Association of Patient Risk Factors, Tumor Characteristics, and Treatment Modality With Poor Outcomes in Primary Cutaneous Squamous Cell Carcinoma

Key Points

Questions

Are patient risk factors and tumor characteristics associated with poor treatment outcomes in patients with primary cutaneous squamous cell carcinoma (cSCC), and which treatment modalities minimize poor outcomes?

Findings

In this systematic review and meta-analysis of 129 studies with more than 125 000 patients with cSCC, several patient risk factors, tumor characteristics, and treatment modalities were associated with poor outcomes. The highest risks for local recurrence and disease-specific death were associated with tumor invasion beyond subcutaneous fat, and the highest risk of metastasis was associated with perineural invasion.

Meaning

The findings of this meta-analysis demonstrate the prognostic value of different risk factors and effectiveness of treatment modalities; as such, these findings can guide prognostication, workup, and treatment of cSCC.

Abstract

Importance

Primary cutaneous squamous cell carcinoma is usually curable; however, a subset of patients develops poor outcomes, including local recurrence, nodal metastasis, distant metastasis, and disease-specific death.

Objectives

To evaluate all evidence-based reports of patient risk factors and tumor characteristics associated with poor outcomes in primary cutaneous squamous cell carcinoma and to identify treatment modalities that minimize poor outcomes.

Data Sources

PubMed, Embase, and SCOPUS databases were searched for studies of the topic in humans, published in the English language, from database inception through February 8, 2022.

Study Selection

Two authors independently screened the identified articles and included those that were original research with a sample size of 10 patients or more and that assessed risk factors and/or treatment modalities associated with poor outcomes among patients with primary cutaneous squamous cell carcinoma.

Data Extraction and Synthesis

Data extraction was performed by a single author, per international guidelines. The search terms, study objectives, and protocol methods were defined before study initiation. A total of 310 studies were included for full-text assessment. Owing to heterogeneity of the included studies, a random-effects model was used. Data analyses were performed from May 25 to September 15, 2022.

Main Outcomes and Measures

For studies of risk factors, risk ratios and incidence proportions; and for treatment studies, incidence proportions.

Results

In all, 129 studies and a total of 137 449 patients with primary cutaneous squamous cell carcinoma and 126 553 tumors were included in the meta-analysis. Several patient risk factors and tumor characteristics were associated with local recurrence, nodal metastasis, distant metastasis, disease-specific death, and all-cause death were identified. Among all factors reported by more than 1 study, the highest risks for local recurrence and disease-specific death were associated with tumor invasion beyond subcutaneous fat (risk ratio, 9.1 [95% CI, 2.8-29.2] and 10.4 [95% CI, 3.0- 36.3], respectively), and the highest risk of any metastasis was associated with perineural invasion (risk ratio, 5.0; 95% CI, 2.3-11.1). Patients who received Mohs micrographic surgery had the lowest incidence of nearly all poor outcomes; however, in some results, the 95% CIs overlapped with those of other treatment modalities.

Conclusions and Relevance

This meta-analysis identified the prognostic value of several risk factors and the effectiveness of the available treatment modalities. These findings carry important implications for the prognostication, workup, treatment, and follow-up of patients with primary cutaneous squamous cell carcinoma.

Trial Registration

PROSPERO Identifier: CRD42022311250

This systematic review and meta-analysis of 129 studies evaluates the predictors and treatment outcomes for primary cutaneous squamous cell carcinoma among a total of 137 449 patients.

Introduction

Cutaneous squamous cell carcinoma (cSCC) is the second most common malignant neoplasm in the US, with an estimated annual incidence of 1.1 million to 1.8 million per year.^1,2,3,4 Although cSCC usually carries an excellent prognosis, a subset of patients develops poor outcomes, including local recurrence (LR), nodal metastasis (NM), distant metastasis (DM), and disease-specific death (DSD). We performed a systematic review and meta-analysis of all electronically available studies of cSCC and reviewed the associations of risk factors and treatments with poor outcomes to identify those with higher risk.

Methods

This systematic review and meta-analysis study was exempt from review, and informed consent was not required because the study used only previously published publicly available reports. The study followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guidelines. The protocol was registered with PROSPERO, and a few amendments were made to the analysis methodology.

Data Sources and Strategy

An electronic literature search was performed on February 8, 2022, using PubMed, Embase, and SCOPUS databases. Articles published from database inception through February 8, 2022, were included in the search. Boolean search terms were cutaneous [AND] squamous cell carcinoma [AND] treatment type (various) [OR] risk factors [AND] recurrence [OR] metastasis [OR] mortality [OR] disease-specific death.

Eligibility Criteria

Study selection was performed using the following inclusion criteria: (1) study type was a randomized clinical trial, case-series report, case-control report, or cohort study; (2) the study results included at least 1 histopathologic finding, tumor morphologic feature, or comorbid medical risk factor for cSCC or the use of a single curative-intent treatment modality; and (3) the study’s outcome of interest was either an LR, NM, DM, DSD, and/or all-cause death or any combination of these outcomes. A study was excluded if it (1) was not published in English; (2) included nonhuman species; (3) was a review article, abstract, errata, poster presentation, conference proceeding, or textbook chapter; (4) included pediatric patients; (5) had a sample size smaller than 10 patients with cSCC; (6) had data on cSCC that could not be extracted from other data; (7) had cSCC data that could not be extracted SCC in situ, noncutaneous SCC, or anogenital SCC data; (8) included patients with genetic disorders predisposing to cSCC; (9) had an outcome of interest present in all patients at study initiation; (10) had no assessment of risk factors or single curative-intent treatment; and/or (11) had no assessment of the outcome of interest.

Study Selection and Data Extraction

Titles and abstracts were assessed on these criteria by 2 independent blinded reviewers (G.A.Z., M.L.S.). If the title or abstract did not include enough information to apply exclusion criteria, the full text was assessed. Reviewers compared results, and discrepancies were resolved by third author arbitration (J.A.C.). The full text of remaining articles was reviewed in detail to determine eligibility based on the inclusion criteria. Data abstraction was performed by an author (G.A.Z.). In cases of study duplicity, the more recent and complete studies were selected for inclusion. Risk of bias was evaluated for each study by one of the authors (A.N.P.) using the Newcastle−Ottawa Scale for non-randomized studies⁵ and the Cochrane Collaboration tool⁶ for randomized studies.

Statistical Analysis

Meta-analysis of risk factor data was performed through 2 strategies based on the format of the reported data—random effects analysis of risk ratio (RR) data (when it was possible to extract or calculate hazard or odds ratios) and random effects analysis of incidence data (when it was possible to calculate incidence proportions). Random effects analysis was performed using Comprehensive Meta-Analysis, version 3, Biostat). Whenever possible, the RRs and 95% CIs were reported for each risk factor per outcome measure. The I² statistic was used to quantify heterogeneity; a value exceeding 50% was considered to be substantial heterogeneity. Forest plots were created for all associations of risk factors and outcomes using Excel, version 16.61.1 (Microsoft). Treatment data were analyzed by random effects analysis for incidence proportions for each treatment modality per outcome.

Results

The initial search identified 5279 articles. After removal of duplicates, the titles and abstracts of 3140 articles were reviewed. After inclusion criteria were applied, the full text of 310 articles were assessed. Of these, 129 studies^{7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135} were included in final analysis (Figure 1). Study characteristics, including risk of bias, are described in eTable 1 in the Supplement. Of these 129 studies, 114 studies^{7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121} assessed risk factors, with a total of 137 449 patients and 126 553 tumors. Of these, 87 studies^{8,10,11,13,14,15,16,17,18,19,21,23,25,27,30,31,33,34,35,37,38,39,40,41,43,44,45,47,48,50,52,53,55,56,57,58,59,60,61,62,63,64,66,69,70,71,72,73,74,75,77,78,79,80,81,83,84,85,86,87,88,90,91,92,93,94,96,97,98,100,101,102,104,106,107,108,109,110,111,112,113,114,115,118,119,120,121} with 122 484 tumors were included in the random effects analysis risk ratio, and 62 studies^{7,8,10,12,13,15,17,18,19,20,22,23,25,26,27,28,29,30,31,32,36,37,38,40,42,46,47,48,50,53,54,55,56,59,63,65,67,68,70,71,73,74,76,79,80,81,82,83,84,85,89,90,92,93,94,95,98,99,101,103,114,116} with 116 958 tumors were included in analysis of incidence data. Treatment modality was assessed by 28 studies^{8,10,27,34,35,40,41,53,59,80,90,96,97,101,122,123,124,125,126,127,128,129,130,131,132,133,134,135} with 10 967 patients and 6118 cSCC tumors.

Figure 1. Search Methodology and Results of Literature Search on Risk Factors Associated With Poor Outcome in Primary Cutaneous Squamous Cell Carcinoma (cSCC).

For the analysis of RR, analyzed histopathologic and clinical risk factors and associations with LR, NM, and DSD are outlined in Tables 1 and 2, respectively. Associations for histopathologic and clinical risk factors and the outcomes described by more than 1 study are outlined in Figures 2 and 3, respectively. All associations with DM, any metastasis (AM), or combined outcomes are outlined in eTable 2 in the Supplement, and those described in more than 1 study are shown in the eFigure in the Supplement.

Table 1. Results of Random Effects Analysis of Risk Ratio for Each Histopathologic Risk Factor for Local Recurrence, Nodal Metastasis, and Disease-Specific Death.

Risk factor	Risk ratio (95% CI)	P value	Studies, No.	I2 statisic, %
Local recurrence
Invasion beyond subcutaneous fat	9.1 (2.8-29.2)	<.001	5	0
Desmoplasia	8.1 (4.1-15.9)	<.001	3	17
Tumor thickness ≥6 mm	4.6 (2.1-9.9)	<.001	3	1
Perineural invasion	4.1 (2.7-6.3)	<.001	12	0
Lymphovascular invasion	2.8 (1.5-5.1)	<.001	6	0
Poor differentiation	2.4 (1.2-4.8)	.02	12	0
Tumor thickness ≥8 mm	2.2 (0.6-7.8)	.23	1	0
Tumor thickness ≥4 mm	2.2 (1.0-4.6)	.05	2	0
Tumor size ≥2 cm	2.0 (1.2-3.4)	.01	10	0
Tumor size ≥5 cm	1.6 (0.4-7.5)	.52	1	0
Tumor thickness ≥2 mm	1.1 (0.5-2.2)	.85	2	0
Nodal metastasis
Tumor thickness ≥5 mm	18.7 (2.5-139)	.004	1	0
Tumor thickness ≥4 mm	7.2 (2.1-24.3)	.002	1	0
Perineural/vascular involvement	6.9 (4.8-9.9)	<.001	3	12
Cytokeratin 19 expression	5.0 (0.7-35.8)	.11	1	0
High PD-L1 expression	3.6 (1.9-7.0)	.001	2	0
LLT-1 expression	3.4 (1.4-8.3)	.007	1	0
Tumor budding	2.9 (2.1-4.2)	<.001	3	50
Desmoplasia	2.8 (2.0-4.0)	<.001	2	0
Perineural invasion	2.7 (2.2-3.2)	<.001	19	0
Poor differentiation	2.6 (2.2-2.9)	<.001	18	29
Membrane E-cadherin	2.2 (1.7-3.0)	<.001	1	0
EGFR+	2.2 (1.4-3.7)	.002	2	0
Lymphovascular invasion	2.1 (1.7-2.6)	<.001	11	0
Cortactin	2.0 (1.1-3.9)	.03	1	0
Tumor size ≥2 cm	2.0 (1.7-2.4)	<.001	13	0
Invasion beyond subcutaneous fat	2.0 (1.4-2.7)	<.001	8	0
Tumor thickness ≥6 mm	1.9 (1.4-2.6)	<.001	5	28
Tumor thickness ≥8 mm	1.8 (1.0-3.1)	.05	2	0
Cytoplasm E-Cadherin	1.3 (0.7-2.3)	.39	1	0
ERBB3+	1.3 (0.7-2.4)	.44	1	0
Tumor size ≥4 cm	1.2 (0.5-2.9)	.71	2	0
Tumor thickness ≥2 mm	1.2 (0.3-4.7)	.82	2	0
Tumor thickness ≥3 mm	1.2 (0.1-9.9)	.88	1	0
FAK expression	1.0 (0.6-1.9)	.95	1	0
Tumor size ≥5 cm	1.0 (0.6-1.9)	.95	1	0
ERBB4+	0.8 (0.4-1.6)	.58	1	0
Podoplanin	0.8 (0.5-1.3)	.38	1	0
Disease-specific death
Invasion beyond subcutaneous fat	10.4 (3.0-36.3)	<.001	5	0
Tumor size ≥5 cm	8.9 (4.7-17.0)	<.001	1	0
Lymphovascular invasion	8.8 (3.9-19.7)	<.001	3	0
Desmoplasia	7.6 (4.9-11.8)	<.001	2	1
Perineural invasion	7.5 (4.5-12.3)	<.001	7	2
Tumor thickness ≥6 mm	6.5 (3.4-12.3)	<.001	1	0
LLT-1 expression	6.2 (1.8-21.3)	.004	1	0
Poor differentiation	5.9 (1.6-21.0)	.007	8	6
Tumor size ≥2 cm	2.9 (0.8-11.1)	.12	6	0
Podoplanin	2.8 (1.0-8.0)	.05	1	0
FAK expression	1.6 (0.8-3.2)	.23	1	0
Cortactin	1.0 (0.5-1.9)	.98	1	0

Abbreviations: EGFR, epidermal growth factor receptor; FAK, focal adhesion kinase; LLT-1, lectin-like transcript 1 expression; PD-L1, programmed death ligand-1.

Table 2. Results of Random Effects Analysis: Risk Ratio for Each Clinical Risk Factor for Local Recurrence, Nodal Metastasis, and Disease-Specific Death.

Risk factor	Risk ratio (95% CI)	P value	Studies, No.	I2statistic, %
Local recurrence
HIV	7.0 (2.1-23.1)	.001	2	0
Organ transplantation	4.0 (2.2-7.0)	<.001	4	5
Chronic lymphocytic leukemia	3.8 (1.5-9.8)	.005	1	0
Age >65 y	2.3 (0.5-9.9)	.26	1	0
Tumor on ear or lip	2.0 (0.6-6.8)	.26	1	0
Tumor on head or neck	1.9 (0.9-3.9)	.09	5	0
Ulceration	1.7 (1.2-2.4)	.001	4	0
Tumor on scalp	1.7 (0.8-3.6)	.21	3	0
Age >70 y	1.6 (0.5-5.0)	.46	2	0
Tumor on ear	1.5 (0.8-3.1)	.23	6	0
Immunosuppression	1.5 (1.1-2.2)	.03	14	0
Growth rate ≥4 mm/mo	1.4 (0.5-4.1)	.50	1	0
Tumor on lip	1.4 (0.6-3.6)	.45	4	0
Male sex	1.2 (0.9-1.6)	.13	10	0
Age >60 y	1.0 (0.4-2.9)	.96	1	0
Age >75 y	0.9 (0.5-1.5)	.91	3	9
Nodal metastasis
Organ transplant	38.6 (21.2-70.3)	<.001	2	0
Chronic lymphocytic leukemia	4.5 (2.1-9.5)	<.001	1	0
Hypothyroidism	4.0 (2.8-5.7)	<.001	1	0
Age >65 y	3.9 (1.7-8.9)	.001	1	0
Growth rate ≥4 mm/mo	3.2 (1.1-9.7)	.04	1	0
Tumor on ear or lip	2.9 (1.4-5.9)	.003	1	0
HIV	2.8 (0.8-10.3)	.11	1	0
Age >70 y	2.7 (1.4-5.0)	.003	1	0
Immunosuppression	2.6 (2.2-3.0)	<.001	8	17
Tumor on scalp	2.5 (1.3-4.6)	.004	4	0
Tumor on lip	2.3 (1.4-3.9)	.002	6	5
Hematologic malignant tumor	2.1 (0.4-11.3)	.38	1	0
Diabetes	1.8 (0.3-12.1)	.53	1	0
Tumor on ear	1.8 (1.4-2.2)	<.001	9	12
Male sex	1.5 (1.1-1.9)	.002	8	3
Tumor on head or neck	1.3 (0.9-2.0)	.14	4	0
Age >75 y	1.3 (0.3-5.9)	.76	2	0
Smoking history	1.1 (0.7-1.7)	.75	1	0
Disease-specific death
Tumor on head or neck	4.4 (1.6-12.2)	.005	1	0
Organ transplant	4.1 (0.7-24.2)	.12	1	0
Immunosuppression	2.7 (1.6-4.5)	<.001	9	36
Ulceration	2.2 (1.2-4.1)	.01	1	0
Male sex	1.8 (0.4-8.6)	.45	3	0
Tumor on ear	1.4 (0.5-3.9)	.58	2	8
Tumor on lip	1.0 (0.7-1.3)	.94	2	0
Tumor on ear or lip	0.9 (0.2-3.8)	.90	1	0
Age >65 y	0.5 (0.2-1.1)	.10	1	0

Figure 2. Forest Plots of Results of Random Effects Analysis for Pairs of Histopathologic Risk Factors and Local Recurrence, Nodal Metastasis, and Disease-specific Death Identified in More Than 1 Study.

Circles denote the composite risk ratio; whiskers, the 95% CIs. EGFR indicates epidermal growth factor receptor, and PD-L1, programmed death ligand-1.

Figure 3. Forest Plots of Results of Random Effects Analysis for Pairs of Clinical Risk Factors and Local Recurrence, Nodal Metastasis, and Disease-specific Death Identified in More Than 1 Study.

Circles denote the composite risk ratio; whiskers, the 95% CIs. For nodal metastasis organ transplant (panel B), the risk ratio is approximately 39.

Several notable tumor factors were associated with poor outcomes. Tumors on the lip, ear, and scalp were associated with a significantly increased risk of NM (lip RR, 2.3 [95% CI, 1.4-3.9]; ear RR, 1.8 [95% CI, 1.4-2.2]; scalp RR, 2.5 [95% CI, 1.3-4.6]) and AM (lip RR, 1.8 [95% CI, 1.4-2.4]; ear RR, 2.8 [95% CI, 1.2-3.2]). Tumors with a diameter of 2 cm or more had a significantly increased risk of LR (RR, 2.0; 95% CI, 1.2-3.4), NM (RR, 2.0; 95% CI, 1.7-2.4), AM (RR, 2.9; 95% CI, 1.6-5.1), and all-cause death (RR, 1.8; 95% CI, 1.3-2.6). Tumors with invasion beyond subcutaneous fat were associated with a statistically increased risk of LR (RR, 9.1; 95% CI, 2.8-29.2), NM (RR, 2.0; 95% CI, 1.4-2.7), AM (RR, 4.0; 95% CI, 2.4-6.5), and DSD (RR, 10.4; 95% CI, 3.0-36.3). Ulcerated tumors were associated with increased risks of LR (RR, 1.7; 95% CI, 1.2-2.4) and DSD (RR, 2.2; 95% CI, 1.2-4.1). Poor differentiation was associated with LR (RR, 2.4; 95% CI, 1.2-4.8), NM (RR, 2.6; 95% CI, 2.2-2.9), DM (RR, 3.9; 95% CI, 1.9-8.2), AM (RR, 2.9; 95% CI, 1.6-5.1), DSD (RR, 5.9; 95% CI, 1.6-21.0), all-cause death (RR, 1.8; 95% CI, 1.1-2.4), and locoregional (RR, 2.6; 95% CI, 1.7-4.0). Desmoplastic stroma was associated with a statistically increased risk of LR (RR, 8.1; 95% CI, 4.1-15.9), NM (RR, 2.8; 95% CI, 2.0-4.0), DM (RR, 17.3; 95% CI, 4.4-68.4), and DSD (RR, 7.6; 95% CI, 4.9-11.8). Tumor budding was associated with a statistically increased risk of NM (RR, 2.9; 95% CI, 2.1-4.2). Perineural invasion was associated with statically significant risks of LR (RR, 4.1; 95% CI, 2.7-6.3), NM (RR, 2.7; 95% CI, 2.2-3.2), DM (RR, 4.8; 95% CI, 2.2-10.4), AM (RR, 5.0; 95% CI, 2.3-11.1), DSD (RR, 7.5; 95% CI, 4.5-12.3), all-cause death (RR, 1.8; 95% CI, 1.4-2.3), and had the highest RR among all risk factors for DM and AM identified in more than 1 study. Lymphovascular invasion had similarly elevated RRs, with statistically significant increased risks of LR (RR, 2.8; 95% CI, 1.5-5.1), NM (RR, 2.1; 95% CI, 1.7-2.6), DM (RR, 9.3; 95% CI, 1.3-66.0), AM (RR, 3.2; 95% CI, 2.1-5.0), DSD (RR, 8.8; 95% CI, 3.9-19.7), and approached statistical significance for all-cause death (RR, 1.9; 95% CI, 1.0-3.6; P = .07).

Several patient factors were associated with poor outcomes. Most notably, patients with immunosuppression had a significantly increased risk of LR (RR, 1.5; 95% CI, 1.1-2.2), NM (RR, 2.6; 95% CI, 2.2-3.0), AM (RR, 1.6; 95% CI, 1.2-2.2), DSD (RR, 2.7; 95% CI, 1.6-4.5), all-cause death (RR, 2.7; 95% CI, 1.5-4.6), and locoregional recurrence (RR, 2.9; 95% CI, 2.0-4.2). When looking at causes of immunosuppression, recipients of solid organ transplantation were at a statistically increased risks of LR (RR, 4.0; 95% CI, 2.2-7.0) and NM (RR, 38.6; 95% CI, 21.2-70.3). We found statistically increased risks of LR (RR, 7.0; 95% CI, 2.1-23.1) in patients with HIV.

Random effects analyses of proportions of each risk factor outcome pairing are described in eTable 3 in the Supplement. For treatment data, random effects analysis of incidence proportion for each treatment modality per outcome is reported in eTable 4 in the Supplement. Mohs micrographic surgery (MMS) had lower incidence than standard excision for each outcome of interest; however, there was overlap of 95% CIs. The eReferences of sources supporting the eTables are available in the Supplement.

Discussion

This study aimed to summarize the literature on patient and tumor factors associated with poor outcomes, providing an update to prior work by Rowe and colleagues.¹³⁶ The meta-analysis of these data quantifies the influence of established and emerging risk factors on patient outcomes. As such, these results have implications on current staging systems, workup protocols, and treatment algorithms of invasive cSCC.

Implications for Staging and Prognostication

The American Joint Committee on Cancer 8 (AJCC)¹³⁷ and Brigham and Women’s Hospital (BWH)⁶⁶ are the current standards in cSCC staging. Both systems include only tumor-specific risk factors; the AJCC is a consensus system based on expert opinion for staging of head and neck cSCC,¹³⁷ whereas the BWH was built on data from 256 tumors.⁶⁶ The findings of this meta-analysis support many of the features used by both systems but suggest others may be of limited prognostic value. We also identified additional risk factors that should be considered in patient prognostication.

Although patient factors are not included in either the AJCC or BWH guidelines, they are present in a number of historical¹³⁸ and contemporary systems.^137,139 Our findings support immunosuppression as a high-risk prognostic feature as defined by the National Comprehensive Cancer Network (NCCN) guidelines.¹⁴⁰ In the present analysis, immunosuppression was a significant risk factor for every individual poor outcome assessed except DM, for which only 2 studies were identified compared with 7 to 18 studies for all other individual outcomes. Given its predictive value for LR, NM, AM, and DSD, omission of immunosuppression in patient prognostication provides an incomplete picture of risk. In addition, our findings suggest a potentially different risk profile based on causes of immunosuppression; however, the statistical analysis was limited by the number of studies. Further investigation is needed in this area to better characterize the distinct risk profiles of each subgroup.

Several tumor characteristics identified as high-risk in this meta-analysis are absent from current staging systems. Among these were histologic and architectural factors that were among the most consistently associated with poor outcomes. Namely, desmoplastic stroma and lymphovascular invasion showed an increased risk of poor outcomes on par with perineural invasion and deserve strong consideration for inclusion in future systems. In addition, poor differentiation was omitted from the most recent AJCC criteria¹³⁷ given concerns of inconsistency in definition and application across the US. In pooling results from multiple different sites across the US and the world, our findings demonstrated that this risk factor was consistently highly predictive of every poor outcome assessed, despite the potential institutional differences in definition and use.

The findings of this study also suggest that the role of tumor size (diameter) in staging should be reexamined. Although the BWH attaches the same risk to all tumors of 2 cm in diameter or larger,⁶⁶ the AJCC designates tumors of 2 to 4 cm as T2 and those of 4 cm or larger as T3.¹³⁷ The findings of the present study indicate that the most important inflection in risk stratification occurs at the 2-cm threshold, thus supporting the BWH classification and challenging the AJCC T3 size stratification. Finally, other factors were associated with some but not all poor outcomes that warrant further research and possible subsequent consideration, including tumor budding, ulceration, and tumor location.

Perioperative Counseling

The results of this meta-analysis may also be used to inform perioperative counseling and work on cSCC. Currently, there is a lack of consensus on which tumors should receive further workup.¹⁴⁰ Per NCCN guidelines, imaging studies are indicated for tumors that are suggestive of extensive disease, defined as deep structural involvement, perineural disease, or deep soft tissue involvement.¹⁴⁰ Our findings suggest that consideration of imaging and more frequent follow-up examinations may also be appropriate in patients with high-risk microscopic features, such as desmoplastic stroma or poor differentiation. Moreover, closer follow-up may be indicated in patients with immunosuppression, especially those with a history of solid organ transplantation.

Treatment Modalities

In addition to addressing the association of risk factors, we explored and compared the association of treatment with poor outcomes. Across outcomes, MMS was found to have the lowest proportions of poor outcomes, although 95% CIs overlapped with those of other treatment modalities. These results support continued use of MMS as first-line therapy for high-risk SCC.

The Appropriate Use Criteria (AUC) for MMS details 270 scenarios for which MMS may be considered and provides recommendations of appropriate, uncertain, and inappropriate based on a synthesis of clinical experience and evidence review; it supports MMS for all histologically aggressive subtypes for all tumor sites and sizes.¹⁴¹ Other tumor factors identified, such as tumor budding and ulceration (Figure 2), require further investigation and may subsequently merit consideration for inclusion in the AUC.

In addition, our study data suggest that MMS should be appropriate for all cSCC in patients with immunosuppression. Although the current AUC designation for MMS is uncertain in smaller nonaggressive cSCC of the trunk and extremities,¹⁴¹ we found that immunosuppression predicts poor outcomes with consistency on par with several AUC aggressive tumor features and size greater than 2 cm.

Finally, our findings support the NCCN designation of perineural invasion greater than 0.1 mm, lymphovascular invasion, desmoplastic stroma, and poor differentiation as the highest-risk features in recommendations for adjuvant radiotherapy; each of these factors were among the most consistently high risk in our analysis.¹⁴⁰ Notably, the NCCN recommendation and BWH data highlight large caliber (>0.1 mm) perineural invasive tumors as being of highest risk⁶⁹; however, we were unable to stratify results by nerve diameter. Our findings also suggest that patients who are immunosuppressed, especially those with a history of solid organ transplantation, may require multidisciplinary discussion for the consideration of radiotherapy when presenting with other high-risk factors.

Histochemical Stains and Tests

In our review, we identified studies assessing risk of poor outcomes with 11 immunohistochemical extracellular molecule stains and 1 gene expression test. Positive stains of many extracellular molecules were associated with a significantly increased risk of poor outcomes in single studies and warrant further research. Stains for epidermal growth factor receptor and programmed death ligand-1, and the 40-GEP test deserve particular attention.

Limitations

Although this analysis included 129 studies, it is possible that pertinent studies may have been omitted given the inherent limitations of database literature searches. Furthermore, data format, study designs, and reporting of follow-up were heterogenous among studies, which may have affected the analysis. Finally, some risk factor−outcome pairs were only assessed by a single study, and we focused primarily on risk factors found by multiple studies.

Conclusions

This systematic review and meta-analysis found several patient risk factors and tumor characteristics associated with LR, NM, DM, DSD, and all-cause death. Several of these—ie, lymphovascular invasion, desmoplasia, and immunosuppression—were associated with a higher risk of poor outcomes; however, they are not currently included in the BWH or AJCC guidelines. These tumor characteristics should be considered in future staging systems, workup protocols, and treatment algorithms. Moreover, we identified several emerging risk factors, immunohistochemical stains, and a prognostic test that warrant further research. Lastly, patients receiving MMS had the lowest proportion of nearly all poor outcomes, supporting the use of MMS as the preferred treatment modality in patients with high-risk factors.

Supplement.

eTable 1. Evaluation of cohort studies – risk factor (level 4 evidence)

eTable 2. Results of random effects analysis: risk ratio for each risk factor and distant metastasis, any metastasis, locoregional recurrence, and combined outcomes

eTable 3. Incidence proportion for each risk factor modality per outcome across all studies

eTable 4. Incidence proportion for each treatment modality per outcome across all studies

eFigure. Forest plots of results of random effects analysis: risk ratio for each risk factor-and distant metastasis, any metastasis, and all-cause death identified in more than 1 study

eReferences