Cutaneous Squamous Cell Carcinoma: Risk Stratification, Staging, and Treatment — A Living Review

Scope, Audience, and How to Use This Review

This clinician-facing review is written for dermatologists, Mohs surgeons, residents, referring clinicians, and medically sophisticated readers. It integrates risk stratification, staging, biopsy interpretation, margin strategy, nodal evaluation, radiation, systemic therapy, immune suppression, prevention, and follow-up in one clinical framework.

For shorter patient guidance, the Skin Cancer Patient Journey translates this evidence into two questions: which treatment fits a newly diagnosed SCC, and what changes when the patient is immunosuppressed.

Evidence is current through July 12, 2026. Individual treatment still depends on the clinical examination, full pathology, patient health, priorities, and professional judgment.

Clinical Thesis: Find the Tumors That Change the Rules

Cutaneous squamous cell carcinoma, or cSCC, is common and usually curable. The danger lies in treating its broad spectrum as though it were one uniform disease.

A small, sharply defined primary tumor on the trunk is not the same problem as a recurrent, deeply invasive tumor near a major nerve in an immunosuppressed patient. The diagnosis may be identical. The probability of subclinical extension, recurrence, nodal spread, treatment failure, and functional harm is not.

What Is Cutaneous Squamous Cell Carcinoma?

Cutaneous squamous cell carcinoma begins in keratinocytes, the cells that make up most of the epidermis. Ultraviolet radiation is the major preventable driver, but the disease is also shaped by age, cumulative sun damage, immune status, chronic inflammation, prior radiation, certain genetic conditions, and the biology of the individual tumor.^1,2

Most cSCCs are cured with local treatment. That reassuring truth should not blur the second truth: cSCC accounts for far more regional spread and disease-specific death than basal cell carcinoma, and its national burden is large because the cancer is so common.² The practical problem is therefore one of distribution. Many tumors are straightforward; a smaller group carries a disproportionate share of recurrence, nodal metastasis, distant metastasis, and death.

The purpose of risk stratification is to find that group early enough to change what happens next.

Risk Stratification and Staging Are Different Jobs

Current NCCN guidance uses low-, high-, and very-high-risk treatment pathways. Size, depth, location, borders, recurrence, histology, nerve involvement, immune status, prior radiation, and other clinical features can change the pathway.

The three major frameworks answer different questions:

• NCCN treatment-risk categories help choose the intensity and method of treatment.
• BWH classification estimates prognosis using a compact set of tumor factors.
• AJCC-8 staging describes anatomic extent within its defined scope.

They should inform one another, but they should not be treated as interchangeable labels.

NCCN treatment risk

NCCN treatment risk is based on the highest-risk feature present. Factors include location, size, borders, whether the tumor is primary or recurrent, immune status, prior radiation or chronic inflammation, rapid growth or neurologic symptoms, differentiation, selected histologic subtypes, depth, perineural invasion, lymphovascular invasion, and bone involvement.¹

Two thresholds are especially useful teaching anchors:

• On the trunk or extremities, a tumor under 2 cm is lower risk by size and location; 2 through 4 cm is high risk; and a tumor over 4 cm is very high risk.¹
• A depth of 2 through 6 mm is high risk. A depth over 6 mm or invasion beyond subcutaneous fat is very high risk.¹

These are examples, not a replacement for the complete risk assessment. A smaller tumor may still be high or very high risk because of location, recurrence, immune suppression, nerve involvement, poor differentiation, or another feature.

The current formal high-risk histologic-subtype list is narrower than many older articles. NCCN now highlights adenosquamous and sarcomatoid cSCC; older lists that automatically included acantholytic or desmoplastic subtypes should not be copied forward as though the terminology never changed.¹

What the outcome gradient looks like

In a retrospective validation cohort of 10,196 tumors, adjusted five-year outcomes separated sharply across the three NCCN tiers:³

These are cohort estimates, not individual predictions. Their value is the gradient: the very-high-risk category identifies tumors whose behavior is not adequately described by the phrase “common skin cancer.”

BWH prognostic classification

The BWH system counts four factors: diameter at least 2 cm, poor differentiation, perineural invasion involving a nerve at least 0.1 mm, and invasion beyond subcutaneous fat.⁴

Invasion beyond subcutaneous fat is one BWH factor. It does not, by itself, make a tumor BWH T3. That error appears in older summaries and materially exaggerates stage.

AJCC-8 stage

AJCC-8 cSCC staging applies to head-and-neck tumors. It does not provide a universal T-staging system for cSCC of the trunk and extremities.⁵ AJCC-8 uses anatomic criteria, including a 4-cm boundary, bone involvement, and defined patterns of deep invasion or perineural involvement. Its T2 and T3 groups have shown outcome overlap in validation studies, which is one reason BWH classification is often added for localized prognostic discussion.⁵

The Biopsy Is a Map Sample, Not the Entire Territory

A diagnostic biopsy may sample part or all of a visible lesion, depending on intent and technique. A pathology report must therefore be read together with the clinical lesion and the sampling method. A reassuring biopsy margin does not prove that no cSCC remains outside the sampled tissue.

Pathology details that can change the plan include depth, differentiation, subtype, perineural involvement, lymphovascular invasion, invasion beyond subcutaneous tissue, and other aggressive features. When the report and clinical behavior do not agree, the disagreement itself deserves attention.

Treatment Options at a Glance

The menu is intentionally broad. The decision becomes useful only after the tumor has been classified.

When Is Mohs or Another Complete-Margin Method Strongest?

Mohs surgery and related forms of peripheral-and-deep en-face margin assessment are designed to examine the complete mapped surgical boundary. Routine vertical sections examine representative cross-sections with intervals between them. Both methods can be appropriate; the value of a continuous map rises when microscopic extension is more likely or when recurrence would be especially costly.

For a well-defined, lower-risk tumor in a forgiving location, standard excision can be an excellent operation. For a recurrent, poorly defined, deeply invasive, nerve-associated, very-high-risk, or tissue-sensitive tumor, complete mapped margin assessment often supplies information that changes both clearance and reconstruction.

Observational studies associate Mohs/PDEMA with better outcomes in high- and very-high-risk cSCC. These comparisons are clinically important but are not randomized proof that one operation prevents every metastasis or death.

When Standard Excision, ED&C, or Nonsurgical Treatment Can Fit

Standard excision

For a selected, well-defined low-risk cSCC, current guidance supports standard excision with a 4- to 6-mm clinical margin and postoperative pathology.^1,2 This can be an excellent operation when the planned margin is anatomically acceptable and the tumor does not require intraoperative mapped feedback.

For high- and very-high-risk tumors, NCCN does not provide one universal numeric excision margin. The margin should be determined case by case, and the reconstructive plan should respect the possibility that routine pathology will identify a positive or uncertain margin after the procedure.¹ In some settings, repair is staged or delayed rather than placing a complex flap over a margin that has not yet been fully adjudicated.

ED&C

Electrodesiccation and curettage can fit a small, primary, lower-risk cSCC in a non-terminal-hair-bearing location. A meta-analysis of appropriately selected lesions reported low recurrence, but the result depends heavily on selection and operator judgment.¹¹ ED&C does not provide the same margin specimen as excision and should not be generalized to recurrent, poorly defined, aggressive, deeply invasive, high-risk, or very-high-risk disease.

SCC in situ and nonsurgical therapy

SCC in situ is confined to the epidermis and has a different treatment menu from invasive cSCC. Excision, ED&C, topical 5-fluorouracil, imiquimod, photodynamic therapy, and other approaches can fit selected surface-confined disease.¹ The exact choice depends on site, size, field cancerization, immune status, adherence, and whether occult invasion is a concern.

Nonsurgical treatment should never borrow the indications or cure expectations of surgery for an invasive higher-risk tumor.

When surgery or Mohs may not be needed

Not every lower-risk cSCC requires Mohs. A well-defined primary tumor on a forgiving site may be treated well with standard excision or, in a narrower setting, ED&C. Selected SCC in situ may be managed with a topical or field-directed approach.

Occasionally, severe frailty, limited life expectancy, competing illness, or the burden of treatment makes observation or symptom-directed care a reasonable shared decision. That is not the same as ignoring the cancer. It is a deliberate choice that weighs tumor risk against what the treatment would ask of the patient.

Perineural and Lymphovascular Invasion

Perineural invasion means tumor is present around or within a nerve. Clinical perineural spread is suggested by pain, numbness, tingling, weakness, or radiating discomfort along a nerve, or by radiologic nerve involvement. Incidental microscopic PNI is a histologic finding without associated clinical symptoms or radiologic evidence. Nerve caliber, depth, extent, symptoms, margin status, and tumor stage all matter.

Lymphovascular invasion is a separate adverse finding associated with worse outcomes. Neither term should be reduced to a binary checkbox without its pathologic and clinical context.

When Should Nodes, Nerves, and Imaging Enter the Plan?

Every cSCC visit should include a clinical examination that matches the tumor risk. For higher-risk disease, that means more than looking at the biopsy site. Regional lymph-node basins, relevant cranial nerves, and symptoms such as pain, numbness, tingling, weakness, or radiating discomfort deserve deliberate attention.¹

Imaging should answer a clinical question:

• Contrast-enhanced CT is commonly used when nodal staging is the primary concern.
• MRI is preferred when perineural spread along a named nerve or toward the skull base is suspected.
• Ultrasound can evaluate a superficial nodal basin and guide needle sampling.
• PET/CT may help in selected advanced cases when regional or distant disease is a concern.

A 2026 expert consensus recommended imaging when the predicted nodal-metastasis risk reaches approximately 15%, while emphasizing that the modality should follow the suspected route of spread.²¹ That threshold is a decision aid, not a substitute for examination and judgment.

Sentinel lymph-node biopsy is not routine standard care for cSCC. NCCN lists it as a category 2B consideration for selected clinically node-negative very-high-risk tumors.¹ The role remains unsettled because no randomized trial has shown that finding microscopic nodal disease through SLNB improves survival.

Radiation and Systemic Therapy

Definitive radiation

Radiation can be an effective primary treatment when surgery is not feasible, is declined, or would create unacceptable functional harm. A systematic review of 46 mostly retrospective studies reported pooled local control of approximately 87%, but the patients, tumors, techniques, and follow-up varied substantially.¹²

It is not defensible to tell every patient that radiation has a lower cure rate than surgery, because there is no randomized head-to-head trial and the observational groups differ in age, health, site, size, and risk. It is equally inappropriate to call the treatments evidence-equivalent. A balanced statement is simpler: surgery is generally preferred when feasible, while radiation is a strong option when surgery is not the right fit, and the exact difference for an individual patient is uncertain.

Adjuvant radiation after clear margins

Postoperative radiation may be considered for clinically significant perineural disease, multifocal nerve involvement, recurrent or very large tumors, selected close-margin situations, and other patterns with high locoregional risk.¹

The evidence is conflicted. A 2025 multicenter cohort associated adjuvant radiation with less local and locoregional recurrence in high-stage disease.¹³ A 2022 systematic review limited to clear-margin cSCC did not find a statistically significant overall benefit across heterogeneous studies.¹⁴ Neither establishes one universal recurrence-reduction percentage, and neither proves a disease-specific-survival advantage. The phase III Spino-RT trial is intended to supply the randomized evidence that the field currently lacks.

Adjuvant cemiplimab

C-POST was a phase III randomized trial of adjuvant cemiplimab versus placebo after surgery and postoperative radiation in selected high-risk cSCC. At 24 months, disease-free survival was 87.1% with cemiplimab and 64.1% with placebo, corresponding to a hazard ratio of 0.32.¹⁶

The adjuvant indication was approved in 2025. This review does not assign a month because the primary regulatory sources used in the audit did not establish one with sufficient confidence. Overall-survival benefit is not established; the mature result is disease-free survival.

C-POST also has an important boundary: it excluded solid-organ transplant recipients, prior stem-cell transplant recipients, and patients requiring substantial systemic immunosuppression. The trial result should not be copied into those populations without a separate multidisciplinary discussion.

Neoadjuvant and advanced-disease PD-1 therapy

Cemiplimab is an anti-PD-1 antibody, not a PD-L1 inhibitor. In a single-arm phase II neoadjuvant study, cemiplimab produced a pathologic complete response in 51% of patients with resectable stage II-IV (M0) cSCC.¹⁷ The result is clinically important but does not mean that surgery can be routinely omitted after an apparent response.

For unresectable locally advanced, regional, or metastatic cSCC, PD-1 therapy has produced durable responses and changed the treatment landscape.¹⁸ Surgery and radiation may still matter; the task becomes sequencing rather than choosing one modality in isolation.

How Immunosuppression Changes the Pathway

Immune suppression can arise from solid-organ transplantation, hematologic disease, stem-cell transplantation, HIV, or immune-modifying medications. These populations are not interchangeable, and one risk estimate should not be applied to all of them.

Immune status should lower the threshold for careful risk assessment, complete pathology review, coordinated surveillance, and multidisciplinary discussion when appropriate. It should not produce fatalism or an automatic rule that every tumor requires the same operation.

Solid-organ transplantation

A 2024 meta-analysis of 67 studies estimated a pooled standardized incidence ratio of 45.9 for cSCC after solid-organ transplantation.²² Individual estimates ranged from roughly fivefold to more than 200-fold. Both extremes can be real in different populations because organ type, latitude, skin phenotype, era, medication regimen, and time since transplantation change the denominator and the biology.

The pooled estimate is useful for showing the magnitude of the problem. It is not an individual prediction and should not be simplified to “every transplant recipient has a 46-fold risk.”

Whether immunosuppression independently worsens the behavior of an individual tumor after adjustment for stage remains unsettled. Studies conflict, although current NCCN treatment-risk stratification includes immune suppression as a risk factor.^1,23 The safest clinical response is neither fatalism nor dismissal: lower the threshold for careful assessment and coordination, then let the tumor's features determine the intensity of treatment.

Other forms of immune suppression

CLL and other hematologic malignancies, stem-cell transplantation, HIV, and chronic immune-modifying medications each carry their own risk patterns.^22,34 They should not inherit one transplant-derived multiplier. The cause and depth of immune suppression, disease control, prior skin-cancer burden, and tumor biology all matter.

Checkpoint inhibitors after transplant

Checkpoint inhibitors can shrink advanced cSCC after solid-organ transplantation, but they can also activate rejection of the transplanted organ. A 2025 individual-participant meta-analysis reported a cSCC-specific objective response rate of 61% and a six-month rejection rate of 23.3%; risks varied by organ and immunosuppression strategy.³⁰

A small phase I kidney-transplant study of cemiplimab given with an mTOR- and steroid-based protocol reported responses with no rejection among 12 participants.²⁹ That is promising, not a universal safety guarantee. The decision requires oncology and transplant expertise, and the value of tumor control must be weighed against the possibility of graft loss or death.

Follow-Up and Prevention

Most recurrences occur early, which is why follow-up is most intensive during the first two years. Current NCCN schedules gradually space visits according to risk: lower-risk tumors may be seen every 3 to 12 months initially, while very-high-risk disease is commonly followed every 3 to 6 months during the first two years and then less often over time.¹

No randomized trial proves one optimal interval.³³ “Every three months indefinitely” is not a universal rule. The schedule should reflect tumor risk, nodal risk, immune status, the number and pace of new tumors, patient reliability, and whether the examination includes the relevant lymph-node basin.

Prevention during immune suppression

The first layer is behavioral and local: ultraviolet protection, self-examination, prompt biopsy of changing lesions, treatment of field cancerization, and regular professional skin examination.^31,32

For patients with a high tumor burden, systemic prevention may be considered:

• Acitretin reduced keratinocyte-cancer burden in small kidney-transplant trials, but benefit wanes after stopping, and mucocutaneous adverse effects, lipid and liver monitoring, and pregnancy constraints limit use.^26,31
• Nicotinamide reduced keratinocyte cancers in high-risk immunocompetent adults, but the 2023 ONTRANS trial found no benefit in solid-organ transplant recipients. The immunocompetent result should not be extrapolated.^24,25
• Sirolimus conversion reduced subsequent cSCC in selected kidney-transplant recipients with prior cSCC, but tolerability was limited and an individual-participant meta-analysis identified increased mortality.^27,28 The finding should not be generalized to every organ or regimen.

No patient should change antirejection or immune-modifying therapy independently. A reduction in one risk may create another, including rejection or loss of control of the disease for which immune suppression was prescribed.

Clinical Decision Framework

1. Confirm the diagnosis and sampling context. Know whether the biopsy was partial and which adverse pathologic features were assessed.
2. Classify treatment risk. Use current guideline features rather than a single superficial characteristic.
3. Add prognostic stage when it changes care. Keep BWH, AJCC-8, and NCCN roles distinct.
4. Choose the needed margin certainty. Match the pathology method to tumor risk, anatomy, and the cost of recurrence.
5. Examine beyond the primary tumor when indicated. Nerves, nodes, imaging, radiation, and systemic therapy enter selectively.
6. Include immune status and competing risks. Treatment should solve the cancer problem without ignoring the person who has it.
7. Make surveillance proportional. The most dangerous tumors need more than a generic annual skin check; lower-risk disease should not inherit the most intensive schedule by default.