Accepted Manuscript

Distinct Patterns of Acral Melanoma Based on Site and Relative Sun Exposure

Alexandra M. Haugh, Bin Zhang, Victor L. Quan, Erin M. Garfield, Jeffrey A. Bubley,
Emily Kudalkar, Anna Elisa Verzi, Kara Walton, Timothy VandenBoom, Emily
A. Merkel, Christina Y. Lee, Timothy Tan, Maria Cristina Isales, Betty Y. Kong,
Alexander T. Wenzel, Christopher G. Bunick, Jaehyuk Choi, Jeffrey Sosman, Pedram
Gerami

PII: S0022-202X(17)32917-2
DOI: 10.1016/j.jid.2017.08.022
Reference: JID 1045

To appear in: The Journal of Investigative Dermatology

Received Date: 17 July 2017

Revised Date: 7 August 2017
Accepted Date: 18 August 2017

Please cite this article as: Haugh AM, Zhang B, Quan VL, Garfield EM, Bubley JA, Kudalkar E, Verzi
AE, Walton K, VandenBoom T, Merkel EA, Lee CY, Tan T, Isales MC, Kong BY, Wenzel AT, Bunick
CG, Choi J, Sosman J, Gerami P, Distinct Patterns of Acral Melanoma Based on Site and Relative Sun
Exposure, The Journal of Investigative Dermatology (2017), doi: 10.1016/j.jid.2017.08.022.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

TITLE PAGE
Distinct Patterns of Acral Melanoma Based on Site and Relative Sun Exposure

Alexandra M. Haugh1*, Bin Zhang1*, Victor L. Quan1, Erin M. Garfield1, Jeffrey A. Bubley1,
Emily Kudalkar1, Anna Elisa Verzi1, Kara Walton1, Timothy VandenBoom1, Emily A. Merkel1,

PT
Christina Y. Lee1, Timothy Tan2, Maria Cristina Isales2, Betty Y. Kong1, Alexander T. Wenzel1,
Christopher G. Bunick3, Jaehyuk Choi1,4, Jeffrey Sosman5,6, Pedram Gerami1, 6

RI
1.
Department of Dermatology, Feinberg School of Medicine, Northwestern University, Chicago, IL
2.
Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL

SC
3.
Department of Dermatology, Yale University, New Haven, CT
4.
Department of Biochemistry and Molecular Genetics, Feinberg School of Medicine, Chicago, IL
5.
Department of Medicine, Feinberg School of Medicine, Northwestern University, Chicago, IL
6.

U
Robert H. Lurie Cancer Center, Feinberg School of Medicine, Northwestern University, Chicago, IL
AN
*Ms. Haugh and Ms. Zhang contributed equally to this article.
M

Corresponding author and reprint requests:

Pedram Gerami, MD
D

Professor of Dermatology and Pathology

Northwestern University, Department of Dermatology
TE

676 N. St. Clair Street, Suite 1765, Chicago, IL 60611

Phone: 312-695-1413
Fax: 312-695-0007
EP

Email: pgerami1@nm.org
C

Word Count: (3500 limit): 3500

Tables: 2
AC

Figures: 4

1
 ACCEPTED MANUSCRIPT

ABSTRACT (200 words)

Acral melanoma is distinct from melanoma of other cutaneous sites, yet there is

considerable variation within this category. To better define this variation, we assessed

melanomas occurring on dorsal (n=21), volar (n=9), and subungual/interdigital (n=13) acral skin

PT
as well as acral nevi (n=24) for clinical, histologic, and molecular features. Melanomas on dorsal

acral surfaces demonstrated clear differences compared to volar and subungual/interdigital

RI
melanomas. The latter two groups exhibited significantly less frequent BRAF mutations (p=.01),

SC
were significantly less likely to have the SSM histologic subtype (p=.01), occurred in older

patients (p=.05), and had more frequent involvement in non-Caucasians (p=.01). These

U
differences can be explained by differing levels of UV exposure. Subungual/interdigital
AN
melanomas had the most diverse group of oncogenic mutations including PIK3CA (2/13),

STK11 (2/13), EGFR (1/13), FGFR3 (1/13), and PTPN11 (1/13). Additionally,
M

subungual/interdigital melanomas had a significantly higher frequency of copy number

D

aberrations (67%) than other subgroups (p=.02), particularly in CDK4 and cyclin D1, and were
TE

less likely to have BRAF mutations or a SSM histologic subtype (p=.05) compared to volar acral

melanomas. Although based on a limited sample size, differences between volar and
EP

subungual/interdigital melanomas in our study may be the result of differing levels of ultraviolet

exposure.
C
AC

2
 ACCEPTED MANUSCRIPT

INTRODUCTION

Melanomas from acral sites are often considered a distinct subgroup with unique clinical,

morphologic and genetic characteristics. However, there can be considerable variation. Some

PT
authors exclusively refer to acral melanoma as cases that occur on glabrous skin of the palms and

soles or on subungual sites. Others include melanomas from the dorsal surfaces of the hands and

RI
feet. Acral melanoma generally occurs at a later age (60 or older), has equal incidence in darker

and lighter-skinned individuals, and has a characteristic WHO histologic pattern known as the

SC
acral lentiginous subtype. Patients often have a poorer prognosis compared to melanomas from

U
other sites (Carrera et al., 2017, Teramoto et al., 2017). Compared to sun-exposed skin, BRAF
AN
mutations are less frequent.

In this study we assessed the clinical, histologic, and molecular features of 43 melanomas
M

from acral skin. Our findings show different genetic and epidemiologic patterns in acral
D

melanomas from the dorsal surface of the hand and foot compared to melanomas from

subungual/interdigital spaces or melanomas from volar surfaces. Melanomas arising in the most
TE

UV-protected sites (subungual/interdigital) are the most genomically diverse, with a variety of
EP

non-BRAF-related mechanisms to activate cellular proliferation.

RESULTS
C

Melanoma from Dorsal Acral Sites

AC

Twenty-one melanomas were identified on dorsal acral skin, with 16 from the foot and 5

from the hand. The mean age was 55 and included 19 women and 2 men. All patients with

available demographic information identified as Caucasian (n=19). Stages of disease included 9

melanoma in situ (MIS) (43%), 7 T1a/b (33%), 3 T2a/b (14%), 1 T3a/b (5%) and 1 T4a/b (5%).

3
 ACCEPTED MANUSCRIPT

WHO histologic subtypes included 16 superficial spreading (SSM) (76%), 1 lentigo maligna

(5%), 2 acral lentiginous (10%), and 2 nodular (5%) melanomas. A majority of cases were

BRAF-mutated (67%, 14/21). Other oncogenic mutations identified included NRAS (4/21), KIT

(2/21), FGFR3 (1/21), and IDH1 (1/21). One case had a mutation in TP53 (1/21). Five of 19

PT
cases demonstrated chromosomal copy number alterations in the targeted loci (5/19, 26%) (Table

RI
1). Copy number gains included TERT, CDK4, AUKRA, CCND1, PAK1 and GAB2. Copy

number deletions included CDKN2A, PTEN, and NF1.

SC
Melanoma from Volar Acral Sites

U
Nine melanomas were included from volar acral surfaces. The mean age was 59 and
AN
included 7 women and 2 men. Ethnicity data was available for eight patients, five of whom

identified as Caucasian (63%), two as Hispanic (25%), and one as African American (13%). The
M

melanoma stages included 2 MIS (22%), 4 T1a/b (44%), and 3 T2a/b (33%). WHO histologic

subtypes included 3 SSM (33%) and 6 acral lentiginous subtype (66%). Mutations in oncogenes
D

included NRAS (3/9), BRAF (2/9), EGFR (1/9), and KIT (1/9). Tumor suppressor mutations
TE

included TP53 (1/9) and ATM (1/9). Copy number aberrations were found in 1 of 9 cases

(11%), which was a KIT-mutated T1a melanoma exhibiting gains in TERT (Table 1).
EP

Melanoma from Subungual or Interdigital Acral Sites

C

Thirteen acral melanomas were found to involve the nail unit or interdigital web spaces.
AC

The mean age was 67 with 8 women and 5 men. Ethnicity data was available for 12 patients.

Seven were Caucasian (54%), two were Hispanic (15%), two were African American (15%) and

one was Asian (8%). The melanoma stages included 3 MIS (23%), 1 T1a/b (8%), 3 T2a/b

(23%), 3 T3a/b (23%), and 3 T4a/b (23%). All cases were of the acral lentiginous (n=11, 85%)

or nodular (n=2, 15%) subtype. Mutations in oncogenes included PIK3CA (2/13), NRAS (2/13),

4
 ACCEPTED MANUSCRIPT

EGFR (1/13), BRAF (1/13), IDH2 (1/13), PTPN11 (1/13), FGFR3 (1/13), and ALK (1/13).

Mutations in tumor suppressor genes included TP53 (2/13), STK11 (2/13), and APC (1/13).

Eight of twelve melanomas tested exhibited copy number alterations (67%). Seven of these eight

cases displayed gains in CCND1 or CDK4 (Table 1).

PT
Acral Nevi

RI
Twenty-four acral nevi were assessed from 16 females and 8 males with a median age of

SC
39 (range 25-70). Nine were found in sun-exposed locations (38%) while 7 were on sun-

shielded surfaces (29%) and 8 were on the digits (33%). Cases included 12 routine compound or

U
intradermal nevi, 7 dysplastic nevi, 4 MANIAC nevi, and 1 compound nevus with a congenital
AN
pattern. Regardless of location or diagnosis, a majority of acral nevi were BRAF-mutated (71%,

n=17). Three cases demonstrated mutations in NRAS, and no mutations were identified in 4
M

cases.
D

Group Comparison
TE

We compared age, race, histologic subtype (SSM vs. non-SSM), frequency of BRAF

mutations, and frequency of chromosomal copy number aberrations between melanomas from
EP

dorsal acral surfaces and those from other acral sites (volar or subungual/interdigital). Patients

with dorsal acral melanomas more frequently had a younger mean age (p=0.05), were Caucasian
C

(p=0.01), had SSM histology (p<0.01), and had BRAF mutations (p=0.01). We compared the
AC

same variables between the volar acral melanomas and the subungual/interdigital melanomas.

Although most volar acral melanomas were of the ALM histologic subtype, there was a

statistically higher likelihood of the SSM (p=0.05) subtype to occur among the volar cases

5
 ACCEPTED MANUSCRIPT

compared to the subungual/interdigital cases. Volar acral melanomas were less likely than

subungual/interdigital melanomas to have chromosomal copy number aberrations (p=0.02).

DISCUSSION
In this study we showed that melanomas from dorsal acral sites are distinct in terms of

PT
epidemiology, morphology, and genetic drivers from melanomas from other acral sites.

RI
Melanomas found on the dorsum of the hand or foot had a younger average age (p=0.05), were

exclusive to Caucasians (p=0.01), and were far more frequent among female patients (19/21),

SC
likely related to more UV exposure with open-toed footwear. BRAF mutations were far more

frequent in this group of acral melanomas (p=0.01), as was the SSM histologic subtype (p<0.01)

U
(Figure 1, A-C). A precursor nevus was found in 5 of 21 cases (24%) in a relatively similar
AN
frequency to melanomas from intermittently sun-damaged skin of the trunk, which has been
M

estimated at 30% (Duman et al., 2015).

This profile suggests melanomas from the dorsal hand and foot are most similar to
D

melanomas from other sites of intermittently sun-damaged skin and likely have a similar UV-
TE

driven pathway of mutagenesis. The predominance of BRAF mutations in acral nevi from all

sites but only in acral melanomas from the dorsum of the foot or hand further supports the
EP

concept that BRAF-mutated melanocytic neoplasms are highly dependent on ultraviolet-

C

mediated mutagenesis for acquisition of additional mutations and progression to melanoma

AC

(Shain et al., 2015).

When compared with melanoma from the dorsal hand or foot, melanomas involving volar

or subungual/interdigital sites exhibited demographic features more consistent with those

typically associated with acral melanoma including a high percentage of non-Caucasian patients

(p=.01), older age (p=.05), acral lentiginous growth pattern, a low frequency of BRAF mutations

6
 ACCEPTED MANUSCRIPT

(p=.01), and infrequent precursor nevi (2/22). While volar acral and subungual/interdigital

melanomas had many similar features, the two groups presented some notable differences. For

example, the SSM histologic subtype which is typical of melanomas in intermittently sun-

exposed skin was seen in 3 of 9 volar cases compared to 0 of 13 subungual/interdigital cases.

PT
Additionally, chromosomal copy number aberrations previously described in melanomas from

RI
acral sites were seen in only a small percentage of our volar acral (11%) melanomas. Conversely,

67% of melanomas from subungual or interdigital sites had chromosomal copy number

SC
aberrations (p=.02), which in 7 of 8 cases included CCND1 or CDK4 (Figure 1, D-F). A high

frequency of chromosomal copy number changes in CCND1 or CDK4 has also been reported in

U
other melanomas from highly sun-protected sites such as sinonasal melanomas (Chraybi et al.,
AN
2013).
M

While our sample size limits the ability to draw definitive conclusions, one possible

explanation for this difference is that while volar skin is relatively sun-protected, it is still
D

possible to have episodes of intense intermittent sun exposure to volar acral surfaces, for
TE

example with indoor or outdoor tanning. Conversely, there are studies which have shown that the

nail plate can filter nearly 100% of UVB and 98% of UVA (Stern et al., 2011). Others have
EP

similarly observed varied frequencies of UV-mutational patterns in volar melanomas (Furney et

C

al., 2012, Furney et al., 2014). Hence while the majority of volar and subungual/interdigital
AC

melanomas are of the ALM subtype, there is a clinical and scientific basis to suspect a higher

likelihood of UV-induced SSM type melanomas occurring on volar acral surfaces compared to

subungual/interdigital sites.

Another factor that may be different in comparing volar acral melanomas and

subungual/interdigital melanomas is the role of trauma. Some investigators have reported

7
 ACCEPTED MANUSCRIPT

preferential involvement of volar melanomas to weight-bearing areas of the foot (Costello et al.,

2017). In our study, 6 of 7 volar melanoma cases with sufficiently detailed site descriptions to

determine the exact volar location occurred in weight-bearing areas of the foot such as the ball or

the heel.

PT
A majority of the volar acral and subungual/interdigital melanoma cases lacked

RI
mutations in BRAF and demonstrated more diversity in oncogenes involved compared to dorsal

acral melanomas, which is further consistent with the supposition that volar acral and especially

SC
subungual/interdigital melanomas are most reliant on non-BRAF and non-UV-mediated

U
pathways to malignancy. Two subungual/interdigital melanomas demonstrated mutations in
AN
PIK3CA, a known oncogenic driver that leads to constitutive activation of the PI3K/AKT

pathway in a wide variety of malignancies. Activating mutations in PIK3CA have been

M

identified in a small proportion of melanoma cases from the trunk and some desmoplastic

melanomas (Janku et al., 2014, Omholt et al., 2006, Shain et al., 2015, Silva et al., 2017). The
D

PIK3CA gene encodes for the catalytic subunit of the Phosphatidylinositol 3-kinase protein,
TE

which acts via AKT to activate a variety of anti-apoptotic factors and simultaneously promote

G1-S cell cycle progression through activation of the mTOR pathway (Figure 2) (Halilovic et
EP

al., 2010). Mutations in mTOR have been shown to be significantly more common in acral
C

melanoma than in other subtypes (Kong et al., 2016). Both mutations in PIK3CA map to the
AC

critical C-terminal alpha-helix of the catalytic/kinase domain (p110 subunit) (Figure 3). The

H1047L variant previously identified in melanoma and the, to our knowledge, previously

unreported M1040V mutation lie in the same critical protein domain, further suggesting that

these variants are true oncogenic drivers. These mutations may be of considerable therapeutic

significance, as there are FDA-approved targeted inhibitors of PIK3CA.

8
 ACCEPTED MANUSCRIPT

Subungual/interdigital cases were the only melanomas found to have mutations in

STK11 (2/13), a gene that acts primarily as a tumor suppressor and is involved in several cell

processes including cell polarity, energy metabolism, checkpoint DNA repair, and protein

translation (Zhou et al., 2014). It acts via the AMPK pathway to decrease the activity of mTOR

PT
(Figure 2). STK11 mutations have been identified in pancreatic, cervical, and non-small cell

RI
lung cancer (NSCLC). Mutations in STK11 are relatively uncommon in melanoma, occurring

in only 10% of tumors (Guldberg et al., 1999, Rowan et al., 1999), yet decreased expression of

SC
STK11 is relatively common (Shackelford and Shaw, 2009). Missense variants at codons 55

and 56 and other substitutions at codon 281 of STK11 have been shown to be pathogenic

U
(Forbes et al., 2015, Weinstein et al., 2013). The E57K and P281R variants identified in this
AN
study are both found in the catalytic domain of the kinase (Figure 3) and to our knowledge

previously unreported.
M

One T3b subungual melanoma exhibited a Q510L mutation in PTPN11. PTPN11

D

mutations have been identified in desmoplastic melanoma, and substitutions at Q510 have been
TE

implicated in several malignancies, including neuroblastoma and AML (Bentires-Alj et al., 2004,

Weinstein et al., 2013). This mutation is located in one of the two Sh2 domains of PTPN11,
EP

which promotes protein binding to downstream receptors (Tartaglia and Gelb, 2011). Activating
C

mutations in PTNP11 lead to constitutive activation of its protein product, Shp2, which directly
AC

activates the MAPK pathway (Bunda et al., 2015).

A T4b interdigital melanoma was found to have a N646D mutation in FGFR3, a gene

shown to be mutated in up to 70% of urothelial carcinomas (Moss et al., 2017). A similar FGFR3

mutation was identified in a dorsal BRAF-mutated melanoma. Both mutations are located in a

region of the kinase domain of the protein (Figure 3) where a high frequency of missense

9
 ACCEPTED MANUSCRIPT

mutations have been identified in other malignancies (Weinstein et al., 2013). FGFR3 activation

decreases apoptosis and leads to activation of the MAPK pathway (Hafner et al., 2010).

Activating mutations in PIK3CA and FGFR3 have been found with high frequency in

solar lentigos while germline mutations in PTPN11 and STK11 can lead to LEOPARD and

PT
Peutz-Jegher Syndromes, two conditions characterized by abnormal lentigines and

RI
mucocutaneous pigmentation (Hafner et al., 2007, Hafner et al., 2009). Interestingly, both early

acral lentiginous melanomas and solar lentigines show increased single melanocytes

SC
predominantly along the basal layer of the epidermis (Figure 1, G-I). Hence, it is plausible that

in some cases this shared early morphology is the result of a common early oncogenic driver

U
mutation in PIK3CA, FGFR3, PTPN11, or loss of STK11. Conversely, melanomas with BRAF
AN
mutations, which predominate in nevi and melanomas from intermittently sun-exposed skin,

have more frequent nesting at earlier stages and more frequent pagetosis (Viros et al., 2008). In
M

our experience assessing many histopathologic samples of melanonychia striata, a

D

histopathologic diagnosis of lentigo or atypical basal melanocytic hyperplasia resembling a solar

TE

lentigo is a considerably more frequent finding than subungual nevus and hence may be a

precursor to subungual melanomas (Cooper et al., 2015).

EP

Mutations in the oncogene EGFR were seen in one volar and one subungual/interdigital
C

melanoma. Recently, Hayward et al. described EGFR mutations in 9% of melanomas (Hayward

AC

et al., 2017). EGFR is a transmembrane receptor involved in MAPK and PI3K pathway

activation. Chromosomal copy number gains in EGFR are not infrequent in melanoma and have

been associated with poorer prognosis and thicker tumors (Boone et al., 2011, Katunaric et al.,

2014). While uncommon in sun-exposed melanoma, activating EGFR mutations are prevalent in

NSCLC. The T790M EGFR mutation, found in a volar melanoma, leads to acquired resistance to

10
 ACCEPTED MANUSCRIPT

tyrosine kinase inhibitors by enhancing receptor ATP affinity in lung cancer but has also been

shown to cause tumorigenesis as a de novo mutation (Liu et al., 2017, Mohammed et al., 2017,

Yun et al., 2008). The L704F is of uncertain significance. As there are known inhibitors to

EGFR, activating mutations may be of therapeutic significance.

PT
Three germline variants of uncertain significance were also identified. Patients with

RI
germline variants in ATM (P604S) and SMO (R400H) were diagnosed with melanoma at ages

significantly younger than the average of all included patients (35 and 42). Heterozygous

SC
missense mutations in ATM, including the variant identified in our study, have been postulated

U
to increase the risk for breast cancer and aggressive Hodgkin’s lymphoma (Liberzon et al., 2004,
AN
Mangone et al., 2015, Offit et al., 2002). An 80-year-old patient was found to have a germline

mutation in MLH1 (S406N), a gene implicated in Lynch syndrome. This mutation is of uncertain
M

significance based on previous studies (Johnston et al., 2012), yet this patient’s tumor tissue

exhibited a significantly higher mutational burden than any other included cases, indicating that
D

the variant likely affected the protein’s DNA repair function and may have played a role in
TE

tumorigenesis.
EP

In summary, within the category of acral melanoma, there is considerable variation

dependent on the degree of sun exposure. Dorsal surface acral melanomas exhibit characteristics
C

of sun-exposed melanomas of other sites. Melanomas from subungual/interdigital web spaces

AC

and volar acral melanomas were characterized by older age of onset, infrequent WHO

classification of SSM, infrequent BRAF mutations, and occurrence in non-Caucasian and

Caucasian populations. Furthermore, our data suggests that subungual/interdigital melanomas

may differ slightly from volar acral melanomas, constituting a class of even greater UV

independence because of UV filtration through the nail plate. These tumors had the greatest

11
 ACCEPTED MANUSCRIPT

oncogenic diversity and the lowest frequency of BRAF mutations (1/13). We identified a variety

of pathogenic oncogenes and tumor suppressor genes and a high frequency of copy number gains

in CCND1 or CDK4 (67%) when compared with the other subgroups. Many of the genes

involved such as PIK3CA, PTPN11, STK11, and FGFR3 have also been associated with

PT
lentigines. Interestingly, early forms of subungual melanoma often have morphologic overlap

RI
with lentigines and do not show early nesting and other features associated with BRAF-mutated

melanomas. Importantly, some mutations found in sun-protected melanomas such as the EGFR

SC
T790M and the PIK3CA mutations have known targeted inhibitors, and thus potential

therapeutic significance. Hence when assessing for potential therapeutic targets, it is especially

U
important in cases of sun-protected acral melanomas to utilize a broad NGS panel, compared to
AN
sun-exposed melanomas where BRAF mutations predominate.
M

MATERIALS AND METHODS

After Northwestern Institutional Review Board approval, we searched our
D

dermatopathology database for melanomas and nevi from acral sites. We identified 90 patients
TE

with melanoma from acral sites and excluded 47 patients who did not have adequate tumor and

control tissue for analysis. Each case was evaluated for clinical, histologic, and genetic features.
EP

Statistical Analysis
C

Student t-test and Fisher’s exact tests (SPSS Statistics 23, IBM Corp., Armonk, NY) were
AC

used to compare dorsal acral and non-dorsal acral melanomas for age, ethnicity (Caucasian vs.

non-Caucasian), histologic subtype (SSM vs non-SSM), frequency of BRAF mutations, and

frequency of chromosomal copy number aberrations. The same variables were compared

between subungual/interdigital and volar acral melanomas. P-values < 0.05 attained statistical

significance.

12
 ACCEPTED MANUSCRIPT

FISH

FISH was performed using probes targeting loci commonly aberrant in acral melanoma including

TERT (5p15.33), AURKA (20q13.2), CCND1 (11q13.3), CDK4 (12q14.1), CDKN2A (9p21.3),

PT
PAK1 (11q14.1), PTEN (10q23.31), NF1 (17q11.2) and GAB2 (11q14.1). Probes were prepared

via the manufacturer’s protocol (Empire Genomics, Buffalo, NY). Slides were examined using

RI
the Olympus BX41 microscope and CellSens Standard for nuclei detection and image

SC
processing. FISH enumeration was performed as previously described, and cases were

considered positive if greater than 30% of enumerated cells showed evidence of copy number

U
gains of a given locus for oncogenes or greater than 30% had deletions for tumor suppressor
AN
genes (Gerami et al., 2009, Gerami et al., 2012).

Next Generation Sequencing

M

We evaluated 50 genes using the Ion Torrent PGM and the Ion Torrent AmpliSeq™Cancer
D

Hotspot Panel v2 (Life Technologies, Grand Island, NY). Sequencing methodology was as
TE

previously described for genital melanoma and nevi cases (Yelamos et al., 2016).

Variant Calling
EP

Data generated by the PGM was analyzed using the Ion Torrent Variant Caller plug-in
C

(Life Technologies, Grand Island, NY). Annotated variants were reviewed using the Ion
AC

Reporter version 5.2 online software (Life Technologies) to evaluate variant effect. Intronic

mutations and/or synonymous mutations were excluded. Non-synonymous exonic variants were

included if they occurred with a frequency of > 1%, were not seen in normal controls, and had at

least 5 reads at the mutant allele. Non-synonymous mutations were visualized using the

Integrative Genomic Viewer (Robinson et al., 2011) to exclude strand bias, homopolymers, and

13
 ACCEPTED MANUSCRIPT

mispriming. A two-tailed Fisher’s exact test was used to compare the read distributions between

tumor and normal samples with a threshold of p<0.05 for statistical significance (Dees et al.,

2012). Statistical analysis was performed using SAS 9.4 (SAS Inc., NC). The pathogenicity of all

mutations identified as somatic variants was analyzed using the SIFT score (Kumar et al., 2009)

PT
as well as the Clinvar (Landrum et al., 2016) and COSMIC databases. Mutations found in both

RI
tumors and normal controls were considered germline variants. Germline variants were evaluated

using the ExAC database (Lek et al., 2016) and those found in greater than 1% of the population

SC
were considered polymorphisms and excluded.

U
CastPCR on EGFR T790M AN
NGS revealed EGFR T790M mutations with a very low allele frequency in 8 cases. Only

1 of 8 samples passed our variant calling criteria above. Considering the potential therapeutic
M

significance, castPCR was performed for validation. 20ng of DNA templates from tumor or
D

normal samples was mixed with 1x Taqman genotyping master mix (Life Technologies, NY),

nuclease-free water, and 1x Taqman EGFR T790M mutant assay or 1x Taqman EGFR reference
TE

assay in a 20ul reaction. Three different wildtype DNA and three technical replicates were used
EP

for ∆Ct cutoff determination. All samples were loaded onto a 96-well plate and underwent real-

time PCR on a QuantStudio 7 Flex Real-Time PCR System (Life Technologies, NY). The PCR
C

reaction started with an incubation step at 95°C for 10min, followed by 5 cycles of 92°C for
AC

15sec and 58°C for 1min and then 40 cycles of 92°C for 15sec and 60°C for 1min.

Amplification plots were generated by the QuantStudio Real-time PCR System v1.3

using 0.2 as CT (Threshold Cycle). ∆CT value from wild type DNA was normalized by the

Mutation Detector Software v2.0 to determine the ∆CT cutoff for other tested samples. The

mutation was detected when the normalized ∆CT was lower than the ∆CT cutoff. In our

14
 ACCEPTED MANUSCRIPT

experiment, samples with CT lower than 37 and ∆CT cutoff lower than 9.61 were designated

EGFR T790M positive. CastPCR resulted in one EGFR T790M positive sample, which was

consistent with our NGS result (Table 2; Figure 4).

PT
CONFLICT OF INTEREST

Dr. Gerami has served as a consultant for Myriad Genomics, DermTech Int., and Castle

RI
Biosciences and has received honoraria for this. All other authors have no relevant conflicts of

SC
interest. This work is original and has not been previously published.

U
ACKNOWLEDGEMENTS
AN
This work was partially supported by the IDP Foundation and the Melanoma Research

Foundation (SP0043559).
M
D
TE
C EP
AC

15
 ACCEPTED MANUSCRIPT

REFERENCES
Bentires-Alj M, Paez JG, David FS, Keilhack H, Halmos B, Naoki K, et al. Activating mutations
of the noonan syndrome-associated SHP2/PTPN11 gene in human solid tumors and adult
acute myelogenous leukemia. Cancer research 2004;64(24):8816-20.
Boone B, Jacobs K, Ferdinande L, Taildeman J, Lambert J, Peeters M, et al. EGFR in melanoma:

PT
clinical significance and potential therapeutic target. Journal of cutaneous pathology
2011;38(6):492-502.
Bunda S, Burrell K, Heir P, Zeng L, Alamsahebpour A, Kano Y, et al. Inhibition of SHP2-

RI
mediated dephosphorylation of Ras suppresses oncogenesis. Nature communications
2015;6:8859.
Carrera C, Gual A, Diaz A, Puig-Butille JA, Nogues S, Vilalta A, et al. Prognostic role of the

SC
histological subtype of melanoma on the hands and feet in Caucasians. Melanoma
research 2017;27(4):315-20.
Chraybi M, Abd Alsamad I, Copie-Bergman C, Baia M, Andre J, Dumaz N, et al. Oncogene
abnormalities in a series of primary melanomas of the sinonasal tract: NRAS mutations

U
and cyclin D1 amplification are more frequent than KIT or BRAF mutations. Human
pathology 2013;44(9):1902-11.
AN
Cooper C, Arva NC, Lee C, Yelamos O, Obregon R, Sholl LM, et al. A clinical, histopathologic,
and outcome study of melanonychia striata in childhood. Journal of the American
Academy of Dermatology 2015;72(5):773-9.
M

Costello CM, Pittelkow MR, Mangold AR. Acral Melanoma and Mechanical Stress on the
Plantar Surface of the Foot. N Engl J Med 2017;377(4):395-6.
Dees ND, Zhang Q, Kandoth C, Wendl MC, Schierding W, Koboldt DC, et al. MuSiC:
D

identifying mutational significance in cancer genomes. Genome research

2012;22(8):1589-98.
TE

Duman N, Erkin G, Gokoz O, Karahan S, Kayikcioglu AU, Celik I. Nevus-Associated versus de

novo Melanoma: Do They Have Different Characteristics and Prognoses?
Dermatopathology (Basel) 2015;2(1):46-51.
Forbes SA, Beare D, Gunasekaran P, Leung K, Bindal N, Boutselakis H, et al. COSMIC:
EP

exploring the world's knowledge of somatic mutations in human cancer. Nucleic acids
research 2015;43(Database issue):D805-11.
Furney SJ, Turajlic S, Fenwick K, Lambros MB, MacKay A, Ricken G, et al. Genomic
characterisation of acral melanoma cell lines. Pigment cell & melanoma research
C

2012;25(4):488-92.
Furney SJ, Turajlic S, Stamp G, Thomas JM, Hayes A, Strauss D, et al. The mutational burden of
AC

acral melanoma revealed by whole-genome sequencing and comparative analysis.

Pigment cell & melanoma research 2014;27(5):835-8.
Gerami P, Jewell SS, Morrison LE, Blondin B, Schulz J, Ruffalo T, et al. Fluorescence in situ
hybridization (FISH) as an ancillary diagnostic tool in the diagnosis of melanoma. The
American journal of surgical pathology 2009;33(8):1146-56.
Gerami P, Li G, Pouryazdanparast P, Blondin B, Beilfuss B, Slenk C, et al. A highly specific and
discriminatory FISH assay for distinguishing between benign and malignant melanocytic
neoplasms. The American journal of surgical pathology 2012;36(6):808-17.

16
 ACCEPTED MANUSCRIPT

Guldberg P, thor Straten P, Ahrenkiel V, Seremet T, Kirkin AF, Zeuthen J. Somatic mutation of
the Peutz-Jeghers syndrome gene, LKB1/STK11, in malignant melanoma. Oncogene
1999;18(9):1777-80.
Hafner C, Di Martino E, Pitt E, Stempfl T, Tomlinson D, Hartmann A, et al. FGFR3 mutation
affects cell growth, apoptosis and attachment in keratinocytes. Experimental cell research
2010;316(12):2008-16.
Hafner C, Lopez-Knowles E, Luis NM, Toll A, Baselga E, Fernandez-Casado A, et al.

PT
Oncogenic PIK3CA mutations occur in epidermal nevi and seborrheic keratoses with a
characteristic mutation pattern. Proceedings of the National Academy of Sciences of the
United States of America 2007;104(33):13450-4.

RI
Hafner C, Stoehr R, van Oers JM, Zwarthoff EC, Hofstaedter F, Landthaler M, et al. FGFR3 and
PIK3CA mutations are involved in the molecular pathogenesis of solar lentigo. The
British journal of dermatology 2009;160(3):546-51.

SC
Halilovic E, She QB, Ye Q, Pagliarini R, Sellers WR, Solit DB, et al. PIK3CA mutation
uncouples tumor growth and cyclin D1 regulation from MEK/ERK and mutant KRAS
signaling. Cancer research 2010;70(17):6804-14.
Hayward NK, Wilmott JS, Waddell N, Johansson PA, Field MA, Nones K, et al. Whole-genome

U
landscapes of major melanoma subtypes. Nature 2017;545(7653):175-80.
Janku F, Hong DS, Fu S, Piha-Paul SA, Naing A, Falchook GS, et al. Assessing PIK3CA and
AN
PTEN in early-phase trials with PI3K/AKT/mTOR inhibitors. Cell reports 2014;6(2):377-
87.
Johnston JJ, Rubinstein WS, Facio FM, Ng D, Singh LN, Teer JK, et al. Secondary variants in
individuals undergoing exome sequencing: screening of 572 individuals identifies high-
M

penetrance mutations in cancer-susceptibility genes. American journal of human genetics

2012;91(1):97-108.
Katunaric M, Jurisic D, Petkovic M, Grahovac M, Grahovac B, Zamolo G. EGFR and cyclin D1
D

in nodular melanoma: correlation with pathohistological parameters and overall survival.

Melanoma research 2014;24(6):584-91.
TE

Kong Y, Si L, Li Y, Wu X, Xu X, Dai J, et al. Analysis of mTOR Gene Aberrations in

Melanoma Patients and Evaluation of Their Sensitivity to PI3K-AKT-mTOR Pathway
Inhibitors. Clinical cancer research : an official journal of the American Association for
EP

Cancer Research 2016;22(4):1018-27.

Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymous variants on
protein function using the SIFT algorithm. Nature protocols 2009;4(7):1073-81.
Landrum MJ, Lee JM, Benson M, Brown G, Chao C, Chitipiralla S, et al. ClinVar: public
C

archive of interpretations of clinically relevant variants. Nucleic acids research

2016;44(D1):D862-8.
AC

Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fennell T, et al. Analysis of
protein-coding genetic variation in 60,706 humans. Nature 2016;536(7616):285-91.
Liberzon E, Avigad S, Yaniv I, Stark B, Avrahami G, Goshen Y, et al. Molecular variants of the
ATM gene in Hodgkin's disease in children. British journal of cancer 2004;90(2):522-5.
Liu Y, Sun L, Xiong ZC, Sun X, Zhang SL, Ma JT, et al. Meta-analysis of the impact of de novo
and acquired EGFR T790M mutations on the prognosis of patients with non-small cell
lung cancer receiving EGFR-TKIs. OncoTargets and therapy 2017;10:2267-79.
Mangone FR, Miracca EC, Feilotter HE, Mulligan LM, Nagai MA. ATM gene mutations in
sporadic breast cancer patients from Brazil. SpringerPlus 2015;4:23.

17
 ACCEPTED MANUSCRIPT

Mohammed AA, El-Tanni H, Alsakkaf MA, Mirza AA, Atiah TA, Atiah AA. Genomic
aberrations in non- small cell lung cancer and their impact on treatment outcome. Journal
of cancer research and therapeutics 2017;13(1):9-15.
Moss TJ, Qi Y, Xi L, Peng B, Kim TB, Ezzedine NE, et al. Comprehensive Genomic
Characterization of Upper Tract Urothelial Carcinoma. European urology 2017.
Offit K, Gilad S, Paglin S, Kolachana P, Roisman LC, Nafa K, et al. Rare variants of ATM and
risk for Hodgkin's disease and radiation-associated breast cancers. Clinical cancer

PT
research : an official journal of the American Association for Cancer Research
2002;8(12):3813-9.
Omholt K, Krockel D, Ringborg U, Hansson J. Mutations of PIK3CA are rare in cutaneous

RI
melanoma. Melanoma research 2006;16(2):197-200.
Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, et al. Integrative
genomics viewer. Nature biotechnology 2011;29(1):24-6.

SC
Rowan A, Bataille V, MacKie R, Healy E, Bicknell D, Bodmer W, et al. Somatic mutations in
the Peutz-Jeghers (LKB1/STKII) gene in sporadic malignant melanomas. The Journal of
investigative dermatology 1999;112(4):509-11.
Shackelford DB, Shaw RJ. The LKB1-AMPK pathway: metabolism and growth control in

U
tumour suppression. Nature reviews Cancer 2009;9(8):563-75.
Shain AH, Garrido M, Botton T, Talevich E, Yeh I, Sanborn JZ, et al. Exome sequencing of
AN
desmoplastic melanoma identifies recurrent NFKBIE promoter mutations and diverse
activating mutations in the MAPK pathway. Nature genetics 2015;47(10):1194-9.
Silva JM, Deuker MM, Baguley BC, McMahon M. PIK3CA-Mutated Melanoma Cells Rely on
Cooperative Signaling through mTORC1/2 for Sustained Proliferation. Pigment cell &
M

melanoma research 2017.

Stern DK, Creasey AA, Quijije J, Lebwohl MG. UV-A and UV-B penetration of normal human
cadaveric fingernail plate. Arch Dermatol 2011;147(4):439-41.
D

Tartaglia M, Gelb BD. PTPN11 (protein tyrosine phosphatase, non-receptor type, 11). Atlas of
Genetics and Cytogenetics in Oncology and Haematology 2011(2).
TE

Teramoto Y, Keim U, Gesierich A, Schuler G, Fiedler E, Tuting T, et al. Acral lentiginous

melanoma - a skin cancer with unfavourable prognostic features. A study of the German
Central Malignant Melanoma Registry (CMMR) in 2050 patients. The British journal of
EP

dermatology 2017.
Viros A, Fridlyand J, Bauer J, Lasithiotakis K, Garbe C, Pinkel D, et al. Improving melanoma
classification by integrating genetic and morphologic features. PLoS medicine
2008;5(6):e120.
C

Weinstein JN, Collisson EA, Mills GB, Shaw KR, Ozenberger BA, Ellrott K, et al. The Cancer
Genome Atlas Pan-Cancer analysis project. Nature genetics 2013;45(10):1113-20.
AC

Yelamos O, Merkel EA, Sholl LM, Zhang B, Amin SM, Lee CY, et al. Nonoverlapping Clinical
and Mutational Patterns in Melanomas from the Female Genital Tract and Atypical
Genital Nevi. The Journal of investigative dermatology 2016;136(9):1858-65.
Yun CH, Mengwasser KE, Toms AV, Woo MS, Greulich H, Wong KK, et al. The T790M
mutation in EGFR kinase causes drug resistance by increasing the affinity for ATP.
Proceedings of the National Academy of Sciences of the United States of America
2008;105(6):2070-5.
Zhou W, Zhang J, Marcus AI. LKB1 Tumor Suppressor: Therapeutic Opportunities Knock when
LKB1 Is Inactivated. Genes & diseases 2014;1(1):64-74.

18
 ACCEPTED MANUSCRIPT

TABLE 1. Variants identified on next generation sequencing and FISH. Oncogenes are shown in green and tumor suppressor genes are
shown in red. FISH results are shown on the far right, with gains in blue and losses in yellow. Melanomas with insufficient tissue for FISH are
shaded grey.
NRA EGF FGF PIK3 PTP STK IDH IDH TP5 PTE TER CDK AUK CCN CDK PAK GAB

PT
A G R T Stage Nevus Site BRAF KIT APC ATM ALK NF1
S R R3 CA N11 11 1 2 3 N T N2A RA D1 4 1 2
N
53 F C SS MIS D V600K G

RI
38 F C SS MIS D V600E

Y
63 F C SS MIS D V600K

SC
R34
65 M LM MIS D L597R
2*
Y
67 F C SS MIS D V600E

N K65

U
52 F C SS MIS D V600E
2M
N
73 F SS T1a D V600E

AN
N
62 F C SS T1a D V600E

Y
35 F C SS T1a D V600E L

M
Y
22 F C SS T2a D V600E

N
60 F C SS T3a D V600E

D
N Q61
28 F SS MIS D
K

TE
N Q61
43 F C SS T1b D
K
Y Q61 R13
61 F C N T2a D
R 2C
N
EP
Q61 L576
52 F C AL MIS D
R P
N
56 M C SS T1a D

N
C

42 F C SS T1a D

N
79 F C N T4b D V600E L L G G G G
AC

N
62 F C SS T1a D V600E G

N
71 F C SS T2a D V600E

N D81
64 F C AL MIS D G G G
6V
Y
45 F C SS T1a V V600E

19
 ACCEPTED MANUSCRIPT

N
70 F C SS T2a V V600E

N Q61 T79
50 F H AL MIS V
K 0M
N Q61
39 F AL T1a V C3Y
R

PT
N Q61
76 M C AL T2a V
R
A N Y55 R33
58 F AL T1a V G
A 7R 7C

RI
N
81 M C AL T2b V

N
62 F C AL MIS V

SC
N
49 F H SS T1a V

Y
73 F C AL T3b I

U
A N
77 F N T4b I V600E L G G G
A
N Q61

AN
71 M C AL T2a I L
R
A N G12 P19
71 M AL T2b SU G G
A D H
N
77 F C AL MIS SU G G G G L

M
N
35 F C AL MIS SU

N R10

D
82 F H AL MIS SU G
Q
*8 N L704 M10 E57 M11 R11 G12
F H AL T1a SU G G G

TE
0 F 40V K 4V 05Q 01R
N
44 M AL T2a SU

N Q51
66 M C AL T3b I G G
0L
EP
N H10
68 F C AL T3b SU
47L
A N
58 M AL T4b I G
s
C

N N64 P28
65 F C N T4b I
6D 1R
AC

A = age, G = gender, R = race, T = subtype, D = dorsal, V = volar, SU = subungual, I = interdigital

*This patient had a germline mutation in MLH1 and had several other mutations in genes not found in other melanomas including MET A1261T, SMO W535R, FGFR2 G305R, HNF1A R272C, CDH1 T73A, ERBB2 D880N and SMAD4 R100G.

20
 ACCEPTED MANUSCRIPT

Table 2. Consistency of results from NGS and CastPCR for EGFR T790M detection.

EGFR T790M Allele Frequency from CastPCR ∆CT

Sample type NGS cutoff CastPCR Normalized ∆CT

PT
Tumor 1.90% 9.61 9.35
Negative
Control n/a 9.61 13.16

RI
SC
FIGURE LEGENDS
Figure 1. Histologic and FISH findings in a superficial spreading MIS, subungual acral lentiginous MIS, and an ALM. Case 83

U
(a, b, c). Superficial spreading MIS (panel b) on the left dorsal foot of a 53-year-old female patient with a BRAF V600E mutation
(panel a) and copy number gains in TERT on FISH (panel c). Case 29 (d, e, f). Subungual acral lentiginous MIS (panel d and e) on the

AN
left great toe of an 82-year-old female patient with no mutations identified on NGS but significant copy number gains in CCND1 on
FISH (panel f). Case 60 (g, h, i). Acral lentiginous melanoma found on the finger of a 68-year-old female with a PIK3CA H1047L

M
mutation (panel g). Notice the acral lentiginous growth pattern with confluent, single atypical melanocytes in the basal layer of the
epidermis without significant pagetosis (panel g and h). Scale bars = 0.2mm in all micrographs.

D
Figure 2. Melanoma signaling pathway depicting degrees of UV-driven mutagenesis and mutational diversity among acral
sites. Schematic showing the RAS/RAF/MEK pathway as well as the PI3K/AKT pathway. Genes or proteins are shown with their

TE
relative mutation distributions among dorsal, volar, and subungual melanomas with dorsal in red, volar in blue, and subungual in
green. The sun symbol indicates ultraviolet predominant pathways. Activation of the PI3K/AKT pathway, primarily through silencing
EP
of PTEN, is a frequent finding in BRAF-mutated melanomas. Mutations in the catalytic subunit of PI3K (PIK3CA) were found with
high frequency in our sun-shielded melanomas. This mutation is relatively uncommon in sun-exposed melanoma, and may represent
C

an alternative means of PI3K/AKT pathway activation in sun-shielded tumors.

AC

Figure 3. Crystal structures showing identified mutations in their respective domains. Structures are labeled and color coded by
their major domains and mutations are represented by red spheres. (a) The crystal structure of PIK3CA (Protein Data Bank Code
4ZOP) is shown with the M1040V and H1047L mutations localizing to the same C-terminal alpha-helix within the catalytic domain of
PIK3CA (gray). (b) The crystal structure of EGFR tyrosine kinase domain (Protein Data Bank Code 2J6M) is shown with the T790M
and L704F mutations. (c) The crystal structure of SHP2 (Protein Data Bank Code 2SHP), encoded by PTPN11 gene, is shown with a

21
 ACCEPTED MANUSCRIPT

Q510L mutation. (d) The crystal structure of STK11 (Protein Data Bank Code 2WTK) is shown with E57K and P281R mutations.
Both mutations are in the serine/threonine-protein kinase domain.
Figure 4. Amplification plot showing signals for tumor sample vs. negative control. Cast-PCR for validating the EGFR T790M

PT
mutation in an acral melanoma patient. Of note, the tumor sample showed a stronger amplification signal than the negative control.

RI
U SC
AN
M
D
TE
C EP
AC

22
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC