Malignant Melanoma: Evolving Practice Management in an Era of Increasingly Effective Systemic Therapies

INTRODUCTION

Melanoma is an aggressive form of skin cancer that, when localized, is highly curable with a simple surgical excision. However, as it grows in the span of just a few millimeters, its lethality increases markedly. More so than many other cancers, the management of melanoma is governed by phase III randomized trials in almost every aspect of treatment. Over the last 10 years, not only have there been several landmark surgical trials but also major advances in systemic treatments using targeted and immunologic approaches to metastatic disease. The availability of effective systemic therapies has led to rapid improvements in survival as well as significant alterations in the traditional approaches to managing this disease. This monograph is organized into 6 sections that provide focused updates on the advances in various aspects of melanoma care. These sections incorporate the latest findings and will help with the management of individual patients. In addition, we also aim to integrate all the approaches to melanoma to help providers better understand the evolving complexity of an integrated multidisciplinary care approach to optimize the care of these patients.

MELANOMA EPIDEMIOLOGY AND GENETICS

Incidence and Mortality

Cutaneous malignant melanoma (CMM) is the fifth most common malignancy in the United States, and it is by far the deadliest cutaneous cancer. The highest-incidence regions worldwide are North America, Australia, and Western Europe (Figure 1).¹ Since the mid-1990s, the yearly incidence in the United States has increased by approximately 0.4% per year (Figure 2).² For most of that period, overall mortality remained the same, with only a slight decline in the last 5 years, from 2.7% to 2.1%. Other nationwide databases in high-incidence areas have demonstrated similar trends,^3,4 which are thought to reflect increased surveillance and detection with some recent improvement in mortality for advanced melanoma. Because most patients are diagnosed with curable stage I disease, 5-year overall survival (OS) for melanoma patients is 92%. However, approximately 6,850 Americans will die from melanoma in 2020, reflecting the deadly nature of advanced melanoma². Despite recent improvements in chemotherapy and immunotherapy, advanced disease remains deadly and carries a 5-year OS of only 19%.⁵

Figure 1.

Melanoma incidence by country (per 100,000 person-years) (with permission from Schadendorf D, van Akkooi ACJ, Berking C, et al. Melanoma. Lancet. 2018;392(10151):971–984).

Figure 2.

Melanoma incidence and mortality in the United States by sex compared to other cancers 1975 to 2017 (with permission from Siegel RL, Miller KD, Fuchs HE, et al. Cancer statistics, 2021. CA Cancer J Clin 2021;71:7–33).

Predisposition

Cumulative research into CMM epidemiology has identified a variety of factors that modify a person’s risk. Risk increases with age, with a notable inflection in incidence rates for both men and women after age 60 years. Men are more likely to be affected overall, an effect that is more pronounced at older ages. By age 65 years, men are 1.55 times more likely to develop melanoma. Perhaps as a result, the average age at diagnosis is lower for women.^{2, 5}

Host Factors

Phenotype

The most prominent phenotypic factors that are associated with development of melanoma are skin pigmentation and presence or absence of melanocytic nevi. Skin pigmentation is largely determined by melanocyte synthesis of eumelanin (ie, dark complexion) or pheomelanin (ie, lighter complexion, blue eyes, and red hair). A pheomelanin-predominant pigmentation phenotype (rather than eumelanin) is associated with increased risk for melanoma. The features of this phenotype and the relative risk (RR) of melanoma are summarized in Table 1.⁶

Table 1.

Relative risk for melanoma associated with phenotypic factors.

	Relative Risk for Melanoma
Phenotypic Factor	(95% Confidence Interval)
Family history	RR = 1.74, 1.41–2.14
Fitzpatrick skin type (I vs. IV)	RR = 2.09, 1.67–2.58
High density of freckles	RR = 2.10, 1.80–2.45
Skin color (fair vs. dark)	RR = 2.06, 1.68–2.52
Eye color (blue vs. dark)	RR = 1.47, 1.28–1.69
Hair color (red vs. dark)	RR = 3.64, 2.56–5.37

Adapted from Gandini S, Sera F, Cattaruzza MS, et al. Meta-analysis of risk factors for cutaneous melanoma: III. Family history, actinic damage and phenotypic factors. Eur J Cancer. 2005;41(14):2040–2059.

Pheomelanin-predominant phenotypes are also associated with sunburn of the skin when exposed to ultraviolet (UV) radiation (UVR), which causes DNA damage. Eumelanin is protective against UVR-mediated sunburn. A personal history of sunburn, a reported tendency to sunburn, and especially a tendency for blistering sunburn are all associated with increased melanoma risk.^{7, 8} Additionally, estimates of total lifetime UV radiation exposure, especially when intermittent, have been shown to be associated with increased risk.⁹

Regarding melanocytic nevi, both the number and quality of nevi seem to convey risk.^{10, 11} There is a steady increase in CMM risk that is proportional to total body nevus count; RR reaches as high as 6.9 for those with more than 100 total nevi.^{10, 11} The characteristics of cutaneous nevi are also relevant. “Atypical” or “clinically dysplastic” are terms used to refer to large (>5 mm), flat nevi with abnormal clinical features, such as poorly defined borders, variegated color, uneven contour, or erythema. Patients with a single atypical nevus have a 2-fold increase in melanoma risk.¹² The presence of at least 10 atypical nevi carries a RR of 12.¹² On the other hand, congenital nevi do not seem to predispose to melanoma.¹²

Medical History

Risk for development of melanoma has also been related to factors in a patient’s personal and medical history. Medical factors, especially prior cutaneous or noncutaneous malignancy, have been described. Also, personal factors such as environmental exposure have been studied in melanoma epidemiology.

History of skin damage predisposes to melanoma. Actinic skin changes are a long-term effect of UVR-mediated DNA damage and are a precursor to skin cancers. Increased risk for melanoma is seen in those with a history of actinic damage alone (RR = 2.02, 95% CI 1.24–3.29), and those with a history of premalignant or malignant nonmelanoma skin lesions (RR = 4.28, 95% CI 2.80–6.55).⁶

Personal history of noncutaneous cancer has also been shown to be a risk factor. Melanoma is an important cause of second primary malignancy in patients who have recovered from common childhood cancers.¹³ Significantly, childhood cancer survivors who develop melanoma are more likely to die of their disease than other primary melanoma patients (hazard ratio = 2.57; 95% CI, 1.55–4.27).¹⁴ In addition to childhood cancer, adult hematologic malignancies such as chronic lymphocytic leukemia (CLL)¹⁵ or non-Hodgkin’s lymphoma¹⁶ are associated with subsequent development of melanoma, possibly due to disruption of normal immune function.

Melanoma is an immunogenic cancer, and a variety of immune system impairments have been associated with increased risk. Specifically, the solid organ transplant (SIR = 2.20, 95%CI 2.01–2.39)^{17, 18} and hematopoietic stem cell transplant (hazard ratio = 5.5, 95% CI, 1.7–17.7)¹⁹ populations have been found to be at risk. Solid organ transplant recipients who develop melanoma have an increased mortality risk of approximately 3–4 times other melanoma patients.^{17, 18} Human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) patients were previously shown to be a high-risk population;²⁰ however, this finding has been disputed.²¹

Genetic Risk

A patient’s genetic risk should also be considered. Familial melanoma is discussed in detail in a subsequent section, but familial cases can often be linked to: (1) a small number of high-penetrance loci (eg, CDKN2A, CDK4); (2) an increasingly large number of low-penetrance loci; (3) a familial cancer syndrome that includes melanoma and other cancers (Cowden syndrome, Li-Fraumeni syndrome, BRCA2); or (4) rare genodermatoses such as xeroderma pigmentosum.²²

Environmental Factors

Environmental exposures can modify patients’ risk for melanoma. These factors invariably relate to lifetime UVR exposure. The UV radiation that is responsible for skin damage is primarily from the UVA (315–400 nm) and UVB (280–315 nm) spectra. UVB radiation is absorbed directly by DNA macromolecules, and primarily causes damage via pyridine cyclodimerization, often affecting adjacent thymine residues.²³ This damage is repaired by the nucleotide excision repair apparatus. UVA radiation represents approximately 95% of sunlight UV radiation and is the primary source of radiation in tanning beds. UVA also creates thymine dimers, but additionally can cause indirect oxidative damage to DNA.²⁴ The characteristic point mutation from UV damage is T->C.

Perhaps the most extensively studied environmental factor is tanning bed use. Once theorized to be potentially protective against UVR-mediated sun damage by inducing melanin production, it is now known that any exposure to tanning beds carries a measurable RR of melanoma.²⁵ Risk appears to be higher among those with many tanning bed exposures, and those who use them at a younger age.^{26, 27} Notably, one study found that those with early-onset melanoma (prior to age 40 years) were much more likely to have used tanning beds at a young age.²⁶

Lifetime sun exposure is also an important factor. Those who live in sunnier climates or low latitudes are at increased risk.^{4, 28} It follows that patients who are exposed to sunlight in their professional or leisure activities would also be at increased risk. This effect has been observed in airline pilots and crew.²⁹

Different patterns in sun exposure may predispose to different molecular pathways for melanoma development. For example, BRAF-mutant cutaneous melanomas are more closely associated with intermittent, intense sun exposure (especially in younger years), and more melanocytic nevi.³⁰ It has been observed that melanoma arising within nevi have the same activating oncogene mutations as adjacent benign nevus cells. Transformation may be the result of acquisition of additional mutations, such as CDKN2A homozygous deletion, resulting in subsequent clonal expansion.^{31, 32} In contrast, TP53-mutant cutaneous melanoma is more often associated with chronic sun exposure, actinic skin damage, and fewer nevi.^{33, 34}

Melanoma Genetics

Molecular Oncogenesis of Melanoma

Melanoma has a very high mutational burden compared to other cancers, reflecting its status as a disease of DNA damage.³⁵ Mutations detected in melanoma average more than 100 per Mb, the highest among common malignancies. The diverse mutational landscape places emphasis on the distinction between driver and bystander mutations. There are 3 main categories of mutation that drive melanoma molecular oncogenesis: 1) driver mutations in cell growth and survival signaling; 2) disruption of canonical cell cycle regulation mechanisms; and 3) pigmentation genes.

Driver mutations in cell growth and survival signaling.

Gain of function mutations in cell growth signaling proteins are common in melanoma. Most external growth factors signal through receptor tyrosine kinases, activating intracellular pathways such as the mitogen-activated protein kinase (MAPK) pathway (Figure 3).³⁶ An activating BRAF mutation, V600E, is present in approximately one half of all melanoma cases.³⁷ Because of its central function in growth signaling, BRAF^V600E is an important therapeutic target for melanoma.³⁸ Other key drivers of cell growth in melanoma are NRAS and PTEN (Figure 3).

Figure 3.

Melanoma oncogenes that promote constitutive cell growth. Most cellular growth factors activate mitogen-activated protein kinase (MAPK) signaling pathways (green) and AKT signaling (yellow) through receptor tyrosine kinases. Variants in BRAF or NRAS (red) are able to function in the absence of external stimuli to activate MAPK signaling. Through another pathway, PTEN inactivation (red) causes accumulation of PIP3, which mediates recruitment of AKT and its activating kinases. AKT activation strongly promotes cell growth. These growth pathways are overactive in melanoma. RAS, rat sarcoma virus (v-Ras) homologs such as KRAS, NRAS, HRAS; RAF, rat accelerated fibrosarcoma (v-RAF) homologs such as BRAF; MEK, MAPK/Erk kinase 1; Erk, extracellular signal-related kinase 1; PI3K, phosphatidylinositol 3-kinase; PTEN, phosphatase and tensin homolog; PIP2/PIP3, phosphatidylinositol 4,5-bisphosphate or 3,4,5-trisphosphate; PDK1, 3-phosphoinositide dependent kinase 1; mTOR, mammalian target of rapamycin; AKT, v-Akt murine thymoma homolog or protein kinase B.

Disruption of canonical cell cycle regulation mechanisms.

Melanoma growth is dependent on escape from typical growth arrest pathways.³⁶ The CDKN2A locus, perhaps the most frequently mutated in melanoma, codes for 2 tumor suppressor genes that regulate cell growth, p16INK4a and p14ARF. The essential anticancer function of these proteins is summarized in Figure 4.

Figure 4.

Canonical cell cycle regulatory pathways that are disrupted by CDKN2A mutation. Transition from the G1 phase of the cell cycle to S phase is controlled by retinoblastoma protein (Rb). Phosphorylation of Rb is required for progression. p16INK4a controls this tightly regulated transition by inhibiting Rb phosphorylation. CDKN2A encodes a second tumor suppressor protein on an alternate reading frame, hence the name ARF. p14ARF regulates the function of MDM2 and MDM4, which inhibit p53 and promote its ubiquitin-mediated degradation. Without p14ARF, p53 is unable to function as a critical regulator of cell proliferation and DNA damage repair. Rb, retinoblastoma; CDK4/6, cyclin dependent protein kinases 4 and 6; INK4a, inhibitor of cyclin-dependent kinase 4; MDM2/4, mouse double minute 2 homolog and mouse double minute 4 homolog.

In summary, the collective function of TP53, MDM2/4, and p14ARF allow the cell to arrest proliferation in the setting of DNA damage. TP53 inactivating mutations are found in only 20% of melanomas; however, most melanomas overexpress MDM2/MDM4, which are downregulators of p53 via ubiquitin-mediated degradation.³⁹ Loss of p53 may be a late event in melanoma biology and may be present in more advanced tumors. As depicted in Figure 4, p16INK4a, Cyclin D, CDK4, and Rb tightly regulate the G1-S cell cycle transition. Dysregulation of this mechanism can lead to unrestricted progression of the cell cycle and is observed in 90% to 100% of melanomas.^{40, 41}

Pigmentation Genes.

MC1R codes for the melanocortin receptor on melanocytes and is a key source of variability in pigmentation between individuals. It is no surprise that MC1R variants are highly associated with melanoma risk.³⁶ Studies have shown that the effect of MC1R extends beyond just pigmentation and photoprotection and may also lead to oncogenesis by non-pigmentation routes.⁴² Additionally, other genes related to melanocyte differentiation and function have been identified as recurrently mutated in melanoma. For example, MITF, a transcription factor required for melanocyte development, may serve as a “lineage survival oncogene” in melanoma, promoting growth by expressing proteins that are important during cell differentiation.⁴³ Dependence on lineage-specific survival mechanisms has been described in melanoma and may convey sensitivity to certain therapeutics.⁴⁴ Transcription factors associated with melanocyte differentiation can be used to subcategorize melanoma and suggest that melanoma can be phenotypically dynamic in response to therapy.

Molecular Genetics of Melanoma Subtypes

Acral lentiginous melanoma and mucosal melanoma have a somewhat different genetic signature than other histological variants of cutaneous melanoma.³⁶ Each represents less than 2% of melanoma cases, but they are notable because they are associated with poorer outcomes. In both subtypes, KIT mutations are particularly common, which may provide a clinically actionable target in the future.⁴⁵ Of note, KIT-mutated tumors usually do not have coexisting BRAF or NRAS mutations. Finally, desmoplastic melanoma is much more likely to harbor an NF1 mutation than other melanoma subtypes.

Modern molecular techniques have also allowed melanoma to be subcategorized by level of differentiation, as detected using expression of transcription factors.⁴⁴ In response to therapy, some tumors develop a de-differentiated phenotype, resembling their neural crest predecessors. This phenomenon may happen in a stepwise pattern with distinct phenotypic progression. Future research will be directed to understanding the genetic and epigenetic events that mediate benign and malignant melanocyte differentiation.

Uveal melanoma is a rare and deadly melanoma subtype that occurs in the eye. Uveal melanoma is associated with mutations in the G_αq (G protein-coupled receptor) signal transduction pathway. Mutations in the BAP1, SF3B1, and EIF1AX genes are associated with aggressive disease. Interestingly, each locus is associated with a different pattern of aneuploidy in chromosomes 3, 6, and 8 that can convey high risk of aggressive disease. Gene expression profiling (GEP) is commonly used to stratify prognosis in uveal melanoma.⁴⁶

Familial Melanoma

Approximately 5% to 12% cases of melanoma occur in a familial pattern.^{47, 48} Almost one half of these cases can be traced to a small number of high-penetrance loci, and 20% to 40% are specifically due to CDKN2A variants. However, 45% of familial cases are linked to low-penetrance variants or do not have an identified cause. Genome-side association studies have begun to elucidate some of the lower-frequency loci.

The CDKN2A locus is responsible for the most famous and well-studied familial melanoma syndrome, familial atypical mole malignant melanoma (FAMMM) syndrome, also known as B-K mole syndrome or dysplastic nevus syndrome (Table 2). However, only 2% of new melanoma diagnoses will harbor a germline CDKN2A variant. Guidelines are available to help clinicians identify high-risk patients who would most benefit from a referral to a genetic counselor. One well-described method, called the “Rule of Twos and Threes” was designed to identify patients with a 10% or higher risk of harboring a CDKN2A germline mutation (Table 3).⁴⁹ Of note, these rules were developed in the Australian melanoma population and should be adapted by clinicians based on local population risk and on clinical suspicion.

Table 2.

Summary of commonly tested genes for familial melanoma.

Gene Locus	Associated Cancers	Syndromes
CDKN2A (NCCN)	CMM, pancreatic cancer, others	FAMMM syndrome
TP53 (NCCN)	Osteosarcoma, CMM, breast cancer, many others	Li-Fraumeni syndrome
CDK4 (NCCN)	CMM	Familial CMM
BAP1 (NCCN)	Uveal melanoma, renal cell cancer, CMM, mesothelioma
MITF (NCCN)	CMM, renal cell cancer
POT1	CMM, glioma, CLL, colorectal cancer
PTEN	Breast, thyroid, uterine, renal, colorectal cancers. CMM. Hamartomatous disease including tricholemmoma and polyps	Cowden Syndrome
RB1	Retinoblastoma, CMM, others
BRCA2	Breast and ovarian cancer, CMM, others	Hereditary breast and ovarian cancer (HBOC)
Other Notable Genes
MCRR (NCCN)	Associated with CMM risk
BRCA1	Breast & ovarian cancer, others.	HBOC
	Possibly melanoma
TERT (NCCN)	Mutated in many cancers
CHEK2	Sarcoma, breast cancer, CMM

The 13 gene loci that are most commonly included in familial melanoma testing panels. NCCN is used to denote the 7 gene loci that are suggested by the National Comprehensive Cancer Network (NCCN) Guidelines. CDKN2A, cyclin dependent kinase inhibitor 2a; TP53, tumor suppressor p53; CDK4, cyclin-dependent kinase 4; BAP1, BRCA-associated protein 1; MITF, microphthalmia-associated transcription factor; POT1, protection of telomeres protein 1; PTEN, phosphatase and tensin homolog; RB1, retinoblastoma; BRCA, familial breast cancer 2; MC1R, melanocortin receptor 1; BRCA1, familial breast cancer 1. TERT, telomerase reverse transcriptase; CHEK2, checkpoint kinase 2; CMM, cutaneous malignant melanoma; CLL, chronic lymphocytic leukemia.

Table 3.

Rule of two’s and three’s for familial melanoma.

• Any patient with three primary melanomas

• A patient with one primary melanoma and two 1st or 2nd degree relatives (same side) with a diagnosis of melanoma or pancreatic cancer

With increasing availability of panel testing, suspicious probands can be tested for a variety of cancer susceptibility syndromes. Some of these feature melanomas as a “subordinate” cancer phenotype. Examples include Cowden syndrome, Lynch syndrome, Li-Fraumeni syndrome, and BRCA 2. More complex guidelines are available to assist the clinician in managing a proband with a dense cancer pedigree.⁵⁰ Referral to a genetic counselor is recommended if there is suspicion of a familial cancer syndrome.

Modern screening panels for germline melanoma predisposition are usually composed of between 7 and 13 high-risk gene loci. A summary of the most commonly tested loci is presented in Figure 2. Genes that are suggested by the National Comprehensive Cancer Network (NCCN) Guidelines are: CDKN2A, CDK4, MC1R, MC1R, TERT, MITF, BRCA2, and BAP1 (especially for uveal melanoma).

Screening for familial melanoma patients should be tailored to each patient’s estimated risk. Patient education is essential because self-examination is the most powerful tool for identifying suspicious lesions. Screening examinations are performed at 3- to 12-month intervals. Those with large numbers of nevi are examined more frequently. Adding images to the medical record helps with tracking of nevi in evolution. Of course, appropriate screening for other malignancies should also be performed if the patient is at risk.

Summary

Melanoma is the deadliest cutaneous malignancy and is a common cause of cancer mortality in North America, Australia, and Western Europe. Those with lighter complexion, increased sun exposure, and immunosuppression are at higher risk. Oncogenesis is dependent on DNA damage from UV radiation, and those with impaired DNA damage repair are predisposed to melanoma. Familial melanoma syndromes have been described and can be detected by targeted gene panels.

MELANOMA EVALUATION AND STAGING

Clinical Suspicion for Melanoma

The diagnosis of melanoma begins with clinical suspicion. Often, patients or their clinicians will notice an abnormal skin lesion on examination. Several methods have been developed to rapidly identify pigmented lesions that should be referred to a specialist for additional evaluation or biopsy. The characteristics of each lesion should be measured within the context of a patient’s overall risk. The ABCDE method⁵¹ (Figure 5) has been well-described and prospectively validated.⁵² Additional methods include the Glasgow 7-point checklist,⁵³ and the “Ugly Duckling” method, which emphasizes cutaneous nevi that differ from others on the patient. The advantage of these criteria is that they can be utilized by dermatologists, other clinicians, and laypersons alike, increasing early detection of disease.^{54, 55} Figure 6 presents a clinical example of an atypical nevus that was later confirmed to contain melanoma.³⁴ Routine screening of average-risk patients is uncommon, but some studies have suggested that it could be cost-effective.⁵⁶ Screening for familial cases is discussed in the section on genetics. Finally, deep convolutional neural networks have been designed to employ machine learning methods for melanoma diagnosis. These interesting new techniques can parse melanoma from other skin lesions based on high-resolution images alone and may play a role in the future of melanoma surveillance.⁵⁷

Figure 5.

American Academy of Dermatology ABCDE criteria (with permission from American Academy of Dermatology Ad Hoc Task Force for the ABCDEs of Melanoma, Tsao H, Olazagasti JM, et al. Early detection of melanoma: reviewing the ABCDEs. J Am Acad Dermatol. 2015;72(4):717–723).

Figure 6.

Melanoma arising within a nevus. Note the characteristics of this lesion, such as variegated color, irregular borders, and distinct appearance compared to neighboring nevi (with permission from Rivers JK. Is there more than one road to melanoma? Lancet 2004;363(9410):728–730).

Evaluation of Suspicious Lesions

Excisional Biopsy

Diagnostic excisional biopsy is the standard of care for suspicious pigmented skin lesions.⁵⁸ Preferred techniques include deep saucerization encompassing the entirety of the lesion, taking care not to transect the deep margin. Full-thickness excisions and punch biopsies are becoming less common because less invasive methods have been shown to be noninferior.⁵⁹ Shallow shave biopsies are appropriate if melanoma is considered unlikely. A summary of biopsy techniques is provided in Table 3. Peripheral margins of 1 to 3 mm of clinically uninvolved skin are preferred to facilitate histopathologic interpretation. Pre-biopsy photographs are particularly helpful if more than 1 area is biopsied. Axial orientation of biopsy is less likely to disrupt lymphatic drainage of the area and future lymphoscintigraphy. If lesion size or location precludes excisional biopsy, full-thickness incisional or punch biopsy may be appropriate. Multiple representative biopsies can be considered for large, pigmented lesions.

Some melanoma diagnoses are made from incisional, shave, or punch biopsies with a positive margin. In these cases, repeat biopsy may be considered if there is suspicion that the tumor thickness was not adequately sampled. Rates of upstaging (increasing T-category) with repeat biopsy range from 6% to 14% depending on original biopsy type. Acral lentiginous and desmoplastic melanoma are most likely to be upstaged. The superficial spreading subtype has only a 1.4% risk of upstage.⁶⁰

If a specimen contains dysplastic melanocytes without melanoma, studies have not shown a benefit to proceeding with surgical excision. In one study, less than 2% of these excisions resulted in a significant change in diagnosis.⁶¹ Some clinicians will discuss excision with the patient if there is high-grade dysplasia.

Evaluation of Biopsies

Biopsy specimens are systematically catalogued and evaluated in keeping with American Academy of Dermatology, the College of American Pathologists (CAP), and the American Joint Committee on Cancer (AJCC).^{58, 62–65} An example of a CAP Synoptic Report is included as a supplementary figure (Figure S1). For histopathologic staging, the essential characteristics of the specimen are the Breslow thickness, reported to 0.1 mm, and the presence or absence of ulceration.⁶⁶ New to the 8th Edition of the AJCC guidelines, the binary measure of mitotic rate greater than or less than 1 mitosis per mm² is no longer considered for histopathologic staging, but remains an important prognostic factor across all thickness categories. Margin status should always be reported. Microsatellitosis, when present, conveys important prognostic information and must be reported.^{67, 68} Other histologic features that may provide prognostic value are the presence of vertical growth phase, angiolymphatic invasion, neurotropism/perineural invasion, regression, and tumor infiltrating lymphocytes (TILs).^69–71 Melanoma has several distinct histologic subtypes; the prognostic importance of subtype has been disputed, with some exceptions.⁷² The lentigo maligna subtype is associated with greater subclinical horizontal spread.⁷³ The desmoplastic subtype is locally aggressive and often has decreased/absent pigmentation. It is associated with increased sensitivity to immunotherapy and decreased risk of lymph node metastasis.⁷⁴ If desmoplastic histology is described, further description of “pure” or “mixed” desmoplastic characteristics provides additional clinical information. Patients with “pure” desmoplastic tumors have an unusually low risk for regional lymph node metastasis and improved 5-year survival.⁷⁵

Initial Evaluation

The initial evaluation for a new diagnosis of primary melanoma should include: (1) a detailed history, with review of environmental and genetic risk factors for melanoma and nonmelanoma skin cancer; (2) physical examination, with attention to local and regional lymph node basins; and (3) a complete skin examination. Abnormal findings should prompt additional investigation as appropriate. For melanoma of unknown primary, an anogenital/pelvic examination and detailed ophthalmologic examination are essential.

Laboratory and Imaging Evaluation

Baseline blood tests are only indicated if there is appropriate clinical suspicion. Routine laboratory studies are not cost-effective. As an example, routine serum lactate dehydrogenase (LDH) is insensitive for metastatic melanoma.⁷⁵ Routine use of chest radiograph (CXR) is also discouraged.^76–79 Regarding advanced imaging techniques, positron emission tomography-computed tomography (PET-CT) and nodal ultrasound (US) have increased sensitivity for metastasis when compared to physical examination, but the routine use of these imaging modalities in melanoma patients leads to high rates of false positive findings. Therefore, cross-sectional imaging has not been universally adopted as part of the initial evaluation for cutaneous melanoma, and nodal US is not a substitute for sentinel lymph node biopsy (SLNB) or sampling of clinically positive nodes.^80–84 Advanced imaging is appropriate for patients with an examination concerning for metastasis.

Cross-sectional imaging has a comparatively higher detection rate of occult distant metastasis in select patients with thick, ulcerated primary melanomas or bulky nodal disease, and may obviate the need for large surgical resection in these patients.^{37, 85–87} Using these criteria as requisites for advanced imaging may increase the detection rate for occult metastasis to 16%.⁸⁶ For thick primary lesions with ulceration (at least T3b), some institutions recommend imaging prior to surgical excision.

Increased understanding of the molecular genetics of melanoma has resulted in the development of molecular testing to identify patients who might benefit from targeted therapeutics. Commonly available tests for melanoma include immunohistochemistry (IHC), GEP analysis via next-gen sequencing, and others.

Specific gene targets for molecular testing include BRAF, NRAS, KIT, and IHC for PD-L1. BRAF, a serine/threonine kinase which is central to cell growth signaling, is the most commonly mutated proto-oncogene in melanoma.⁸⁸ Missense mutation of BRAF at codon V600 is responsible for 80% of activating BRAF mutations and conveys sensitivity to small molecule inhibitors such as vemurafenib and dabrafenib.⁸⁸ IHC and sequencing tests for BRAF^V600 are widely available, and some clinicians choose to include these tests as a part of the initial evaluation for patients who are considered at high risk for recurrence, such as those with category T2 or T3 primary tumors. Mutations in KIT, a receptor tyrosine kinase, are rare in cutaneous melanoma but present in 10% to 15% of mucosal and acral lentiginous melanomas. KIT mutants are detectable with molecular testing and provide a potential drug target for patients with these unusual but deadly melanoma subtypes.⁴⁵ IHC positivity for PD-L1 greater than 1% is correlated with response to immune checkpoint blockade, but has poor predictive performance in melanoma.⁸⁹ Thus, many clinicians will choose a trial of checkpoint blockade immunotherapy even in patients with low PD-L1 IHC staining.

GEP tests, for example DecisionDX-Melanoma (Castle Biosciences, Friendswood, Texas), classify melanoma patients as low- or high-risk based on a panel of transcripts detected in the tumor.⁹⁰ It has been suggested that GEP tests can be used at diagnosis to supplement risk-stratification for early-stage patients with a negative SLNB. Because most new melanoma diagnoses happen at an early stage, and more than one half of deaths occur in patients who initially had a negative SLNB, identification of high-risk patients would be valuable.⁹¹ However, GEP tests are limited by low positive predictive value and a lack of strong evidence in large cohorts of patients to support their routine use at this time.^{92, 93} Currently, GEP tests are not recommended for treatment decisions in average-risk patients with cutaneous melanoma.⁹⁴ They may be used in patients who are judged to be at high risk for recurrence or in the context of a clinical trial.

Staging

Complete histopathologic staging after wide local excision (WLE) +/− SLNB +/− completion lymph node dissection (CLND) is recommended. The AJCC staging method involves determination of the T, N, and M category of each tumor. The principles of TNM categorization for melanoma are depicted in Tables 4, 5, and 6. Tables 7 and 8 describe the assignment of stage based on TNM category. All recommendations reflect the AJCC 8th edition guidelines on melanoma, published in 2017.^{63, 64}

Table 4.

Definition of N category for American Joint Committee on Cancer (AJCC) TNM staging.

	Extent of Regional Lymph Node and/or Lymphatic Metastasis
N Category	No. of Tumor-Involved Regional Lymph Nodes	Presence of In-transit, Satellite, and/or Microsatellite Metastases
NX	Regional nodes not assessed (ie, SLNB not performed, regional nodes previously removed for another reason); Exception: pathological N category is not required for T1 melanomas, use clinical N information	No
N0	No regional metastases detected	No
N1	One tumor-involved node or any number of in-transit, satellite, and/or microsatellite metastases with no tumor-involved nodes	No
N1a	One clinically occult (ie, detected by SLNB)	No
N1b	One clinically detected	No
N1c	No regional lymph node disease	Yes
N2	Two or 3 tumor-involved nodes or any number of in-transit, satellite, and/or microsatellite metastases with one tumor-involved node
N2a	Two or 3 clinically occult (ie, detected by SLNB)	No
N2b	Two or 3, at least 1 of which was clinically detected	No
N2c	One clinically occult or clinically detected	Yes
N3	Four or more tumor-involved nodes or any number of in-transit, satellite, and/or microsatellite metastases with 2 or more tumor-involved nodes, or any number of matted nodes without or with in-transit, satellite, and/or microsatellite metastases
N3a	Four or more clinically occult (ie, detected by SLNB)	No
N3b	Four or more, at least 1 of which was clinically detected, or the presence of any number of matted nodes	No
N3c	Two or more clinically occult or clinically detected and/or presence of any number of matted nodes	Yes

Adapted from Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma of the skin. In: Amin MB, Edge SB, Greene FL, et al, eds. AJCC Cancer Staging Manual, 8^th ed., Springer International Publishing; 2017:563–585.

SLNB, sentinel lymph node biopsy.

Table 5.

Definition of T category for American Joint Committee on Cancer (AJCC) TNM staging.

T Category	Thickness	Ulceration Status
TX: Primary tumor thickness cannot be assessed (eg, diagnosis by curettage)	Not applicable	Not applicable
T0: No evidence of primary tumor (eg, unknown primary or completely regressed melanoma)	Not applicable	Not applicable
Tis (melanoma in situ)	Not applicable	Not applicable
T1	≤1.0 mm	Unknown or unspecified
T1a	<0.8 mm	Without ulceration
T1b	<0.8 mm	With ulceration
	0.8–1.0 mm	With or without ulceration
T2	>1.0–2.0 mm	Unknown or unspecified
T2a	>1.0–2.0 mm	Without ulceration
T2b	>1.0–2.0 mm	With ulceration
T3	>2.0–4.0 mm	Unknown or unspecified
T3a	>2.0–4.0 mm	Without ulceration
T3b	>2.0–4.0 mm	With ulceration
T4	>4.0 mm	Unknown
T4a	>4.0 mm	Without ulceration
T4b	>4.0 mm	With ulceration

With permission from Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma of the skin. In: Amin MB, Edge SB, Greene FL, et al., eds. AJCC Cancer Staging Manual, 8^th ed. Springer International Publishing; 2017:563–585.

Table 6.

Definition of M category for the American Joint Committee on Cancer (AJCC) TNM staging.

M Category	M Criteria
	Anatomic Site	LDH Level
M0	No evidence of distant metastasis	Not applicable
M1	Evidence of distant metastasis	See below
M1a	Distant metastasis to skin, soft tissue including muscle, and/or nonregional lymph node	Not recorded or unspecified
M1a(0)		Not elevated
M1a(1)		Elevated
M1b	Distant metastasis to lung with or without M1a sites of disease	Not recorded or unspecified
M1b(0)		Not elevated
M1b(1)		Elevated
M1c	Distant metastasis to non-CNS visceral sites with or without M1a or M1b sites of disease	Not recorded or unspecified
M1c(0)		Not elevated
M1c(1)		Elevated
M1d	Distant metastasis to CNS with or without M1a or M1c sites of disease	Not recorded or unspecified
M1d(0)		Not elevated
M1d(1)		Elevated

With permission from Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma of the skin. In: Amin MB, Edge SB, Greene FL, et al, eds. AJCC Cancer Staging Manual 8^th ed. Springer International Publishing; 2017:563–585.

LDH, lactate dehydrogenase; CNS, central nervous system.

Table 7.

Assignment of stage based on T, N, and M category: clinical stage group.

When T is…	And N is…	And M is…	Then the clinical stage group is…
Tis	N0	M0	0
T1a	N0	M0	IA
T1b	N0	M0	IB
T2a	N0	M0	IB
T2b	N0	M0	IIA
T3a	N0	M0	IIA
T3b	N0	M0	IIB
T4a	N0	M0	IIC
ANY T, TIS	≥N1	M0	III
ANY T	ANY N	M1	IV

With permission from Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma of the skin. In: Amin MB, Edge SB, Greene FL, et al, eds. AJCC Cancer Staging Manual 8^th ed. Springer International Publishing; 2017:563–585.

Table 8.

Assignment of stage based on T, N, and M category: pathological group.

When T is…	And N is…	And M is…	Then the pathological stage group is…
Tis	N0b	M0	0
T1a	N0	M0	IA
T1b	N0	M0	IA
T2a	N0	M0	IB
T2b	N0	M0	IIA
T3a	N0	M0	IIA
T3b	N0	M0	IIB
T4a	N0	M0	IIB
T4b	N0	M0	IIC
T0	N1b, n1c	M0	IIIB
T0	N2b, N2c, N3b or N3c	M0	IIIC
T1a/b-T2a	N1a or N2a	M0	IIIA
T1a/b-T2a	N1b/c or N2b	M0	IIIB
T2b/T3a	N1a-N2b	M0	IIIB
T1a-T3a	N2c or N3a/b/c	M0	IIIC
T3b/T4a	Any N >N1	M0	IIIC
T4b	N1a-N2c	M0	IIIC
T4b	N3a/b/c	M0	IIID
Any T, Tis	Any N	M1	IV

With permission from Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma of the skin. In: Amin MB, Edge SB, Greene FL, et al, eds. AJCC Cancer Staging Manual 8^th ed. Springer International Publishing; 2017:563–585.

Summary of AJCC Staging

Stages IA through IIC are principally determined by the T category of the primary lesion (ie, Breslow thickness and ulceration). Mitotic rate is no longer a criterion for T category but has prognostic value across all tumor thicknesses. T1 lesions (traditional “thin melanoma”) are up to 1 mm thick. T2 refers to 1 to 2 mm thickness, T3 to 2 to 4 mm thickness, and T4 to “thick melanoma” of Breslow thickness greater than 4 mm. T categories are modified with “a” or “b” to reflect the presence or absence of ulceration. Any nodal disease places the patient in stage III.

Stages IIIA-IIID refer to node-positive disease of any T category. N1 refers to one positive node, N2 to 2 or 3 nodes, and N3 to 4 or more nodes with detectable melanoma. The NXa & NXb designation denotes nodal disease that is clinically occult (“a” designation) or clinically evident (“b” designation). NXc denotes micro/macrosatellites or in-transit metastasis in addition to tumor-involved lymph nodes. Stage III is divided into 4 subcategories based on differential outcomes (Table 9). As usual, stage IV is reserved for distant metastasis. M categories all result in stage IV designation but are subdivided based on the anatomic location of the metastasis. M1a refers to skin/soft tissue and nonregional lymph nodes, M1b refers to lung metastasis, M1c refers to non-central nervous system (CNS) visceral metastasis, and finally M1d refers to CNS metastasis. Each M category is subdivided based on presence/absence of elevated serum LDH, which is an independent adverse predictor of survival across all M categories. LDH is also a useful indicator of disease response and recurrence in select patients.

Table 9.

Description of stage III subgroups.

With permission from Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma of the skin. In: Amin MB, Edge SB, Greene FL, et al, ed. AJCC Cancer Staging Manual 8th Ed. Springer International Publishing; 2017:563–585.

Stage-Specific Survival and Treatment Applications

Stage I melanoma has an excellent prognosis, with 5-year disease-specific survival (DSS) of 99% and 97% for stages IA and IB. Stage II is subdivided to IIA, IIB, and IIC, which have disease-specific 5-year survival of 94%, 87%, and 82%, respectively. Stage IIIA survival is 93%, which is higher than stages IIB and IIC. Patient reclassification with the 8th edition of the AJCC Guidelines resulted in an increase in survival for both of these groups, possibly due to the Will Rogers phenomenon⁹⁵ (ie, when more accurate staging causes “stage migration” of a subset of patients, the average survival of both stages is increased). Stage IV survival is not subdivided, but new M category designations were designed to facilitate future research into the heterogeneity of metastatic melanoma.

Stage-specific disease management is a topic of current investigation. Although resection remains the standard for curable disease, there has been increasing interest in neoadjuvant therapy for patients presenting with stage III melanoma.⁹⁶ Additional research is needed, but neoadjuvant therapy will be a central question in the future of melanoma management. Adjuvant therapies such as BRAF/MEK inhibitors (for BRAF-positive tumors) and immune checkpoint blockade are currently approved for stage IIIB/C/D patients and resected stage IV patients. Some stage IIIA patients are also eligible. These treatment modalities are also being studied in stage IIB/C disease, considering the poor 5-year survival in these patients.⁹⁷ Intratumoral therapies are also an area of active research.⁹⁸

CURRENT CONSIDERATIONS IN THE MANAGEMENT OF EARLY-STAGE MELANOMA

Early stage melanoma includes stage 0, I, and II melanomas as defined by the AJCC 8^th edition staging system.⁹⁹ These are essentially melanomas of any T classification with a negative lymphatic basin and no systemic metastasis. Early stage melanoma has a generally favorable prognosis, with 10-year survival ranging from 98% for stage IA disease to 75% for stage IIC disease.¹⁰⁰ The mainstay of treatment is surgery, which includes resection of the primary tumor with adequate margins, and an evaluation of the tumor’s lymphatic basin using SLNB. The necessity and extent of each of these surgical procedures is guided by the histopathological evaluation of a biopsy taken from these lesions. Here, we review the current considerations and debates regarding each of these elements of treatment, from the biopsy itself, to the mode of excision and its necessary margins, as well as the indications and implications of a SLNB. We also briefly discuss novel molecular techniques and their potential role in these decisions.

Technique of Biopsy

The Breslow thickness of a melanoma has been shown to be the strongest predictor of prognosis.^{101, 102} Since the AJCC staging manual’s 6^th edition, when the anatomical Clark level was abandoned, Breslow thickness has become the dominant variable, together with ulceration, defining the T classification.¹⁰³ As a result, an accurate assessment of melanoma thickness is central to deciding on the clinical and surgical management of patients. Current NCCN guidelines, as well as the American Academy of Dermatology guidelines, recommend an excisional biopsy or at least a full thickness incisional biopsy (punch biopsy) to accurately assess Breslow thickness. They discourage the use of shave biopsy, as it can limit the assessment of depth, but allow its consideration in cases where the index of suspicion is low or for the assessment of lentigo maligna.^{104, 105}

Shave biopsies have been proven to understage the Breslow thickness more than any other biopsy technique. Zager and colleagues⁵⁹ show that residual tumor can be found upon final excision in up to 22% of cases. This residual tumor, left behind during shave biopsy, will cause an upgrading of the T classification upon final excision in 3% to 12.5% of cases. But interestingly enough, studies have shown that this upgrading rarely influences surgical recommendations and it does not seem to affect tumor recurrence or DSS.^{59, 106–108} Consequently, there have been calls for a change in biopsy recommendations.⁵⁹ These calls have been fortified by data showing that almost two thirds (62.6%) of lesions diagnosed as melanoma are biopsied by dermatologists; and although the preferred mode of sampling for surgeons is an excisional biopsy and that of primary care physicians is a punch biopsy, dermatologists, as shown by both retrospective analyses and surveys, categorically prefer shave biopsies, rendering shave biopsy the leading means of melanoma diagnosis.^{106, 109} This understanding has led many, including NCCN panelists, to comment that “any diagnosis is better than none even if microstaging may not be complete,” vindicating the use of shave biopsy in the diagnosis of melanoma.¹⁰⁵ Further discussion of the evaluation of malignant melanoma, including indications for systemic evaluation, can be found elsewhere in this monograph.

Excision of the Primary Tumor

WLE of the primary tumor has been the standard of care in melanoma surgery since the beginning of the 20^th century. The aim of excision is removal of the tumor in its entirety with sufficiently clear microscopic margins to ensure the best chances for cure and local, regional, and systemic control. The historically accepted surgical margins, advocated by Handley¹¹⁰ in 1907, were 5 cm. These generous excision margins often led to the need for skin grafts and complex reconstruction, considerable wound complications, and eventual long-term morbidity. Led by the motivation to preserve function as well as aesthetics, and avoid the disfiguring nature of these procedures, a series of trials were published in the 1980s challenging this 5 cm paradigm.^{110, 111} The major randomized controlled trials (RCTs) that influence current margin recommendations are presented in Table 10. They can be generally divided into trials addressing excision margins in melanomas with a Breslow thickness of up to 2 mm and those which addressed thicker melanomas. Most of these trials exclude melanomas of the hands, feet, head, and neck, and the anogenital areas, as wider excisions in these anatomical locations are more morbid and less practiced.

Table 10.

Randomized controlled trials investigating thin resection margins for invasive melanoma.

	Group	Number of patients	Melanoma thickness	Excision margins	Follow up	Survival	Recurrence	Excluded	Comments
Veronesi et al., NEJM, 1988120 Veronesi et al., Arch Surg, 1991119	WHO melanoma program trial 10	612	≤2 mm	1 vs 3 cm	90 months	8-year DFS 81.6% vs 84.4% 8-year OS 89.6% vs 90.3%	14% vs 13% Local 1.3% vs 0% Lymphatic 6.9% vs 7.8% Distant 5.6% vs 4.6%	Facial, fingers, toes	No local recurrence when thickness <1 mm
Ringborg et al., Cancer, 1996113 Cohn-Cedermark et al., Cancer, 2000112	1st Swedish melanoma group trial	989	0.8–2 mm	2 vs 5 cm	11 years	10-year DFS 81% v. 83% 10-OS 79% vs 76%	21% vs 19% Local 0.6% vs 1% Lymphatic 15% vs 12% Distant 15% vs 14%	Head & neck, hands, feet, vulva
Khayat et al., Cancer, 2003114	French cooperative group trial	337	<2.1 mm	2 vs 5 cm	16 years	10-year DFS 85% vs 83% 10-year OS 87% vs. 86%	Median time to recurrence 43 vs. 37.6 months 13.6% vs 20% recurrence Local 0.6% vs 2.4%, Lymphatic 8.1% vs 6.7% Distant 2.5% vs 6.1%	Toe, nail, finger, acral-lentiginous	LND not performed
Balch et al., Ann Surg, 1993116 Balch et al., Ann Surg Oncol, 2001115	Intergroup melanoma surgical trial	468	1–4 mm	2 vs 4 cm	72 months	5-year OS 79.5% vs 83.7% 10-year OS 70% vs 77%	Local 2.1% vs 2.6% Distant 10.9% vs 8.5%		Primary closure in 89% vs 54% (p<.001)
Gillgren et al., Lancet, 2011365 Utjés et al., Lancet, 2019366	2nd Swedish melanoma group trial (Scandinavian Baltic trial)	936	>2 mm	2 vs 4 cm	19.6 years	Melanoma specific mortality 41% vs 43.5% Overall mortality 65% vs 67%	41.7% vs 42.5% Local 4.3% vs 1.9 Regional 4.1% vs 3.2% Lymphatic 21.5% vs 24.2% Distal 8.2% vs 11.5%	hands, feet, head & neck, anogenital region	9% underwent SLNB
Thomas et al., NEJM, 2004118 Hayes et al., Lancet Oncol, 2016117	UK Melanoma Study Group, the British Association of Plastic Surgeons and the Scottish Cancer Therapy Network	900	>2 mm	1 vs 3 cm	8.8 years	Melanoma specific mortality 42.8% vs 36.9% (p=0.041). Overall mortality 55.8% vs 53.9% (p=0.14)	Local 3.3% vs 2.9% Locoregional 37% vs 31.8% (p=0.05) Distal 8.4% vs 6.7%		General anesthesia in 32.1% vs 66.4% of patients

DFS, disease-free survival; OS, overall survival; WHO, World Health Organization; LND, lymph node dissection; SLNB, sentinel lymph node biopsy.

Questioning the 5 cm margin, the Swedish Melanoma Group trial and the French Cooperative Group trial compared a 2 cm excision to a 5 cm excision in melanomas of up to 2 mm thickness. The Swedish study, including 769 patients, published its first findings in 1996, demonstrating a regional lymph node recurrence rate of 12.1% versus 8.3%, favoring the wider 5 cm excision. Nevertheless, both groups were comparable in overall recurrence rates (20.4% versus 20.5%); but more importantly, it showed a comparable 5-year OS of 86.4% and 88.7%, that was preserved upon long-term follow-up, with a 10-year OS of 79% and 76%, for a 2 cm versus 5 cm margin, respectively.^{112, 113} The French Cooperative Group trial, performing a second randomization of its 337 patients, assigning them to receive adjuvant isoprinosine or no adjuvant therapy, demonstrated similar findings. After a follow-up period of 19 years, the authors reported a recurrence rate of 13.6% versus 20% and a mortality rate of 19.8% versus 17.5% for 2 cm and 5 cm margins, respectively.¹¹⁴ These findings helped to establish a 2 cm margin as adequate for melanomas of up to 2 mm thickness.

This 2 cm margin was further investigated and shown to be adequate in melanomas of greater than 2 mm thickness. The Intergroup Melanoma Surgical Trial, led by Charles Balch, dealt with intermediate thickness melanomas of 1 to 4 mm, and randomized 486 patients to either a 2 cm or 4 cm excision margin. It showed comparable results, in terms of recurrence rates and 5- and 10-year OS, between the 2 groups; and although a narrow resection margin did not confer a negative prognostic effect, the authors demonstrated that both tumor thickness (RR=1.74, p=0.0008) and ulceration (RR=2.21, p=0.0095) conferred a statistically significant negative effect on survival. Moreover, they were able to demonstrate the negative effect a wider excision had on patient outcomes, in terms of a longer hospital stay (7.0 ± 5.1 versus 5.2 ± 4.5 days) and the ability to perform primary closure of the surgical site (54% versus 89%).^{115, 116}

Noticing that the mean thickness of melanomas in the Intergroup trial was 1.96 mm, 2 groups investigated the feasibility of narrow margins in melanomas greater than 2 mm thickness. The 2^nd Swedish Melanoma Group trial randomized 936 patients to a 2 cm versus 4 cm excision margin. With a follow-up of more than 19 years, they were able to show comparable recurrence rates (41.7% vs 42.5%), melanoma specific mortality (41% vs 43.5%), and overall mortality rates (55.8% vs 53.9%) between the 2 groups. The United Kingdom Melanoma Study group, together with the British Association of Plastic Surgery and the Scottish Cancer Therapy Network challenged the 2 cm margin, randomizing 900 patients with melanomas of more than 2 mm thickness to a 1 cm excision margin versus a 3 cm margin. They demonstrated similar overall mortality rates of 55.8% versus 53.9% between the 2 groups. Nevertheless, patients undergoing a narrow excision with 1 cm margins exhibited a higher locoregional recurrence rate (37% vs 31.8%, p=0.05) and worse melanoma specific survival (MSS, 42.8% vs 36.9%, 0=0.041).^{117, 118} These findings helped to establish the recommended margins for T3 and T4 melanomas as 2 cm.

As for thinner melanomas (≤2 mm), the World Health Organization (WHO) trial, led by Veronesi and Cascinelli, showed a narrow 1 cm excision margin to be equivalent to a wide 3 cm excision margin, in terms of 8-year OS (89.6% vs 90.3%). Disease-free survival (DFS), as well as recurrence rates, were likewise comparable between the 2 groups, with a recurrence rate of 14% and 13%, respectively. The authors were also able to show superior results in terms of DFS and OS in patients with 0.1 to 1 mm melanomas compared to those with 1.1 to 2 mm lesions, and that no patient with a melanoma less than 1 mm thickness recurred locally after excision with a 1 cm margin.^{119, 120} These results, in conjunction with findings showing that a margin narrower than 1 cm in these thin (≤1 mm) melanomas confers higher recurrence rates,¹²¹ established a margin of 1 cm as the recommended margin for up to 1 mm thick melanomas. These left melanomas of a 1 to 2 mm thickness in a controversial gray zone, with an ambiguous recommended margin of 1 to 2 cm. The MelmarT trial, an international multicenter RCT aims to answer this question, and is randomizing patients with melanomas of more than 1 mm thickness into 1 cm versus 2 cm resection margins.¹²² It is estimated to be completed in 2026 and will provide further guidance.

There are no similar RCTs for melanoma in situ, with the recommended excision margins being 0.5 to 1 cm.^{58, 105} Most authorities recognize a 0.5 mm margin as adequate for most melanoma in situ cases, but others advocate a wider excision. ^{58, 105, 123–125} One prospective database analysis showed, using Mohs micrographic surgery (MMS), that a 6 mm excision margin confers a 86% clear margin rate, while a 9 mm margin allows for a 98.9% margin rate.¹²⁴ Several other studies show similar results for both melanoma in situ and lentigo maligna, advocating margins up to 1.2 cm for head and neck tumors. Agreement does exist regarding the need for a wider margin than 0.5 cm for lentigo maligna, a subtype of melanoma that has a tendency for subclinical peripheral tumor extension.¹²⁶ Table 11 summarizes the current recommendations for margins in primary cutaneous melanoma.

Table 11.

Recommended surgical margins for primary cutaneous melanoma.

Tumor thickness	Surgical margins*
In situ	0.5–1 cm**
≤1 mm	1 cm
1–2 mm	1–2 cm
>2 mm	2 cm

^*Surgical margins are considered as measured during surgery from the edge of the lesion or post-excisional biopsy site.

^**For large or poorly defined melanoma in situ, lentigo maligna type, a surgical margin or more than 0.5 cm can be considered.

Consider histologic margin assessment prior to reconstruction and closure (Adapted from the Cutaneous Melanoma NCCN Guidelines, Version 4.2020).¹⁰⁵

Surgical Technique

Modern surgical techniques and excision margins usually allow for primary wound closure. More complex reconstructive techniques, such as skin grafting or tissue flaps, may be required for larger defects or those in challenging areas such as the face, scalp, or neck. In these cases, it is recommended to use grafts from the contralateral side in order to minimize the risk of local recurrence due to microscopic in-transit disease.¹¹¹ It is recommended that excision is carried down to the depth of the underlying muscular fascia, but not necessarily including the fascia.⁵⁸ No randomized trials have addressed the question of depth of resection or whether excision to the deep adipose layer is adequate.

The use of MMS in melanoma surgery, commonly complemented by melanoma antigen recognized by T-cells 1 (MART-1) immunostaining, is advocated by some authorities as an alternative for WLE, mostly in the dermatological literature. It is advocated not only for melanoma in situ and lentigo maligna, but also for invasive melanoma of up to 1 mm thickness. Several retrospective studies show MMS to be equivalent to WLE in terms of OS, MSS, and local recurrence in these patients, with some studies claiming superiority in terms of survival, especially in head and neck melanomas.^127–130 It has also been shown that MMS is used more frequently than WLE in the face as well as in the scalp and neck areas, and several authors advocate its use in these areas where tissue conservation is critical.^{130, 131}

Lymph Node Evaluation

The subject of immediate elective dissection of regional lymph nodes (ELND), in clinically node-negative melanoma patients, was an issue of controversy throughout the 1960s and 1970s. Although most authorities agreed that dissection was not necessary for patients with thin tumors with low-risk features, the definitions of thin and low-risk were debated. A prospective clinical trial conducted by the WHO, comparing ELND with follow-up and lymphatic dissection upon clinically detectable nodal involvement failed to show a difference in survival between the 2 groups, deeming observation a viable option in node-negative patients. This study also showed that only 20% of patients had lymphatic involvement, with lymph node involvement being directly related to melanoma thickness. This finding stressed that 80% of patients were unnecessarily exposed to the potential morbidity of lymphatic dissection, most notably lymphedema, with no benefit to their survival.^132–134 Nevertheless, a survival advantage of 11% was found with ELND in the subgroup of patients with intermediate Breslow thickness lesions who had lymphatic involvement, either discovered upon ELND or diagnosed clinically during observation.^{134, 135}

The need to identify this subset of patients with occult lymphatic disease led to the implementation of SLNB into the management of stage I and II melanoma. The technique, introduced by Morton and colleagues¹³⁵ in 1992, incorporated preoperative planar lymphoscintigraphy with technetium-labeled dextran in ambiguously located tumors, such as the trunk, to designate the target lymphatic draining basin. Intraoperatively, blue dye was injected into the area of the tumor and the lymphatic channels were followed using meticulous dissection in the expected lymphatic basin until the sentinel lymph nodes were identified and excised. Later, the use of 99mTc-labeled albumin for intraoperative radiolymphoscintigraphy was added to the technique, enabling dual tracer identification of sentinel nodes. Morton was able to show the accuracy of the technique, with a false negative rate of 1%, establishing SLNB as a central part of the management of stage I and II melanoma. The technique was incorporated into the AJCC staging system in 2001.^{103, 135}

The first Multicenter Selective Lymphadenectomy Trial (MSLT-1) investigated the potential contribution of SLNB to the detection of occult lymph node metastases in patients with intermediate thickness melanoma (1.2 to 3.5 mm). This multicenter trial randomized patients to undergo WLE with SLNB or WLE with observation. In the SLNB arm, if the biopsy was positive, the patient underwent CLND. In the observation arm, if the patient developed clinically evident nodal disease, they then also returned to the operating room for CLND. The 5-year DFS rate was higher in the SLNB group (78.3% vs 73.1%, respectively, p=0.009), with similar 5-year MSS rates (87.1% vs 86.6%, respectively), findings that persisted in the final analysis. Nevertheless, much like the WHO trial, within the subgroup of patients with nodal metastases, patients who underwent SLNB with CLND had improved 5-year survival compared to those who had lymph node dissection (LND) upon clinical recurrence (72.3% vs 52.4%, respectively, p=0.004). In addition, SLNB was shown to be an important staging modality, as patients with nodal metastases upon SLNB were restaged as stage III and had significantly worse survival compared to those with a negative SLNB (72.3% vs 90.2%, respectively, p<0.001).¹³⁶ Thick (>3.5 mm, in this study) melanomas were also evaluated, demonstrating similar 10-year DFS (50.7% vs 40.5%, in the SLNB+CLND vs. observation, p=0.03). Again, patients with positive SLNB had worse 10-year MSS compared to negative SLNB (62.1 vs 85.1%, p<0.001), further reinforcing the role of SLNB for staging purposes.¹³⁷

These findings led international clinical and surgical oncology societies to recommend the use of SLNB in intermediate thickness melanomas, as well as in thick melanomas.^138–141 The utility of SLNB in thin melanomas (≤1 mm) remains controversial. Although the incidence of nodal involvement in this group is low (approximately 5%), the majority (70%) of melanomas diagnosed in the United States are thin. As a result, a considerable number of melanoma deaths occur in patients who originally were diagnosed with thin melanoma. Subgroup analysis has demonstrated that the main predictor of SLNB positivity in thin melanoma is tumor thickness. Among patients with a tumor thickness of less than 0.75 mm, the rate of SLNB positivity is 2.7%, whereas for those with a thickness of 0.75 to 1 mm, it is 6.2%.¹⁴² Other risk factors do not predict sentinel node positivity as reliably. Among these risk factors are ulceration, mitotic rate, lymphovascular invasion (LVI), Clark level, tumor vertical growth phase, TILs, tumor regression, age, and others. The recommendation of most professional societies is to consider SLNB in melanomas 0.8 to 1 mm thick and those less than 0.8 mm thick with high-risk features, including ulceration, high mitotic rate and LVI, especially in younger patients.^{58, 105, 140, 141} Table 12 describes the recommendations for SLNB.

Table 12.

Recommendations for sentinel lymph node biopsy.

Breslow thickness + other characteristics	NCCN recommendation*	ASCO + SSO recommendations**	% positive SLN
Stage IA (T1a): <0.8 with no ulceration	Not recommended, unless uncertain about adequacy of biopsy staging	Not recommded	<5%
Stage IB (T1b): <0.8 with ulceration or 0.8–1 mm with no ulceration T1a with adverse features (high mitotic index≥2/mm2 [in setting or young age], LVI, or a combination)	Discuss with patient and consider SLNB	Discuss with patient and consider (does not include T1a tumors with adverse features)	5–10%
Stage IB (T2a) or II: >1 mm	Offer SLNB (unless non-mitogenic or older patients, with a lower probability of positive SLNB)	Recommended for T2 and T3 tumors For T4 tumors, may be recommended	>10%

SLN, sentinel lymph node; SLNB, sentinel lymph node biopsy; LVI, lymphovascular invasion; NCCN, National Comprehensive Cancer Network; ASCO, American Society of Clinical Oncology; SSO Society of Surgical Oncology.

If patient is unfit or unwilling to act based on SLNB findings, reasonable to forgo SLNB.

^*Adapted from the Cutaneous Melanoma NCCN Guidelines, Version 4.2020.¹⁰⁵

^**Adapted from the ASCO and SSO clinical practice guideline update.¹⁴⁰

Completion Lymph Node Dissection in Patients with a Positive Sentinel Lymph Node Biopsy

With the implementation of SLNB in the management of stage I and II melanoma, 5% to 40% of patients will consequently be upgraded to stage III disease. MSLT-I demonstrated that these patients benefited from a SLNB complemented by CLND more than they did from observation. However, within the MSLT-I cohort, 88% of patients who had a single positive sentinel lymph node had no additional positive nodes upon CLND.⁴⁹

In light of these findings, 2 major randomized trials set out to understand the value of CLND after a positive SLNB. First was the German Dermatologic Cooperative Oncology Group DeCOG-SLT trial, which was ended early due to low accrual and low event rate. It randomized 483 patients with at least 1 mm thick melanomas and a positive SLNB to undergo CLND or observation. Its results, published in 2016, showed a similar 3-year distant metastasis-free survival (DMFS) between the 2 groups (77.0% in the observation group and 74.9% in the CLND group, p=0.87). Three-year OS and recurrence-free survival (RFS) were likewise similar. Furthermore, although the observation group had no serious adverse events, 14% of the CLND group suffered from grade 3 or 4 events. Therefore, the authors concluded that CLND can be omitted in these patients; but as 66% of the cohort had only up to 1 mm micrometastases in the sentinel lymph node, the authors limited their recommendations to such patients.¹⁴³

The second such trial was the MSLT-II trial, which also randomized patients with a positive SLNB (including micrometastatic disease identified with molecular techniques) to CLND or observation with serial nodal US. Intention-to-treat analysis results were similar to those of the DeCOG-SLT trial, with comparable 3-year MSS of 86% in both groups. The authors noted that CLND conferred a favorable DFS, related to an improved rate of disease control in the regional nodes, but lymphedema rates were considerably higher in the CLND arm (24.1% vs 6.3%). The authors concluded that although CLND improves regional control and provides additional staging information, it comes at the cost of an increased lymphedema rate, and confers no benefit in MSS. Therefore, observation with serial nodal US in SLNB-positive patients is a safe option for patients who can reliably undergo such surveillance. One caveat to these findings is that CLND identified additional positive nodes in 11.5% of patients, a finding that was a strong predictor of recurrence. This subset of patients derives benefit from CLND.¹⁴⁴ The results of these 2 trials led to the overwhelming acceptance by both the NCCN and American Society of Clinical Oncology (ASCO)/Society of Surgical Oncology (SSO) of observation as an acceptable treatment option for low-risk patients with a positive SLNB.^{105, 141} Patients potentially excluded from these recommendations are those with what can be considered as high-risk features, such as extracapsular extension of their lymph node metastases, significant tumor burden in the sentinel lymph node exceeding 10 mm, microsatellitosis of the primary tumor, more than 3 involved nodes, more than 2 involved nodal basins, and patients on immunosuppression.

Adjuvant Therapy

Surgery remains the mainstay of treatment in early-stage cutaneous melanoma. However, patients with stage IIA through IIC disease are noted to have a 10-year MSS of 88% to 75%. In comparison, stage IIIA melanoma has a 10-year MSS of 88%, and stage IIIB a 77% MSS.⁶⁴ With the tremendous effect that the introduction of immune checkpoint inhibitors and BRAF-directed therapy has had on the treatment paradigms and outcomes of metastatic melanoma and stage III disease, the role of these novel agents in the adjuvant setting for stage II melanoma is being investigated.^145–148 Several RCTs are currently investigating this question. The largest is KEYNOTE-716 (Safety and Efficacy of Pembrolizumab Compared to Placebo in Resected High-Risk Stage II Melanoma), an international, randomized, phase III trial testing the efficacy of 1 year of pembrolizumab in stage IIB and IIC patients.^{97, 149}

Gene Expression Profiling in Melanoma

In recent years, several commercial GEP tests have become available. Much like the genomic assays for breast cancer, they aim to differentiate low clinical risk patients into low and high genomic risk patients, helping to more accurately discern these patients’ actual risk of recurrence and using that information to guide treatment and surveillance recommendations. The 2 available GEP tests are the 31-gene expression profile DecisionDx-Melanoma^™ (Castle Biosciences; TX, USA), which is the most commonly available assay in the United States, and the 8-gene expression profile MelaGenix (NelaCare; Koln, Germany). There are several other GEPs which are still in clinical development.

The DecisionDx-Melanoma uses 28 signature genes and 3 control genes to classify tumors into low-risk (class 1) or high-risk (class 2) for recurrence, with a sub-classification into A and B subgroups.¹⁵⁰ Several studies have demonstrated the association between a class 2(B) score and worse RFS and DMFS, compared to a class 1(A) score.^151–154 This association has been shown to be statistically significant and independent of the traditional clinical variables of Breslow thickness, ulceration, mitotic rate, and SLNB status. One trial has gone further and shown an association between DecisionDx score and SLNB-positivity, showing 55 to 64 and ≥65-year-old patients with a class 1A score to have 4.9% and 1.6% rate of SLNB-positivity, respectively, and class 2B patients to have 30.8% and 11.9% SLNB-positivity, respectively. These findings point to the potential of GEPs to affect treatment decisions, identifying patients in which SLNB might be avoided or recommended.¹⁵⁵

In spite of these early findings, the NCCN and the American Academy of Dermatology discourage the use of any GEP outside of a clinical trial.^{58, 105} Nevertheless, an estimated of 5% to 10% of cutaneous melanomas are already being subjected to GEP testing.¹⁵⁰ One survey shows that more than 20% of “pigmented-lesion experts” are using one of the GEPs, with 63% indicating that their management of patients is affected by the results.¹⁵⁶ Questioning this premature use of GEPs, a meta-analysis conducted by Marchetti and colleagues¹⁵⁷ points to the limited ability of both DecisionDx-Melanoma and MelaGenix to predict recurrence in localized, stage I and II melanoma. It demonstrates that the GEPs were able to accurately classify patients with recurrence in 76% to 82% of stage II and only 29% to 32% of stage I cases.¹⁵⁷ Another criticism regards the clinical application of some of these recent findings regarding DecisionDx, emphasizing the low sensitivity and positive predictive value of the test for T1 disease, resulting in only 1% of T1 patients potentially benefiting from the test, while 13% will be falsely identified as either low- or high-risk.⁹² Many authorities further question the clinical implication of GEP results, even for accurately identified high-risk patients, pointing to the limitation of diagnostic imaging in detecting pre-symptomatic recurrence, the unproven benefit of treating an asymptomatic recurrence, and the lack of evidence regarding benefit of adjuvant systemic treatment in these patients.

Nevertheless, with time and further research, these assays have the potential to help identify the 2% of T1a patients who will relapse and die of their disease, as well as the 5% of patients with a negative SLNB who will experience a recurrence. They might eventually assist in management decisions regarding these patients, allowing for escalation or de-escalation of treatment according the GEP score, guiding wider or narrower margins for excision, deciding upon the necessity of SLNB, as well as informing decisions concerning adjuvant systemic treatments and patient surveillance.

In summary, there have been great advances in the surgical treatment of early-stage melanoma in recent decades. There has been considerable de-escalation both in the extent of surgery for the primary tumor, with the narrowing of recommended resection margins, as well as in the extent of surgery of the lymphatic basin, with the introduction of SLNB and the selective omission of CLND from a subset of lymph node positive patients. Recommendations continue to evolve as further subsets of patients are identified whose risk for recurrence calls for more or less aggressive surgical and/or systemic therapy. Promising adjuncts to this process include the emerging GEP tests, as well as tumor molecular testing. By characterizing tumors on the molecular level, we might be able to better identify patients who will benefit from further escalation or de-escalation of treatment, as well as to assign targeted therapies to selected patients.

TREATMENT OF REGIONALLY ADVANCED MELANOMA

The hallmark of stage III melanoma is nodal involvement. Previously divided into 3 groups, the 8^th edition of the AJCC staging for melanoma now classifies stage III melanoma into 4 subsets (IIIA, IIIB, IIIC, and IIID) based on histopathologic factors: the nodal involvement (number of nodes and microscopic versus macroscopic disease) and tumor characteristics (depth and the presence of ulceration).⁹⁹ Stage III disease represents a heterogeneous group of patients with regional microscopic or macroscopic nodal disease with or without in-transit or satellite metastasis. As a result of this marked heterogeneity of disease, the outcomes also encompass a wide range of DSS, from 88% 10-year DSS in stage IIIA to 24% for stage IIID disease.⁶⁴ Stage III subcategorization introduced by the AJCC 8^th edition now allows for more accurate evaluation of outcomes and prognosis for patients with stage III disease. Appropriate risk stratification is essential, because patients with macroscopic nodal disease had very poor outcomes prior to the introduction of modern systemic therapy. For example, these patients previously had a greater than 70% chance of recurrence or death at 5 years, and a 9% 5-year OS rate.^{101, 115, 158}

Stage IIIA encompasses non-ulcerated tumors less than 2 mm thick or ulcerated tumors less than 1 mm thick, with microscopic nodal disease in 3 or fewer lymph nodes. Stage IIIB disease includes non-ulcerated tumors less than 4 mm thick, ulcerated tumors less than 2 mm thick with nodal disease in 3 or fewer lymph nodes with at least 1 clinically palpable node, or satellite or in-transit tumors. Stage IIIC disease is comprised of non-ulcerated tumors less than 4 mm thick or ulcerated tumors less than 4 mm thick with either: in-transit disease plus microscopic nodal disease, microscopic nodal disease in 4 or more nodes, 2 or more palpable lymph nodes, or clumped lymph nodes. Stage IIID disease involves ulcerated tumors greater than 4 mm thick with lymph node involvement in more than 3 nodes, clumped lymph nodes, palpable lymphatic disease in 2 or more nodes, or nodal disease plus in-transit disease.

Therapeutic Lymph Node Dissection

The 8th edition of the AJCC staging breaks down lymph nodes into microscopic and macroscopic (ie, occult versus clinically detectible). This distinction is important since the management of the 2 entities is diverging. Although it was previously believed that CLND improved survival and outcomes for all node positive patients, more recent studies have shown that in the setting of microscopic disease there is no survival benefit to CLND.^{143, 144} However, those patients with high-risk features such as primary tumor microsatellitosis, extracapsular extension, greater than 2 lymph node basins involved, and greater than 3 involved nodes, potentially should be considered for CLND to help with locoregional disease control.¹⁴¹ In summary, CLND may be used selectively based on risk stratification in patients with micrometastatic disease.

For patients with microscopic disease only, 2 recent trials—the German Dermatologic Cooperative Oncology Group (DeCOG-SLT) and the Multicenter Selective Lymphadenectomy Trial-II (MSLT-II)—have been crucial in the movement from CLND to observation alone.^{143, 144} The DeCOG-SLT is a multicenter, randomized, phase III clinical trial that looked at survival in CLND versus observation alone in patients with positive SLNB.^{143, 159} At median follow-up of 72 months, there was no significant difference in the 5-year RFS, DMFS, OS.¹⁵⁹ Similarly, the MSLT-II trial, an international, randomized, phase III trial, also evaluated the 3-year MSS, DFS, OS, and DMFS between CLND and observation alone.¹⁴⁴ The MSLT-II trial showed no significant differences in mean 3-year MSS, DMFS, or OS between the 2 groups. However, unlike the DeCOG-SLT trial, there was a significant difference in the 3-year DFS (68% in CLND group vs. 63% in the observation group, p=0.05). Furthermore, the investigators also noted a higher rate of non-sentinel lymph node (NSLN) recurrence in patients in the observation arm, and further analysis revealed this to be an independent prognostic factor for melanoma-related death.¹⁴⁴ As such, MSLT-II supports the role of CLND for diagnosing NSLN metastasis, the presence of which serves as a prognostic factor for melanoma-related death. It does not, however, show any additional survival benefit in those patients diagnosed with NSLN disease (despite achieving a 70% relative decrease in NSLN recurrence). In short, for patients with metastatic disease identified on SLNB, there is a role for observation alone, involving frequent clinical follow-up with visits every 4 months for 2 years, then every 6 months between years 3 and 5, with annual US nodal evaluation for 5 years. However, the option for observation in lieu of CLND is not recommended for those who are unable to keep this intensive follow-up regimen, or those who have clinically evident nodal disease. Of note, both DeCOG-SLT and MSLT-II trials excluded patients with in-transit disease, satellite disease, and locoregional metastasis.

For those patients with clinically evident nodal disease usually consisting of a tumor deposit of 10 mm or greater, CLND is still recommended as standard practice, although this recommendation is in evolution given some early promising results with neoadjuvant treatment approaches in this group of patients (see below). For patients with positive axillary nodes, the recommendation is excision of levels 1 through 3. For positive inguinal lymph nodes, there remains some controversy about superficial versus deep groin dissection. A 2011 retrospective study by van der Ploeg and colleagues¹⁶⁰ showed no difference in DFS between patients undergoing superficial versus deep groin dissection, with greater incidence of chronic lymphedema in the deep groin dissection group. In this study, CT had high negative predictive value for the detection of pelvic nodal involvement (91%). Thus, the current recommendation is for superficial groin dissection for those patients without pathologic iliac nodes on CT. Deep inguinal dissection should be reserved for patients with positive deep nodes on CT or multiple positive nodes on superficial dissection.

Neoadjuvant Therapy—Systemic Immunotherapy

Stage III disease can be conceptually divided into resectable or unresectable disease. For those with resectable disease, the current treatment guidelines include surgical resection with or without adjuvant therapy consisting of immunotherapy and/or targeted kinase inhibitors in patients with BRAF-mutant disease.¹⁶¹ Recently, there have been several studies suggesting that a neoadjuvant treatment approach with immunotherapy and/or targeted therapies may improve outcomes. The use of neoadjuvant therapy in melanoma may provide several advantages, including reducing surgical morbidity by reducing the disease burden, as well as providing the opportunity to evaluate pathologic response to therapy, which can be predictive recurrence risk. Several trials have shown changes in expression of PD-1, PD-L1, and CD8 that are associated with tumor response to therapy and recurrence.^162–165 Furthermore, the use of immunotherapy prior to resection of the tumor (and its associated neoantigens) may potentially induce a stronger tumor-specific T-cell response.

Several phase I studies have demonstrated the potential efficacy of neoadjuvant therapy in stage III melanoma using pembrolizumab, nivolumab, and ipilimumab, either alone or in combination.^163–165 The OpACIN phase Ib trial randomized patients with stage III melanoma to adjuvant or neoadjuvant therapy using combination ipilimumab and nivolumab.¹⁶⁵ Of the 10 patients in the neoadjuvant treatment arm, 7 patients had a pathologic response (3 with pathologic complete response [pCR], 4 with partial response). Of note, none of the patients who achieved a pCR had recurred by 30 months, whereas the 2 patients in the neoadjuvant arm who did not have a pCR had recurrent disease. At 30 months, the RFS was 80% in the neoadjuvant group and 60% in the adjuvant group, with a 30-month OS of 90% for neoadjuvant therapy and 67% for adjuvant therapy. Adverse events were common in this study, with 18 of 20 (90%) patients experiencing 1 or more grade 3 or 4 toxicity events.¹⁶³

The success of early studies paved the way for subsequent phase II studies, which have also shown promising results with use of neoadjuvant therapy in stage III melanoma.^{162, 166–168} In a phase II randomized study, Amaria and colleagues¹⁶² showed that neoadjuvant immune checkpoint blockade with combination ipilimumab and nivolumab resulted in high response rates, with pCR of 45%. Single agent therapy with nivolumab resulted in pCR of 25%, but with fewer adverse events. As with the OpACIN phase Ib study, there was a high rate of toxicity with combination therapy, with 73% of patients experiencing grade 3 or higher events, as opposed to the 8% seen with single agent therapy.¹⁶² Based on the high rate of toxicity observed in their initial studies,¹⁶³ the OpACIN-neo trial, a multicenter, phase II RCT identified the optimal dosing and schedule of ipilimumab plus nivolumab to provide effective but less toxic therapy.¹⁶⁸ This study identified the regimen of ipilimumab 1 mg/kg and nivolumab 3 mg/kg as optimal with regard to response rates and rates of adverse events.¹⁶⁸

In addition to studying the use of neoadjuvant immunotherapy, there are several studies that have evaluated the use of neoadjuvant targeted kinase inhibitors. Two trials looking at the use of neoadjuvant dabrafenib (BRAF-inhibitor) and trametinib (MEK-inhibitor) in patients with BRAF-mutated melanoma have shown positive outcomes.

First, NeoCombi, a single-arm phase II study, reported outcomes of 35 patients with BRAF-mutated melanoma receiving combination dabrafenib plus trametinib for 12 weeks followed by resection and subsequent 40 weeks of adjuvant therapy. At the time of surgery, all 35 patients had some pathologic response, with 49% having pCR. The 1-year RFS was 82% in those patients who achieved a pCR, compared to 63% in those who did not achieve pCR.¹⁶⁹

In a second single center, phase II randomized trial, Amaria and colleagues¹⁷⁰ compared the use of neoadjuvant and adjuvant combination dabrafenib plus trametinib with the standard of care (surgery followed by adjuvant therapy). In the neoadjuvant group, patients received 8 weeks of neoadjuvant therapy followed by surgery and then up to 44 weeks of adjuvant therapy. This trial was stopped early after the interim safety analysis revealed significantly longer survival in patients who received neoadjuvant therapy, and then continued as a single arm study. At a median follow-up of 18.6 months, 71% of the 14 patients treated with neoadjuvant therapy were alive without disease progression, and none of the 7 patients treated with standard of care were alive.¹⁷⁰

To better characterize the impact of neoadjuvant therapies, the International Neoadjuvant Melanoma Consortium performed a meta-analysis of 6 trials of neoadjuvant therapy (4 with immunotherapy, and 2 with targeted therapy). After pooling the data, there was a significant difference in outcomes seen with those patients who achieved pCR versus those who did not; there was a 12-month RFS of 95% in those with pCR versus 62% in those without, and 2-year RFS of 89% in those with pCR versus 48% in those without. Overall, the response associated with neoadjuvant immunotherapy appeared to be more durable than that seen with targeted therapy. Although a similar percentage of patients achieved pCR (38% with immunotherapy and 47% with targeted therapy), patients with pCR following targeted therapy had a 25% recurrence rate at 2 years, whereas those with immunotherapy had 0% recurrence. Furthermore, RFS was superior in patients treated with immunotherapy compared to targeted therapy, at 83% versus 45% at 2 years, respectively.¹⁷¹ In summary, this study identified pCR as the most important prognostic factor after neoadjuvant therapy. Responses with immunotherapy were more durable than those achieved with targeted therapy.

Treatment of In-transit Disease and Locally Advanced Unresectable Disease

Some patients with stage III melanoma have unresectable disease, thus presenting a unique therapeutic challenge. The prognosis of patients with non-nodal locoregional metastasis is similar to that of patients with clinically detectable nodes.⁶⁴ In-transit disease, which occurs in 3% to 10% of melanomas, is disease that occurs between the primary site and the regional nodal basins.¹⁷² It most commonly manifests as multiple tumor nodules but may also be present as a single site of disease. In-transit disease is most common in the extremities, but also occurs on the trunk, head, and neck. Location, depth of invasion, and ulceration are all factors that are known to contribute to the risk of development of in-transit disease. For example, the lower extremity is a common location.¹⁷³

The approach to treatment must consider the overall condition of the patient, the anatomic site, extent of locoregional disease, and the presence or absence of disseminated metastases. There are several options for the treatment of in-transit disease that range from surgical excision to regional treatment and now also include several effective systemic therapies. Part of the challenge is that the definition of resectable disease had not been clearly established. Many surgeons feel that the presence of 3 or more lesions or resection of a single lesion after which the skin cannot be closed primarily should merit consideration of nonsurgical therapy as a first step. Given the significant advances and effectiveness of systemic therapy, the role of regional therapy and surgery in this group of patients is rapidly changing. Although regional treatments such as limb perfusion and limb infusion were once destination therapies, they have now been replaced as front-line treatments by systemic and/or intralesional treatments. Intralesional therapy may be conceived of as an adjunct to systemic therapy, with the aim to help attain locoregional control and/or to alter the regional tumor microenvironment to help augment systemic immunotherapy. Isolated limb perfusion (ILP) and isolated limb infusion (ILI) are currently reserved for salvage therapy or for patients with no other treatment options.

Surgery

Surgery can have a role up front to remove small volume lesions if morbidity is felt to be minimal. Most of these patients should then be considered for adjuvant systemic therapy in the form of immunotherapy based on the results of the checkmate 238 trial¹⁴⁸ or targeted BRAF/MEK inhibition based on the results of the COMBI-AD trial.¹⁴⁷ Surgery may also have a role in the removal of refractory metastatic lesions when there is evidence of response to systemic therapy elsewhere in the body. Palliative procedures may be considered for a lesion that is large, necrotic, and draining fluid.^{174, 175}

When regional in-transit disease is deemed to be unresectable or recurrent, multidisciplinary discussion is indicated to consider optimal sequencing of systemic therapy versus regional chemotherapy (ILP or ILI) or intralesional treatment (talimogene laherparepvec [T-VEC]). Increasingly, a systemic neoadjuvant approach is being applied to these patients based on preliminary studies that suggest high response rates for both targeted therapies¹⁷⁰ and immunotherapies.^{162, 176} Use of regional chemotherapy with either ILP or ILI using melphalan requires a center with adequate expertise in this technique. There are data to suggest that a combination of regional therapy followed at some point by checkpoint blockade immunotherapy has better regional and systemic disease control than regional chemotherapy alone.^{177, 178}

Isolated Limb Perfusion

One of the first therapies for the treatment of in-transit disease was hyperthermic isolated limb perfusion. First described in 1958 by Creech and colleagues,¹⁷⁹ ILP involves the dissection and cannulation of the major vessels in either the upper or lower extremity, followed by isolation of the extremity from systemic circulation using a tourniquet and the instillation of heated chemotherapy. The heated chemotherapy (most often melphalan) is then circulated through the extremity at 38° to 42° C at a rate of 300 to 500 L/min.¹⁸⁰ A meta-analysis from 2010 evaluated the outcomes in 2,018 hyperthermic ILPs performed between 1990 and 2008. This analysis showed clinical response rates between 64% and 100% (median, 90%), with complete response (CR) rates ranging from 25% to 89% (median, 58%), with a limb salvage rate of 95%. Patients in this study had a median OS of 36.7 months, with a median 5-year survival of 36.5%.¹⁸¹ Additional European studies showed improved responses with the addition of TNF-alpha to melphalan, however a 2006 phase III trial (ACO-SOG Trial Z0020) showed no significant difference between melphalan alone and melphalan plus TNF-alpha in either overall response rate (ORR) or CR at 3 months, but did show a higher complication rate with the addition of TNF-alpha.¹⁸² Therefore, TNF-alpha is not routinely used in the United States. Although highly effective, hyperthermic ILP does carry significant toxicity as measured on the Wieberdink toxicity scale (toxicity ranges from grade I - no change in the affected limb, to grade V - severe tissue damage requiring amputation). Although the rates of severe toxicity (grade IV or higher) are relatively low (with limb loss reported in <3.3% of patients), the long-term morbidity including limb edema, muscle atrophy, stiffness, and functional impairment, is as high as 44%.^183–186

Isolated Limb Infusion

Similar to hyperthermic ILP, ILI also involves isolation of the limb with a tourniquet followed by infusion of hot melphalan. However, unlike hyperthermic ILP, it does not involve dissection of the vessels, and does not involve perfusion of the limb. Instead, it is performed without an oxygenator and, as such, the extremity becomes acidotic and hypoxic during drug circulation. First described by Thompson and colleagues¹⁸⁷ in 1998, ILI has been shown to have less toxicity than ILP, with studies showing 97% of patients having grade III toxicity or less, with comparable response rates.¹⁸⁸ A multicenter analysis performed by Miura and colleagues¹⁸⁹ in 2019 evaluated long-term outcomes for patients with stage IIIB or IIIC melanoma who underwent first-time ILI, across 9 different Australian and US tertiary centers, between 1992 and 2018. Out of the 687 first time ILIs performed during the study period, only 3.9% of patients developed grade IV toxicity and there were no grade V toxicities noted. Favorable response to ILI was associated with significantly longer progression-free survival (PFS) (53.6 vs. 12.7 months; p<0.0001) and OS of 46.5 vs. 24.4 months.¹⁸⁹ Furthermore, because it does not involve dissection of the vessels, ILI can be repeated multiple times, unlike ILP. For patients with recurrence within 3 months of ILI, ILP is recommended as an escalation of care. For patients with recurrence more than 3 months after ILI, either repeat ILI or ILP are acceptable options.¹⁹⁰

Intralesional Therapy

Although ILI and ILP are useful for extremity lesions, in-transit and locally advanced disease can also occur on the trunk, head, and neck, where isolation of the vasculature is not feasible. Intralesional therapy is useful for disease not located on an extremity and for patients who are ineligible for operative resection. There are multiple different intralesional therapies that have been used over the years to treat in-transit disease and this review is by no means inclusive. One of the first intralesional therapies was bacillus Calmette-Guerin (BCG). First described by Zbar and colleagues¹⁹¹ in 1971 as having an effect on tumor growth, this live, attenuated strain of M. bovis was one of the most commonly used treatments for in-transit disease. Early studies by Morton and colleagues¹⁹² showed that 90% of lesions injected with BCG had a CR, and reported regression in 17% of uninjected lesions, suggesting an abscopal response.¹⁹³ Although early studies reported the efficacy of BCG injections in the treatment of melanoma,^193–195 later studies did not show the same promise. The Eastern Cooperative Oncology Group (ECOG) phase III randomized trial (E1673) in 2004 looked at the effect of BCG on resected stage I, II, and III melanoma and found no significant difference in DFS or OS.¹⁹⁶ In addition to questionable impact on OS, BCG injections carry a 12% complication rate and the use of BCG injections requires close monitoring and pre-injection prophylaxis with isoniazid. The toxicity of BCG ranges from induration at the injection site to severe sepsis and even death.¹⁹⁷ For these reasons, BCG is now rarely used.

The history of intralesional therapy for melanoma also includes injection of IL-2. A glycoprotein affecting growth, proliferation, and activation of cytotoxic T lymphocytes, NK, and lymphokine-activated killer cells, IL-2 was originally approved for systemic therapy in metastatic renal cell carcinoma and melanoma.^{198, 199} However, systemic administration of IL-2 resulted in significant toxicities and the response rates only ranged from 10% to 15%.^{198, 200} Intralesional use of IL-2 was based on several phase II trials^{201, 202} and a subsequent systematic review,²⁰³ which showed CR rates in 62% to 70% of injected lesions, with a durable response of at least 6 months with minimal toxic effects. There was, however, no abscopal effect observed with intralesional IL-2.

PV-10 (Rose Bengal) is another intralesional therapy that has been used for the treatment of in-transit disease. A xanthene dye most commonly used for detection of ophthalmic injury, it was first used for melanoma in a phase I trial in 2008 by Thompson and colleagues²⁰⁴ In this small study of 11 patients, injection of melanoma lesions with PV-10 resulted in CR in 36% of injected lesions and 15% of non-injected lesions. A subsequent phase II study in patients with refractory metastatic melanoma showed a CR rate of 26% and a partial response rate of 25% after 4 injections, with 33% of patients experiencing an abscopal response in untreated lesions.²⁰⁵

The newest therapy for intralesional injections is talimogene laherparepvec (T-VEC), the only oncolytic viral therapy approved by the FDA for treatment of locoregionally advanced melanoma. Derived from a genetically modified herpes simplex virus (HSV-1), it has 2 putative mechanisms of action: selectively infecting tumor cells and promoting immune-mediated destruction.²⁰⁶ Intratumoral injection of T-VEC in a phase II trial comprised of patients with stage IIIC/IV melanoma showed an ORR of 26% in both injected and non-injected lesions, with a durable response of 7 to 31 months in 92% of patients who responded.²⁰⁷ The OPTiM trial, a phase III trial by Andtbacka and colleagues,²⁰⁸ compared intratumoral injections of T-VEC and granulocyte-macrophage colony-stimulating factor (GM-CSF). This trial showed a higher durable response rate in the T-VEC group compared to the GM-CSF group, with a longer median OS (23.3 months vs. 18.9 months, p=0.051).²⁰⁸

Adjuvant Therapy

Immunotherapy

Following resection, the current standard of care involves the use of adjuvant therapy in stage IIIA melanoma and higher. Similar to neoadjuvant therapy, adjuvant therapy involves the use of immunotherapy and targeted therapies. Ipilimumab, nivolumab, and pembrolizumab are all FDA approved for use in resected stage III melanoma, as is combination dabrafenib and trametinib for patients with BRAF-mutant disease. Ipilimumab, an anti-CTLA-4 monoclonal antibody, was the first to be approved, following the completion of the international phase III European Organisation for Research and Treatment of Cancer (EORTC) 18071 trial, which compared adjuvant high-dose ipilimumab to placebo in patients with completely resected stage III melanoma. In this study, they showed a significant difference in median RFS of 26.1 versus 17.1 months with ipilimumab versus placebo, as well as a difference in 3-year RFS (46.5% in ipilimumab group vs. 34.8% in placebo group).²⁰⁹ Further analysis showed significantly improved RFS and OS at 5 years in the ipilimumab group compared to placebo (40.8% vs. 30.3%, p<0.001 for RFS and 65.4% vs. 54.4%, p=0.0013 for OS).²¹⁰

Nivolumab and pembrolizumab are anti-PD-1 antibodies also approved for use in the adjuvant setting in stage III and IV melanoma. Several studies have been performed confirming the efficacy of nivolumab in the adjuvant setting, including the CheckMate 066 trial, which evaluated the OS in patients treated with nivolumab or dacarbazine. At 1 year, there was a significant difference in OS between the nivolumab and dacarbazine groups (73% vs. 42%, p<0.001).²¹¹ However, nivolumab was not approved for use until after the CheckMate 238 trial, a randomized, double-blind, phase III trial with a primary endpoint of RFS. This study compared patients with high-risk resected stage IIIb, IIIC, or IV melanoma treated with either ipilimumab or nivolumab in the adjuvant setting and followed for up to 1 year or until disease recurrence. Their 12-month RFS was 70.5% versus 60.8% for nivolumab and ipilimumab, respectively, with lower rates of grade III or IV adverse events in the nivolumab group.¹⁴⁸ Pembrolizumab is the most recent immunotherapy approved for adjuvant use in stage III melanoma. Approval was granted following the European Organisation for Research and Treatment of Cancer (EORTC) 1325/KEYNOTE-054 trial, a phase III randomized trial comparing pembrolizumab to placebo in patients with high-risk stage III melanoma. There was a significant difference in 1-year RFS between pembrolizumab and placebo (75.4% vs. 61%, HR 0.57, CI 95%, 0.43–0.74, p<0.001).¹⁴⁵ Similar to nivolumab, pembrolizumab has also been found to be superior to ipilimumab with regard to response rates and toxicity.²¹² There have also been several studies looking at the combination of CTLA-4 and PD-1 inhibitors, demonstrating significant improvements in PFS and OS. The CheckMate 067 trial is what led to the approval of combination ipilimumab plus nivolumab for use in unresectable or metastatic disease. Patients treated with combination therapy had median PFS of 11.5 months, compared with 2.9 months with ipilimumab alone, and 6.9 months with nivolumab alone. Patients with PD-L1 positive tumors had median PFS of 14 months with nivolumab monotherapy; however, in those with PD-L1 negative tumors the response was superior with combination therapy (11.2 months vs. 5.3 months with combination vs. nivolumab monotherapy). Combination therapy resulted in a higher percentage of grade III or IV adverse events (55% in combination vs. 16.3% in nivolumab monotherapy or 27.3% in ipilimumab monotherapy).²¹³ In summary, combination checkpoint blockade therapy improves survival but results in more adverse events.

Targeted Therapies

Targeted therapies have also been approved for adjuvant therapy in stage III melanoma. There are 3 combinations of BRAF and MEK inhibitors currently approved for the adjuvant setting: vemurafenib and cobimetinib, dabrafenib and trametinib, and encorafenib and binimetinib. The COMBI-AD trial is a double-blind, placebo-controlled phase III study in patients with resected BRAF-mutated stage III melanoma comparing RFS in those treated with dabrafenib plus trametinib versus placebo. Initial results showed a significant difference in RFS and 3-year OS between groups. The RFS was 58% versus 39% in combination versus placebo at a median follow-up of 2.8 years, with higher rates of DMFS and RFS in the combination therapy group.¹⁴⁷ More recently, the investigators reported the RFS at a median follow-up of 44 months. The current 3- and 4-year RFS rates were 59% versus 38% in the dabrafenib plus trametinib and placebo groups, respectively. Subgroup analysis performed at that time also favored the combination of dabrafenib plus trametinib over placebo irrespective of baseline factors including stage, nodal disease burden, and tumor ulceration.¹⁴⁶

Combination Immunotherapy and Intralesional Therapy

More recently, there have been studies looking at the combination of immunotherapy with intratumoral therapies, specifically T-VEC. It is believed that T-VEC can help expose tumor-associated antigens and thus prime the anti-tumor immune response.²¹⁴ Several studies have been conducted looking at the combination of T-VEC with ipilimumab as well as pembrolizumab.

Chesney and colleagues²¹⁵ evaluated the combination of T-VEC plus ipilimumab versus ipilimumab alone in a patient with unresectable stage IIIB-IV melanoma. They found an objective response in 38 of 98 (39%) patients treated with T-VEC plus ipilimumab versus 18 of 100 (18%) patients treated with ipilimumab alone, without any significance difference in adverse events.²¹⁵ The MASTERKEY-265 trial, an open-label single-arm phase Ib trial, focused on the combination of T-VEC plus pembrolizumab in patients with unresectable stage IIIB-IV melanoma. At a median follow-up of 36.8 months, the ORR was 67% with a CR of 43% (9/21 patients).^{216, 217} The larger randomized phase III trial (KEYNOTE-034, NCT02263508) comparing T-VEC plus pembrolizumab versus pembrolizumab alone is currently underway.

Radiation Therapy

For patients who are high risk for recurrence, radiation therapy can be useful to help prevent local recurrence, but has not been shown to have an effect on overall long-term survival.²¹⁸ The ANZMTG trial, a prospective multicenter phase III study evaluated the difference in locoregional recurrence in patients treated with regional nodal basin radiation versus observation following lymphadenectomy for a palpable lymph node field relapse. Patients were treated with adjuvant radiotherapy (48Gy in 20 fractions given over a maximum of 30 days) or observation. Radiation did significantly reduce the risk of locoregional recurrence. At a median follow-up of 73 months, there was a significantly lower rate of recurrence in the adjuvant group compared to the observation group (21% vs. 36%, p=0.023). However, OS and RFS did not differ.²¹⁹ Thus, current recommendations are for the use of radiation therapy in patients considered high risk for locoregional recurrence and for patients after failure of previous regional LNDs as salvage therapy. Radiation is also useful as salvage therapy for those patients who are not good surgical candidates. In this situation, radiation therapy is aimed at symptom relief and management.

Conclusions

The heterogeneity of stage III melanoma lends itself to multiple different avenues of therapy. The mainstay remains surgical resection, however newer systemic therapies including immunotherapy and targeted therapies provide benefit in both the neoadjuvant and adjuvant settings. As such, all patients with stage III disease would benefit from presentation at a multidisciplinary tumor board to make sure the optimal treatment strategy is developed based on the patient’s individual situation and the treatment portfolio of the treating institution.

METASTATIC MELANOMA

The last decade has seen a revolution in the treatment of metastatic melanoma. After a quarter century of stagnation and limited therapeutic options that conferred minimal survival benefit, a new understanding of the molecular behavior of melanoma has led to multiple breakthroughs. The development of immunotherapy and targeted therapy has resulted in an improved outlook for patients with stage IV melanoma. Surgery continues to provide a benefit for patients with resectable metastases, and the addition of the new systemic agents in the neoadjuvant or adjuvant setting has shown initial promise. New therapeutic modalities under investigation are exploring ways to harness the power of the immune system through vaccination, novel cytokines, and the administration of autologous tumor-specific immune cells. Multiple ongoing trials seek to build on the promise of the last decade.

Presentation and Prognosis

Cutaneous melanoma is considered to have metastasized and become systemic when it has spread beyond regional lymph node basins. The presence of distant metastases is classified as stage IV disease.²²⁰ The most recent edition of the AJCC staging guidelines for melanoma breaks down metastatic disease into substages based on location of distant metastasis. M1a disease includes metastasis to distant skin, subcutaneous tissue, and non-regional lymph nodes. M1b disease represents lung metastases. M1c disease includes visceral metastases outside of the CNS. Finally, M1d disease represents metastatic disease in the CNS.⁶⁴ The LDH level has proven to be a significant historical prognostic factor in patients with metastatic disease and is again included in the staging system.²²¹ The rates of 1- and 2-year OS in patients with normal LDH levels are twice that of patients with elevated LDH (65% and 40%, vs 32% and 18%, respectively).⁹⁵ The prognostic effect of LDH in patients with stage IV melanoma has persisted even after treatment with modern systemic therapy.^{222, 223} In the most recent AJCC staging, M-stage disease has been further amended within each substage to specify patients with baseline (0) or elevated (1) LDH.⁶⁴ See Table 6 for a summary of the AJCC 8^th edition M-staging criteria.

Of all patients who are diagnosed with cutaneous melanoma, roughly 4% have metastatic disease at the time of presentation.²²⁴ Of those who present with local disease, 11% will ultimately develop distant metastases.²²⁵ The median time to development of distant metastases is just over 2 years.^{225, 226} The only variable that has an influence on survival in those who develop distant metastases is the stage of the primary tumor, suggesting that prognosis and the behavior of the disease is driven by the underlying biology of the primary, even after complete resection.²²⁷ For patients who ultimately go on to develop metastases after presenting with stage I disease, median OS after discovery of metastasis is 29.5 months. The outlook is markedly worse for patients who initially presented with stage II or stage III disease, as their median OS after diagnosis of distant metastases is 15 months.²²⁵ Unlike in other malignancies, the time interval to distant recurrence does not affect survival.²²⁶

For patients who developed metastatic melanoma in the latter half of the 20^th century, the outlook was grim. Of those who qualified for clinical trials, only 12% had a response, and the median OS was 6 to 7 months.^{228, 229} A long-term survey of patients with metastatic melanoma showed that historic site-specific median OS for metastases to the skin, distant lymph nodes, and the gastrointestinal (GI) tract was 12.5 months, with 14% of patients surviving 5 years. This was in agreement with other analyses of national-level data, which showed 5-year OS of 16% for stage IV disease.²³⁰ The prognosis was worse for those who presented with pulmonary disease, with a median OS of 8.3 months and 5-year OS of just 4%. Patients with liver, brain, or bone metastases fared even worse, with a median OS of 4.4 months and 5-year OS rate of 3%. These figures did not change significantly over the 22-year period of study from 1971 to 1993.²²⁸ The most recent sub-stage-specific analysis performed by the AJCC, for the 7th edition of their staging manual, showed 1-year OS of 62% for patients with M1a disease. Patients with M1b and M1c disease (defined per AJCC 7th edition staging as non-pulmonary visceral metastases with normal serum LDH or distant metastases at any site with an elevated serum LDH) had 1-year survival of 53% and 33%, respectively.⁹⁵

The revolution in systemic therapy for melanoma that occurred in the second decade of the 21^st century has rearranged the map for patients with stage IV disease. Although data from individual studies is incredibly promising and it is clear that the development of immunotherapy and targeted therapy improve DFS and OS in patients with metastatic melanoma, the long-term survival data are not yet mature, making it difficult to calculate meaningful stage-specific prognostic information and to chart a course for patients diagnosed with stage IV disease. A broad analysis of national-level survival data recently found that between 2013 and 2016 mortality due to melanoma decreased by 17.9%, due almost entirely to the impact of new systemic therapies that have become available since 2011.²³¹ See Figure 7 for the striking change in mortality rate that occurred shortly after the introduction of immunotherapy. In the context of an incidence rate that has been rising steadily over the last several decades,²³² such a sharp decline in mortality represents one of the great breakthroughs in the history of oncology. So profoundly has the paradigm shifted that for the most recent edition of the AJCC staging manual for melanoma, the AJCC opted not to update their analysis for patients with metastatic disease, given that the results would not be reflective of the therapeutic options currently available and the stage-specific survival data would not reflect the impact of these newer agents.⁶⁴ The stage-specific survival figures from the AJCC 7^th edition⁹⁵ were retained, with the caveat that the continued adoption and proliferation of modern systemic therapy will markedly improve the prognosis for patients with stage IV disease.

Figure 7.

Trends in mortality rate for cutaneous melanoma. (With permission from Berk-Krauss J, Stein JA, Weber J, Polsky D, Geller AC. New Systematic Therapies and Trends in Cutaneous Melanoma Deaths Among US Whites, 1986–2016. Am J Public Health. 2020;110(5):731–733).

Evaluation

Patients who present with concern for metastatic disease should be thoroughly evaluated in concordance with the most recent guidelines.⁹⁴ Given its prognostic importance for stage IV disease, a serum LDH should be drawn. A biopsy should be obtained for confirmation of the diagnosis. An excisional biopsy is preferred. For a distant skin recurrence, punch biopsy is acceptable if an excisional biopsy would be anatomically prohibitive.⁵⁸ If molecular testing for BRAF V600 mutant status has not already been performed on the primary tumor, this should be assessed to determine the patient’s eligibility for targeted therapy. Immunohistochemistry is a readily available method that can be utilized with a high degree of sensitivity and specificity.²³³ Additional genetic testing for BRAF mutations other than V600E should be done with PCR-based methods (eg, for V600K, which also responds to targeted therapy) and for KIT mutations for acral and mucosal melanomas, given the availability of targeted therapy for these otherwise less immunotherapy-responsive melanoma subtypes. Imaging should be obtained to assess for the full extent of distant metastatic disease. No specific imaging modality is recommended, but a meta-analysis has shown that PET-CT is more sensitive and specific than CT alone in the evaluation of metastatic melanoma.²³⁴ Imaging should be driven by most likely sites of metastasis and evaluation of symptoms. PET-CT covering skull vertex to mid-thigh is sufficient except in the case of melanomas arising in the distal lower extremity.

Treatment

Upon completion of a thorough, guideline-directed evaluation, the patient with metastatic melanoma may proceed down one of several treatment pathways depending on the burden and location of the metastases.

Patients with disease that is not amenable to surgical resection should be initiated on systemic therapy. It is developments within this modality that has revolutionized metastatic melanoma over the last decade. Prior to 2011, the primary drug was the traditional chemotherapy agent dacarbazine. It was initially approved in 1975 and persisted not because of its effectiveness (in fact, it confers no long-term response or survival benefit²³⁵) but rather because of the lack of success finding a better alternative.

Immunotherapy

The first wave of immunotherapy began with approval of the cytokine interferon in 1996. The use of this agent has been associated with a marginal RFS benefit in several trials in the adjuvant setting but has limited evidence of improving OS. The high rate of serious side effects has also been a limiting factor in its widespread adoption.²³⁶ IL-2 is a cytokine that was approved by the FDA in 1998. Although there was no difference in ORR compared to cytotoxic chemotherapy, high-dose IL-2 did result in a single-digit percentage of durable responses with improvement in long-term DFS.^{237, 238} Although associated with a high rate of toxicity events, the small possibility of durable CR made this risk more palatable.

A new era of systemic therapy for metastatic melanoma began in early 2011 when the CTLA-4 inhibitor ipilimumab was approved. CTLA-4 is a transmembrane protein that downregulates T-cell activity after binding its ligand B7 and inhibits the immune system’s response to foreign epitopes, such as those associated with cancer cells. When bound by a checkpoint inhibitor like ipilimumab, the inhibitory activity generated by CTLA-4-mediated signaling is lost, thus activating T-cells and allowing them to proliferate and wage a prolonged response against the encroaching malignant cells.²³⁹ Ipilimumab improves survival in metastatic melanoma compared to dacarbazine,²⁴⁰ with a meta-analysis showing 3-year OS of 22%.²⁴¹ A phase III trial comparing ipilimumab monotherapy to both the gp100 tumor-associated antigen vaccine and ipilimumab vaccine combination therapy demonstrated that, in patients with unresectable stage III or IV melanoma that had failed previous therapy, ipilimumab monotherapy resulted in superior 1-year and 2-year OS (45.6% and 23.5%, respectively), with a median OS of 10 months.²⁴² The response rate was low (11%), but 60% of the patients who responded maintained their response for at least 2 years. The overall toxicity rate was high (60%), but only 15% of the adverse events were grade III or higher. A unique feature of ipilimumab’s pharmacodynamic profile is a relatively delayed tumor response, sometimes taking more than 3 to 4 months to show its effects. If the patient undergoes an imaging study within that interval, the mass may actually look to have initially progressed due to an influx of inflammatory cells that have been recruited to the area (termed “pseudoprogression”), before later shrinking.²⁴³ This prolonged development of a response makes ipilimumab a suboptimal choice for patients who present with symptomatic metastases (eg, biliary or bowel obstruction).

Programmed cell death protein 1 (PD-1), a T-cell surface protein, also downregulates the immune system when it binds the ligand PDL-1, a protein that is found on many normal cells as well as expressed by multiple human cancers as a mechanism to evade immune defenses.²⁴⁴ The mechanism of action of pembrolizumab, similar to other checkpoint inhibitors, involves binding PD-1, thereby allowing the immune response to continue and proliferate against cancer cells. Pembrolizumab was approved by the FDA in 2014 for patients who had failed ipilimumab and BRAF mutants who had progressive disease on a BRAF inhibitor. The KEYNOTE-006 trial compared pembrolizumab with ipilimumab in patients with unresectable stage III-IV melanoma. Pembrolizumab was superior across all categories, with a response rate of 33.6% (compared to 11.9% for ipilimumab), PFS at 6 months of 47.3% (compared to 26.5%), 12-month OS of 74.1% (compared to 58.2%), and a rate of grade III or higher adverse events of 13.3% (compared to 19.9%).²¹²

Nivolumab, another checkpoint inhibitor that acts on PD-1, was approved shortly after pembrolizumab. In a phase III trial comparing nivolumab and dacarbazine in treatment-naïve patients with metastatic BRAF wild-type melanoma, nivolumab demonstrated an improved response rate compared to dacarbazine (40% vs 13.9%), superior DFS of 5.1 months (vs 2.2 months), superior 1-year OS of 72.9% (vs 42.1%), and a decreased rate of grade III or higher adverse events at 11.7% (vs 17.6%).²¹¹

Combination therapy with ipilimumab and nivolumab has been shown to provoke objective responses in more than 50% of patients²⁴⁵ and to improve PFS (median 11.5 months) compared to nivolumab (6.9 months) or ipilimumab (2.9 months) monotherapy.²¹³ This, however, is accompanied by a greatly increased incidence of grade III or IV adverse events, at 55.0% (vs 16.3% and 27.3%, respectively). Five-year follow-up results of the CheckMate 067 trial show ipilimumab/nivolumab combination therapy continuing to outperform either of the drugs alone in unresectable stage III or stage IV disease. OS was 52% in the combination therapy group, 44% for the nivolumab group, and 26% for the ipilimumab group. Patients with BRAF mutations who received combination therapy had a 5-year OS rate of 60%, compared to 48% for patients with BRAF wild-type disease. There were fewer substantial discrepancies between BRAF mutant and BRAF wild-type patients in the nivolumab and ipilimumab monotherapy groups. Median OS has not been reached in the combination therapy group at 5 years follow-up, whereas it was 36.9 months and 19.9 months for the nivolumab and ipilimumab groups, respectively. Combination therapy was associated with 5-year PFS of 36%, compared to 29% for nivolumab and 8% for ipilimumab. Median PFS was superior in the combination therapy group as well, at 11.5 months, compared to 6.9 months for nivolumab and just 2.9 months for ipilimumab. The rate of grade III or IV adverse events was considerably higher than in either of the monotherapy groups, at 59% for nivolumab with ipilimumab, 28% for ipilimumab, and 22% for nivolumab.²⁴⁶

Targeted Therapy

In cutaneous melanoma, as well as many other human cancers, the mitogen-activated protein kinase pathway is a key contributor to the loss of cell-cycle control and the development of malignancy.²⁴⁷ Refer to Figure 8 for a schematic drawing of this pathway, with the points of action for targeted therapy agents indicated.²⁴⁸ A gene that produces a protein component of this pathway—the BRAF gene—is the most commonly mutated gene in melanoma, with reported frequency of mutation between 42% and 66%.^{37, 249–251} Of the BRAF mutations, the overwhelming majority are of the V600E type. There is also a not insignificant frequency of V600K mutation. Patients with either mutation are candidates for treatment with BRAF-inhibitor therapy. BRAF-inhibitors exert their effect by inhibiting the mutant B-raf protein, interrupting the transmission of signals down the MAPK pathway, thereby halting the replication of malignant cells. The first drug in this category became available in 2011, with the approval of vemurafenib. In the phase III, randomized BRIM-3 trial, which compared vemurafenib and dacarbazine in treatment-naïve patients with BRAF-mutant unresectable stage III and stage IV melanoma, vemurafenib was associated with a 73% relative reduction in the risk of disease progression or death compared to dacarbazine. The response rate was nearly 50% in those patients with V600E mutations, compared to just 5% in the dacarbazine group. Six-month OS was 84% in the vemurafenib group, compared to 64% in the dacarbazine group.²⁵² Extended follow-up of this study demonstrated a median OS of 13.6 months in the vemurafenib group, compared to 9.7 months in the dacarbazine group. However, the survival benefit diminished over time, as 18-month OS was 39% in the vemurafenib group and 34% in the dacarbazine group. Significant rates of secondary skin malignancies were noted in the vemurafenib group, with 19% of patients developing squamous cell carcinomas and 10% developing keratoacanthomas. This effect, interestingly, is abrogated with the addition of MEK inhibitors, as detailed later in this section.²⁵³

Figure 8.

The MAP kinase pathway. Points of action for current targeted therapy agents are indicated. (With permission from Kainthla R, Kim KB, Falchook GS. Dabrafenib for treatment of BRAF-mutant melanoma. Pharmgenomics Pers Med. 2014;7:21–29).

The development of dabrafenib, another BRAF-inhibitor, closely followed vemurafenib, and this drug was approved by the FDA in 2013. Dabrafenib and vemurafenib induce similar response rates, however there are different rates of adverse events that accompany each drug. Dabrafenib is associated with an increased rate of pyrexia, but a much lower incidence of secondary cutaneous skin cancers compared to vemurafenib.²⁵⁴ A randomized phase III trial (BREAK-3) comparing dabrafenib and dacarbazine showed a greatly increased response rate with dabrafenib (50%) compared to dacarbazine (6%), as well as increased PFS with dabrafenib (median PFS, 6.7 months) compared to dacarbazine (2.9 months). Only 6% of patients in the dabrafenib arm developed squamous cell carcinomas or keratoacanthomas, which represents a substantial decrease compared to the incidence of these lesions in vemurafenib trials.²⁵⁵ Using a standardized health-related quality of life measurement tool, patients in the dabrafenib arm of the trial reported improved quality of life metrics over the course of the treatment period.²⁵⁶ However, the dramatic effect of dabrafenib appears to be only temporary. Long-term follow-up of patients on dabrafenib trials shows 5-year OS rates just above 20%, essentially equivalent to dacarbazine.²⁵⁷ The pattern of successful initial response, followed by eventual relapse and return to the mean suggests the prospect of acquired resistance. Indeed, a later review showed that approximately one half of patients developed resistance to BRAF-inhibitors after less than a year of therapy.²⁵⁸

Acting on the premise that the MAP kinase pathway was circumventing the BRAF blockade and re-engaging downstream to reactivate the pathway via alternative signaling pathways, the MEK-inhibitor trametinib was introduced in an attempt to combat this phenomenon. Trametinib monotherapy offered superior response and survival benefit compared to historical monotherapy protocols. For instance, in comparison with dacarbazine or paclitaxel, the response rate was 22% with trametinib (vs. 8%). Median PFS was 4.8 months in the trametinib group (vs. 1.5 months), and 6-month OS was 81% (vs. 67%).²⁵⁹ On the basis of this finding and other data, trametinib was approved in early 2013. Unfortunately, despite acting further down the MAPK pathway, trametinib had a lower response rate and worse PFS compared to BRAF-inhibitors.²⁵⁹ However, the true value of trametinib is due to its synergistic relationship with dabrafenib. A phase II trial of dabrafenib/trametinib combination therapy demonstrated increased response rate with combination therapy versus monotherapy (76% vs. 54%). Median PFS was also significantly improved compared to monotherapy (9.4 months vs. 5.8 months). Adverse event profiles were also notable, with just 7% of patients on combination therapy compared to 19% of patients on monotherapy developing additional malignant cutaneous lesions. A quarter of patients on monotherapy developed pyrexia, compared to 71% of patients on combination therapy. However, only 5% of patients developed grade III pyrexia.²⁶⁰ Multiple phase III trials have demonstrated superior objective response rates compared to monotherapy (64–68% vs. 45–51%).^{211, 259–264} A review of long-term outcomes shows that combination therapy provides improved OS at 5 years (34%) compared to monotherapy. For those patients who experienced a CR, this was associated with a dramatic increase in 5-year OS, to 71%.^{259, 262, 263, 265, 266} Combination therapy was associated with lower rates of secondary cutaneous malignancies than monotherapy.^{261, 262, 264} The combination of dabrafenib and trametinib for metastatic melanoma with BRAF V600 mutation was approved by the FDA in January, 2014. Hope remains that among the population of patients who present with low metastatic disease burden and normal LDH, some patients may be in effect cured by combination targeted therapy, as 55% of patients in this group are alive at 5 years.²⁶⁵ Recently published long-term data on a cohort of patients with BRAF-mutant stage III disease status after resection and 12 months of dabrafenib trametinib further reinforces the near-curative effect of combination therapy in some patients, as 52% of the cohort were disease-free at 5 years and 65% had not had distant recurrence.²⁶⁷

An additional option for metastatic disease that is unresectable but anatomically accessible is intralesional therapy. Most prominent is talimogene laherparepvec (T-VEC), a modified herpes virus which, when injected into a melanoma lesion, generates a local response by increasing the population of dendritic cells, resulting in increased presentation of tumor antigens to T-cells. Distant metastases that are not injected have also been noted to respond to T-VEC therapy in more than 30% of cases, indicating a systemic effect of this local intervention.²⁶⁸

Symptomatic but unresectable metastatic melanoma may be treated with surgical debulking or with radiation therapy. Larger doses per fraction have been associated with increased likelihood of durable response.²⁶⁹

Brain Metastasis

Patients with melanoma that has metastasized to the CNS are categorized as having M1d disease. These patients have historically been excluded from many clinical trials, leaving physicians with limited evidence upon which to base their decisions.²⁷⁰ An analysis of patients excluded from phase III clinical trials reports outcomes not dissimilar from other patients with metastatic disease. For patients with normal LDH levels, ECOG status 0–1, and asymptomatic brain metastases, median OS was 16.4 months, with 35.1% of patients surviving 3 years. For patients with the same characteristics, but LDH 250–500 U/L, median OS drops to 9.9 months and 3-year OS to 21.0%. The prognosis becomes incrementally worse with declining ECOG status, rising LDH, and the presence of symptoms related to the metastases. Patients with ECOG scores of 2 or greater, with LDH greater than 500 U/L, and symptomatic brain metastases only have a median OS of 3.4 months, with just 0.3% of patients alive at 3 years. BRAF mutation status was not associated with survival in this cohort. The conclusion of this study was that the presence of brain metastases did not preclude long-term survival in select patients with aggressive treatment and that patients with brain metastasis could benefit from inclusion in phase III trials.²⁷¹

The treatment of all patients with brain metastases should begin with discussion at a multidisciplinary conference involving representatives from surgical oncology, medical oncology, radiation oncology, and neurosurgery. Although all disciplines may not be involved in the patient’s care at the outset, their input is often needed eventually upon progression of disease.

If patients present with asymptomatic disease or symptomatic brain metastases in the setting of considerable extracranial disease burden, it is recommended they initiate systemic therapy. Combination therapy with ipilimumab and nivolumab has been shown to be safe and to generate reasonable response rates^{272, 273} that are superior to ipilimumab and nivolumab monotherapy.²⁷⁴ Pembrolizumab has also been shown to generate a response, however less than ipilimumab/nivolumab combination therapy.²⁷⁵ Patients with a BRAF mutation who are not candidates for immunotherapy may benefit from combination BRAF/MEK inhibition. A phase II trial showed response rates of 44% to 58% in patients with asymptomatic brain lesions, although these responses were short-lived, as the median PFS interval was just under 6 months.²⁷⁶

In the scenario of a patient with known brain metastases being treated with systemic therapy, the patient should undergo serial brain magnetic resonance imaging (MRI) every 6 to 8 weeks, per NCCN guidelines, to monitor for progression of disease. If the disease progresses on systemic therapy, alternate regimens or modalities of therapy should be considered.

Should the patient present with symptomatic brain metastases, administration of steroids is a prudent temporizing measure to decrease inflammation and minimize symptoms. If the metastasis is large, singular, and resectable, neurosurgical intervention is recommended.

Systemic therapy or radiation may be considered in the adjuvant setting if the patient is free of residual disease. Stereotactic radiosurgery is the modality of choice, as it improves local control, although it does not improve OS compared to observation.²⁷⁷ Although whole-brain radiation (WBRT) provides a slight advantage in local control, there is no OS benefit, and WBRT is associated with significant negative effects on cognitive function.²⁷⁸

Resectable Disease

For patients who present with metastases amenable to complete surgical excision, resection is recommended. Despite the recent development of significantly more effective systemic therapy, surgical resection still offers a DSS benefit compared to systemic therapy alone.¹⁷⁵ Those patients who have small volume disease with the potential for complete resection, and who are stable on systemic therapy, should be taken to the operating room.²⁷⁹ Even in patients who develop recurrent disease after metastasectomy, serial resection confers a survival benefit.²⁸⁰

Metastasis to distant skin, subcutaneous tissue, muscle, or lymph nodes (stage M1a disease) represents 20% of metastases.²⁸¹ Aggressive resection with 2 cm margins is indicated for what is often symptomatic disease, and in cohorts that undergo complete resection, median OS can reach 60 months.²⁸²

The lung is the most common site of distant metastasis in melanoma. Pulmonary metastases are usually asymptomatic and diagnosed via routine imaging during the follow-up period. Patients from the MSLT-I trial who developed pulmonary metastases, followed by metastasectomy, experienced a clinically significant, although not statistically significant, improvement in survival (median OS, 17.9 months) compared to patients treated with systemic therapy alone (9.1 months).²⁸³ Stereotactic body radiation therapy (SBRT) is also an option for nonsurgical candidates to treat a limited number of small pulmonary metastases in the appropriate clinical setting.²⁸⁴

Patients with non-pulmonary visceral metastasis (stage M1c disease) also benefit from surgical resection. Again reviewing MSLT-I patients, those with resected visceral metastases had a median OS of 15.0 months, compared to 6.3 months in those who received only systemic therapy.²⁸³ Forty percent of patients with M1c disease will have liver metastases.²⁸⁵ A systematic review reported median interval of 54 months to the development of liver metastases.²⁸⁶ Two single institution reviews demonstrated OS beyond 2 years after metastasectomy, far superior to 8 months in patients who received systemic therapy alone.^{287, 288} Radiofrequency ablation of hepatic lesions is an alternative for patients who are not surgical candidates.²⁸⁹ Of those with stage M1c disease, 20% will have metastases to the GI tract, primarily the small bowel.²⁸⁵ Metastasis to other abdominal organs, including the adrenals,²⁹⁰ spleen,²⁹¹ and pancreas²⁹², is less common, but all benefit from surgical resection compared to nonoperative management.²⁸⁵

Patients who undergo complete resection may initiate adjuvant therapy. For patients who are BRAF wild-type, data from a phase III RCTs suggest that nivolumab is associated with improved DFS at 1 year (63.0%) compared to ipilimumab (57.5%), with substantially fewer adverse events necessitating the cessation of therapy (9.5% vs 42.6%).¹⁴⁸ Given pembrolizumab’s superiority to ipilimumab,²¹² pembrolizumab may also be used in the adjuvant setting. However, if the patient has previously failed PD-1 inhibitors, then ipilimumab may be used. More than 40% of patients who develop metastatic disease harbor BRAF V600 mutations,²⁹³ and these patients may be treated with combination BRAF/MEK inhibitors for 1 year or, as above, with immunotherapy with a PD-1 checkpoint inhibitor for 1 year, depending on comorbidities (eg, preexisting autoimmune disease may preclude the use of immunotherapy) and clinician discretion in the absence of a head-to-head comparison of these agents.

Patients with residual disease after metastasectomy have, by definition, disseminated disease not amenable to resection and should be treated with systemic therapy per NCCN guidelines. Details of systemic therapy regimens are discussed earlier in this section.

Neoadjuvant Therapy for Metastatic Melanoma

In parallel with the evolution of neoadjuvant therapy for other malignancies, the administration of neoadjuvant systemic therapy prior to resection of melanoma metastases has come to the fore in recent years. One phase II study (NeoCombi) investigating neoadjuvant dual BRAF/MEK inhibition in resectable stage IIIB and IIIC melanoma may have relevance to stage IV patients. The authors reported 46% of patients with a CR, 40% with a partial response, and no patients with progression of disease. Of these patients, nearly 50% had a cPR.¹⁶⁹ A randomized phase II trial comparing neoadjuvant dabrafenib trametinib combination therapy followed by surgery, with surgery subsequently followed by adjuvant targeted therapy in patients with resectable metastatic melanoma, was halted early due to a statistically significant survival benefit conferred by neoadjuvant therapy. At the point of cessation, PFS was 19.7 months in the neoadjuvant group compared to 2.9 months in the surgery-first group,¹⁷⁰ making a convincing argument for neoadjuvant systemic therapy prior to resection of oligometastatic disease.

Metastasis from Unknown Primary

A curious problem in cutaneous melanoma is the discovery of distant metastatic disease in the absence of an identifiable primary lesion. Of all patients with cutaneous melanoma, 3.2% are diagnosed with a metastasis with unknown primary (MUP).²⁹⁴ Although often disconcerting to the patient in not knowing the initial origin of the tumor, patients with nodal disease from an unknown primary who undergo therapeutic LND have a prognosis that is superior to that of patients with stage III disease. Multiple studies report 5-year OS of 55% to 58% for MUP compared to 40% to 42% for stage III disease with a known primary.^{295, 296} Superior survival is also observed for patients with visceral metastases from an unknown primary compared to those patients with a known primary that has metastasized to visceral organs.²⁹⁷ As a result of the positive prognostic effect conferred by MUP, an aggressive surgical and systemic approach with curative intent is indicated.

There are multiple hypotheses as to the development of MUP and the mechanism responsible for its relatively good prognosis. Perhaps the most agreed upon is that a previously extant primary lesion was eradicated by the patient’s immune system.^296–298 A robust immune response would also suggest better control of distant metastasis, contributing to relatively improved survival even when distant disease was present. Another hypothesis is that melanoma arising de novo in the axillary nodal tissue does not necessarily represent loss of local control of a cutaneous primary, but rather the presence of ectopic nevus cells in a nodal basin that develop into a primary malignancy.²⁹⁷

Patients who present with a metastatic lesion and no known primary should have their history taken followed by a complete examination, including commonly overlooked locations such as the eyes, oral cavity, nasopharynx, anus, rectum, and interdigital webbing. A whole-body PET-CT is also indicated to screen for additional foci of metastatic disease. Finally, although still investigational, commercially available tumor genome sequencing can provide clues as to the origin of the tumor if a characteristic mutation is present (eg, BRAF V600E mutation would essentially exclude a uveal primary, whereas a GNAQ/11 mutation would strongly support a uveal primary).

Future Directions and Ongoing Clinical Trials

After a decade of discovery that has revolutionized the landscape of melanoma, there is a great deal of optimism and excitement for what the next decade holds. Although it is unlikely that quantum leaps of equivalent magnitude to the development of immunotherapy and targeted therapy will occur again, several areas offer promise for continued improvements in the care of patients with cutaneous melanoma.

Given the success of immunotherapy for melanoma, it is logical to use an immune checkpoint backbone with additional agents in attempts to engender an increased response. Current trials pursuing this line of inquiry are summarized in Table 13.

Table 13.

Clinical trials investigating combination therapy with a backbone of BRAF/MEK inhibition.

Clinicaltrials.gov identifier	Study name	Phase / Status	Description of trial
NCT03595683 367	Pembrolizumab and EDP1503 in Advanced Melanoma	2 – Active, not recruiting	Investigation of effect of biologic agent (EDP1503) on response to pembrolizumab in patients with advanced melanoma.
NCT04493203 368	Nivolumab Plus Axitinib in Patients with Anti-PD1 Refractory Advanced Melanoma	2 – Not yet recruiting	Combination PD1-inhibition and tyrosine kinase inhibitor therapy in patients with unresectable stage III or IV melanoma that have failed on previous PD1-inhibition and/or CTLA4-inhibitor therapy.
NCT02965716 369	Talimogene Laherparepvec and Pembrolizumab in Treating Patients with Stage III-IV Melanoma	2 – Recruiting	Investigation of synergistic effect of intralesional anti-tumor vaccine and PD1-inhibitor in patients with stage III-IV melanoma.
NCT03897881 309	Personalized Cancer Vaccine mRNA-4157 and Pembrolizumab Versus Pembrolizumab Alone After Complete Resection of High-Risk Melanoma (KEYNOTE-942)	2 – Recruiting	Phase 2 study comparing adjuvant immunotherapy with pembrolizumab / personalized vaccine and pembrolizumab alone after resection of high-risk melanoma.

A similar path is being followed with targeted therapy. Combination BRAF and MEK inhibition led to significant survival benefit in patients with advanced melanoma, and now ongoing trials are investigating the addition of PD-1 inhibitors and other agents. See Table 14 for ongoing trials of combination therapy with a BRAF/MEK inhibitor backbone.

Table 14.

Clinical trials investigating combination therapy with a backbone of immunotherapy.

Clinicaltrials.gov	Study name	Phase / Status	Description of trial
NCT02967692 370	A Study of the Anti-PD1 Antibody PDR001, in Combination with Dabrafenib and Trametinib in Advanced Melanoma [COMBI-I trial]	3 – Active, not recruiting	Randomized, double-blinded trial comparing combination BRAF/MEK inhibitor therapy with BRAF/MEK inhibition and the addition of PD1-inhibition in patients with advanced melanoma.
NCT04527549 371	Testing the Combination of Dabrafenib and Trametinib With or Without Hydroxychloroquine in Patients with Stage IIIC or IV BRAF V600E/K Melanoma	2 – Not yet recruiting	Determining if hydroxychloroquine-induced autophagy improves survival when that agent is added to BRAF/MEK inhibitor therapy for patients with BRAF-mutant stage IIIC or IV melanoma.

Neoadjuvant therapy, as discussed previously, is a topic under intense investigation for patients with nodal or systemic disease that may be amenable to resection.²⁹⁹ There are multiple ongoing trials in this realm, and a group has coalesced around the subject—the International Neoadjuvant Melanoma Consortium (INMC). This group has published guidelines for trial design in an attempt to streamline the many investigative efforts underway.⁹⁶ Several representative clinical trials investigating neoadjuvant therapy are summarized in Table 15.

Table 15.

Neoadjuvant clinical trials.

Clinicaltrials.gov identifier	Study name	Phase / Status	Description of trial
NCT02036086 372	Neoadjuvant Vemurafenib Plus Cobimetinib for BRAF Mutant Melanoma with Palpable Lymph Node Metastases	2 – Active, not recruiting	Neoadjuvant vemurafenib/cobimetinib in patients with BRAF V600 mutant stage IIIB-C melanoma
NCT01972347 373	Neoadjuvant Dabrafenib/Trametinib for Stage IIIB-C BRAF V600 Mutation Positive Melanoma	2 – Active, not recruiting	Phase 2 study of neoadjuvant dabrafenib/trametinib in stage IIIB-C BRAF V600 mutant melanoma.
NCT04495010 374	Neoadjuvant Nivolumab plus Ipilimumab Followed by Adjuvant Nivolumab or Neoadjuvant Nivolumab plus Ipilimumab Followed by Adjuvant Observation Compared with Adjuvant Nivolumab in Treatment-Naïve High-risk Melanoma (CheckMate 7UA)	2 – Not yet recruiting	Neoadjuvant nivolumab/ipilimumab followed by adjuvant nivolumab versus neoadjuvant nivolumab/ipilimumab followed by observation versus adjuvant nivolumab in treatment-naïve high-risk melanoma

The concept of using autologous tumor-specific immune cells to attack malignant cells has been around for several decades, but there has been increased interest in TILs in recent years. The immunogenic nature of melanoma makes it an ideal candidate for the generation of immune cells that recognize melanoma-specific antigens, collection, ex vivo proliferation, and subsequent reintroduction into the patient. ³⁰⁰ Although the collection and proliferation process are complex with imperfect yield, and few centers have TIL programs, those patients who successfully receive this therapy have high rates of response (approaching 50% in several studies^{301, 302}) and these responses are often durable. A current line of investigation involves TIL administration after a course of modern immunotherapy.^{303, 304}

Cytokines such as IL-2 were some of the first agents that ushered in the era of immunotherapy for melanoma. Efforts continue to develop novel cytokine equivalents with anti-tumor activity. IL-15 acts by upregulating the activity of natural killer cells, which act directly against tumor cells and also amplify other cytokine responses.³⁰⁵ IL-21 plays a key role in potentiating the proliferation of memory T-cell precursors.³⁰⁶ The implications of this activity in developing memory T-cells for adoptive cellular therapy, such as TIL therapy mentioned previously in this section, may be significant.

Other attempts to direct the power of the immune system against melanoma include vaccination with custom peptides (NEO-PV-01 trial)³⁰⁷ or mRNAs (KEYNOTE-942 trial)^{308, 309}, representing predicted immunogenic neoantigens from a given specific patient’s tumor exome. Another approach involves using dendritic cells harvested from the patient, matured and proliferated ex vivo, then reinfused to the patient. Studies have shown dendritic cell vaccines to induce a response, improve survival, and to do so with acceptable toxicity rates.³¹⁰ The administration of dendritic cell vaccines in concert with PD1-inhibitors has increased the response to those agents in mice,³¹¹ offering the prospect of synergistic combinations of vaccines with modern systemic therapy regimens.

Conclusions

Advances in systemic therapy have defined the last decade in melanoma and have irrevocably changed the horizon for both patients with metastatic disease and the researchers and clinicians who seek to help them. Improved prospects for long-term survival created by developments in immunotherapy and targeted therapy encourage more ambition in the pursuit of new therapies and justify optimism that the development and implementation of novel modalities can again revolutionize the field.

STAGE-SPECIFIC FOLLOW-UP FOR CUTANEOUS MELANOMA

Once a patient with a primary cutaneous melanoma undergoes successful resection and there is no evidence of residual disease, they enter into a period of surveillance. After the usual initial postoperative visits to monitor healing of the surgical wound, the patient returns to clinic regularly over the next several years in order to monitor for the development of recurrent disease, or the presentation of a second primary skin cancer. For patients with a previous melanoma, the risk of developing a second skin cancer is 2 to 3 times that of the general population. As each patient’s risk of recurrence is related to the stage of their primary disease, guidelines for frequency and duration of follow-up visits differ by stage. Beyond a thorough history and physical examination, the components of a follow-up clinic visit also may differ depending on the stage of the patient’s primary tumor, as some patients will undergo serial imaging for a period of time after resection. A prolonged period of surveillance also allows for the provision of emotional support to patients as they navigate various survivorship issues. This long-term relationship facilitates an ongoing dialogue between the patient and their oncologists wherein questions and concerns can be addressed, and positive behaviors reinforced.

Rationale for Follow-up After Treatment of Primary Cutaneous Melanoma

There are multiple objectives in the follow-up period after a patient has completed initial therapy for a primary cutaneous melanoma. Of chief importance is the detection of recurrent disease or a second primary skin cancer.

Recurrence

It is well-documented that the diagnosis of a primary melanoma at an earlier stage is associated with an improved prognosis.⁹⁵ There also appears to be a survival benefit associated with earlier detection of metastatic disease,^{139, 312} and this benefit has been shown to be independent of lead-time bias.³¹³ As therapeutic options for metastatic melanoma have become more effective, the timing of detection of metastases has gained additional relevance. Patients with a lower burden of metastatic disease at presentation are more likely to be candidates for complete surgical resection and those with a smaller tumor burden may have a better response to systemic therapy,³¹⁴ although this is an extrapolation from the literature on primary melanoma.

Of all patients diagnosed with a cutaneous melanoma who underwent resection, overall recurrence rates have been reported between 5.6% and 22.9%.^{312, 315–318} However, this rate is very heterogeneous depending on the stage of the primary lesion and on certain high-risk features that confer a higher risk of recurrence. These features include the presence of ulceration in the primary lesion, a primary located on the head or neck, mitotic rate greater than 3 per square millimeter, and thick primary melanomas.^{315, 317, 318}

The overall rate of recurrence for patients with stage I melanoma is 6.8%.³¹⁹ Patients with stage IA primary lesions have reported recurrence rates of 5.2%, with stage IB patients developing recurrent disease in 18.4% of cases.³²⁰ Stage II disease has an overall recurrence rate of approximately 29%,^{321, 322} but with a substantially increased risk of recurrence from stage IIA (28.7%) compared to stage IIB (40.6%) and IIB disease (44.3%).³²⁰ Please see Table 16 for a summary of stage-specific recurrence rates.

Table 16.

Stage-specific recurrence rates for cutaneous melanoma.

Stages	Recurrence rate158, 323–325, 331, 344, 346, 348, 349
Stage I	6.8% overall
IA	5.2%
IB	18.4%
Stage II	30 – 47% overall
IIA	21.6 – 28.7%
IIB	23.4 – 40.6%
IIC	44.3 – 46%
Stage III	48.2 – 62.2% overall
IIIA	40.0 – 48%
IIIB	38.6 – 71%
IIIC	50 – 85%

Range of stage-specific recurrence rates reported across multiple studies.

In patients who recur, locoregional recurrence is most common in those with a primary of lower stage, occurring in 70.2% to 76% of cases.^{318, 323} These recurrences are distributed between local/in-transit recurrence (20.8–30%)^{158, 315, 323, 324} and regional nodal recurrence (21–46%).^{158, 315, 318, 323, 324} The rate of initial recurrence manifesting as distant disease increases with stage of the primary. Rates of distant recurrence have been reported in the range of 24% to 30% across all stages of disease.^{315, 318, 323} Park and colleagues³²⁵ report that in stage II disease, 50% of first recurrences were distant, whereas two thirds of first recurrences were distant in patients with a stage III primary. First recurrence as distant metastasis is associated with worse outcomes.¹⁵⁸ The most common methods by which recurrences are detected align with stage-specific patterns of recurrence. Patients with stage IIA and IIB disease are more likely to suffer locoregional recurrence, with nearly 80% of recurrences detected on history and physical examination.³²⁶ Distant metastasis becomes more common with increasing stage of the primary lesion, and more than 60% of recurrences in patients with stage III disease are asymptomatic and detected on surveillance imaging.^{327, 328}

The vast majority of recurrences happen within the first 2 to 4 years after the primary diagnosis.³²⁹ Romano, and colleagues¹⁵⁸ report that the rate of locoregional recurrence after 3 years in stage IIIA disease is less than 5%. For stage IIIB and IIIC disease, the rate of locoregional recurrence drops below 5% at 2 years and 7 months, respectively, after initial diagnosis. The rate of distant metastasis follows a similar pattern, with the rate dropping below 5% after 32 months for stage IIIA disease, 40 months for stage IIIB disease, and 21 months for stage IIIC disease. In a subset of patients, recurrence can occur after a very short postoperative interval. For instance, Bloemendal and colleagues³³⁰ published data from a cohort in which 18% of patients with stage IIIB and IIIC disease developed recurrence within 12 weeks of resection, with a median time to recurrence of 7.4 weeks. In a similar population, Lim and colleagues reported a median time to recurrence of 10.1 months.³²⁸ Some have interpreted these results as an argument for neoadjuvant therapy in patients with high-risk disease in an attempt to select those who will obtain long-term benefit from an operation.³³¹ In contrast, for other patients recurrence can occur over a decade after initial diagnosis. Faries and colleagues³³² detail a cohort of patients in which 6.9% had a recurrence more than 10 years after a primary lesion was treated. These recurrences are more likely to be distant.

Second Primary

Patients who have been diagnosed with a primary cutaneous melanoma have an increased risk for developing a second primary skin cancer and merit an increased intensity of surveillance on these grounds. Recent analyses of United States and Swedish national databases, respectively, have reported rates of 3.6%³³³ (for patients with stage I and II primary melanoma) and 4.5%³³⁴ (no staging breakdown mentioned) for the development of a second primary melanoma. Zimmer and colleagues³³⁵ reported a 5.9% rate of second primary at 10 years in a cohort of patients with stage III and IV primary melanomas. This is in comparison to a lifetime risk of just more than 2% for the entire United States population.²³² A deeper analysis of the United States cohort seems to show the benefit of rigorous follow-up and patient education as, of those diagnosed with a second primary, almost all had a lesion at the same stage or lower than their initial primary. Forty-eight percent had a second primary with a lower stage than their first, and 50% were diagnosed at an equivalent stage.³³³ Similar findings have been reported from a cohort of patients with multiple primary melanomas from Memorial Sloan Kettering Cancer Center, with an 11.4% risk of developing a second primary within 5 years.³³⁶ Sixty percent of the patients in that cohort developed the second primary within 1 year of their initial diagnosis. The Swedish study showed an increasing rate of second primary development over the last half century. For patients initially diagnosed in the 2000s, 6.4% of women and 7.9% of men in this cohort developed a second primary melanoma within 10 years of their initial diagnosis. An increasing incidence of developing additional primary melanomas after the initial diagnosis has implications for follow-up and may lead to recommendations for increased intensity of surveillance in future guidelines.

Psychosocial Support

In addition to monitoring for recurrence or the onset of a new primary lesion, follow-up clinic visits can be utilized as an opportunity to address other patient-related issues that may be dealt with most effectively in a face-to-face manner. For instance, reinforcing the importance of behaviors that minimize sun exposure, and encouraging sun protection, including the use of sunscreen, which has been shown by a RCT to reduce the incidence of melanoma.³³⁷ The demonstration of self-examination techniques for the skin and lymph node basins can also be effectively accomplished in the clinic setting. Assuring adequate time for patients to ask questions can help ease anxiety and can be an opportunity for the physician to clarify or correct information the patient may have encountered in their own searches but is confusing or incorrect. Multiple studies have demonstrated that patients value psychosocial support regarding cancer diagnosis and treatment, and significant patient-reported rates of anxiety (12.1%) and depression (11.6%) exist in a large German study population. However, in that same cohort, documented rates of anxiety (0.2%) and depression (1.7%) were discrepant, indicating an unaddressed need.³³⁸ There are few validated survey instruments with which clinicians may reliably collect quality-of-life data pertinent to patients with melanoma.³³⁹ This knowledge gap may contribute to clinicians’ inability to identify and address areas of concern in patients being followed after the diagnosis and treatment of a cutaneous melanoma. A more deliberate approach to the evaluation of a patient’s psychosocial well-being could help clinicians more effectively address these issues.³⁴⁰ Although cancer surgeons may be well placed to help inform, guide, and reassure patients about clinical matters in the follow-up period, referral to ancillary specialists may be more appropriate for other issues in which other team members have more expertise. Lacking appropriate pathways down which to direct patients may result in physicians ordering superfluous and non-indicated laboratory tests or imaging studies in attempts to relieve patient anxiety.^{341, 342} Clinicians may also feel the need to schedule more frequent follow-up appointments with certain patients, beyond the frequency recommended by clinical guidelines, in an attempt to deal with psychosocial issues surrounding cancer survivorship.³⁴³ This is not only an inefficient use of resources, but also merely a temporary solution to a psychosocial problem that will persist without appropriate management. Faced with both an aging population with a larger pool of people at an age typical for development of disease, and a rising incidence of melanoma, oncologists may face unmanageable clinical volumes. To offset these volumes, it may be helpful for a subset of patients with low-risk disease but with psychosocial needs necessitating frequent office visits beyond what is indicated for clinical surveillance to be directed to more appropriate providers in order to address the psychosocial component of their recovery.³⁴³ Other health systems may provide more readily accessible resources for the melanoma patient in the follow-up period. For example, in Germany melanoma patients are routinely routed to counselors or physical therapists depending on their needs.³⁴⁴ In Scotland, a RCT showed that melanoma follow-up visits with a general practitioner were given positive reviews by the patients involved.³⁴⁵

These programs may be more difficult to implement in the United States where private insurers offer differing levels of access and coverage and a declining number of Americans maintain a relationship with a primary care provider.

Components of Follow-up After Treatment of Primary Cutaneous Melanoma

For all patients who are in the surveillance period after resection of a primary cutaneous melanoma, the NCCN⁹⁴ has baseline recommendations for what should be included in a follow-up visit, regardless of the nature of the primary lesion.

History and Physical Examination

A thorough history should be taken, with a focus on new skin lesions, changing skin lesions, or the development of systemic symptoms worrying for metastasis. A complete physical examination should be performed, including interdigital spaces, the anogenital region, the oral cavity, and lymph nodes. For patients with stage IIB-IV melanomas, it is recommended that these follow-up visits occur every 3 to 6 months for the first 2 years after resection, then every 3 to 12 months until the patient is 5 years out from resection. At that point, the patient should return to care at least annually for surveillance, as there remains a risk of delayed recurrence and they are also at increased risk of developing another primary skin cancer. Patients with melanoma in situ may be followed annually, while patients with stage IA-IIA disease should be seen every 6 to 12 months during the first 5 years after resection before transitioning to annual follow-up. Although routine to all physicians, the history and physical examination can be a powerful tool, with studies showing that 47% of all recurrences were detected on history and physical examination, compared to just 21% of all recurrences detected via imaging.³⁴⁶ As the majority of primary melanomas are lower stage lesions, which tend to recur locally, this increased proportion of recurrences detected on examination is consistent with the preponderance of local recurrence in early-stage disease. As the stage of the primary increases, the likelihood of distant recurrence also increases. The findings of Kurtz and colleagues³²⁶ reflect this, with their paper reporting 79% of recurrences in patients with stage IIA or IIB primaries detected on examination while 50% of recurrences in stage IIC through IIIC patients were asymptomatic and discovered on imaging. At our institution, after the initial postoperative visit, the patient’s surgeon and the patient’s dermatologist will alternate staggered biannual visits, so the patient is being seen every 3 months. After 2 years, the frequency of the visits will decrease to every 6 months, with the surgeon and the dermatologist each seeing the patient once per year. After 5 years, patients may stop seeing the surgeon but are encouraged to follow up with their dermatologist annually for the rest of their lives. Throughout the follow-up period, the importance of sun-safe behavior is emphasized, and proper techniques for self-examination are reinforced with encouragement to perform self-examinations regularly.

The European Society for Medical Oncology (ESMO) recommends post-resection follow-up but did not reach a consensus position on the frequency or duration of such visits.³⁴⁷ Australian guidelines recommend a 10-year duration of follow-up, citing a less than 1% rate of recurrence beyond that period. For stage I disease, they recommend annual visits. Patients with stage IIA disease are recommended to return for clinic visits every 6 months for 2 years, before returning annually for the remainder of the 10-year follow-up period. Patients with stage IIB-IIC disease should be seen in follow-up every 3 to 4 months for the first 2 years postoperatively, then every 6 months for 1 year, before transitioning to annual visits. Patients with stage III disease are recommended to return for follow-up every 3 months for 2 years, every 6 months for 1 year, then annually for the remaining 5 years.³⁴⁸

Imaging

The primary role of imaging is to investigate the suspicion of distant metastasis. Should the patient report symptoms concerning for visceral metastasis, relevant CT images should be obtained with intravenous (IV) contrast. In patients with stage IIC disease and above, routine imaging is indicated even in the absence of symptoms. A cost analysis of postoperative surveillance imaging concluded that biannual CT of the chest, abdomen, and pelvis is cost-effective over the first 4 years of follow-up in patients with stage IIC and stage III disease, and for the first 3 post-resection years in patients with stage IIB disease.³⁴⁹ Chest radiographs have been used historically to screen for pulmonary metastasis but have proven to be ineffectual and that practice has been discarded. If there is concern for metastasis to the brain, the patient should undergo MRI of the brain. Romano and colleagues¹⁵⁸ from Memorial Sloan Kettering Cancer Center published data showing that 36% of patients with a first recurrence in the brain presented with seizures. They proposed that routine brain MRI, at least in the first year after resection, would prevent some number of seizures in this group of patients. The addition of immunotherapy and targeted therapy to the armamentarium and the associated survival benefit for patients with metastatic disease has enhanced the value of routine imaging during the follow-up period. Having an efficacious treatment modality increases the importance of early diagnosis of distant metastatic disease. According to NCCN guidelines, patients with stage IIB or greater primary melanomas should undergo serial imaging to screen for recurrence. This may be performed every 3 to 12 months for the first 2 years post-resection, and then every 6 to 12 months up until 5 years post-resection, at the physician’s discretion. As most recurrences occur within the first 2 to 4 years postoperatively, empiric imaging should not be continued after 5 years.⁹⁴ The recommendation for routine surveillance imaging for the first several years after treatment of a primary melanoma in otherwise asymptomatic patients is justified by several studies which show that many recurrences are both early and asymptomatic, discovered by surveillance imaging alone. Lim and colleagues³²⁸ report that in a cohort of patients with primary stage IIIB and IIIC melanomas who experienced a recurrence, the majority (66%) were asymptomatic and had a short median time to recurrence. Lewin and colleagues³⁵⁰ reported similar rates of asymptomatic recurrence, detected only by imaging. Their study also highlighted the role of PET-CT in screening for recurrent disease, as did a study by Leon-Ferre and colleagues,³²⁷ which reported 60% of all recurrences were detected by PET-CT within the first year after resection. Although PET-CT has been shown to effectively detect distant recurrence in high-risk stage IIB-IIIB disease before symptoms arise, this early detection is not associated with a survival benefit.³⁵¹ Several other studies report high rates (66–96.5%) of distant recurrence detected initially by imaging.^{325, 328, 338, 350, 352, 353} Despite the intuitive appeal of initiating systemic therapy while the burden of metastatic disease is low, there is no evidence that early discovery through surveillance and subsequent intervention for distant recurrence is associated with a survival benefit.^354–356

For those patients with a positive sentinel lymph node who did not undergo completion LND, US surveillance of the regional lymph node basin has equivalent distant RFS¹⁴³ and MSS,¹⁴⁴ according to 2 recent multicenter trials. If the patient elects to pursue observation, the NCCN recommends US every 4 months for the first 2 years, then every 6 months until 5 years post-resection. Serial US examination may also be used in the surveillance of regional lymphatic basins in higher risk patients who were unlikely to tolerate general anesthesia and were deemed poor candidates for SLNB. Although regional control suffers with this approach, there is no impact on distant RFS or DSS.³⁵⁷ US can also be utilized in the setting of a patient who develops worrisome regional adenopathy during the follow-up period. US has been shown to be superior to other imaging modalities in the evaluation of lymphatic basins, with high sensitivity (96%) and specificity (99%).^{358, 359}

Australian guidelines mirror United States guidelines, recommending routine imaging every 3 to 12 months for the first 3 years post-resection of stage IIC and stage III melanoma.³⁴⁸ ESMO did not reach consensus on imaging recommendations, deferring to national guidelines, but noted that serial imaging in high-risk patients could lead to earlier detection of recurrence, although a survival benefit has yet to be borne out.³⁴⁷

Laboratory Tests

Laboratory tests are only relevant in melanoma in the setting of LDH holding prognostic value in stage IV disease. A representative study showed that recurrent disease was only identified through blood tests in 2.5% of cases.³⁵³ Laboratory tests are not recommended for inclusion in follow-up protocols.

Patient Education

Follow-up visits should be used to demonstrate self-examination skills that patients can utilize at home, including monitoring of cutaneous lesions using the “ABDCE” criteria and monitoring of regional lymph node basins. The importance of minimizing sun exposure should be emphasized. Patients should be counseled on avoiding sun exposure during peak hours, wearing protective clothing if they anticipate prolonged exposure, and using a broad-spectrum, high sun protection factor (SPF) sunblock when they are outdoors for prolonged periods. This suite of measures is consistently endorsed in recent guidelines released by major professional bodies across the United States, Europe, and Australia. Please refer to Table 17 for a comparison of guidelines between the NCCN (United States), ESMO (Europe), and Cancer Council Australia.

Table 17.

Comparison between US, European, and Australian guidelines for follow-up in patients with resected cutaneous melanoma.

	National Comprehensive Cancer Network (NCCN)63	European Society for Medical Oncology (ESMO)66	Cancer Council Australia67
History and physical examination	-Stage 0: annually -Stage IA-IIA: every 6–12 mo for 5 y, then annually -Stage IIB-IV: every 3–6 mo for 2 y, then every 3–12 mo for 3 y, then annually -Consider office-based imaging technologies in patients with numerous moles and/or atypical nevi	No consensus, but follow-up is generally recommended. Defer to national guidelines. Range of recommendations: -every 3 mo for 3 y, then every 6–12 mo -return visits as needed	-Stage I: annually for 10 y -Stage IIA: every 6 mo for 2 y, then annually for 8 y -Stage IIB-C: every 3–4 mo for 2 y, then every 6 mo for 1 y, then annually for 5 y -Stage IIIA-C: every 3 mo for 2 y, then every 6 mo for 1 y, then annually for 5 y -Those with risk factors (eg, family history, dysplastic nevus syndrome) may continue follow-up beyond 10 y
Imaging	Indications: -Equivocal lymph node examination -Positive SLN, but did not undergo CLND: every 4 mo for 2 y, then every 6 mo for 3 y -Stage IIB-IV disease: every 3–12 mo for 2 y, then every 6–12 mo for 3 y	-Not recommended for patients with thin melanomas -No consensus, but patients with high-risk disease may undergo serial imaging to aid early identification of recurrence, although no survival benefit	-Stage I-IIB: do not recommend routine imaging -Stage IIC-III: every 3–12 mo for 3 y -CT/MRI when indicated by clinical findings
Genetic testing	Consider additional genetic testing/genetic counselling if family history of multiple melanomas, or combination of melanoma, astrocytoma, pancreatic cancer	Test for actionable mutations at time of diagnosis No recommendations on further testing	Consider genetic testing for CDKN2A mutations/genetic counselling if family history of 3 or more melanomas in first- or second-degree relatives with high-risk features (eg, early onset, multiple primaries, history of pancreatic cancer)
Laboratory testing	Routine labs not recommended.	Routine labs not recommended	Routine labs not recommended.
Patient education	-Sun-safe behavior (avoiding exposure during peak hours, protective clothing, sunscreen) -Self-examination techniques	-Sun-safe behavior (avoid sunburns, avoid extended unprotected exposure) -Self-examinations for life -Educate regarding increased risk to family members	-Sun-safe behavior -Self-examination techniques

CLND, completion lymph node dissection; CT, computed tomography; MRI, magnetic resonance imaging; SLN, sentinel lymph node.

Future Directions

Recent advances in systemic therapy have resulted in significant improvements in survival for patients with melanoma, ensuring larger populations of patients who will need to be followed after successful treatment of their primary lesion. Moving forward, approaches to follow-up may place a greater emphasis on factors unique to individual patients—characteristics of the primary tumor, personal and family history, and genetic characteristics of the primary disease, not simply blanket recommendations based on tumor stage.

GEP holds promise for the more accurate placement of patients into high- or low-risk groups, and postoperative imaging and follow-up can be planned accordingly. However, at this point, the implications of test results and the appropriate response based on these results remain to be determined.³⁶⁰ Multiple reports evaluating a commercially available GEP tool verify its efficacy as a predictor of recurrence, distant metastasis, and survival.³⁶¹ Studies suggest that the combination of tumor characteristics and information gleaned from GEP can be synergistic in the creation of a follow-up regimen. Ferris and colleagues³⁶² report on a cohort in which 21% of patients had discrepant risk stratification between an AJCC prediction tool and that generated by GEP. Of the deaths in this discordant group, 85% were predicted to be high-risk by GEP but were classified as low risk per AJCC prediction tool. Using these tools in concert may improve identification of high-risk patients, allowing for intensified surveillance.

Measurement of circulating cell-free tumor DNA has been shown to indicate the development of progressive disease before it becomes apparent clinically, offering a potential tool for early detection of recurrence.³⁶³

Experiences with telemedicine during the SARS-CoV-2 pandemic have illuminated a possible role for telemedicine in the follow-up of early-stage melanoma patients. As the majority of recurrences in these patients are locoregional, and many are detected by patients on self-examination, it may be reasonable to incorporate video visits in an attempt to minimize clinical volume for the physician and travel burden for the patient. Other technology-based interventions include smartphone apps that leverage artificial intelligence to analyze photos of lesions submitted by patients. When a lesion is classified as worrisome, the patient is encouraged to see a physician. Although success rates at identifying lesions that eventually turned out to be malignant were only between 20% and 30%,³⁶⁴ this concept holds promise for earlier detection of skin cancers and increasing access to expert evaluation for patients who live in rural areas or may otherwise face barriers to attending office-based consultations.

Conclusions

After the diagnosis and resection of a primary cutaneous melanoma, patients are at risk for recurrence as well as the development of a second primary skin cancer. The risk of recurrence increases with increasing stage of the primary lesion, and intensity of follow-up is augmented accordingly. As the majority of recurrences occur within the first several years after diagnosis and resection of the primary, clinical examinations and imaging studies are performed with decreasing frequency over the period of follow-up. The occurrence of one primary melanoma is associated with an increased risk of a developing another primary skin cancer. This risk persists and, as such, patients are followed for life. In the future, genetic testing will help clinicians to more precisely characterize a patient’s risk for recurrence and will contribute to the development of personalized follow-up protocols.

Supplementary Material

2

Figure S1. Synoptic report according to College of American Pathologists (CAP) standards. This is an example of a synoptic report for reporting on a skin biopsy of a melanoma. It includes all criteria that are reported according to CAP standards (with permission from Shon W, Frishberg DP, Gershenwald JE, et al. Protocol for examination of excision specimens from patients with melanoma of the skin. College of American Pathologists website. https://documents.cap.org/protocols/cp-skin-melanoma-excision-19-4100.pdf.