What is the role of adjuvant radiotherapy in the treatment of cutaneous squamous cell carcinoma with perineural invasion?
Anne Han, Désireé Ratner  Cancer 2007;109:1053

Perineural invasion (PNI) in cutaneous squamous cell carcinoma (SCC) is infrequent, occurring in 2.5% to 14% of patients, but it is important prognostically, because it carries an increased risk of recurrence and metastasis. Although both excision and Mohs micrographic surgery (MMS) are used to treat SCC with PNI, postoperative radiation therapy (XRT) often is recommended to minimize the risk of recurrence. To date, the effectiveness of adjuvant XRT in this setting has not been determined definitively.

METHODS.The authors evaluated the effectiveness of adjuvant XRT in treating SCC with PNI by performing a thorough literature review.

RESULTS.For SCC with PNI, the local control rate after MMS with or without XRT was from 92% to 100% compared with a control rate from 38% to 100% after standard excision with or without XRT. A better prognosis was associated with negative pretreatment magnetic resonance imaging or computed tomography findings than with positive radiographic evidence of PNI.    Primary SCC with PNI was associated with better local control than recurrent SCC with PNI. When treatment outcomes were stratified by PNI type, SCC with microscopic PNI and SCC with extensive PNI had local control rates from 78% to 87% and from 50% to 55%, respectively. Adjuvant XRT was associated in selected patients with 100% local control.

CONCLUSIONS.Few studies addressed the effectiveness of adjuvant XRT in patients who have SCC with PNI. Although XRT has been established as an adjuvant treatment for selected patients, the extent of nerve involvement by tumor, particularly in the setting of other high-risk features, may be helpful in defining its role. In the future, a multicentered, prospective, randomized clinical trial will be needed to assess the true efficacy of adjuvant XRT in the treatment of patients with SCC and PNI.

Perineural invasion (PNI) occurs by contiguous spread along the potential space that exists inside the nerve sheath surrounding a nerve. This space has minimal resistance for extension to occur and may serve as a potential pathway for skin carcinomas of the head and neck region to enter the subarachnoid space of the cranial cavity. Perineural invasion usually occurs in an antegrade (centripetal) fashion, because tumor is more likely to enter the nerve sheath at 1 of the smaller peripheral nerve branches. In some patients, retrograde (centrifugal) extension has been observed and may result when perineural tumor encounters a ganglion and reverses direction along another peripheral branch of the ganglion. Typically, squamous cell carcinoma (SCC) has a higher incidence of PNI than basal cell carcinoma, with reported rates ranging from 2.5% to 14% in SCC compared with 1% in basal cell carcinomas.The presence of PNI in patients with SCC often is associated with aggressive clinical behavior, an increased risk of recurrence and metastasis, and a poorer prognosis. Risk factors for PNI in SCC include male sex, tumor size >2 cm, midfacial tumor location, recurrent tumor, less well differentiated histologic subtypes, and significant subclinical extension.

Patients with SCC who have risk factors for PNI require an organized approach to their evaluation, because early detection of PNI maximizes the chance of effective treatment. Pretreatment evaluation of these high-risk patients consists of a thorough history and physical examination. Detecting PNI requires a high index of suspicion, because clinical findings often are subtle, and the natural history is sometimes long and unpredictable. Some patients with extensive PNI show physical evidence of disease without any findings on imaging studies. These patients should be asked about symptoms, including numbness, tingling, pain, paralysis, or formication (the sensation of bugs crawling on the skin).  During the physical examination, any tumor mass near a major nerve trunk should be noted. A sensory examination should evaluate deficits in the trigeminal nerve distribution. Paralysis of the facial nerve can be tested by using the Schirmer test, taste test, and impedance audiometry. Ptosis, ophthalmoplegia, or visual disturbances may suggest involvement of cranial nerves II, III, IV, or VI. In addition to physician assessment of the signs and symptoms of PNI, preoperative imaging studies serve as important adjunctive tools in the initial evaluation of the high-risk SCC patient, as discussed below.

PNI can be classified into microscopic or extensive invasion. Microscopic PNI is often an incidental, routine finding that may be identified in small peripheral nerves (<1 mm) of the reticular dermis in clinically asymptomatic patients. It may be detected only by histopathology and often is noted in Mohs micrographic surgical (MMS) sections. It cannot be detected on radiographic studies and is associated with minimal morbidity or mortality. It has been estimated that from 60% to 70% of all patients with PNI have microscopic findings,which have also been termed incidental findings in the literature. Because both microscopic and widespread evidence of PNI may be identified at the time of surgery, in this article, we have classified minimal PNI as microscopic rather than incidental. The current literature on PNI uses the term clinical PNI to refer to both signs and symptoms of PNI and radiographic evidence of PNI. In this article, we have expanded the classification by referring to widespread disease that is found clinically, radiologically, or during surgery as extensive PNI.

A patient who exhibits a cranial nerve deficit, gross radiographic evidence of tumor invasion within the tract of a nerve, or evidence of widespread PNI at the time of surgery is considered to have extensive PNI. Often, PNI is difficult to detect by diagnostic imaging and, thus, may present initially at a clinically advanced stage. Radiographic imaging of PNI most often is performed with magnetic resonance imaging (MRI) or computed tomography (CT) scanning. MRI offers the advantage of detecting the extent of macroscopic disease through the presence of nerve enlargement or enhancement or the obliteration of the normal fat plane surrounding a nerve. CT scanning can detect bone erosion that results from tumor invasion of the foramina associated with cranial nerves. A patient may have positive imaging studies without neurologic manifestations of disease. The importance of detecting PNI in SCC as early as possible cannot be underestimated, because appropriately aggressive treatment must be initiated.

The therapeutic effect of radiotherapy (XRT) results from DNA damage inflicted by x-rays, which cause replicating cancer cells to undergo mitotic death. XRT can be low-energy, which treats superficial lesions, or high-energy, which can spare the skin and penetrate to deep-seated lesions more than a few centimeters deep. The latter is particularly useful in tumors that have invaded the nerve with extension toward the cranial cavity. The role of XRT can be definitive, adjunctive, or palliative. Definitive XRT often is given alone as primary treatment with curative intent and may provide excellent cosmetic and functional outcomes, particularly in cosmetically sensitive locations, such as the lip, nasal vestibule, and ear. When a cure is not possible, palliative XRT can help improve both symptoms and survival.

Recommendations for postoperative XRT generally are made on the basis of concerns relating to the morbidity and mortality associated with PNI and concerns regarding the reliability of surgical margins. Adjuvant XRT is used as a secondary treatment modality to decrease the risk of local or regional recurrence and metastasis and to increase disease-free and overall survival. XRT can be effective in treating microscopic deposits of cancer cells that remain after surgical removal of the tumor mass, and it may be used for tumor destruction in places that are difficult to reach by surgical excision, such as the skull base or the cranial cavity. Adjuvant XRT may be recommended in patients who have SCC with involvement of multiple or large nerves, perineural tumor extension through a craniofacial foramen, and in SCC with metastases found at the skull base or within the cranial cavity.

The objective of the current study was to confirm the rationale for adjunctive XRT in patients with SCC who have clinical PNI and to determine the circumstances under which patients with SCC who have microscopic PNI should be considered for adjuvant XRT.The key aspects of the reviewed studies are summarized in Tables 1 through 4. Table  1 compares SCC local control rates after standard surgical excision or MMS with or without adjunctive XRT. Patients who underwent standard excision with or without receiving XRT had local control rates that ranged from 38% to 87%. Patients who underwent MMS with or without XRT achieved superior local control rates of 92% to 100%. Early studies, which consisted almost entirely of patients who were symptomatic or who had imaging-positive disease (ie, clinical PNI) demonstrated local control rates at the lower end of the spectrum, ranging from 38% to 53%,with 1 reported 5-year survival rate of 33%.The results reported by Ampil et al, who performed aggressive surgery with or without XRT for patients with extensive PNI, differed somewhat from the other studies, with a local control rate of 66.7% and 100% 5- and 10-year disease-free survival rates. Later studies, which divided patients into groups with microscopic or extensive PNI, demonstrated that patients who had microscopic PNI had local control rates of 78% to 87% after treatment compared with patients who had extensive PNI (local control rates, 50-55%)

Table 3 summarizes local control rates of patients who had SCC with PNI based on their imaging results. Patients who had positive imaging of extensive PNI had 5-year local control rates of 25% to 57% after treatment with XRT with or without wide local excision or MMS compared with 5-year local control rates of 76% to 78% in patients who had negative imaging results. The 5-year cause-specific and absolute survival rates ranged from 50% to 61% in imaging-positive patients and from 86% to 100% in imaging-negative patients. Patients who had macroscopic or central disease noted on imaging studies had the poorest local control rate of 25%. Nevertheless, the 5-year cause-specific and absolute survival rates for these patients (61% and 58%, respectively) were roughly comparable to and even slightly better than those for patients with minimal to moderate imaging-positive disease, who had cause-specific and absolute survival rates of 50% and 58%, respectively.

Table 4 summarizes the local control rates of patients who had primary and recurrent SCC with PNI. Patients with primary SCC and PNI had better overall outcomes, exhibiting local control rates from 50% to 75% regardless of whether they underwent surgery with XRT or received XRT alone. In contrast, patients with recurrent SCC with PNI exhibited control rates from 11% to 50%, again irrespective of their treatment regimen. When the extent of PNI within primary and recurrent SCCs was stratified into microscopic versus extensive disease, the local control rate for patients with microscopic PNI was 79%, and the local control rate of patients with extensive PNI was from 45% to 47% regardless of whether the disease was primary or recurrent.

The use of standard excision alone may place SCC patients with PNI at high risk for local recurrence because of incomplete tumor removal. XRT often is recommended thereafter as postoperative adjuvant therapy. There are no studies that stratify their data based on XRT alone versus XRT plus standard excision/MMS. Although most studies reviewed here included between 9 and 135 patients, only 1 report was a meta-analysis of 70 studies.  The disparate methodologies of these cited articles render their results and conclusions difficult to validate or compare. Because of the disproportionate power of a meta-analysis, the findings from the study by Rowe et al (ie, 53% local control with standard excision with or without XRT and 100% local control with MMS with or without XRT) are more precise and, thus, carry more weight compared with the range of local control rates described above.

The difference between the local control rates achieved with MMS versus standard excision with or without XRT for SCC with PNI is striking, with reported local control rates from 92% to 100% for MMS compared with control rates from 38% to 87% for standard excision. Because MMS examines the entire peripheral and deep margin of the surgical specimen, it provides the greatest likelihood of complete tumor excision. It is important to note, however, that, if a nerve is surrounded only partially by tumor, then a twist or turn in the specimen during tissue sectioning may result in a false-negative margin or a skip area, leaving undetected residual tumor cells behind. An inflammatory response around a nerve branch may be a clue for the Mohs surgeon to continue tracing out the nerve either until microscopic tumor cells are found or until inflammation no longer is observed. Even if inflammation is not detected, skip lesions have been found on subsequent stages to occur at a distance of up to 14 cm beyond the initial tumor site.For this reason, some Mohs surgeons prefer to treat their patients with adjuvant XRT when there is extensive or microscopic evidence of PNI, whereas others take an additional safety margin for frozen or permanent section evaluation even after a histologically clear margin has been obtained.

The National Comprehensive Cancer Network (NCCN) released a set of 2006 practice guidelines noting that the role of XRT for nonmelanoma skin cancers probably was the single largest source of disagreement in their panel of experts. Despite the lack of consensus regarding a strategy for using adjuvant XRT to improve morbidity and mortality, the NCCN guidelines do specify that, when properly applied, XRT can achieve very good rates of cure and excellent cosmesis. Criteria listed in the NCCN guidelines for the consideration of XRT include the size and location of the tumor (tumors in high-risk locations, ie, mask areas of the face - central face, eyelids, eyebrows, periorbital, nose, lips, etc - up to 15 mm in greatest dimension; and tumors in middle-risk locations, ie, cheeks, forehead, scalp, and neck, up to 20 mm in greatest dimension). Furthermore, the guidelines mention that postoperative XRT should be considered when there is evidence of substantial PNI (ie, involvement of more than just a few small sensory nerve branches).

There is no question that patients with extensive PNI require aggressive therapy to minimize the risk of tumor recurrence and metastasis, given the significant differences in local control, cause-specific survival, and absolute survival rates reported in the literature between imaging-positive and imaging-negative patients. Adjuvant XRT is an important treatment option for patients who have PNI that extends close to or into the cranial cavity, because it can target areas where surgical manipulation is not feasible. Tumor that extends along a cranial nerve to the skull base often is detected at the time of surgery and may be treated thereafter by XRT. Tumor that extends past the cranial foramen to seed the brain may be seen more often on pretreatment radiographs; however, it is noteworthy that such tumors no longer are amenable to surgery, and that the chance of curing such patients is unlikely.

Even in patients who have very advanced SCC with PNI, it has been demonstrated that aggressive surgery with or without adjuvant XRT is effective in controlling the extent of tumor involvement and in maximizing disease-free survival. All 9 of the patients reported in the study by Ampil et al achieved local control for advanced SCC with PNI after aggressive treatment with surgery with or without XRT. Although 3 of 9 patients in that study had recurrent tumors after initial treatment, what was different in this study was that 100% of these patients received aggressive treatment with successful salvage. Although the authors cited survival rates of 33% at 5 years and 22% at 10 years, it is important to note that no patient died of recurrent cancer and that no patient who died had any residual skin cancer, for an overall disease-free survival rate of 100%.

The study by Ampil et al also was unique, because it stratified treatment outcomes based on the use of adjuvant XRT. The decision to treat with XRT was related to the size of the tumor (1 patient had a tumor that measured 12.5 cm, the largest in the series) and to evidence of tumor found at the deep margin of resection close to the skull base (2 patients had tumor found in the infraorbital area of the skull). All 3 of the patients in that study who received adjuvant XRT had 100% local control with no evidence of disease at follow-up between 18 months and 201 months. All 6 patients who did not receive adjuvant XRT in that study also had 100% local control, suggesting that aggressive surgical treatment of advanced SCC patients, with or without adjuvant XRT, may achieve long-term disease control as long as the tumor is recognized and treated adequately prior to its entry into the cranial cavity.

The role of adjuvant XRT for microscopic PNI is not defined as clearly in the literature as it is for extensive PNI, and it has not yet been determined definitively whether or not treatment of microscopic PNI must be equally aggressive. It is important to separate these 2 groups, because patients who have microscopic PNI have a better prognosis than patients who have extensive PNI, with 78% to 87% local control rates compared with 50% to 55% local control rates after aggressive treatment with XRT with or without surgery.  Despite better outcomes, microscopic PNI has been linked with subsequent proximal recurrence along the nerve and poor salvage rates in some patients. For this reason, McCord et al recommend treating all patients who have microscopic PNI with XRT, and their approach has yielded a 77% overall local control rate.  Similarly, Garcia-Serra et al used adjuvant XRT to treat 93% of SCC lesions with microscopic PNI and achieved an overall local control rate of 87% at 5 years.  These results suggest that adjunctive XRT can be effective in improving local control rates in patients with microscopic PNI.

If microscopic PNI is found in or near a major nerve branch, then the risk of disease progression is significant. Because PNI most frequently involves the trigeminal and facial nerves, any tumor found in the distribution of these branches should be considered for adjuvant radiation.  Microscopic PNI can have minimal focal findings or consist of more pervasive, diffuse spread within the tumor mass. If the Mohs surgeon or pathologist finds widespread rather than focal microscopic PNI in the excised tumor, then adjuvant XRT likely is indicated. Recurrent SCCs have a 6.9% incidence of PNI compared with a 4.7% incidence in primary SCCs and are associated with a poorer prognosis. Recurrent tumors, therefore, are more likely than primary tumors to benefit from adjuvant therapy. Finally, other high-risk features, such as size >2 cm, tumor depth, poorly differentiated histologic subtype, advanced patient age, and male sex, constitute important considerations for additional therapy with radiation in the setting of PNI.

It is important to recognize that not all patients with PNI should receive adjuvant XRT. Some cannot tolerate the side effects, which may include redness, irritation and dryness of the mouth and skin, sore throat, loss of taste, trouble swallowing, nausea, earache, hair loss, and jaw stiffness. Patients with recurrent lesions close to a site of prior irradiation cannot be treated with XRT because of the risk of necrosis.  XRT also has been implicated in orbital and central nervous system damage as well as bone exposure, fistula formation, and wound infection.  Patients with connective tissue diseases and those with genetic conditions that predispose them to skin cancer, such as xeroderma pigmentosum, should not receive radiation treatment because of their increased sensitivity to the effects of radiation.  The risk of developing XRT-induced cutaneous malignancy also may need to be taken into consideration, particularly in younger patients. Because patients who receive adjuvant XRT must adhere to a schedule of visits, noncompliant patients may be poor candidates for this treatment modality. Finally, the lack of radiation oncologists or facilities in a particular geographic region may pose a barrier to providing comprehensive care for certain high-risk patients.

In conclusion, few studies have addressed the effectiveness of adjuvant XRT in SCC with PNI. Although adjuvant XRT is established as a postoperative treatment modality for patients with clinical PNI, the use of this modality in the setting of microscopic PNI is not as clear cut. Features associated with aggressive SCC, such as tumor location, recurrent SCC, and the extent of peripheral nerve involvement by tumor, may be helpful in developing guidelines for the use of adjuvant XRT in patients with microscopic PNI. It also is important to recognize that long-term cure with surgery, with or without XRT, may be possible for selected patients with aggressive SCC and PNI as long as the tumor is recognized and treated adequately prior to entry into the cranial cavity. These findings give hope to those patients who do not choose or are unable to tolerate adjunctive XRT and who otherwise may be considered to have a dismal prognosis. In the future, a well-designed, multicenter, prospective, randomized clinical trial will be needed to assess definitively the efficacy of adjuvant XRT in treating patients who have SCC with PNI.