The size of primary lesions in CSCC has been described by many authors as being an important predictor of lymph node metastasis.12,22–25

Data from the first prospective study of CSCC, involving over 1000 patients, showed that horizontal tumor size was an independent risk factor for metastasis. Specifically, the risk of metastatic spread was 0.01% in lesions measuring 2cm or less in diameter and 10% in larger lesions. In the second group, 7% of patients with a tumor size of between 2 and 5cm developed metastasis compared with 20% of those with a tumor size of over 5cm.26–29

Based on our experience and on data from the literature, we consider a horizontal size of 2cm to be the cutoff for increased risk of lymph node metastasis in CSCC. A smaller size, by contrast, would exert a protective effect, meaning that there would not be a risk of distant metastasis in immunocompetent patients with a tumor diameter of less than 2cm.

Lesion sites with the highest incidence (20%-30%) of metastatic CSCC are the external ear (Fig. 1) and the nonglabrous lip (Fig. 6).19,30

Middle-risk sites include the scalp (mainly the temple), the perineal and genital areas, and acral sites (hands and feet).6,31 It is also important to consider that areas not exposed to sunlight, such as the perineum, the sacral region, and the soles of the feet, have a proportionally higher rate of metastasis than chronically sun-exposed areas.32

Tumor recurrence tends to be associated with poor prognosis in skin cancers.

Comparative analysis of lymph node metastasis in recurrent CSCC (15%) and nonrecurrent CSCC (2%) (P<.001) has led to the conclusion that tumor recurrence is an important risk factor in CSCC.33

Clayman et al.6 observed an association between CSCC recurrence and tumor size, and reported that large tumors were associated with a significantly higher rate of recurrence (2.4vs 1.5cm, P<.0001). They also found recurrent lesions to be associated with a higher rate of perineural invasion (PNI) (24% vs 10%), lymphovascular invasion (17% vs 8%), and subcutaneous tissue invasion (30% vs 10%).6

Recurrence has also been significantly associated with positive margins in surgically excised CSCC, with recurrent tumors—and consequently increased risk of metastasis—observed in up to 50% of patients with positive margins.33

β-Human papillomaviruses (HPVs) are the most common type of HPVs involved in CSCC. Numerous studies have demonstrated a relationship between β-HPVs and CSCC, above all in immunosuppressed patients, although these viruses may also act as a cofactor with UV radiation in immunocompetent patients. Nevertheless, because β-HPVs appear to be involved in the etiology and pathogenesis of CSCC and not in metastatic spread, they are not considered prognostic factors.α-HPVs associated with CSCC of the genital region, the head and the neck, and acral sites might be associated with a higher risk of metastasis as they alter regulatory mechanisms, such as p53 and the retinoblastoma gene/p16. Most studies of p16 in CSCC have reported loss of p16 expression to be associated with the transformation of in situ CSCC to invasive CSCC but not with a higher risk of metastatic spread. As mentioned previously, only p16-positive cases associated with the presence of α-HPVs would carry an increased risk of metastasis.34–39

Degree of tumor differentiation is another important prognostic factor in CSCC and other neoplastic diseases.

In a study of 571 patients with CSCC, a significant difference was observed for the rate of metastasis between lesions with a high degree of differentiation and other lesions (17% vs 4%, P=.004),22 and in another study involving a large number of patients, high-grade CSSS was associated with a greater risk of malignant transformation than other types of CSCC (44% vs 5%, P<.01).12

Tumor differentiation is also associated with an increased risk of early recurrence. Poorly differentiated CSCCs have a 2.9-fold increased risk of distant metastasis and death compared with well-differentiated CSCCs, although well-differentiated tumors may also be associated with the development of advanced disease.12,40

Finally certain histologic subtypes of CSCC (acantholytic, adenoid, isolated cell pattern) should be considered as a risk factor in combination with tumor differentiation (Fig. 3).

Incomplete tumor excision—and consequently—disease persistence, is a predictor of poor prognosis in CSCC. Disease, and with it an increased risk of metastasis, recurs in up to 50% of patients with histologically positive margins following surgical excision.33 Recurrence following tumor excision appears to be related to a risk of subclinical tumor progression, which would, in turn, favor metastasis.41

The decision to take a watch and wait approach with patients with incompletely excised CSCCs, i.e., with a pathology report showing the involvement of 1 or more margins, should be weighed up carefully given the high rate of lymph node disease in recurrent CSCC. Several studies have shown a history of disease recurrence in between 45% and 51% of patients with CSCC and lymph node involvement.12,22,42,43

Consequently, all patients with CSCC should undergo surgery until disease-free margins are achieved,44 and if this is not possible, other treatments, mainly radiation therapy, should be considered.

PNI occurs in approximately 5% to 10% of patients with CSCC and is usually detected as an incidental finding.6,45 Nonetheless, histologic evidence of PNI appears to be associated with a significant increase in disease recurrence and distant metastasis rates.6,45 A study performed at the Anderson clinic in Texas, United States, showed that compared with CSCC patients without PNI, those with PNI had a significantly increased frequency of regional metastasis (35% vs 15%, P<.005) and distant metastasis (15% vs 3.3%, P<.005).46PNI is important not only because of the risk of locoregional spread, but also because of disease caused by perineural spread through the cranial nerves, mostly the facial and the trigeminal nerves (Fig. 4) and because PNI is associated with worse 3-year survival in CSCC (64% in patients with PNI vs 91% in those without, P=.002).47–49

The evaluation of PNI risk in CSCC should vary depending on the thickness of the nerves affected and the presence of clinical and/or radiologic signs of invasion. Infiltration of nerves with a diameter of less than 0.1mm, for instance, is associated with a low risk of local or distant complications, while invasion of nerves measuring more than 0.1mm in diameter has been associated with poor short-term and long-term prognosis (CSCC-specific death of 0% in individuals with PNI of nerves<0.1mm vs 32% in those with PNI of nerves>0.1mm, P=.003).50

PNI can manifest as an incidental finding on histology, with or without accompanying symptoms. Symptoms include pain on palpation, regional paresthesia, and acute intermittent or shooting pain. Based on data from the University of Florida College of Medicine in the United States, it has been suggested that patients with asymptomatic PNI not visible on radiography have a better prognosis than those with clinical or radiological evidence of PNI (5-year local control rate of 87% vs 55%, P=.006).51

Recent studies have suggested that lymphovascular invasion may increase the risk of metastasis in CSCC. Moore et al.43 defined lymphovascular invasion as an independent predictor of lymph node metastasis in a multivariate analysis (OR, 7.54, P.<.00001), and reported that 40% of patients with metastasis had lymphovascular invasion, compared with just 8% of those without. The prognostic significance of lymphovascular invasion, however, has been questioned by some authors.12,50

The implications of CSCC in dermal lymph vessels, which has been rarely described, are unknown, but it may increase the risk of recurrence and explain in-transit metastasis (Fig. 5).

Other factors that have been proposed as possible prognostic factors in CSCC are peritumoral actinic keratosis,37–39 Clark level,37–39 Ki67 expression, desmoplasia,26,38 and the presence of a tumor inflammatory response with mainly eosinophils and plasma cells. The true prognostic value, however, of these factors, is still a matter of debate and needs to be investigated in further studies.

Tumors that overexpress EGFR tend to be associated with more advanced disease, a greater risk of lymph node metastasis, and higher rates of early recurrence and shorter survival in various malignant disorders, including squamous cell carcinoma of the mucosa of the upper aerodigestive tract.57–62

Just 1 small study has analyzed the importance of EFGR mutations in the prognosis of CSCC. In an analysis of 15 cases of metastatic CSCC in the head and neck region, EGFR overexpression was significantly associated with metastatic potential, with strong overexpression found in 79% of patients with metastatic disease and in just 36% of those without.62,63 Overexpression was independent of gene amplification. Alternative mechanisms that would explain the increase in EGFR expression include increased messenger RNA (mRNA) transcription, activating receptor mutations, increased levels of receptor ligands, and increased expression of heterologous receptors, such as Her-2.62 With respect to Her-2, abnormal Her-2 expression and alterations in the gene encoding Her-2 (chromosome 17) have been analyzed in patients with EGFR overexpression.63 While the authors did not observe Her-2 overexpression in any of the 27 cases they analyzed, they did detect Her-2 polysomy, which, like in breast cancer, was not associated with increased overexpression. The absence of overexpression leads to treatment failure with anti-Her2 drugs, such as trastuzumab.64 Although the significance of this last finding has not yet been clarified, the detection of Her-2 polysomy may have an impact on predicting therapeutic response to tyrosine kinase inhibitors.61,62 Nevertheless, EGFR overexpression has only been observed in up 65% to 75% of metastatic CSCCs, adding strength to the hypothesis that multiple factors are involved in the etiology of this form of CSCC.62,63

Data from several studies support a correlation between p16 overexpression and degree of malignancy, suggesting that p16 expression (like p53 expression) might be a biomarker of tumor progression.64,65 Other authors, however, have reported a correlation between loss of p16 expression and high-risk CSCC.66,67 In 1 of these studies, Chang et al.67 detected a correlation between loss of p16 expression and the development of metastasis, suggesting that loss of this protein might be a predictor of poor prognosis.

The explanation behind why elevated levels of p16 might be associated with poor prognosis could be related to the coexistence of HPV infection. However, with the exception of epidermodysplasia verruciformis in immunocompromised individuals and CSCC in areas other than the head and neck, the role of HPV in the development of CSCC is still a matter of debate.

Another possible explanation for increased p16 levels might be p16 UV-induced changes. The presence of a mutated p16, with a long half-life but with a loss of anti-oncogenic function, would explain the increased risk of malignancy.

The CKS1B gene encodes the cyclin-dependent kinases regulatory subunit 1. The protein binds to the catalytic subunit of cyclin-dependent kinases and play an essential role in their biologic function. CKS1B mRNA appears to be expressed in different patterns throughout the cell cycle of HeLa cells, indicating a specific role for the encoded protein. The peptide plays a role in cell-cycle regulation by interacting with other proteins, mainly SKP2 and CDKN1B.68,69

CKS1B appears to have a critical role in tumor progression in CSCC. In a study of 43 CSCCs and 26 actinic keratoses, Salgado et al.70 analyzed CKS1B gene and protein status using fluorescence in situ hybridization and immunohistochemical analysis, and reported that chromosome 1 polysomy was a frequent event in both CSCC (30 of 43 cases) and actinic keratosis (13 of 23 cases). CKS1B amplification, observed in 4 cases (9.3%), was associated in all cases with aggressive tumor behavior (PNI, lymph node spread, and CSCC in transplant recipients).

In conclusion, CKS1B amplifications might be a marker of high-risk CSCC.

CSCC is one of the most common cancers in the world and, as such, responsible for many deaths.

The ability to clearly differentiate between high-risk and low-risk CSCC would appear to be key in improving survival and optimizing management of the disease according to risk.

The American Joint Committee on Cancer (AJCC) recently modified its staging system for CSCC.71 The main changes are summarized in Table 1. One of the most significant changes was the introduction of a list of high-risk clinical and histologic features that modify the T designation, regardless of tumor size. These features include a tumor thickness of more than 2mm, a Clark level of IV or more, location on the external ear or nonglabrous lip, PNI, bone involvement, and poor tumor differentiation.71 Based on our experience, however, this list is not sufficient to accurately define high-risk CSCC. While anatomic location and PNI are good prognostic predictors, the situation with tumor differentiation and Clark level is less clear. Furthermore, the establishment of a cutoff of 2mm for differentiating between low-risk and high-risk CSCC deserves special mention. A high proportion of CSCCs are thicker than 2mm, meaning that this cutoff has high sensitivity but very poor specificity in terms of predicting risk. Finally, the new AJCC staging system does not include important factors such as immune system status, tumor recurrence, or lymphovascular invasion.

In its latest update, published in 2010, the National Comprehensive Cancer Network (NCNN) proposed that the treatment of CSCC should be guided by a series of variables. It lists a set of risk factors and considers 2 possible scenarios. First, it recommends that all CSCCs with any of the risk factors shown in Table 2 should be surgically excised with safety margins of 10mm or with Mohs micrographic surgery. And second, it recommends that patients with 3 or more of these factors should receive special attention. In our opinion, the NCCN guidelines also have shortcomings. A high percentage of patients have at least 1 of the factors shown in Table 2, meaning that suggesting treatment modifications based on this alone would appear to be quite a broad recommendation. The second scenario is even more confusing, as the NCNN considers that patients with 3 or more of the risk factors shown in Table 2 should receive special treatment. However, there is no mention of exactly how treatment or follow-up should be modified.

Additional management strategies are necessary in patients with high-risk CSCC given the high associated risk of lymph node invasion and mortality.

In 2006, Ross et al.74 concluded that sentinel lymph node biopsy (SLNB) was associated with a reliable diagnosis of locoregional invasion and with low morbidity in CSCC provided that the surgeon had sufficient experience. This is also the case with cutaneous melanoma.

Indeed, SLNB provides better results in CSCC than in melanoma as the early detection of lymph node involvement in high-risk CSCC leads to a significant reduction in mortality.

We therefore believe that SLNB is justified in CSCCs defined as high-risk according to our provisional definition (Table 3 and Fig. 6).

High-risk CSCC is associated with a higher risk of recurrence and lymph node metastasis than low-risk CSCC in the first 5 years after treatment. The early detection of recurrence and lymph node metastasis is of supreme importance, as has been shown in multiple studies. Of particular relevance in this respect is a study by Ebrahimi et al.75 that showed that in patients with CSCC and lymph node involvement, a single diseased lymph node measuring 3cm or less in diameter without extracapsular nodal spread had a low risk of distant metastasis and mortality.

We therefore propose a follow-up approach based on risk (Table 4), whereby patients with high-risk CSCC should be followed very closely during the first 5 years after surgery. Close monitoring is particularly important during the first 24 months, which is when the risk of lymph node invasion is highest.