Comprehensive Genomic Profiling of Acral Melanoma: Insights From the AACR Project GENIE Database

Abstract

Background:

Acral melanoma (AM) is a rare but aggressive melanoma subtype that arises on palmoplantar surfaces and nail units. It disproportionately affects individuals with darker skin tones and is frequently diagnosed at advanced stages. Limited genomic data have hindered the development of effective targeted therapies.

Methods:

A retrospective genomic analysis of AM was conducted using the American Association for Cancer Research Project Genomics Evidence Neoplasia Information Exchange repository, evaluating 212 tumor samples from 203 patients for somatic mutations, copy number alterations, and mutational patterns across demographic and clinical variables. Co-occurrence, mutual exclusivity, and survival analyses were also performed.

Results:

NRAS (21.2%), BRAF (18.3%), and KIT (9.0%) were the most common mutations. CDKN2A and CDKN2B deletions occurred in over 20% of the samples, along with recurrent amplifications in CDK4, CCND1, and TERT. Significant comutation patterns included NF1-PTPRT and KRAS-TERT. Mutation frequencies varied across sex and racial groups, and NAB2 mutations were exclusive to metastatic tumors.

Conclusion:

This study provides a comprehensive genomic overview of AM, highlighting recurrent alterations in the MAPK and cell cycle pathways, and potential demographic-specific molecular signatures. These findings support the need for expanded molecular profiling to improve prognostic accuracy and identify targets for future therapy.

INTRODUCTION

Acral melanoma (AM), also referred to as acral lentiginous melanoma, is a rare subtype of cutaneous melanoma that arises on glabrous (hairless) skin, most commonly the palms, soles, and nail beds.¹ According to the World Health Organization, “AM” describes the anatomical location, whereas “acral lentiginous melanoma” refers to the histopathological subtype.² Clinically, AM typically presents as asymmetric brown macules with irregular borders and variable pigmentation.³ It is associated with poorer prognosis and lower survival rate than non-AM, likely due to delayed diagnosis and greater tumor thickness at presentation.^4,5

Although AM comprises only 2%–3% of all melanoma cases, it accounts for a disproportionately higher percentage of individuals with darker skin tones, including those of African, Asian, and Hispanic descent.⁶ Survival outcomes are significantly poorer among Hispanic White and Asian/Pacific Islander patients.⁷ Incidence is comparable between males and females,⁸ although female sex has been associated with a more favorable prognosis.⁶ The average age at diagnosis is approximately 62.8 years,⁹ although increased age does not appear to correlate with worse outcomes.⁴ Unlike most cutaneous melanomas, AM is not associated with ultraviolet radiation, likely due to its predilection for sun-protected areas.¹⁰ Other proposed risk factors include trauma or mechanical pressure on the acral sites, although the underlying pathogenesis remains poorly understood.¹¹

The diagnostic workup for AM typically begins with a thorough clinical and dermoscopic examination, followed by an excisional or punch biopsy for histopathologic confirmation.¹² On dermoscopy, characteristic features such as a parallel-ridge pattern and irregular, diffuse pigmentation suggest melanoma on volar (thick) skin, including the palms and soles.¹³ Magnetic resonance imaging (M) has also emerged as a useful noninvasive tool for differentiating AM from benign pigmented lesions.¹⁴ Staging is performed using the American Joint Committee on Cancer criteria, which incorporate tumor thickness and ulceration.¹⁵ Surgical excision remains the mainstay of treatment, either through wide local excision or Mohs micrographic surgery, often with sentinel lymph node evaluation.¹⁶ In advanced stages, immune checkpoint inhibitors, including ipilimumab, pembrolizumab, and nivolumab, are used alone or in combination.¹⁷ Due to the low frequency of BRAF mutations in AM, targeted therapies, such as BRAF inhibitors, have shown limited efficacy.¹⁸

AM displays a distinct molecular profile, characterized by low single-nucleotide variant burden but a high frequency of somatic mutations and copy number alterations (CNAs).¹⁹ Common alterations include mutations and amplifications in KIT and TERT.^20,21 Notably, approximately 50% of cases are triple wild type, lacking mutations in the major melanoma driver genes: BRAF, NRAS, or NF1.²¹ Loss or deletions of tumor suppressor genes, particularly CDKN2A, are frequently observed, whereas TP53 mutations are relatively uncommon.^20–22 Compared with other melanoma subtypes, the molecular landscape of AM remains incompletely understood, limiting the development of effective targeted therapies. Broader genomic profiling is essential for improving the diagnostic accuracy and identifying novel therapeutic targets for this rare and aggressive malignancy.

Despite advancements in melanoma research and treatment, the genomic characteristics of AM remain poorly defined, particularly in large and diverse populations. This multi-institutional genomic analysis aimed to identify recurrent somatic alterations that may inform prognosis and highlight potential therapeutic targets, contributing to a broader understanding of the molecular features underlying this rare and understudied melanoma subtype.

METHODS

This study utilized clinical and genomic data from the American Association for Cancer Research (AACR) Project Genomics Evidence Neoplasia Information Exchange (GENIE) database. Data available from 2017 onward were accessed on June 9, 2025, through the cBioPortal platform (v17.0-public). The study was exempt from institutional review board approval by Creighton University (Omaha, NE) based on its use of deidentified, publicly available data.

The GENIE consortium aggregates genomic data from 19 international cancer centers, producing a diverse and heterogeneous dataset with variability in the sequencing platforms. Samples were processed using one of three approaches: targeted gene panels (comprising 50–555 genes; 80% cases), whole-exome sequencing (WES; 15%), and whole-genome sequencing (5%). The sequencing depth varied by platform, with targeted panels exceeding 500× coverage, WES at a median depth of 150×, and whole-genome sequencing with approximately 30× coverage. In sample composition, 65% were tumor-only specimens, whereas 35% were matched tumor-normal pairs, enabling the exclusion of germline variants during analysis.

Despite institutional variation in sequencing platforms, data were standardized through the harmonization framework, Genome NEXUS. To ensure consistency, the data were processed using the Genome Analysis Tool Kit and ANNOVAR before inclusion in the GENIE dataset. Although it includes therapeutic response and clinical outcomes for select cancer types, treatment information is not available for AM.

Patients included in this study had a confirmed pathological diagnosis of AM. The tumor samples were classified as either primary or metastatic. Differences in gene-level mutation frequencies between primary and metastatic tumors were evaluated using χ² tests.

The dataset included tumor-derived somatic mutation data, and clinical and pathological information, such as age, sex, race, and histologic subtype. Most institutions included well-known cancer-associated genes such as TP53, KMT2C/D, and KRAS, whereas genes lacking established clinical or biological relevance were excluded. The structural variants were not included in this study.

Copy number alterations were examined to identify recurrent amplification and homozygous deletions. Tumor mutational burden was calculated as the number of somatic mutations per megabase of sequenced DNA, normalized by the size of the targeted panel. To estimate the WES-equivalent tumor mutational burden, these normalized values were then adjusted using GENIE-provided regression models, which account for panel size and other related variables.

All statistical tests were performed on the cBioPortal Web site or R (R Foundation for Statistical Computing, Boston, MA) using RStudio. Continuous variables are reported as mean ± SD, and categorical variables are reported as frequencies and percentages. Normally distributed variables were analyzed using a two-sided Student t test, and non-normally distributed variables were analyzed using the Mann–Whitney U test. Associations between categorical variables were evaluated using the χ² test. Where applicable, P-values were adjusted for multiple comparisons using the Benjamini–Hochberg false discovery rate method. Statistical significance was defined as P < 0.05.

Somatic mutation analysis was restricted to nonsynonymous variants, including frameshift, missense, nonsense, and splice-site mutations. To be included, mutations were required to have a minimum variant allele frequency of 5% and sequencing depth of 100×. Synonymous mutations and those of unknown clinical significance were excluded. All mutation data were obtained from the mutation annotation format files provided by the GENIE consortium, which ensured consistent variant nomenclature across participating institutions.

To evaluate the association between genetic mutations and patient survival, Kaplan–Meier curves were generated. The 15 most frequently mutated genes were analyzed, and for each gene, patients were divided into two groups: mutated versus wild type. Differences in survival between the two groups were assessed using log-rank tests, implemented with the survival package (version 3.7.0). Because the GENIE database does not include the time of diagnosis, the age at sample sequencing was used as a proxy. The survival time was calculated from the sequencing date to either death or the most recent follow-up. For patients with multiple samples, mutation data were consolidated across all specimens. A gene was considered mutated for a given patient if the total number of alterations exceeded one across all available samples.

RESULTS

Patient Demographics of AM

The final genomic dataset included 212 tumor samples from 203 patients. To address the limited sample size of AM in the genomic datasets, we combined the primary (43.4%) and metastatic tumor (54.2%) samples. A breakdown of patient demographics is shown in Table 1.

TABLE 1.

Patient Demographics of Acral Melanoma Cohort

Demographics	Category	n (%)
Sex	Male	106 (52.2)
	Female	97 (47.8)
Age category	Adult	202 (99.5)
	Pediatric	1 (0.5)
Ethnicity	Non-Spanish/Non-Hispanic	138 (68.0)
	Unknown/not collected	41 (20.2)
	Spanish/Hispanic	24 (11.8)
Race	Asian	11 (5.4)
	White	133 (65.5)
	Black	15 (7.4)
	Other	11 (5.4)
	Unknown/not collected	33 (16.2)
Sample type	Primary	92 (43.4)
	Metastasis	115 (54.2)
	Unspecified/not collected	5 (2.4)

Of the 203 patients, 52.2% were male. In ethnicity, the majority classified as non-Spanish/non-Hispanic (68%), whereas 11.8% were Spanish/Hispanic. Racial identification within the cohort was as follows: 65.5% were White, 7.5% were Black, 5.4% were Asian, and 5.4% other. Regarding age, most of the patients were adults, with only one pediatric case.

Most Common Mutations and Alterations

NRAS mutations were the most prevalent somatic mutations within the AM cohort and were identified in 45 cases (21.2%). Mutations in BRAF followed closely, occurring in 39 samples (18.3%), whereas KIT and TERT mutations were detected in 19 (9.0%) and 13 (6.1%) samples, respectively. Other altered genes include PTPRT and NF1 (12 cases each; 5.7%), KRAS (nine cases; 4.2%), FAT1, HRAS, ERBB4 (seven cases each; 3.3%), and LRP1B (six cases; 2.8%). Overall, NRAS, BRAF, and KIT were the most frequently mutated genes in this population, and Figure 1 highlights the most observed mutations.

FIGURE 1.

Oncoprint of most frequent alterations in acral melanoma cohort. An Oncoprint, provided by Project GENIE, helps to visualize gene mutations. Alterations included are NRAS, BRAF, KIT, TERT, PTPRT, NF1, LRP1B, GRM3, KRAS, and FAT1.

Beyond somatic mutations, 154 tumor samples revealed CNAs, with loss of heterozygosity being the most common. These loss of heterozygosity events predominantly affected key tumor suppressor genes, most notably CDKN2A and CDKN2B, with 38 (24.2%) and 34 (21.4%) deletions, respectively. Amplification events were also observed for several oncogenes, including CCND1 (30 cases; 19.1%), CDK4 (28; 19.9%), and TERT (28; 19.9%). Other amplified genes included members of the FGF family—FGF19 and FGF4 with 26 cases each (17.4%) and FGF3 (25; 16.8%), and PAK1 (25; 18.5%) and MDM2 (23; 14.6%).

Sex and Race Stratifications

Certain mutations that were found exclusively in males included PIK3C2G (n = 10; P = 0.002), KRAS (n = 16, P = 0.003), and RAD52 (n = 6; P = 0.03). Single occurrences of mutations in AFF3, CDH23, and ITGA10 were detected only in males. In contrast, AKT3 mutations appeared more frequently in females than in males, with four cases versus zero cases, respectively (P = 0.04). Table 2 illustrates the sex-stratified analysis that revealed distinct patterns of mutational enrichment between the male and female patients.

TABLE 2.

Statistically Significant Differences by Sex

Gene	Male, n (%)	Female, n (%)	P
PIK3C2G	10 (11.90)	0 (0.0)	0.001882
KRAS	16 (14.29)	3 (3.0)	0.003849
AFF3	1 (100.0)	0 (0.0)	0.009901
CDH23	1 (100.0)	0(0.0)	0.009901
ITGA10	1 (100.0)	0 (0.0)	0.009901
MTR	1 (100.0)	0 (0.0)	0.009901
MYH11	1 (100.0)	0 (0.0)	0.009901
PDE4DIP	1 (100.0)	0 (0.0)	0.009901

Distinct mutational profiles were observed across racial groups within this AM cohort. Several mutations were found exclusively in White patients, each occurring in a single case (P < 0.001). These include AFF3, CDH23, DNM2, ITGA10, KIAA1549, MTR, MYH11, PDE4DIP, PGAP3, RNF21, and ITPKB. Among the Black patients, exclusive mutations were also identified (P < 0.001 for all three): PIK3CD (n = 3), SDHAF (n = 2), and EWSR1 (n = 1). NAV3 and CASP8 mutations were exclusively found within Asian patients (n = 1; P < 0.001). Table 3 illustrates the differences in mutations between White, Black, and Asian patients, and Table 4 depicts the mutational differences in White versus non-White patients.

TABLE 3.

Statistically Significant Differences by Race (White, Black, Asian)

Gene	White, n (%)	Black, n (%)	Asian, n (%)	P
EWSR1	0 (0.0)	1 (100.0)	0 (0.0)	0.000000504
NAB2	2 (100.0)	0 (0.0)	0 (0.0)	0.000000832
AFF3	1 (100.0)	0 (0.0)	0 (0.0)	0.000001371
NAV3	0 (0.0)	0 (0.0)	1 (100.0)	0.00001670
CASP8	0 (0.0)	1 (0.8)	1 (14.29)	0.00005253
SDHAF2	0 (0.0)	2 (15.38)	0 (0.0)	0.0001072
CDH23	1 (0.0)	0 (0.0)	0 (0.0)	0.000001371

TABLE 4.

Statistically Significant Differences by Race (White vs. Non-White)

Gene	White, n (%)	Non-White, n (%)	P
NAB2	2 (100.0)	0 (0.0)	0.002646
GNAS	3 (2.21)	5 (19.23)	0.002956
AURKA	2 (1.54)	4 (19.05)	0.000001371
RAD50	1 (0.79)	3 (14.29)	0.009114
FGF19	18 (14.29)	8 (40.00)	0.0102
SDHAF2	0 (0.0)	2 (10.00)	0.0227
PIK3CD	2 (1.89)	3 (15.79)	0.0248

Co-occurrence and Mutual Exclusivity Patterns

Analysis of the interactions between mutations revealed several co-occurrence patterns. PTPRT mutations were significantly associated with NF1 mutations, appearing together in three out of seven cases (P < 0.001). Similarly, TERT and KRAS co-occurred in 7 of 28 KRAS-mutated tumors (P = 0.048). Although NRAS and TERT mutations were found together in 12 of the 29 cases, the difference was not statistically significant (P = 0.06). A mutually exclusive relationship was found between NRAS and BRAF, with a comutation occurring in one case; however, the result was also not statistically significant (P = 0.07). Table 5 illustrates the co-occurrence and mutually exclusive interactions between the various mutations.

TABLE 5.

Statistically Significant and Nonstatistically Significant Interactions Between Mutations

Gene A	Gene B	Neither	A Not B	B Not A	Both	Log2 Odds Ratio	P	q-value	Tendency
NRAS	BRAF	85	33	33	1	<−3	<0.001	0.101	Mutual exclusivity
BRAF	KIT	100	34	18	0	<−3	0.013	0.501	Mutual exclusivity
KRAS	ERBB4	133	12	4	3	>3	0.022	0.501	Co-occurrence
NRAS	KRAS	103	34	15	0	<-3	0.024	0.501	Mutual exclusivity
BRAF	KRAS	103	34	15	0	<−3	0.024	0.501	Mutual exclusivity
TERT	NF1	89	35	12	0	<−3	0.036	0.590	Mutual exclusivity
PTPRT	NF1	112	7	8	3	2.585	0.039	0.590	Co-occurrence
TERT	KRAS	94	28	7	7	1.747	0.048	0.627	Co-occurrence
NRAS	TERT	82	19	23	12	1.171	0.066	0.775	Co-occurrence
NRAS	KIT	101	33	17	1	−2.474	0.077	0.804	Mutual exclusivity
BRAF	FAT1	109	31	2	3	2.399	0.084	0.804	Co-occurrence
KIT	PTPRT	107	13	7	3	1.819	0.107	0.939	Co-occurrence
LRP1B	KRAS	13	1	0	1	>3	0.133	1.000	Co-occurrence
LRP1B	TP53	13	1	0	1	>3	0.133	1.000	Co-occurrence
PLCG2	FAT1	125	5	3	1	>3	0.169	1.000	Co-occurrence

Primary Versus Metastatic Tumor Mutations

Eighty nine primary and 111 metastatic tumors from the cohort were available for comparative genomic analysis, providing balanced group sizes and reducing the risk of sampling bias. The NAB2 mutation was detected exclusively in metastatic tumors (n = 2; P < 0.001) and was absent in the primary samples. In addition, a set of mutations appeared only once in the metastatic tumor samples and was absent in primary tumor samples. These include AFF3, CDH23, ITGA10, MTR, MYH11, PDE4DIP, and RNF213 (P = 0.01). In contrast, mutations in ASXL1 (n = 3) were observed exclusively in the primary tumor samples and were not detected in metastatic tumors (P = 0.04).

Mutations Associated With Reduced Survival

Kaplan–Meier analysis identified three genes with statistically significant associations with overall survival, as shown in Table 6. RAF1 mutations were observed in six patients and were associated with reduced survival (P = 0.003). Similarly, mutations in LRP1B, also found in six patients, correlated with worse survival outcomes (P = 0.026). In contrast, NRAS mutations, present in 42 patients, were associated with improved overall survival (P = 0.048). These findings suggest that RAF1 and LRP1B may serve as markers of poor prognosis, whereas NRAS alterations may reflect a distinct biological subset with more favorable outcomes.

TABLE 6.

Mutations Correlated With Reduced or Prolonged Survival

Mutation	Differential Survival	Number of Patients With Mutation	Number of Patients Without Mutation	P	Gene Function
RAF1	Reduces survival	6	179	0.003	MAP kinase pathway
LRP1B	Reduces survival	6	179	0.026	LDL receptor
NRAS	Prolongs survival	42	143	0.048	Ras family

DISCUSSION

Demographic Characteristics and Mutational Differences

This study aimed to characterize the genomic landscape of AM using the AACR Project GENIE database. Our analysis revealed a high prevalence of mutations in NRAS, BRAF, and KIT, and frequent CNAs involving CDKN2A, CDK4, and TERT. Variation in mutational patterns were also observed across demographic subgroups, including sex, race, and ethnicity.

The cohort included both primary and metastatic tumor samples from a diverse population, with a slight male predominance (52.2%) and the majority were White (65.5%). Consistent with previous reports, this study supports the observation that although AM comprises only 2%–3% of all melanoma cases, it disproportionately affects individuals with darker skin tones, including those of African, Asian, and Hispanic descent.⁶ Notably, prior studies have documented poorer survival outcomes among Hispanic White and Asian/Pacific Islander patients, underscoring the importance of studying this melanoma subtype within racially diverse populations.⁷

In the genomic analysis, several mutation patterns differed according to sex and race. PIK3C2G, KRAS, and RAD52 mutations occurred more often in males, whereas AKT3 mutations were found exclusively in females. Racially stratified analysis revealed differences in the frequency of select alterations across groups: PIK3CD and SDHAF2 were found only in Black patients, whereas NAV3 and CASP8 occurred exclusively in Asians. However, these findings were based on very small numbers and should be interpreted cautiously. As many of the observed alterations are not currently targetable, this analysis is intended to be descriptive and hypothesis-generating rather than clinically directive, highlighting areas for future study in larger, racially diverse cohorts.

Common Mutations and Molecular Pathways

The most frequently observed mutations in our cohort were NRAS (21.2%), BRAF (18.3%), KIT (9.0%), and TERT (6.1%), which is consistent with previous studies that highlighted these as common genetic alterations in AM.²¹ CNAs were also prevalent, with deletions of the tumor suppressor genes CDKN2A and CDKN2B occurring in 24.2% and 21.4% of cases, respectively. Amplifications involving oncogenes such as CDK4 (19.9%), CCND1 (19.1%), and TERT (19.9%) were also common. These findings underscore the significant disruption of pathways critical for regulating cell growth and survival, consistent with the established evidence of MAPK pathway activation and cell cycle dysregulation in AM.^19,21

Functionally, activating mutations in NRAS and BRAF stimulate the MAPK signaling cascade, promoting unchecked cell proliferation and survival.²³ Concurrent loss of tumor suppressors such as CDKN2A/B, combined with amplification of cell cycle promoters, likely facilitated unregulated cell division. In addition, elevated TERT expression may contribute to tumor progression by maintaining telomere length and enabling cellular immortality.^20,24 Unlike many other melanoma subtypes largely influenced by UV-induced mutations, AMs' unique molecular profile reflects alternative oncogenic mechanisms that may influence responses to current targeted therapies.

MAPK Pathway

The MAPK signaling pathway plays a central role in cell proliferation, differentiation, and survival.²³ Activating mutations in genes such as NRAS and BRAF are well-established oncogenic drivers of melanoma. Unlike UV-driven cutaneous melanomas, AM exhibits a unique mutation pattern with less frequent BRAF mutations but relatively higher NRAS involvement. This suggests an alternative activation mechanism for the MAPK pathway in AM. Trametinib, a MEK inhibitor, has shown promise in managing MAPK-driven tumors, particularly in NRAS-mutant cases.²⁵ Ongoing clinical trials are investigating combination regimens that target multiple components of this pathway to overcome resistance and improve the outcomes in this molecularly distinct melanoma subtype.²⁶

CDKN2A/B and Cell Cycle Dysregulation

CDKN2A and its neighboring gene CDKN2B are key tumor suppressors involved in regulating the G1 to S phase transition of the cell cycle via cyclin-dependent kinases. Deletions and loss-of-function alterations in these genes were observed in approximately one-quarter of our cohort, reflecting a common mechanism for cell cycle dysregulation in AM.²¹ Their loss leads to unchecked cell proliferation and contributes to tumor progression. In addition, frequent amplifications of cyclin-dependent kinases and CCND1 further drive aberrant cell cycle progression. Therapeutic strategies targeting CDK4/6, such as palbociclib or abemaciclib, have shown efficacy in melanoma, particularly when used in combination with BRAF or MEK inhibitors to enhance treatment efficacy.²⁷

Patterns of Co-occurring and Mutually Exclusive Mutations

Several notable co-occurrence patterns were observed in this cohort. PTPRT and NF1 mutations significantly co-occurred (three out of seven cases; P < 0.001), suggesting a potential cooperative role in MAPK pathway dysregulation, as both genes are involved in the negative regulation of intracellular signaling.²³ Interestingly, alterations in PTPRT and NF1 have been associated with improved response to immune checkpoint inhibitors in melanoma.^28,29 In addition, TERT and KRAS mutations co-occurred at a statistically significant level (P = 0.048), supporting a possible interaction between proliferative signaling and telomere maintenance in tumor progression. A trend toward mutual exclusivity was observed between NRAS and BRAF mutations, both canonical drivers of the MAPK pathway, with only one case harboring both mutations (P = 0.07), likely reflecting functional redundancy.²³

These patterns reflect known oncogenic signaling dynamics, particularly the mutual exclusivity of NRAS and BRAF mutations due to overlapping downstream effects within the MAPK cascade. Although coalteration of TERT with MAPK pathway mutations is less well-characterized in AM, it has been reported in other melanoma subtypes and may represent a mechanism by which tumors enhance their replicative potential.²⁴ The co-occurrence of PTPRT and NF1, although rarely described in the literature, may point to novel cooperative interactions unique to the molecular landscape of AM. These findings highlight the importance of further investigation of pathway-level alterations that may inform combination treatment strategies or serve as predictive biomarkers.

Distinct Molecular Profiles in Primary Versus Metastatic Tumors

A comparison between primary and metastatic AM samples revealed both overlapping features and distinct genomic alterations. Although many mutations were shared across both groups, some were exclusive to a single tumor type. NAB2 mutations were only found in metastatic tumors (n = 2; P < 0.001). Also known as the melanoma-associated delayed early response gene, NAB2 has been implicated in tumor progression and metastatic potential in melanoma.³⁰ In contrast, ASXL1 mutations were observed solely in the primary tumors (n = 3; P = 0.04). Although more commonly associated with myelodysplastic syndromes, the ASXL family members have demonstrated both tumor suppressive and oncogenic roles in melanoma.³¹ However, ASXL1 is also frequently implicated in clonal hematopoiesis, and given the tumor-only sequencing design of many GENIE samples, these alterations may not reflect true tumor-specific events. Accordingly, differences between primary and metastatic tumors based on rare alterations should be interpreted cautiously.

Survival Analysis

The survival associations identified in this study provide insight into the prognostic relevance of specific genomic alterations in AM. Mutations in RAF1 and LRP1B were associated with worse overall survival. RAF1, a key component of the MAPK pathway, may contribute to enhanced proliferative signaling, which is consistent with prior studies that associate it with aggressive tumor behavior and poor outcomes.³² In contrast, LRP1B is a well-established tumor suppressor that has been linked to a favorable prognosis in multiple cancers, including melanoma.³³ Its association with worse survival in this cohort may reflect increasing resistance to immune checkpoint inhibitors, which has been observed in other contexts. Interestingly, NRAS mutations are typically considered a negative prognostic marker in cutaneous melanoma but were associated with improved survival in this AM cohort.³⁴ This unexpected result may reflect biological differences between melanoma subtypes. Further studies in larger, clinically annotated cohorts are needed to validate these findings and assess their potential use in risk stratification and therapeutic decision making for AM.

Limitations

This study was subject to limitations inherent to the structure and content of the AACR Project GENIE database. A primary constraint is the absence of transcriptomic data, which limits the ability to examine the downstream effects of genetic mutations on gene expression and signaling activity. Another key limitation was the lack of treatment-related data in the GENIE repository. Without this information, it was not possible to assess the relationship between mutation patterns and clinical responses or to analyze how prior treatments may have influenced tumor evolution. Survival metrics such as overall survival or progression-free survival were not available, limiting our ability to associate specific mutations with clinical outcomes. Genomic data were compiled from multiple institutions using heterogeneous sequencing technologies, introducing the possibility of technical variability. The sample size, although one of the largest available for genomic AM analysis, was still modest and may have limited the statistical power to detect less frequent but clinically meaningful alterations or to confirm associations with demographic variables. Finally, the lack of proteomic data, including immunohistochemical markers, precluded any correlation between genomic alterations and protein expression in either tumor cells or the surrounding immune microenvironment. Acknowledging these limitations, this analysis still provides meaningful insights into the genomic landscape of AM.

CONCLUSION

This study highlights the involvement of key oncogenic pathways such as MAPK, cell cycle regulation, and telomere maintenance in AM and identifies potential novel targets for future therapeutic development. Given the distinct molecular profile of AM relative to other cutaneous melanomas, ongoing comprehensive molecular profiling is essential for improving risk stratification, guiding targeted therapies, and ultimately advancing precision medicine for this rare and challenging subtype.