of spread, and the difficulties in determining accurate histologic margins.2 Several series were published recently on radiotherapy alone, or in combination with surgical excision, for skin tumors with PNI. These series included a mixed population of SCC and BCC, and so preclude definite conclusions on the cure rate for SCC.7,18,19 McCord et al.7,18 found that local control rate for patients with histologic evidence of PNI (3 BCC and 31 SCC) was 78% for radiotherapy alone or combined with surgical excision, whereas in another series of patients with clinical PNI (16 BCC and 46 SCC), the recurrence rate was even higher (50%). Mendenhall et al16 analyzed 25 patients with cutaneous SCC of the head and neck with clinical PNI who were treated with radiotherapy alone or in combination with operation. They found a 5-year local control rate of 20% for radiotherapy alone, and 38% for the combined treatment. The high sensitivity of MMS in detecting PNI makes it the preferred mode of treatment for SCC with PNI.2,20,21 Rowe et al21 reviewed all studies since 1940 on the prognosis of SCC. They found that the local recurrence rate for 72 cases of SCC and PNI treated with surgical excision was 47.2%, whereas no cases of recurrence were noted in 17 cases treated with MMS. Cottel14 reported one cases of metastasis (5.9%) and no cases of local recurrence in 17 patients with SCC and PNI who were treated with MMS and followed up for a period of 1 to 42 months. There were 9 patients in their series (53%) who were treated with adjunctive radiotherapy. In another report by Lawrence and Cottel,15 of the 44 cases of SCC with PNI treated with MMS, 3 (6.8%) had recurrences. More than 50% of the cases of PNI in our series were treated with adjunctive radiotherapy.
266 Leibovitch et al
J AM ACAD DERMATOL AUGUST 2005
This was decided by the MMS surgeon, based on the clinical and histologic findings. The use of adjunctive radiotherapy has been advocated in cases with perivascular or regional lymphatic spread by the tumor, involvement of multiple nerves, large nerves, or when a perineural tumor extended through a craniofacial foramen.22 The 5-year recurrence rate of 8% for patients with PNI in our series is lower than with other non-MMS modalities and is comparable with the outcome in the above-mentioned reports on MMS. Nevertheless, this 5-year recurrence rate was significantly higher than for patients without PNI (3.7%). A possible explanation for this relatively higher recurrence rate is ‘‘apparent skip lesions.’’ These are areas in the nerve that are microscopically free of tumor while there is still tumor in the more proximal or distal parts of the nerve.12 These areas are thought to represent artifacts introduced by twists or turns of the specimen during tissue manipulation and cutting.8 These apparent skip lesions preclude identification of tumor spread and may result in false-negative margins. Some authors believe that to overcome this possible false negative in MMS, an additional level should be taken beyond the point where sections are free of tumor for all cases of PNI.12 Only 35.7% of the patients with PNI in our series completed the 5-year follow-up period. Although these missing data are certainly important, we found no significant differences between the clinical and histologic parameters of patients who completed the follow-up and those who did not, and we believe that the information obtained after 5 years is valuable. In conclusion, although PNI is an uncommon feature of cutaneous SCC, it was found to be an important risk factor that is associated with more aggressive tumor behavior. Tumors with PNI were larger, had significant subclinical extension, and had larger postoperative defects compared with tumors without PNI. Although MMS offers the highest cure rate, it still results in a significant recurrence rate, emphasizing the importance of excision with margin control and long-term monitoring of these patients. The authors wish to thank the SCFA and the participating MMS surgeons for their generosity in providing the data for this research. The MMS surgeons involved were Drs Phillip Artemi, John Coates, Brian De’Ambrosis, Timothy Elliott, Gregory Goodman, Irene Grigoris, Dudley Hill, Shyamala Huilgol, Michelle Hunt, David Leslie, Robert Paver, Shawn Richards, William Ryman, Robert Salmon, Margaret Stewart, Howard Studniberg, Carl Vinciullo, and Perry Wilson. The authors also wish to thank Emmae Ramsay (statistician), Department of Public Health University of Adelaide, for her help and advice in the statistical analysis of data.
REFERENCES 1. Cruveilhier J. Maladies des nerfs. In: Cruveilhier J. Anatomie pathologique du corps humain. 2nd ed, part 35. Paris: JB Bailliere; 1835. p. 3. 2. Feasel AM, Brown TJ, Bogle MA, Tschen JA, Nelson BR. Perineural invasion of cutaneous malignancies. Dermatol Surg 2001;27:531-42. 3. Ballantyne AJ, McCarten AB, Ibanez ML. The extension of cancer of the head and neck through peripheral nerves. Am J Surg 1963;106:651-67. 4. Mendenhall WM, Amdur RJ, Williams LS, Mancuso AA, Stringer SP, Mendenhall NP. Carcinoma of the skin of the head and neck with perineural invasion. Head Neck 2002;24:78-83. 5. Mohs FE, Lathrop TG. Modes of spread of cancer of skin. Arch Dermatol Syph 1952;66:427-39. 6. Chen-Tsai CP, Colome-Grimmer M, Wagner RF Jr. Correlations among neural cell adhesion molecule, nerve growth factor, and its receptors, TrkA, TrkB, TrkC, and p75, in perineural invasion by basal cell and cutaneous squamous cell carcinomas. Dermatol Surg 2004;30:1009-16. 7. McCord MW, Mendenhall WM, Parsons JT, Amdur RJ, Stringer SP, Cassisi NJ, et al. Skin cancer of the head and neck with clinical perineural invasion. Int J Radiat Oncol Biol Phys 2000;47:89-93. 8. Matorin PA, Wagner RF Jr. Mohs micrographic surgery: technical difficulties posed by perineural invasion. Int J Dermatol 2002;31:83-6. 9. Goepfert H, Dichtel WJ, Medina JE, Lindberg RD, Luna MD. Perineural invasion in squamous cell carcinoma of the head and neck. Am J Surg 1984;148:542-7. 10. Larson DL, Rodin AE, Roberts DK, O’Steen WK, Rapperport AS, Lewis SR. Perineural lymphatics: myth or fact. Am J Surg 1966;112:488-92. 11. Dodd GD, Dolan PA, Ballantyne AJ, Ibanez ML, Chau P. The dissection of tumors of the head and neck via the cranial nerves. Radiol Clin North Am 1970;8:445-61. 12. Hanke CW, Wolf RL, Hockman SA, O’Brian JJ. Chemosurgical reports: perineural spread of basal cell carcinoma. J Dermatol Surg Oncol 1983;9:742-7. 13. Barrett TL, Greenway HT Jr, Massullo V, Carlson C. Treatment of basal cell carcinoma and squamous cell carcinoma with perineural invasion. Adv Dermatol 1993;8:277-304. 14. Cottel WI. Perineural invasion of squamous cell carcinoma. J Dermatol Surg Oncol 1982;8:589-600. 15. Lawrence N, Cottel WI. Squamous cell carcinoma of the skin with perineural invasion. J Am Acad Dermatol 1994;31:30-3. 16. Mendenhall WM, Parsons JT, Mendenhall NP, Brant TA, Stringer SP, Cassisi NJ, et al. Carcinoma of the skin of the head and neck with perineural invasion. Head Neck 1989;11:301-8. 17. Richmond J, Devie RM. The significance of incomplete excision in patients with basal cell carcinoma. Br J Plast Surg 1987;40:63-7. 18. McCord MW, Mendenhall WM, Parsons JT, Flowers FP. Skin cancers of the head and neck with incidental perineural invasion. Int J Radiat Oncol Biol Phys 1999;43:591-5. 19. Garcia-Serra A, Hinerman RW, Mendenhall WM, Amdur RJ, Morris CG, Williams LS, et al. Carcinoma of the skin with perineural invasion. Head Neck 2003;25:1027-33. 20. Shriner DL, McCoy DK, Goldberg DJ, Richard F, Wagner RF Jr. Mohs micrographic surgery. J Am Acad Dermatol 1998;39:79-97. 21. Rowe DE, Carroll RJ, Day CL. Prognostic factors for local recurrence, metastasis and survival rates in squamous cell carcinoma of the skin, ear and lip. J Am Acad Dermatol 1992; 26:976-90. 22. Osguthorpe JD, Abel CG, Lang P, Hochman M. Neurotropic cutaneous tumors of the head and neck. Arch Otolaryngol Head Neck Surg 1997;123:871-6.