1

Introduction

The definition and pathogenesis of perineural invasion (PNI) in head and neck carcinoma have been evolving for over a century. It has historically been defined as “invasion in, around, and through peripheral nerves .” In 2009, Leibig et al. defined perineural invasion as “finding tumor cells within any of the 3 layers of the nerve sheath or tumor foci outside of the nerve with involvement of ≥ 33% of the nerve’s circumference .” The pathogenesis of PNI was initially thought to be an extension of lymphatic metastases and later thought to occur due to the neural sheaths providing a “path of least resistance” for tumor spread. Recent studies have demonstrated signaling interactions between the tumor and the nerve that may explain the pathogenesis of PNI . PNI is determined to be incidental based on evidence of microscopic invasion of the nerve detected only by histopathology. With incidental PNI, there is no evidence of a cranial neuropathy on physical examination or radiographic evidence of gross tumor involvement along the tract of a nerve .

PNI is a significant prognostic indicator of poor outcome in cutaneous basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) of the head and neck, and is associated with increased rates of local recurrence and metastases . As noted in 2005 by Leibovitch et al., PNI is relatively uncommon and occurs in 2.74% of BCCs and 5.95% of SCCs that were treated with Mohs micrographic surgery (MMS) .

Patients with cutaneous BCC and SCC of the head and neck are typically treated with MMS or wide local excision of the primary lesion. After the diagnosis of incidental PNI is made on histopathology, patients are often treated with postoperative radiotherapy (RT). Due to the 15% to 20% risk of subclinical nodal disease in patients with SCC with PNI, elective treatment of clinically negative regional lymph nodes is recommended .

The current guidelines adapted by Medicare for reimbursement include PNI as one of the indications for treatment with MMS . The main goal of MMS compared to wide local excision in the treatment of high-risk cutaneous cancers is improved margin control . MMS has been shown to lower the number of recurrences compared to wide local excision when used in recurrent BCC . In 2010, Pugliano-Mauro et al. showed MMS to be an effective first step in the multidisciplinary approach to the treatment of high-risk SCC, including those with PNI . The operative cost of MMS may be more than wide local excision , but it becomes more cost-effective in the treatment of cancers where there is a high risk of incomplete excision and recurrence .

The purpose of this study is to update the University of Florida experience using postoperative RT to treat patients with cutaneous SCC and BCC of the head and neck with incidental PNI. We compare outcomes between the patients who received Mohs surgery and those who did not.

2

Materials and methods

The charts of all patients treated with RT for cutaneous SCC and BCC at the University of Florida from June of 1987 to February of 2009 were retrospectively reviewed under an institutional review board-approved protocol. Pathology was reviewed at the time of consultation to determine the presence of PNI. Patients were excluded if they had symptoms of cranial nerve deficits on history, signs of cranial nerve deficits on physical exam, or radiographic evidence of gross perineural invasion. Patients who received prior RT to the same location were excluded. A total of 848 patients treated with curative intent were identified: 36 (4.2%) had incidental PNI detected with MMS and were treated with postoperative RT. These 36 patients were compared to a post-publication analysis of 82 patients treated with postoperative RT at the University of Florida who were found to have incidental PNI after resection other than MMS .

Tumor site distribution of the MMS patients is described in Table 1 . The forehead, temple, and scalp were the most common sites of disease. Patient and tumor characteristics of the MMS patients are described in Table 2 . Most lesions were SCC (89%) and the remaining lesions were BCC (11%). The differentiation was moderate to poor in 44% of lesions; differentiation was not otherwise specified in 39% of lesions. Primary lesions were clinically staged according to the 1997 American Joint Committee on Cancer (AJCC) staging system . All patients who received MMS and postoperative RT were clinically node-negative at the time of RT and no patients were treated with preoperative RT or received adjuvant chemotherapy. Twenty-one of the MMS patients (58%) had previously untreated lesions, 8 (22%) were recurrent for the first time, and 7 (20%) had multiple recurrences. Postoperative skin margins were positive in 6 of the MMS patients (16%), close (< 5 mm) in 2 patients (6%), and unknown in 3 patients (8%). Two of the MMS patients were immunosuppressed. Intensity-modulated RT (IMRT) has been used for patients with head and neck cancers at our institution since 2001 and 18 (50%) patients were treated after January 1, 2001. One MMS patient received an interstitial implant as part of the treatment in addition to external-beam RT. The median follow-up for all of the MMS patients was 4.2 years (range, 0.2 to 21.5 years), and for all living MMS patients it was 6.3 years (range, 2.6 to 21.5 years). The median patient age was 67.5 years (range, 33 to 92) and the male-to-female ratio was approximately 6:1 in the MMS group.

The median RT dose with external-beam RT in the MMS group was 63 Gy (range, 15 to 77.2 Gy). Patients were treated with either once-daily fractionation (86%) or twice-daily fractionation (14%). Radiotherapy techniques included orthovoltage X-rays, megavoltage X-rays, electrons, or a combination of these. Forty-four percent of patients were treated with local fields only (primary site plus a margin); 44% of patients were treated with extended fields (> 2-cm margin); and 12% of patients were treated to the base of skull. Regional nodes were electively treated in 17 (47%) patients, all of whom had SCC.

All statistical calculations were accomplished with SAS and JMP software (SAS Institute, Cary, NC). A local failure was defined as a treatment failure either (1) at the site of the primary skin lesion, (2) along the course of an at-risk nerve (including the skull base), or (3) in the central nervous system (defined to include the cerebrospinal fluid, brainstem, and brain) if the tumor reached the brain by direct extension. A regional failure was defined as a recurrence in a lymph node region of the head and neck (including the parotid area). Distant failures were defined as recurrences elsewhere . Hematogenous brain metastases were considered distant metastases. The Kaplan–Meier product limit method provided estimates of local control, local–regional control, distant metastases, cause-specific survival, and overall survival . The log-rank test provided estimates of statistical significance between strata of selected prognostic factors for each end point. A multivariate analysis was performed using Cox regression for each end point for the following variables: sex, histology, disease status, and T stage .