Thickness has been related to prognosis of cSCC in several studies,11,12,14–17,21,22,25,69–72 and it has even been identified as the most important predictive factor for metastasis.16 The probability of metastasis is almost nonexistent if the thickness of the tumor is less than 2mm, approximately 6% in tumors of 2 to 6mm, and 16% in those more than 6mm thick.16 Thickness should be measured from the stratum granulosum of normal skin adjacent to the tumor to the last invasive cell.23 This approach may be difficult when measuring thickness of exophytic lesions.

Tumor invasion beyond subcutaneous fat is another prognostic factor that has received attention in recent years.10,17,18 Presence of such invasion has been associated with a 7-fold greater risk of recurrence and an 11-fold greater risk of metastasis.22 Assessment of invasion beyond fatty tissue may, however, be more useful compared with tumor depth in some aspects. On one hand, it is useful in Mohs micrographic surgery, which assesses horizontal cuts in the surgical piece, in which the tumor thickness in millimeters cannot be assessed. In addition, the difficulties associated with assessing tumor depth in exophytic lesions are obviated and, furthermore, this assessment can justify the poor prognosis in some lesions that invade muscle despite having a moderate thickness.

Poor differentiation has been associated with poor prognosis of cSCC in several studies.11,16,17,19–22,25–27,69–79 According to published series, between 8%76 and 25%16 of cSCC are poorly differentiated. Differentiation has been associated with a 3-fold greater risk of local recurrence and a 5-fold greater risk of metastasis,22 as well as earlier recurrence.

PNI is present in 2.5% to 10% of cSCC and is associated with other histopathological features of poor prognosis.14 Studies have consistently recognized its implication in poor prognosis and it has been associated with a 4-fold greater risk of recurrence and a 5-fold greater risk of metastasis.22 More recently, it has been established that PNI should be considered as high risk when nerves with a caliber ≥0.1mm are involved.10,17,18 Moreover, PNI of deep nerves (located below the dermis) and extensive PNI are considered PNI features also associated with poor prognosis.80 It is necessary to differentiate between PNI and perineural dissemination. The former corresponds to histological detection whereas the second, by definition, is a more advanced stage that can be detected clinically and radiologically and is the pathway for invasion of the central nervous system by the tumor.81

Other histopathological risk factors in cSCC are lymphovascular invasion,25 infiltrative growth pattern,82 desmoplasia,16,83,84 defined as least 30% of the desmoplastic stroma associated with the tumor, and budding,85,86 defined by the presence of nests of 4–5 cells at the invasive front, although a clear consensus is lacking on this point.87 In addition, worse prognosis has been associated with certain histopathological variants of cSCC, particularly acantholytic cSCC,88,89 and adenosquamous cSCC,90 although recent series seem to suggest that the prognostic relevance of these features could be limited.91 The presence of an eosinophil-rich infiltrate and plasma cells has also been assessed,26 but more studies are needed to characterize inflammatory infiltrate associated with cSCC and relate the findings more reliably to prognosis. Some markers, such as EGFR,92,93 D2-40,94,95 and epithelial to mesenchymal transition markers96 have been associated with unfavorable prognosis in cSCC. Recently, expression of PD-L197–99 has been linked to a higher risk of lymph node metastases.

The staging systems use a combination of risk factors to define the different risk groups in. For cSCC, the 8th edition of the American Joint Committee on Cancer (AJCC) staging system and the alternative system of the Brigham and Women's Hospital are used. The characteristics of these systems are shown in Table 1.

In practical terms, the presence of one of the factors of poor prognosis is what defines a cSCC as being of high risk (Table 2). However, the prognostic relevance of each factor is not the same and the combination of several is known to confer a greater risk than the presence of each by itself.18,100–102 Thus, a better definition of high-risk cSCC would require a better understanding of the prognostic relevance of each of the risk factors on one hand and their combination on the other. According to the current staging system23 and the alternative BWH system,100 high-risk cSCC would be considered as T3/T4 tumors as per the AJJCC-8 system or T2b/T3 as per the alternative BWH staging system. However, many tumors staged according to these systems are not going to be particularly aggressive.

There is no standard definition of locally advanced cSCC, although all definitions do have points in common. These tumors have been defined as those not potentially curable or that are unlikely to be cured with surgery, radiotherapy, or combination treatment (surgery and radiotherapy) after consultation in a multidisciplinary committee.103 This definition is routed in clinical practice and has also been taken as a reference in some clinical trials in other types of nonmelanoma skin cancer, specifically basal cell carcinoma.104

The criteria for considering cSCC as locally advanced in a trial of the anti-PD1 drug recently approved for cSCC was that patients were not candidates for surgery for the following reasons: recurrence after 2 or more surgical procedures and curative resection was considered unlikely or because surgical treatment would lead to substantial complications or deformity.105 It is more difficult in theory to define which patients with cSCC cannot be considered candidates for conventional treatment. Those cases with extensive bone involvement, in view of the availability of new drugs, could be considered to benefit. Table 3 shows some clinical scenarios that would allow classification of cSCC as locally advanced.

Metastatic cSCC is defined as cSCC that has spread from its primary site in the skin. In most cases, metastasis is to lymph nodes or the parotid gland, and to a lesser extent to other organs. In transit involvement in the skin in cSCC, defined as a presence of foci of tumors differentiated from the primary tumor and before the first draining lymph node,106 could be considered a form of locoregional metastatic cSCC, by analogy with melanoma, although no change to the N category occurs in staging of the cSCC.107

Approximately 5% of cSCC are thought to progress to metastasis,108 although to date no exhaustive registers have been available to derive a true estimate. In 2013, in England, the automated registries included cSCC and metastatic cSCC, thereby enabling study of data at the population level for the first time.109 During 15.5 months of follow-up, 2.1% of the patients with cSCC developed metastatic disease,109 although the authors considered that the rate could be an underestimate given that International Classification of Diseases, 10th version, did not include specific codes for metastatic cSCC. Of these metastatic cSCC, metastasis occurred within 3 years of diagnosis of the primary tumor in 90% of cases.109

In practice, the terms advanced cSCC and aggressive cSCC have been used to refer to more general clinical situations. Thus, the term advanced cSCC has been used to group locally advanced cSCC and metastatic cSCC.105 The term aggressive cSCC, in contrast, has been used to refer to metastatic cSCC and cSCC that leads to death due to local progression.45 The terms advanced and aggressive, however, may be nonspecific and, therefore, when they are used, it is appropriate to define the clinical context to which they refer. The following terms are related to poor prognosis in cSCC: aggressive cSCC, locally advanced cSCC, metastatic cSCC, advanced cSCC, and aggressive cSCC (Table 4).