Cutaneous Melanoma: Diagnosis, Staging, and Treatment — A Living Review

Scope, Audience, and Clinical Thesis

This review is written for dermatologists, Mohs surgeons, dermatopathologists, residents, referring clinicians, and medically sophisticated readers who need a coherent map from biopsy report to local surgery, nodal staging, systemic therapy, and surveillance.

The thesis is that melanoma care is a sequence of linked decisions. The pathology report determines the staging questions. Stage and anatomy determine the local operation and the value of sentinel-node staging. Resectability, nodal burden, mutation status, recurrence risk, medical fitness, and patient goals determine whether systemic therapy should precede surgery, follow it, or treat advanced disease.

No single procedure is the answer to “melanoma.” Wide local excision, Mohs or staged excision, sentinel-node biopsy, nodal surveillance, immunotherapy, targeted therapy, radiation, and observation each belong to defined forks in the pathway.

Pathology Features That Control the Pathway

Breslow thickness and ulceration

Breslow thickness and ulceration determine the AJCC-8 primary-tumor category and remain major prognostic variables.^1,2 Thickness is measured from the top of the granular layer—or the base of true ulceration—to the deepest invasive melanoma cell and rounded to 0.1 mm.

True ulceration requires loss of the full epidermis with a host reaction. Traumatic or biopsy-related surface loss should not be staged as ulceration. That distinction matters because ulceration changes the T subcategory and can change stage, sentinel-node discussion, adjuvant options, and surveillance.

Mitotic rate, lymphovascular invasion, and neurotropism

Mitotic rate was removed from AJCC-8 T1 substaging but remains reportable and prognostic.^1,3 It should not be silently reinserted as a staging criterion. When a current guideline includes mitotic activity among several adverse features for a selected thin-melanoma sentinel-node discussion, the complete feature context must be preserved.

Lymphovascular invasion and neurotropism are reportable adverse findings but are not independent AJCC stage categories. They can alter concern for regional disease, local control, or radiotherapy in selected cases, yet the strength of evidence is largely observational.³

Microsatellites

Microsatellites are discontinuous microscopic melanoma deposits separated from the primary tumor by normal stroma. They change the N category and establish at least stage IIIB regional disease even without a positive lymph node.^1,3 A finding that looks small under the microscope can therefore produce a major stage transition.

Regression

Histologic regression is not an AJCC staging criterion. Contemporary observational data do not justify using regression by itself to mandate sentinel-node biopsy. Its associations vary by stage and population, so the safest use is as contextual pathology rather than a binary favorable or unfavorable label.^1,3

Desmoplastic melanoma

Desmoplastic melanoma should be classified as pure or mixed because sentinel-node positivity differs between groups. A partial biopsy can misclassify mixed disease as pure; final excision pathology may change the nodal conversation.^3,4

Pure desmoplastic melanoma often has lower sentinel-node positivity, making routine SLNB controversial, while local control, neurotropism, and margin definition may be more prominent concerns. “Desmoplastic” is not one uniform biology.

AJCC Stage Describes Extent, Not One Treatment

AJCC-8 cutaneous melanoma staging uses thickness and ulceration for T category; nodal burden, method of nodal detection, and nonnodal regional deposits for N; and distant metastatic site plus lactate dehydrogenase for M.^1,3

Stage is a language of extent and prognosis. It does not independently choose the operation or drug. Two patients with the same stage may differ in anatomy, resectability, BRAF status, immune-related risk, pace of disease, and goals.

Historical stage-survival tables remain useful for epidemiology but should not be presented as one person's destiny after modern therapy. Tumor biology, treatment era, competing health risks, and subsequent options all modify outcomes.

Wide Local Excision Margins

Wide local excision remains the standard operation for most primary invasive cutaneous melanomas when guideline margins can be obtained with acceptable function and reconstruction.^2,3

These are clinical surgical margins, not the microscopic clearance distance on a pathology report. The purpose is to remove the primary tumor and the surrounding field in which microscopic extension is most likely while limiting functional and reconstructive harm.^2,3

Wider is not automatically better

In a multicenter randomized trial of melanomas thicker than 2 mm, a 2-cm versus 4-cm excision margin produced a whole-trial melanoma-specific-survival hazard ratio of 0.95 at a median follow-up of 19.6 years.⁵ The trial supports 2 cm rather than routine 4-cm excision for that population.

It does not prove that every thickness and every margin pairing has been randomized. The defensible conclusion is population-specific: doubling the standard margin from 2 to 4 cm did not improve melanoma-specific survival in the studied thick-melanoma population and would predictably increase reconstructive burden.

Melanoma in Situ Is Not One Margin Problem

No randomized trial defines the optimal margin for melanoma in situ. A small, sharply defined lesion on a lower-risk trunk or extremity site differs from lentigo maligna spreading across chronically sun-damaged facial skin.^2,3,6

When five millimeters may be enough

A 2024 case series supports a 5-mm margin for carefully selected small melanoma in situ on lower-risk sites, with low recurrence after extended follow-up.⁷ That evidence should not be generalized to lentigo maligna, head-and-neck, acral, large, recurrent, or poorly defined disease.

Why head-and-neck melanoma in situ often needs mapping

In a single-center series of 846 head-and-neck melanoma-in-situ cases treated with Mohs, 62% cleared at 5 mm and 97% required up to 15 mm.⁸ The cohort was selected and referred; it does not mean every head-and-neck lesion needs 15 mm. It demonstrates why a fixed narrow margin can underperform when subclinical extension is broad and asymmetric.

The practical choice is between sacrificing more normal tissue upfront and mapping the actual extension in stages. Anatomy, border definition, lentigo-maligna pattern, prior treatment, and the laboratory's ability to interpret melanocytic margins all matter.

Where Mohs and Staged Excision Fit

Staged excision or Mohs with complete peripheral and deep margin assessment can be considered for lentigo maligna, large or poorly defined melanoma in situ, acral disease, and selected minimally invasive T1a melanoma in anatomically constrained sites.^2,3,9

The evidence boundary is essential: no prospective randomized trial compares wide local excision, Mohs, and staged excision for melanoma. Meta-analyses reporting lower local recurrence with margin-controlled techniques are retrospective and subject to differences in age, site, thickness, selection, processing, and follow-up.^9,10

Margin mapping is the intervention

The defining feature is not a smaller scalpel. It is the ability to map the peripheral and deep margin and return to a specific positive area before final reconstruction. This is particularly useful when the visible border poorly predicts microscopic extension or when tissue loss carries high functional cost.

Immunostains improve visibility, not judgment

MART-1, SOX10, and PRAME provide complementary information:

• MART-1 is sensitive for melanocytes and useful in frozen-section workflows, but it can over-highlight dendritic melanocytes on chronically sun-damaged skin and performs poorly for desmoplastic melanoma.
• SOX10 gives a nuclear stain and can be valuable in spindle-cell or desmoplastic disease, but it also stains Schwannian elements and does not distinguish benign from malignant melanocytes.
• PRAME can help separate melanoma from benign background melanocytes on permanent sections, but sensitivity and specificity are imperfect and a universally validated rapid frozen protocol is not established.

MART-1, SOX10, MITF, and PRAME are complementary tools with different strengths and failure modes; no single stain reliably classifies every melanocyte as benign or malignant. No stain replaces expert integration of morphology, sun-damaged background, clinical map, permanent pathology, and the patient's tumor subtype.^11–14

Protect staging before reconstruction

When Mohs or staged excision is used for apparent melanoma in situ or thin invasive disease, the central debulk specimen must undergo representative permanent-section pathology. Upstaging can change Breslow depth, T category, and sentinel-node options.^15,16

Complex reconstruction should not erase a staging opportunity. The sequence should preserve the specimen and, when relevant, the lymphatic mapping decision before anatomic planes are substantially altered.

When Sentinel-Node Biopsy Adds Value

Sentinel lymph-node biopsy samples the first draining nodal basin to identify microscopic regional disease. It provides staging and prognostic information and can determine eligibility for adjuvant therapy or surveillance intensity.^1,3

A practical threshold framework

• Uncomplicated T1a under 0.8 mm without ulceration: routine SLNB is not recommended.
• T1b and selected adverse-feature T1a disease: discuss the probability of positivity, what the result would change, procedural morbidity, and uncertainty from incomplete microstaging.
• Melanoma thicker than 1 mm: generally offer or discuss SLNB when the result will alter staging or management.

The decision is not purely mathematical. Age, medical fitness, lymphatic-basin morbidity, pathology certainty, patient goals, and whether a result would change therapy all matter.

What MSLT-I proved—and did not prove

MSLT-I established the strong prognostic value of sentinel-node status. The randomized sentinel-biopsy strategy did not produce a statistically significant overall melanoma-specific-survival advantage for the entire study population.¹⁷ Analyses restricted to people who became node-positive condition on a post-randomization event and should not be treated as clean randomized proof that the procedure itself is therapeutic.

The patient-facing explanation is therefore straightforward: SLNB is primarily a staging procedure. It can change what is known and what is offered next, but it is not the same as removing a known therapeutic target.

A Positive Sentinel Node Does Not Mandate Completion Dissection

MSLT-II and DeCOG-SLT changed nodal management. Immediate completion lymph-node dissection did not improve melanoma-specific survival compared with active nodal observation, while the larger operation caused more lymphedema.^18,19

In MSLT-II, three-year melanoma-specific survival was 86% in both groups. Lymphedema occurred in 24.1% after completion dissection and 6.3% with observation.¹⁸

Active observation is not “doing nothing.” It requires clinical and ultrasound surveillance of the nodal basin, timely evaluation of changes, and a plan for regional recurrence. Completion dissection may still fit selected circumstances, but it is no longer the automatic consequence of a positive sentinel node.

Adjuvant Therapy After Complete Resection

Adjuvant therapy aims to reduce recurrence after all known disease has been removed. The value of that reduction must be balanced against immune-related toxicity, treatment burden, competing risk, and the maturity of overall-survival evidence.

Stage IIB and IIC

In KEYNOTE-716, pembrolizumab improved 48-month recurrence-free survival from 58.3% to 71.3%, with hazard ratio 0.62, after complete resection of stage IIB or IIC melanoma.²⁰ Overall-survival benefit has not been demonstrated.

That distinction should remain visible. Recurrence-free survival is clinically meaningful, but it does not answer every patient's question about living longer. A lower absolute baseline recurrence risk, immune toxicity, endocrine consequences, comorbidity, and willingness to accept preventive treatment all belong in shared decision-making.

Stage III

Adjuvant anti-PD-1 therapy improves recurrence-free survival after resection of stage III melanoma. Available mature trial data have not demonstrated a statistically significant overall-survival benefit to date.^20,21

BRAF V600-mutant stage III disease also has a targeted-therapy option in appropriate patients. The choice between anti-PD-1, BRAF/MEK therapy, observation, or a trial depends on mutation status, recurrence risk, toxicity profile, autoimmune history, reproductive considerations, and patient preference.

Neoadjuvant Therapy Changes the Sequence

For selected resectable macroscopic stage III melanoma, systemic therapy before surgery can expose the intact tumor and nodes to treatment, provide an early pathologic-response signal, and potentially guide postoperative therapy.

In NADINA, neoadjuvant ipilimumab plus nivolumab with response-adapted postoperative management produced 12-month event-free survival of 83.7% versus 57.2% with adjuvant nivolumab, hazard ratio 0.32.²²

The result compares complete strategies: combination therapy before surgery, surgery, and response-adapted postoperative management versus surgery followed by nivolumab. It is not a pure test of timing and cannot be generalized to every neoadjuvant drug or to omission of postoperative therapy after any apparent response.

Neoadjuvant management belongs in a multidisciplinary melanoma program with coordinated pathology, imaging, surgery, medical oncology, and toxicity management.

Advanced Melanoma: Durability, Toxicity, and Sequence

Advanced melanoma treatment is built around immune checkpoint therapy, BRAF/MEK targeted therapy when a BRAF V600 alteration is present, tumor-infiltrating lymphocyte therapy in selected later-line settings, radiation or surgery for selected local problems, and clinical trials.

Checkpoint therapy

In CheckMate 067, 10-year overall survival was 43% with nivolumab plus ipilimumab, 37% with nivolumab, and 19% with ipilimumab.²³ The combination also caused substantially more grade 3 or 4 treatment-related toxicity.

The comparison is not “effective” versus “ineffective.” It is the probability of durable control weighed against the probability, reversibility, and long-term consequence of immune toxicity. Disease tempo, CNS involvement, autoimmune disease, organ function, performance status, and patient priorities change the balance.

BRAF-mutant disease

For eligible patients with untreated BRAF V600-mutant advanced melanoma, DREAMseq supports starting with nivolumab plus ipilimumab rather than BRAF/MEK targeted therapy.²⁴ Rapidly threatening disease, contraindications to combination immunotherapy, prior therapy, or other clinical constraints can change the sequence.

Targeted therapy remains highly active and can produce rapid responses. The sequencing lesson is about durability across a treatment course, not the absence of value in BRAF/MEK inhibition.

Tumor-infiltrating lymphocyte therapy

TIL therapy can produce meaningful responses after checkpoint therapy in selected patients, but treatment is intensive and comparative evidence differs by product and trial. The M14TIL randomized trial compared a TIL strategy with ipilimumab; later lifileucel evidence is largely single-arm and should not be merged into one comparative claim.²⁵

Surveillance Should Match Stage

Follow-up should detect recurrence when intervention could matter, identify additional primary melanomas, manage treatment effects, and reinforce skin and nodal awareness without turning low-yield testing into routine harm.

Routine imaging is not recommended for asymptomatic stage IA through IIA disease. Time-limited imaging may be considered for selected stage IIB through IV patients with no evidence of disease, with modality and interval tied to stage, symptoms, treatment, and the actionable consequence of a finding.³

Most recurrences cluster in the first several years, but recurrence pattern varies substantially by stage. Follow-up is generally more intensive early and then spaced over time. No surveillance schedule should override new symptoms.

Lifelong skin surveillance

Melanoma survivors retain a continuing risk of a second primary melanoma. A large SEER cohort reported Kaplan-Meier incidence estimates of 3.9% at five years and 6.7% at ten years.²⁶ These are population estimates influenced by surveillance and ascertainment, not individual forecasts.

The practical point is durable: a treated melanoma does not end the need for skin examination. Patients should understand both recurrence surveillance and the possibility of a new primary melanoma.

Special Clinical Contexts

Pregnancy

Definitive local surgery generally should not be delayed solely because of pregnancy. Sentinel-node timing, tracer choice, anesthesia, imaging, and systemic therapy require individualized maternal-fetal and multidisciplinary planning.^2,3

The reassuring message is not that every intervention is risk-free. It is that timely control of the primary melanoma remains important and a coordinated plan can preserve both oncologic and obstetric priorities.

Solid-organ transplantation

Transplant recipients have higher melanoma incidence and mortality. Checkpoint blockade can threaten the transplanted organ through rejection or graft loss, while observational associations involving specific immunosuppression regimens are not prescriptions.^27,28

Any immunosuppression change requires the transplant team. Cancer control and graft preservation are simultaneous goals, and the relative value of surgery, radiation, checkpoint therapy, targeted therapy, and clinical trials must be individualized.

Acral and mucosal boundaries

Acral melanoma on palms, soles, and nails is biologically and surgically distinct. Site-specific margin control, reconstruction, nodal staging, and systemic options may be required. Digit-sparing treatment can fit selected in-situ or thin tumors but is not a universal rule.^2,3

Mucosal melanoma is biologically and clinically distinct from cutaneous melanoma. Cutaneous stage, margin, and systemic-therapy evidence should not be applied automatically to mucosal disease.

A Practical Treatment Map

Related Patient Guides

The following guide presents a focused treatment roadmap for patients.