- Email: [email protected]
Skin cancer of the head and neck with incidental microscopic perineural invasion
Int. J. Radiation Oncology Biol. Phys., Vol. 43, No. 3, pp. 591–595, 1999 Copyright © 1999 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/99/$–see front matter
PII S0360-3016(98)00474-X
CLINICAL INVESTIGATION
Skin
SKIN CANCER OF THE HEAD AND NECK WITH INCIDENTAL MICROSCOPIC PERINEURAL INVASION MARK W. MCCORD, M.D.,* WILLIAM M. MENDENHALL, M.D.,* JAMES T. PARSONS, M.D.,* FRANKLIN P. FLOWERS, M.D.†
AND
*Department of Radiation Oncology and †Division of Dermatology, University of Florida College of Medicine, Gainesville, Florida Purpose: To address outcomes in clinically asymptomatic patients in whom the unexpected finding of microscopic perineural invasion is noted at the time of surgery. Methods and Materials: The 35 patients included in this study had skin cancers of the head and neck treated with curative intent between January 1965 and April 1995 at the University of Florida. All patients were without clinical or radiographic evidence of perineural invasion but, at the time of biopsy or surgical excision, had the incidental finding of microscopic perineural invasion. Definitive therapy consisted of radiotherapy alone after lesion biopsy (3 patients) or surgical excision preceded (2 patients) or followed (30 patients) by radiotherapy. All patients had follow-up for at least 1 year, 13 patients (37%) had follow-up for at least 5 years. Results: The 5-year local control rate was 78%. The 5-year local control rate for the few patients treated with radiotherapy alone was statistically similar to that for patients treated with surgery and radiotherapy (100% vs. 77%, p 5 0.4). Multivariate analysis for factors affecting local control included sex, histology, age, treatment group, clinical T stage, initial histologic differentiation, and previously untreated vs. recurrent tumors, none of which was found to be significant. Conclusions: Both surgery plus radiotherapy and radiotherapy alone provide a relatively high rate of local control for patients with incidentally discovered perineural invasion secondary to skin cancer. © 1999 Elsevier Science Inc. Head and neck neoplasms, Local neoplasm recurrence, Recurrence, Skin neoplasms, Treatment outcome.
INTRODUCTION
METHODS AND MATERIALS
The limited but growing literature examining the treatment and outcome of skin cancers of the head and neck with perineural invasion reveals two different patient populations. In one group of clinically asymptomatic patients, the unexpected finding of microscopic perineural invasion is noted upon pathologic review of the surgically resected specimen. Such patients composed most of the population in the study by Cottel (1) of the role of Mohs’ chemosurgery in the treatment of perineural invasion by squamous cell carcinoma. Patients in the other group described by Ballantyne et al. (2, 3) and Goepfert et al. (4) have either clinical symptoms or signs of perineural invasion, radiographic evidence of perineural extension, or the finding of gross nerve involvement at the time of surgical resection. This study addresses outcomes in the former patient group, those in whom microscopic perineural extension is noted incidentally at the time of lesion biopsy or surgical excision. Treatment results in the latter patient population, one that has a markedly worse prognosis, were previously reported by Lee et al. (5).
The 35 patients included in this analysis were treated at the University of Florida with curative intent between January 1965 and April 1995. All patients (either at the time of biopsy or at the time of surgery) had the pathologic finding of microscopic perineural extension. Patients were excluded if they had clinical signs or symptoms of perineural involvement or radiographic evidence of perineural extension, or if they were found to have a grossly involved nerve at the time of surgery. Individuals with evidence of perineural extension from a lymph node metastasis were also excluded from this study. The patient population consisted of 30 men and 5 women. All were white. Ages ranged from 32 to 88 years (median, 69 years). No patient was lost to follow-up. All patients had follow-up for at least 1 year, 21 patients (60%) had follow-up for at least 2 years, and 13 patients (37%) had follow-up for at least 5 years. All patients had a primary lesion of the skin, located in the head and neck region (including the lip and nasal vestibule). The primary tumor sites are listed in Table 1.
Address correspondence to: William M. Mendenhall, M.D., Department of Radiation Oncology, University of Florida Health
Science Center, P.O. Box 100385, Gainesville, FL 32610-0385. Accepted for publication 9 October 1998. 591
592
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 43, Number 3, 1999
Table 2. Patterns of relapse (35 patients with 11 recurrences)
Table 1. Tumor site Site
No. patients
Forehead/brow Ear Postauricular Nose Cheek Nasolabial fold Temple Preauricular Scalp Maxilla Nasal vestibule Lip Periorbital skin
1 5 4 2 4 2 6 1 1 1 1 6 1
Thirty-one patients had squamous cell carcinomas, 3 patients had basal cell carcinomas, and 1 patient had a metatypical basal cell carcinoma. Ten patients had previously untreated lesions, and 25 patients underwent therapy for locally recurrent tumors. Information regarding histologic differentiation of the lesions was available for 20 of the 35 patients. Of these 20 patients, 7 had well-differentiated lesions, 7 had moderately differentiated lesions, and 6 had poorly differentiated lesions. Tumors were clinically staged based on the 1998 American Joint Committee on Cancer (AJCC) system (6), in which the presence of microscopic perineural invasion does not alter the clinical stage. Four patients had clinical Tx lesions, 9 had T1 lesions, 9 had T2 lesions, 3 had T3 lesions, and 10 had T4 lesions. Twenty-eight patients had NO neck disease, 4 had N1 disease, and 3 had N2B disease. Therapy consisted of surgery followed by radiotherapy in 30 patients, 3 patients had definitive radiotherapy after biopsy alone, and 2 patients had preoperative radiotherapy followed by surgical excision. Radiotherapy fields were designed to encompass only the primary skin cancer with an adequate margin in 18 patients. In 17 patients, the nerve pathways to the base of the skull were included to a minimum dose of 50 Gy. Patients treated to the primary site alone usually were those with negative margins, whereas those who received irradiation along the course of the nerve to the skull base usually had either positive margins and/or invasion of a named nerve. Thirty-two patients received once-a-day radiotherapy and 2 patients were treated twice a day. One patient was treated with implant alone. Patients treated once a day received from 44.8 Gy to 75 Gy, with a median dose of 60 Gy and a median fraction size of 1.9 Gy. Patients treated twice a day received from 60 Gy to 76.8 Gy, with a median fraction size of 1.2 Gy and a minimum 6-hour interfraction interval. Radiotherapy fields consisted of a single en face field for 15 patients, “wedge-pair” arrangement for 4 patients, a three-field arrangement for 4 patients, an ipsilateral combination of photon- and electron-beam fields for 4 patients, opposed lateral fields for 6 patients, and a combination of techniques for 1 patient. External-beam
Local recurrence Local recurrence 1 nodal Local recurrence 1 distant metastasis Local recurrence 1 nodal 1 distant metastasis Nodal only Nodal 1 distant metastasis Distant metastasis only
6 (17%) 1 (3%) 0 1 (3%) 2 (6%) 1 (3%) 0
irradiation was given with orthovoltage X-rays in 8 patients and megavoltage electron and/or photon beams in the remainder. Two patients received ipsilateral elective neck radiotherapy, and 9 patients received bilateral elective neck treatment. Patients with involved cervical lymph nodes underwent neck treatment with radiotherapy alone (3 patients), surgery and radiotherapy (3 patients), or surgery alone (1 patient). The rates of local control, cause-specific survival, and absolute survival were calculated actuarially according to the Kaplan-Meier product-limit method (7, 8). Significance levels between curves were calculated using the log-rank test. (8, 9). An assessment of the significance of sex, histology (squamous cell carcinoma vs. basal cell carcinoma), age, treatment (radiotherapy alone vs. combined modality therapy), T stage, initial histologic differentiation, and previously untreated vs. locally recurrent tumor was performed using a multivariate analysis. This analysis used a forward step-wise progression of chi-squares for the log-rank test (8, 10). RESULTS Patterns of failure Treatment failure patterns were examined according to treatment approach and site of failure. The treatment approach was analyzed on the basis of treatment with radiotherapy alone or with surgery and adjuvant radiotherapy, for previously untreated or recurrent disease; furthermore, patients given therapy to the primary lesion alone were compared with those who additionally received elective treatment of an at-risk nerve pathway to the base of the skull. Local recurrence was defined as treatment failure at the site of the primary skin lesion, along the course of an at-risk nerve (including the cavernous sinus), or in the central nervous system (defined to include the cerebrospinal fluid, brain stem, and brain). Regional failure was defined as recurrence in a lymph node region of the head and neck. Recurrences elsewhere were defined as distant failures. Patterns of failure are shown in Table 2. Local control The 5-year ultimate local control rate for the small group of patients (n 5 3) treated with radiotherapy alone was not statistically different from that for patients treated with surgery and radiotherapy (100% vs. 77%, p 5 0.4). The 5-year ultimate local control rate for patients treated for
Skin cancer with perineural invasion
● M. W. MCCORD et al.
593
Table 4. Incidence of recurrence according to margin status (32 patients); no. recurred/no. treated Site of recurrence Margin status* Skin Positive Close (,5 mm) Negative Nerve Positive Negative Not reported
recurrent disease was equivalent to that for patients undergoing therapy for previously untreated lesions (79% vs. 79%, p 5 0.9). The group of patients with radiotherapy fields limited to the primary lesion had a statistically better 5-year ultimate local control rate than the group of patients whose treatment included irradiation of the at-risk nerves to the base of the skull (Fig. 1). The failure sites for 8 patients who had local recurrence are shown in Table 3; the predominant site of local recurrence was within the original skin primary lesion. For 1 patient who received irradiation to the nerve pathway, the tumor recurred along the nerve at the field edge in the cavernous sinus. One patient had a recurrence within the brain after irradiation of the nerve. Patterns of failure based on skin margin and nerve margin status are shown in Table 4. Skin margin status was known for all patients. Nerve margin status was known for 17 patients. The 3 patients treated by biopsy alone before definitive radiotherapy were not included in Table 4. Recurrence rates in the primary skin lesion as well as in the at-risk nerves and central nervous system (CNS) are shown in Table 5 according to clinical T stage. Seven of the 8 recurrences occurred within 2 years after treatment (median time, 11 months). Multivariate analysis was undertaken to examine factors that might influence local control. Patient sex, histology, age, therapy (radiotherapy alone vs. surgery and adjuvant radiation), clinical T stage, histologic differentiation, and prior treatment (previously untreated vs. locally recurrent lesions) were examined as predictors of local control. None of these variables significantly influenced the likelihood of local failure (Table 6). Two patients with local recurrence underwent salvage surgery; one remained free of recurrence. Table 3. Site of local recurrence (8 patients)
* Cerebrospinal fluid, brain stem, brain.
Nerve or CNS†
1/11 (9%) 1/3 4/18 (22%)
2/11 (18%) 0/3 0/18
0/3 3/14 (21%) 3/15 (20%)
1/3 0/14 1/15 (7%)
* Does not include 3 patients treated with radiotherapy alone. † Cerebrospinal fluid, brain stem, brain.
Fig. 1. Local control versus extent of irradiation (RT) fields.
Skin only Nerve and cavernous sinus CNS* only
Skin
6 1 1
Regional disease control Twenty-eight patients had no evidence of neck disease at the time of initial treatment; 27 (96%) remained free of recurrence in the neck. For the single patient with recurrence in the neck, salvage therapy consisting of surgery and radiotherapy was unsuccessful. Seven patients had clinically positive necks. Three of the 7 patients (43%) remained free of recurrence in the neck. Three underwent salvage therapy, which was successful in none. Distant metastases Two patients had distant metastases at the time of local and/or regional failure. No patient developed distant metastasis as the only site of failure. Cause-specific and absolute survival The 10-year cause-specific survival rate for the entire study population of 35 patients was 74%. The 10-year cause-specific survival rates for the 18 patients whose radiotherapy was limited to the primary site was 75%, compared with 72% for the 17 patients whose radiotherapy fields extended to the base of skull. The 10-year absolute survival rate for the entire study population of 35 patients was 25%.
Table 5. Local recurrence according to clinical T stage (35 patients); no. recurred/no. treated Site of local recurrence Clinical stage
Skin
Nerve or CNS*
Overall
Tx T1 T2 T3 T4
1/4 2/9 2/9 0/3 1/10
0/4 0/9 1/9 0/3 1/10
1/4 2/9 3/9 0/3 2/10
* Cerebrospinal fluid, brain stem, brain.
594
I. J. Radiation Oncology
●
Biology
●
Physics
Table 6. Multivariate analysis of local control Variable
Significance
Sex Histology Age Treatment T stage Initial histologic differentiation Previously untreated vs. locally recurrent
0.14 0.32 0.41 0.42 0.70 0.84 0.91
Complications Three patients experienced treatment-related complications: soft tissue necrosis, bone exposure, and radiationinduced encephalopathy, which was managed conservatively. Surgical intervention was not required in any of the three patients. DISCUSSION It is difficult to establish the prognostic significance of the presence of microscopic perineural extension for patients with skin cancers of the head and neck. However, review of the outcome of the patients included in this study, as well as the limited literature addressing skin cancer of the head and neck with perineural extension, identifies three distinct patient populations. The natural history and treatment approach for each group is discussed. Minimal microscopic perineural extension This group includes patients who have minimal, focal microscopic perineural extension within an asymptomatic early-stage lesion. Patients with such findings at our institution are treated either with surgery and radiotherapy or radiotherapy alone. The radiotherapy field is designed to include an additional margin of 2 to 3 cm beyond the field that would be treated in the absence of perineural disease. No dose adjustment is made from the guidelines previously described by Mendenhall et al. (11) for treatment of the primary skin cancer. The local control rate of this group is represented by those receiving radiotherapy to the primary alone, with a 10-year local control rate of 93% (see Fig. 1). Perineural extension, even microscopic, is an indication at our institution for postoperative radiotherapy, because salvage rates are poor for patients with documented perineural extension and subsequent proximal relapse along the nerve. Extensive microscopic perineural extension A second group of asymptomatic patients demonstrates more diffuse microscopic perineural extension within the primary skin lesion. These patients have perineural extension to the surgical margin and/or involvement of a named nerve pathway (usually cranial nerves V or VII). Cottel (1) describes the application of Mohs’ chemosurgery in the treatment of patients with perineural extension from primary skin cancers of the head and neck. His article
Volume 43, Number 3, 1999
emphasizes the distance that carcinoma can travel along an involved nerve and the potential clinical underestimation of a lesion’s true pathologic extent. Cottel advocates excision of the involved nerve to the point where there is no evidence of tumor or inflammatory response. Cottel regards an inflammatory response as a potential indication of “skip” lesions. Ampil et al. (12) agree that “whenever an inflammatory response is observed around a nerve branch, the nerve should be traced microscopically until the perineural reaction is no longer observed.” Despite the aggressive surgical approach recommended by Cottel, 9 of 17 patients described in his series received postoperative radiotherapy, which was added for “the most difficult cases.” Di Gregorio et al. (13) express particular concern regarding “invasion of the perineural space along the larger nerves” and recommend wide surgical excision and adjuvant radiotherapy. Ampil et al. (12) suggest that postoperative radiotherapy to the skull base should be strongly considered if there is evidence of perineural invasion of a major nerve trunk. We recommend extension of the radiation portals to include the primary skin lesion and involved nerves with a margin of several centimeters for patients with diffuse distal cutaneous nerve involvement in whom negative surgical margins are obtained. Ampil et al. (12) conclude with similar recommendations for treating the “tumor bed, including treating the histologically demonstrated area of distal perineural extension to a dose of at least 50 Gy given in 25 fractions.” In our series, we observed no recurrences along the nerve pathway among the 14 patients who had histologic evidence of negative nerve margins after surgery and who received subsequent radiotherapy. Patients who have positive margins involving cutaneous nerves undergo postoperative radiotherapy to the primary site with inclusion of the at-risk nerve to the base of the skull, unless the morbidity associated with such extendedfield treatment is prohibitive. If involvement of a named nerve pathway is demonstrated, we recommend radiotherapy to the primary lesion and the involved nerve to the base of skull, regardless of margin status. The clinician should recognize the potential for retrograde involvement of other cranial nerves, as described by Carter et al. (14), in planning the treatment of patients with perineural invasion to major nerve trunks. Magnetic resonance imaging (MRI) has been demonstrated to be of value in defining the extent of gross perineural disease (15). Bourne (16) suggests a minimum tumor dose of at least 50 Gy over 4 weeks to include “as much of the distribution of the nerve as possible” for patients undergoing postoperative irradiation. The expected local control rate in asymptomatic patients with diffuse perineural extension or involvement of a named nerve pathway is best represented by the group receiving radiotherapy fields encompassing the primary lesion and base of skull (see Fig. 1). These patients had a 10-year local control rate of 62%.
Skin cancer with perineural invasion
● M. W. MCCORD et al.
595
graphically or at the time of surgery, also portends a poor prognosis. The treatment approach and outcomes of this patient group are addressed elsewhere. The markedly different outcomes for patients treated at our institution with microscopic, clinical, or gross evidence of perineural extension are shown in Fig. 2. The 35 patients included in this series are those in the “incidental microscopic” curve with a 78% 10-year local control rate, compared with 50% for patients with neurologic symptoms secondary to tumor. CONCLUSION
Fig. 2. Overall local control stratified by extent of perineural involvement.
Clinical and/or gross evidence of perineural extension Patients with clinical symptoms (pain, anesthesia, paresthesia) of perineural extension have a significantly worse prognosis, compared with asymptomatic patients (2, 4, 16). Gross perineural tumor extension, identified either radio-
Patients with incidentally observed, microscopic perineural invasion have a significantly better prognosis than those who are symptomatic or have gross evidence of perineural disease on radiographic studies or at the time of surgery. Although there are no randomized data comparing surgery alone with combined surgery and irradiation for those with incidental microscopic invasion, we recommend irradiation in order to improve the likelihood of local control because of the relatively poor salvage rate and morbidity of salvage therapy in those with recurrence proximally along a major nerve trunk.
REFERENCES 1. Cottel WI. Perineural invasion by squamous-cell carcinoma. J Dermatol Surg Oncol 1982;8:589 – 600. 2. Ballantyne AJ, McCarten AB, Ibanez ML. The extension of cancer of the head and neck through peripheral nerves. Am J Surg 1963;106:651– 667. 3. Ballantyne AJ. Perineural invasion by SCC (letters). J Dermatol Surg Oncol 1984;10:502–504. 4. Goepfert H, Dichtel WJ, Medina JE, Lindberg RD, Luna MD. Perineural invasion in squamous cell skin carcinoma of the head and neck. Am J Surg 1984;148:542–547. 5. Lee WR, Mendenhall WM, Parsons JT, Million RR. Radical radiotherapy for T4 carcinoma of the skin of the head and neck: A multivariate analysis. Head Neck 1993;15:320 –324. 6. American Joint Committee on Cancer. AJCC Cancer Staging Handbook., 5th ed. Philadelphia: Lippincott-Raven 1998. p. 147–151. 7. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457– 481. 8. SAS Institute Inc. SAS Technical Report P-179, Additional SAS/STAT Procedures, Release 6.03 ed. Cary, NC: SAS Institute Inc; 1988. p. 49 – 89. 9. Lawless JF. Statistical models and methods for lifetime data. New York: Wiley; 1982. p. 420 – 422.
10. Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. New York: Wiley; 1980. p. 1–321. 11. Mendenhall WM, Parsons JT, Mendenhall NP, Brant TA, Stringer SP, Cassisi NJ, Million RR. Carcinoma of the skin of the head and neck with perineural invasion. Head Neck 1989; 11:301–308. 12. Ampil FL, Hardin JC, Peskind SP, Stucker FJ. Perineural invasion in skin cancer of the head and neck: A review of nine cases. J Oral Maxillofac Surg 1995;53:34 –38. 13. DiGregorio C, Gebbia V, Florena AM, Franco V, Moschella F. Perineural infiltration by cutaneous squamous cell carcinomas of the head and neck. Anticancer Res 1995;15:1107– 1115. 14. Carter RL, Tanner NS, Clifford P, Shaw HJ. Perineural spread in squamous-cell carcinomas of the head and neck: A clinicopathological study. Clin Otolaryngol 1979;4:271–281. 15. Kolin ES, Castro D, Jabour BA, Lufkin RB, Hanafee WN. Imaging case study of the month. Perineural extension of squamous cell carcinoma. Ann Otol Rhinol Laryngol 1991; 100:1032–1034. 16. Bourne RG. The Costello Memorial Lecture: The spread of squamous cell carcinoma of the skin via the cranial nerves. Australas Radiol 1980;24:106 –114.
PII S0360-3016(98)00474-X
CLINICAL INVESTIGATION
Skin
SKIN CANCER OF THE HEAD AND NECK WITH INCIDENTAL MICROSCOPIC PERINEURAL INVASION MARK W. MCCORD, M.D.,* WILLIAM M. MENDENHALL, M.D.,* JAMES T. PARSONS, M.D.,* FRANKLIN P. FLOWERS, M.D.†
AND
*Department of Radiation Oncology and †Division of Dermatology, University of Florida College of Medicine, Gainesville, Florida Purpose: To address outcomes in clinically asymptomatic patients in whom the unexpected finding of microscopic perineural invasion is noted at the time of surgery. Methods and Materials: The 35 patients included in this study had skin cancers of the head and neck treated with curative intent between January 1965 and April 1995 at the University of Florida. All patients were without clinical or radiographic evidence of perineural invasion but, at the time of biopsy or surgical excision, had the incidental finding of microscopic perineural invasion. Definitive therapy consisted of radiotherapy alone after lesion biopsy (3 patients) or surgical excision preceded (2 patients) or followed (30 patients) by radiotherapy. All patients had follow-up for at least 1 year, 13 patients (37%) had follow-up for at least 5 years. Results: The 5-year local control rate was 78%. The 5-year local control rate for the few patients treated with radiotherapy alone was statistically similar to that for patients treated with surgery and radiotherapy (100% vs. 77%, p 5 0.4). Multivariate analysis for factors affecting local control included sex, histology, age, treatment group, clinical T stage, initial histologic differentiation, and previously untreated vs. recurrent tumors, none of which was found to be significant. Conclusions: Both surgery plus radiotherapy and radiotherapy alone provide a relatively high rate of local control for patients with incidentally discovered perineural invasion secondary to skin cancer. © 1999 Elsevier Science Inc. Head and neck neoplasms, Local neoplasm recurrence, Recurrence, Skin neoplasms, Treatment outcome.
INTRODUCTION
METHODS AND MATERIALS
The limited but growing literature examining the treatment and outcome of skin cancers of the head and neck with perineural invasion reveals two different patient populations. In one group of clinically asymptomatic patients, the unexpected finding of microscopic perineural invasion is noted upon pathologic review of the surgically resected specimen. Such patients composed most of the population in the study by Cottel (1) of the role of Mohs’ chemosurgery in the treatment of perineural invasion by squamous cell carcinoma. Patients in the other group described by Ballantyne et al. (2, 3) and Goepfert et al. (4) have either clinical symptoms or signs of perineural invasion, radiographic evidence of perineural extension, or the finding of gross nerve involvement at the time of surgical resection. This study addresses outcomes in the former patient group, those in whom microscopic perineural extension is noted incidentally at the time of lesion biopsy or surgical excision. Treatment results in the latter patient population, one that has a markedly worse prognosis, were previously reported by Lee et al. (5).
The 35 patients included in this analysis were treated at the University of Florida with curative intent between January 1965 and April 1995. All patients (either at the time of biopsy or at the time of surgery) had the pathologic finding of microscopic perineural extension. Patients were excluded if they had clinical signs or symptoms of perineural involvement or radiographic evidence of perineural extension, or if they were found to have a grossly involved nerve at the time of surgery. Individuals with evidence of perineural extension from a lymph node metastasis were also excluded from this study. The patient population consisted of 30 men and 5 women. All were white. Ages ranged from 32 to 88 years (median, 69 years). No patient was lost to follow-up. All patients had follow-up for at least 1 year, 21 patients (60%) had follow-up for at least 2 years, and 13 patients (37%) had follow-up for at least 5 years. All patients had a primary lesion of the skin, located in the head and neck region (including the lip and nasal vestibule). The primary tumor sites are listed in Table 1.
Address correspondence to: William M. Mendenhall, M.D., Department of Radiation Oncology, University of Florida Health
Science Center, P.O. Box 100385, Gainesville, FL 32610-0385. Accepted for publication 9 October 1998. 591
592
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 43, Number 3, 1999
Table 2. Patterns of relapse (35 patients with 11 recurrences)
Table 1. Tumor site Site
No. patients
Forehead/brow Ear Postauricular Nose Cheek Nasolabial fold Temple Preauricular Scalp Maxilla Nasal vestibule Lip Periorbital skin
1 5 4 2 4 2 6 1 1 1 1 6 1
Thirty-one patients had squamous cell carcinomas, 3 patients had basal cell carcinomas, and 1 patient had a metatypical basal cell carcinoma. Ten patients had previously untreated lesions, and 25 patients underwent therapy for locally recurrent tumors. Information regarding histologic differentiation of the lesions was available for 20 of the 35 patients. Of these 20 patients, 7 had well-differentiated lesions, 7 had moderately differentiated lesions, and 6 had poorly differentiated lesions. Tumors were clinically staged based on the 1998 American Joint Committee on Cancer (AJCC) system (6), in which the presence of microscopic perineural invasion does not alter the clinical stage. Four patients had clinical Tx lesions, 9 had T1 lesions, 9 had T2 lesions, 3 had T3 lesions, and 10 had T4 lesions. Twenty-eight patients had NO neck disease, 4 had N1 disease, and 3 had N2B disease. Therapy consisted of surgery followed by radiotherapy in 30 patients, 3 patients had definitive radiotherapy after biopsy alone, and 2 patients had preoperative radiotherapy followed by surgical excision. Radiotherapy fields were designed to encompass only the primary skin cancer with an adequate margin in 18 patients. In 17 patients, the nerve pathways to the base of the skull were included to a minimum dose of 50 Gy. Patients treated to the primary site alone usually were those with negative margins, whereas those who received irradiation along the course of the nerve to the skull base usually had either positive margins and/or invasion of a named nerve. Thirty-two patients received once-a-day radiotherapy and 2 patients were treated twice a day. One patient was treated with implant alone. Patients treated once a day received from 44.8 Gy to 75 Gy, with a median dose of 60 Gy and a median fraction size of 1.9 Gy. Patients treated twice a day received from 60 Gy to 76.8 Gy, with a median fraction size of 1.2 Gy and a minimum 6-hour interfraction interval. Radiotherapy fields consisted of a single en face field for 15 patients, “wedge-pair” arrangement for 4 patients, a three-field arrangement for 4 patients, an ipsilateral combination of photon- and electron-beam fields for 4 patients, opposed lateral fields for 6 patients, and a combination of techniques for 1 patient. External-beam
Local recurrence Local recurrence 1 nodal Local recurrence 1 distant metastasis Local recurrence 1 nodal 1 distant metastasis Nodal only Nodal 1 distant metastasis Distant metastasis only
6 (17%) 1 (3%) 0 1 (3%) 2 (6%) 1 (3%) 0
irradiation was given with orthovoltage X-rays in 8 patients and megavoltage electron and/or photon beams in the remainder. Two patients received ipsilateral elective neck radiotherapy, and 9 patients received bilateral elective neck treatment. Patients with involved cervical lymph nodes underwent neck treatment with radiotherapy alone (3 patients), surgery and radiotherapy (3 patients), or surgery alone (1 patient). The rates of local control, cause-specific survival, and absolute survival were calculated actuarially according to the Kaplan-Meier product-limit method (7, 8). Significance levels between curves were calculated using the log-rank test. (8, 9). An assessment of the significance of sex, histology (squamous cell carcinoma vs. basal cell carcinoma), age, treatment (radiotherapy alone vs. combined modality therapy), T stage, initial histologic differentiation, and previously untreated vs. locally recurrent tumor was performed using a multivariate analysis. This analysis used a forward step-wise progression of chi-squares for the log-rank test (8, 10). RESULTS Patterns of failure Treatment failure patterns were examined according to treatment approach and site of failure. The treatment approach was analyzed on the basis of treatment with radiotherapy alone or with surgery and adjuvant radiotherapy, for previously untreated or recurrent disease; furthermore, patients given therapy to the primary lesion alone were compared with those who additionally received elective treatment of an at-risk nerve pathway to the base of the skull. Local recurrence was defined as treatment failure at the site of the primary skin lesion, along the course of an at-risk nerve (including the cavernous sinus), or in the central nervous system (defined to include the cerebrospinal fluid, brain stem, and brain). Regional failure was defined as recurrence in a lymph node region of the head and neck. Recurrences elsewhere were defined as distant failures. Patterns of failure are shown in Table 2. Local control The 5-year ultimate local control rate for the small group of patients (n 5 3) treated with radiotherapy alone was not statistically different from that for patients treated with surgery and radiotherapy (100% vs. 77%, p 5 0.4). The 5-year ultimate local control rate for patients treated for
Skin cancer with perineural invasion
● M. W. MCCORD et al.
593
Table 4. Incidence of recurrence according to margin status (32 patients); no. recurred/no. treated Site of recurrence Margin status* Skin Positive Close (,5 mm) Negative Nerve Positive Negative Not reported
recurrent disease was equivalent to that for patients undergoing therapy for previously untreated lesions (79% vs. 79%, p 5 0.9). The group of patients with radiotherapy fields limited to the primary lesion had a statistically better 5-year ultimate local control rate than the group of patients whose treatment included irradiation of the at-risk nerves to the base of the skull (Fig. 1). The failure sites for 8 patients who had local recurrence are shown in Table 3; the predominant site of local recurrence was within the original skin primary lesion. For 1 patient who received irradiation to the nerve pathway, the tumor recurred along the nerve at the field edge in the cavernous sinus. One patient had a recurrence within the brain after irradiation of the nerve. Patterns of failure based on skin margin and nerve margin status are shown in Table 4. Skin margin status was known for all patients. Nerve margin status was known for 17 patients. The 3 patients treated by biopsy alone before definitive radiotherapy were not included in Table 4. Recurrence rates in the primary skin lesion as well as in the at-risk nerves and central nervous system (CNS) are shown in Table 5 according to clinical T stage. Seven of the 8 recurrences occurred within 2 years after treatment (median time, 11 months). Multivariate analysis was undertaken to examine factors that might influence local control. Patient sex, histology, age, therapy (radiotherapy alone vs. surgery and adjuvant radiation), clinical T stage, histologic differentiation, and prior treatment (previously untreated vs. locally recurrent lesions) were examined as predictors of local control. None of these variables significantly influenced the likelihood of local failure (Table 6). Two patients with local recurrence underwent salvage surgery; one remained free of recurrence. Table 3. Site of local recurrence (8 patients)
* Cerebrospinal fluid, brain stem, brain.
Nerve or CNS†
1/11 (9%) 1/3 4/18 (22%)
2/11 (18%) 0/3 0/18
0/3 3/14 (21%) 3/15 (20%)
1/3 0/14 1/15 (7%)
* Does not include 3 patients treated with radiotherapy alone. † Cerebrospinal fluid, brain stem, brain.
Fig. 1. Local control versus extent of irradiation (RT) fields.
Skin only Nerve and cavernous sinus CNS* only
Skin
6 1 1
Regional disease control Twenty-eight patients had no evidence of neck disease at the time of initial treatment; 27 (96%) remained free of recurrence in the neck. For the single patient with recurrence in the neck, salvage therapy consisting of surgery and radiotherapy was unsuccessful. Seven patients had clinically positive necks. Three of the 7 patients (43%) remained free of recurrence in the neck. Three underwent salvage therapy, which was successful in none. Distant metastases Two patients had distant metastases at the time of local and/or regional failure. No patient developed distant metastasis as the only site of failure. Cause-specific and absolute survival The 10-year cause-specific survival rate for the entire study population of 35 patients was 74%. The 10-year cause-specific survival rates for the 18 patients whose radiotherapy was limited to the primary site was 75%, compared with 72% for the 17 patients whose radiotherapy fields extended to the base of skull. The 10-year absolute survival rate for the entire study population of 35 patients was 25%.
Table 5. Local recurrence according to clinical T stage (35 patients); no. recurred/no. treated Site of local recurrence Clinical stage
Skin
Nerve or CNS*
Overall
Tx T1 T2 T3 T4
1/4 2/9 2/9 0/3 1/10
0/4 0/9 1/9 0/3 1/10
1/4 2/9 3/9 0/3 2/10
* Cerebrospinal fluid, brain stem, brain.
594
I. J. Radiation Oncology
●
Biology
●
Physics
Table 6. Multivariate analysis of local control Variable
Significance
Sex Histology Age Treatment T stage Initial histologic differentiation Previously untreated vs. locally recurrent
0.14 0.32 0.41 0.42 0.70 0.84 0.91
Complications Three patients experienced treatment-related complications: soft tissue necrosis, bone exposure, and radiationinduced encephalopathy, which was managed conservatively. Surgical intervention was not required in any of the three patients. DISCUSSION It is difficult to establish the prognostic significance of the presence of microscopic perineural extension for patients with skin cancers of the head and neck. However, review of the outcome of the patients included in this study, as well as the limited literature addressing skin cancer of the head and neck with perineural extension, identifies three distinct patient populations. The natural history and treatment approach for each group is discussed. Minimal microscopic perineural extension This group includes patients who have minimal, focal microscopic perineural extension within an asymptomatic early-stage lesion. Patients with such findings at our institution are treated either with surgery and radiotherapy or radiotherapy alone. The radiotherapy field is designed to include an additional margin of 2 to 3 cm beyond the field that would be treated in the absence of perineural disease. No dose adjustment is made from the guidelines previously described by Mendenhall et al. (11) for treatment of the primary skin cancer. The local control rate of this group is represented by those receiving radiotherapy to the primary alone, with a 10-year local control rate of 93% (see Fig. 1). Perineural extension, even microscopic, is an indication at our institution for postoperative radiotherapy, because salvage rates are poor for patients with documented perineural extension and subsequent proximal relapse along the nerve. Extensive microscopic perineural extension A second group of asymptomatic patients demonstrates more diffuse microscopic perineural extension within the primary skin lesion. These patients have perineural extension to the surgical margin and/or involvement of a named nerve pathway (usually cranial nerves V or VII). Cottel (1) describes the application of Mohs’ chemosurgery in the treatment of patients with perineural extension from primary skin cancers of the head and neck. His article
Volume 43, Number 3, 1999
emphasizes the distance that carcinoma can travel along an involved nerve and the potential clinical underestimation of a lesion’s true pathologic extent. Cottel advocates excision of the involved nerve to the point where there is no evidence of tumor or inflammatory response. Cottel regards an inflammatory response as a potential indication of “skip” lesions. Ampil et al. (12) agree that “whenever an inflammatory response is observed around a nerve branch, the nerve should be traced microscopically until the perineural reaction is no longer observed.” Despite the aggressive surgical approach recommended by Cottel, 9 of 17 patients described in his series received postoperative radiotherapy, which was added for “the most difficult cases.” Di Gregorio et al. (13) express particular concern regarding “invasion of the perineural space along the larger nerves” and recommend wide surgical excision and adjuvant radiotherapy. Ampil et al. (12) suggest that postoperative radiotherapy to the skull base should be strongly considered if there is evidence of perineural invasion of a major nerve trunk. We recommend extension of the radiation portals to include the primary skin lesion and involved nerves with a margin of several centimeters for patients with diffuse distal cutaneous nerve involvement in whom negative surgical margins are obtained. Ampil et al. (12) conclude with similar recommendations for treating the “tumor bed, including treating the histologically demonstrated area of distal perineural extension to a dose of at least 50 Gy given in 25 fractions.” In our series, we observed no recurrences along the nerve pathway among the 14 patients who had histologic evidence of negative nerve margins after surgery and who received subsequent radiotherapy. Patients who have positive margins involving cutaneous nerves undergo postoperative radiotherapy to the primary site with inclusion of the at-risk nerve to the base of the skull, unless the morbidity associated with such extendedfield treatment is prohibitive. If involvement of a named nerve pathway is demonstrated, we recommend radiotherapy to the primary lesion and the involved nerve to the base of skull, regardless of margin status. The clinician should recognize the potential for retrograde involvement of other cranial nerves, as described by Carter et al. (14), in planning the treatment of patients with perineural invasion to major nerve trunks. Magnetic resonance imaging (MRI) has been demonstrated to be of value in defining the extent of gross perineural disease (15). Bourne (16) suggests a minimum tumor dose of at least 50 Gy over 4 weeks to include “as much of the distribution of the nerve as possible” for patients undergoing postoperative irradiation. The expected local control rate in asymptomatic patients with diffuse perineural extension or involvement of a named nerve pathway is best represented by the group receiving radiotherapy fields encompassing the primary lesion and base of skull (see Fig. 1). These patients had a 10-year local control rate of 62%.
Skin cancer with perineural invasion
● M. W. MCCORD et al.
595
graphically or at the time of surgery, also portends a poor prognosis. The treatment approach and outcomes of this patient group are addressed elsewhere. The markedly different outcomes for patients treated at our institution with microscopic, clinical, or gross evidence of perineural extension are shown in Fig. 2. The 35 patients included in this series are those in the “incidental microscopic” curve with a 78% 10-year local control rate, compared with 50% for patients with neurologic symptoms secondary to tumor. CONCLUSION
Fig. 2. Overall local control stratified by extent of perineural involvement.
Clinical and/or gross evidence of perineural extension Patients with clinical symptoms (pain, anesthesia, paresthesia) of perineural extension have a significantly worse prognosis, compared with asymptomatic patients (2, 4, 16). Gross perineural tumor extension, identified either radio-
Patients with incidentally observed, microscopic perineural invasion have a significantly better prognosis than those who are symptomatic or have gross evidence of perineural disease on radiographic studies or at the time of surgery. Although there are no randomized data comparing surgery alone with combined surgery and irradiation for those with incidental microscopic invasion, we recommend irradiation in order to improve the likelihood of local control because of the relatively poor salvage rate and morbidity of salvage therapy in those with recurrence proximally along a major nerve trunk.
REFERENCES 1. Cottel WI. Perineural invasion by squamous-cell carcinoma. J Dermatol Surg Oncol 1982;8:589 – 600. 2. Ballantyne AJ, McCarten AB, Ibanez ML. The extension of cancer of the head and neck through peripheral nerves. Am J Surg 1963;106:651– 667. 3. Ballantyne AJ. Perineural invasion by SCC (letters). J Dermatol Surg Oncol 1984;10:502–504. 4. Goepfert H, Dichtel WJ, Medina JE, Lindberg RD, Luna MD. Perineural invasion in squamous cell skin carcinoma of the head and neck. Am J Surg 1984;148:542–547. 5. Lee WR, Mendenhall WM, Parsons JT, Million RR. Radical radiotherapy for T4 carcinoma of the skin of the head and neck: A multivariate analysis. Head Neck 1993;15:320 –324. 6. American Joint Committee on Cancer. AJCC Cancer Staging Handbook., 5th ed. Philadelphia: Lippincott-Raven 1998. p. 147–151. 7. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457– 481. 8. SAS Institute Inc. SAS Technical Report P-179, Additional SAS/STAT Procedures, Release 6.03 ed. Cary, NC: SAS Institute Inc; 1988. p. 49 – 89. 9. Lawless JF. Statistical models and methods for lifetime data. New York: Wiley; 1982. p. 420 – 422.
10. Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. New York: Wiley; 1980. p. 1–321. 11. Mendenhall WM, Parsons JT, Mendenhall NP, Brant TA, Stringer SP, Cassisi NJ, Million RR. Carcinoma of the skin of the head and neck with perineural invasion. Head Neck 1989; 11:301–308. 12. Ampil FL, Hardin JC, Peskind SP, Stucker FJ. Perineural invasion in skin cancer of the head and neck: A review of nine cases. J Oral Maxillofac Surg 1995;53:34 –38. 13. DiGregorio C, Gebbia V, Florena AM, Franco V, Moschella F. Perineural infiltration by cutaneous squamous cell carcinomas of the head and neck. Anticancer Res 1995;15:1107– 1115. 14. Carter RL, Tanner NS, Clifford P, Shaw HJ. Perineural spread in squamous-cell carcinomas of the head and neck: A clinicopathological study. Clin Otolaryngol 1979;4:271–281. 15. Kolin ES, Castro D, Jabour BA, Lufkin RB, Hanafee WN. Imaging case study of the month. Perineural extension of squamous cell carcinoma. Ann Otol Rhinol Laryngol 1991; 100:1032–1034. 16. Bourne RG. The Costello Memorial Lecture: The spread of squamous cell carcinoma of the skin via the cranial nerves. Australas Radiol 1980;24:106 –114.