ORIGINAL
ARTICLE
Primary dermal melanoma: A unique subtype of melanoma to be distinguished from cutaneous metastatic melanoma A clinical, histologic, and gene expressioneprofiling study Michael Sidiropoulos, MD, MS,a Roxana Obregon, BA,a Chelsea Cooper, BA,a Lauren Meldi Sholl, MS,a Joan Guitart, MD,a,b and Pedram Gerami, MDa,b Chicago, Illinois Background: Primary dermal melanoma (PDM) is a subtype of melanoma confined to the dermis that may be morphologically impossible to distinguish from cutaneous metastatic melanoma (CMM). Objective: We sought to better characterize PDM by describing the clinical, histologic, and molecular features of 49 cases of PDM and determine whether a gene expressioneprofiling test could help distinguish PDM from CMM. Methods: We describe 49 cases of PDM and determined whether any clinical or histopathologic features had a statistically significant relationship with outcome. Secondly, we performed a melanoma gene expressioneprofiling test on a subset of the PDM and CMM cases. Results: Overall recurrence was infrequent and seen in 9 of 49 cases. Six patients had locoregional recurrences and 3 patients had distant metastasis. None of the clinical or histologic parameters showed a statistically significant relationship with recurrence. There was a statistically significant association of a class I signature by DecisionDx-Melanoma assay (Castle Biosciences Inc, Friendswood, TX) for PDM whereas CMM were more frequently class II (P value = .023). Limitations: The mean follow-up time was 26 months. Conclusions: Most conventional staging parameters used for prognosis in cutaneous melanoma have limited applicability to PDM. The melanoma prognostic assay may be a useful tool for distinguishing PDM from CMM. ( J Am Acad Dermatol http://dx.doi.org/10.1016/j.jaad.2014.07.051.) Key words: American Joint Committee on Cancer; cutaneous metastatic melanoma; fluorescence in situ hybridization; gene expression profile; melanoma; primary dermal melanoma.
rimary dermal melanoma (PDM) is a variant of melanoma confined to the dermis and/or subcutaneous tissue without an epidermal component.1-3 The main histologic differential diagnosis for PDM is cutaneous metastatic melanoma
P
(CMM). Diagnosis of PDM requires excluding the possibility of metastasis from a separate primary melanoma. There are few case series in the literature characterizing PDM, but the existing literature suggests a remarkably good, long-term survival
From the Department of Dermatologya and Robert H. Lurie Cancer Center,b Feinberg School of Medicine, Northwestern University. Supported by the Irene D. Pritzker Foundation and partially supported by Castle Biosciences Inc. Disclosure: Dr Gerami has served as a consultant to Castle Biosciences Inc, Myriad Genetics, and DermTech Inc, receiving honoraria. Dr Guitart has served as a consultant to Castle Biosciences Inc, receiving honoraria. The other authors declared no conflicts of interest. Dr Sidiropoulos and Ms Obregon contributed equally to this article.
Accepted for publication July 29, 2014. Reprint requests: Pedram Gerami, MD, Department of Dermatology, Feinberg School of Medicine, Northwestern University, 676 N Saint Clair St, Suite 1765, Chicago, IL 60611. E-mail: [email protected]. Published online September 23, 2014. 0190-9622/$36.00 Ó 2014 by the American Academy of Dermatology, Inc. http://dx.doi.org/10.1016/j.jaad.2014.07.051
1
J AM ACAD DERMATOL
2 Sidiropoulos et al
compared with similarly staged conventional and 15 cases of CMM were identified in our dermatocutaneous melanomas.1-9 pathology database at Northwestern University. Differentiating PDM from CMM has significant All lesions were evaluated by 3 dermatopathologists staging and therapeutic implications. Although (J. G., P. G., and M. S.). Clinical histories and demost patients with PDMs including those with scriptions of lesions were obtained from medical higher Breslow depths have excellent long-term surrecords. Fourteen cases were diagnosed within our vival (80%-100%), the presence of CMM institution and 35 cases were external consultation indicates stage IIIC disease cases. In addition to a histowith estimated 10-year surpathologic diagnosis consisvival of 24%.10,11 Patients CAPSULE SUMMARY tent with PDM, inclusion with stage IIIC disease may criteria required a thorough Primary dermal melanoma has a far be candidates for adjuvant physical examination better prognosis than cutaneous therapy or more aggressive excluding the possibility of a metastatic melanoma, although these imaging protocols. Alterseparate primary melanoma, entities may in some cases be natively for patients with ulceration, or regression. indistinguishable. PDM, excision or excision The following clinical with sentinel lymph node bifeatures were assessed and Most conventional prognostic markers opsy (SLNB) and clinical documented for each PDM used for cutaneous melanoma have follow-up may be adequate case: patient’s age, sex, site of limited applicability to primary dermal therapy. At the histologic the lesion, and clinical appearmelanoma. level, this distinction is at times ance as initially described by The melanoma prognostic assay may impossible. The vast majority the submitting dermatologist help distinguish primary dermal of cases can be confidently (Tables I and II). A number of melanoma from cutaneous metastatic diagnosed based on the clinhistopathologic parameters melanoma. ical history and examination noted in Tables I and II were findings; however, there will recorded for each case. We remain a proportion of cases also subtyped the cases into 1 that, despite adequate clinical information, cannot be of 4 morphologic categories: Spitzoid; blue nevuselike resolved. The goal of the current study is to describe in morphology; associated with an intradermal nevus depth the clinical, histologic, immunohistochemical (IDN); and none of the above (not otherwise specified (IHC), and molecular characteristics of a cohort of [NOS]). PDMs and determine whether a novel molecular In 33 of 49 PDM cases and in 25 of 49 PDM cases prognostic test for melanoma can assist in differentifluorescence in situ hybridization (FISH) and IHC ating this entity from CMM. studies were performed, respectively. The results The prognostic test for melanoma, also known of the FISH findings and the IHC stains are reported as DecisionDx-Melanoma (Castle Biosciences Inc, in Table II. FISH was performed using probes Friendswood, TX), is a gene expressioneprofiling targeting chromosome 6p25 (RREB1), chromosome (GEP) test that looks at the messenger RNA 6q23 (MYB), chromosome 11q13 (CCND1), the expression profile of 28 genes. The data are centromeric portion of chromosomes 6 (CEP6), reported as either a class I (low risk) or class II chromosome 9p21 (CDKN2A) and 8q24 (MYC), as (high risk) signature. The assay has been shown previously described.14 to be an independent prognostic marker for In 13 PDM cases there was sufficient tissue cutaneous melanoma in multivariate analysis.12,13 available for messenger RNA GEP testing. Fifteen Because PDM has a good prognosis, we exCMM cases were evaluated as control specimens. pected a considerable proportion of PDMs to Thirteen PDM and 15 CMM cases with formalin-fixed have a class I signature whereas most CMMs paraffin embedded melanoma samples were would be class II. Hence, we studied a subset of macrodissected and analyzed blindly by the the PDMs that had adequate tissue for the GEP DecisionDx-Melanoma GEP signature assay (Castle testing and a control group of CMMs to evaluate Biosciences Inc).12,13,15 Fisher exact test was used to this hypothesis. compare the frequency of class I and class II results in the PDM and CMM cases. Clinical follow-up was available for 48 of the 49 METHODS reported PDM cases. All patients had a physical After obtaining approval from the Northwestern examination excluding a separate primary University Cancer Center, Chicago, IL, and internal melanoma. In 22 of 49 patients, imaging studies institutional review board, 49 cases of PDM d
d
d
J AM ACAD DERMATOL
Abbreviations used: CGH: CMM: FISH: GEP: IDN: IHC: NOS: PDM: SLNB:
comparative genomic hybridization cutaneous metastatic melanoma fluorescence in situ hybridization gene expression profiling intradermal nevus immunohistochemical not otherwise specified primary dermal melanoma sentinel lymph node biopsy
were performed to exclude metastatic disease. Clinical data were retrieved from medical records and patient interviews (Table III). None of the patients had clinically palpable lymph node disease at the initial presentation. Frequency tables were generated to summarize the categorical variables between PDM cases with and without recurrence. The Fisher exact test was used for comparison. Two-sample t test was used to compare continuous variables between the 2 groups. Comparisons between the groups were considered statistically significant at a P value less than .05.
RESULTS The clinical findings for the 49 PDM cases are summarized in Table I. The database review of 6000 patients resulted in 49 patients (0.82%) with a diagnosis of PDM. The 49 PDM included 25 (51%) male and 24 (49%) female patients, ranging in age from 8 to 83 years (mean 51.2 years). The most common anatomic site of involvement was the trunk (19/49; 39%), followed by the head and neck (16/49; 33%), extremities (11/49; 22%) and acral area (3/49; 6%). The histopathologic and IHC findings are summarized in Tables II and III. Histopathologically, all 49 cases (100%) of PDM were characterized by a dermalbased melanocytic neoplasm (Fig 1). There was no evidence of an overlying in situ component or regression in any of the cases analyzed. The mean Breslow depth was at least 3.01 mm (range 0.5 to $ 10.05 mm). All cases demonstrated a Clark level equal to or greater than III: Clark level III (2/49; 4%), Clark level IV (34/49; 69%), and Clark level V (5/49; 11%). In 8 cases the base of the lesion was transected and a definitive assessment of the Clark level could not be made (8/49; 16%). The mean mitotic rate was 4.3 mitoses/mm2 (range 1-32/mm2). An associated benign IDN was identified in 14 cases (14/49; 29%) (Fig 2). Angiotropism (5/49; 10%), angioinvasion (2/49; 4%), and perineural invasion (3/49; 6%) were uncommonly identified. Morphologically, the cases could be roughly categorized into 4 patterns: Spitzoid (13/49; 26%), blue nevuselike (3/49; 6%), PDM in
Sidiropoulos et al 3
association with an IDN (14/49; 29%), and NOS (19/ 49; 39%). Among the cases with a clinical description, the most common differential diagnosis was melanocytic nevus or melanoma (16/28; 57%), nonmelanoma skin cancer (6/28; 22%), dermatofibroma (2/ 28; 7%), cyst (2/28; 7%), and scar (2/28; 7%). Of 49 cases, 33 were tested by FISH and criteria for FISH were met in 28 of 33 cases analyzed (85%) (Table II). The following copy number aberrations were noted: gains at 6p25 (19/33; 58%), gains at 11q13 (8/33; 24%), deletions of 6q23 (4/33; 12%), homozygous deletions of 9p21 (7/27; 26%), and gains at 8q24 (7/11; 64%). Among the Spitzoid PDM, the most common copy number aberration was a homozygous deletion of 9p21 seen in 6 of 10 cases (60%). Among the blue nevuselike PDM, 2 of 3 cases (67%) showed copy number gains at 6p25 and 2 of 3 cases (67%) had copy number gains at 11q13. Among the PDM associated with an IDN, the most common copy number aberration was gain of 6p25 seen in 7 of 12 cases (58%). Among the NOS cases, 6 of 8 (75%) showed copy number gains at 6p25 and of the 5 tested for 8q24 change, 5 (100%) showed copy number gains at 8q24. Comparison of morphologic and American Joint Committee on Cancer parameters between PDM and CMM cases (Table IV) demonstrated statistical significance with the presence of an associated nevus in PDM (P = .03), Clark level (P = .0001), and age (P \.0001). GEP using the DecisionDx-Melanoma assay (Castle Biosciences Inc) in 13 PDM cases demonstrates class I (low risk) expression in 11 cases and class II (high risk) expression in only 2 cases (Table IV). Among the 15 cases of CMM, the assay demonstrated class I expression in 6 cases and class II expression in 9 cases. The findings are summarized in Table IV. The Fisher exact test gave a P value of .023 showing a statistically significant association of class I signature with PDM. At least 2 months of clinical follow-up was available in 48 of 49 patients (Table III). The average follow-up time was 26 months and it ranged from 2 to 108 months. In 22 of 48 patients (46%), imaging studies were performed and 3 of 22 cases (14%) showed evidence of distant metastasis. Overall, 9 of 48 (19%) patients developed recurrence despite re-excision with clear margins from the primary melanoma. Three patients (3 of 48; 6%) developed distant metastatic disease, 2 to the lung and liver and 1 with isolated lung metastasis. Locoregional recurrence of melanoma was present in 6 patients (6 of 48; 13%), 5 of whom developed in transit metastasis and 1 with regional lymph node disease. SLNB was performed in 34 of 48 (71%) patients and
J AM ACAD DERMATOL
4 Sidiropoulos et al
Table I. American Joint Committee on Cancer characteristics of primary dermal melanoma cases Case
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
Age, y
Sex
Site
Site code
60 75 70 17 55 71 30 54 83 41 32 29 57 48 45 80 52 53 51 59 28 51 75 39 72 68 8 63 33 66 74 26 63 51 63 58 24 62 47 35 55 72 58 58 38 14 62 17 66
F M M F M F F M M F F M M M F M F M F M F F M F F F M M M M M F M M F F F F M F F F F F M M M M M
Eye, right Forehead Ear, left Arm, left Nose Scalp, right Shoulder, left Buttock, right Sideburn, right Thigh, left Nose, tip Neck, posterior Back, right Ear, left Ear, right Back, mid Scalp, mid Ear, left Thigh, left upper Thigh, left Scalp, crown Abdomen, mid Breast, right Ankle, left Cheek, left Forehead, medial Back Scapula, left Calf, left Arm, left Ear, right Thigh, left Back, upper Back, left upper Back, right Thigh, right Foot, left Back, left Umbilicus Arm, right Back, left Great toe, right Arm, right Arm, right Chest, mid Back, right upper Back, mid Back Chest wall, posterior
1 1 1 3 1 1 2 2 1 3 1 1 2 1 1 2 1 1 3 3 1 2 2 4 1 1 2 2 3 3 1 3 2 2 2 3 4 2 2 3 2 4 3 3 2 2 2 2 2
Breslow depth, mm
At
At At At
At At At At At
At
At
At
At
At At
3.3 1.05 4.8 least 7.5 0.5 6.5 least 0.9 least 5.0 least 0.9 4.5 0.85 least 1.0 least 3.2 least 0.9 least 1.16 1.2 least 2.4 1.1 1.5 2.7 3.5 0.65 6.7 least 1.1 1.33 1.1 least 2.8 7.5 0.6 1.7 least 2.5 7 7.6 10.05 0.53 least 5.6 1.9 1.5 9.3 least 1.45 least 1.7 3.8 1.2 1.4 3.3 3.9 0.9 3.5 2.2
Clark level
Ulceration
Mitoses, /mm2
AJCC T stage
4 4 4 5 4 5 4 4 least 4 4 least least 4 least 4 4 4 4 4 4 4 4 least 4 4 4 4 3 4 4 4 4 5 3 least 4 4 5 least least 4 4 4 4 4 4 4 4
No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No
1 2 6 4 2 3 4 30 3 3 2 2 8 1 1 1 3 1 5 4 3 2 32 2 4 3 6 8 3 5 1 1 16 10 1 2 1 1 2 1 1 1 2 3 2 1 2 2 2
T3a T1b T4a T4a T1b T4a T1b T4a T1b T4a T1b T1b T3a T1b T2a T1b T3a T2a T2a T3a T3a T1b T4a T2a T2a T2a T3a T4a T1b T2a T3a T4a T4a T4a T1b T4a T2a T2a T4a T2a T2a T3a T2a T2a T3a T2a T1b T3a T3a
At
At At At
At
At
At At
4
4 4 4
4
4
4 4
AJCC, American Joint Committee on Cancer; F, female; M, male.
was positive in 6 (6 of 34; 18%) patients. A positive SLNB specimen did have a trend toward adverse prognosis according to Cox proportional hazards with a P value of .07.
None of the parameters studied including age, sex, anatomic site, Breslow depth, Clark level, mitotic rate, ulceration, regression, angioinvasion, angiotropism, perineural invasion, SLNB, Spitzoid
J AM ACAD DERMATOL
Sidiropoulos et al 5
Table II. Histopathologic characteristics of primary dermal melanoma cases Case
Clinical presentation
1
Histologic subtype
NOS
2
BCC vs SCC
NOS
3
8-mm Smooth, pink, pedunculated papule with crusted surface; fibrolipoma vs BCC vs SCC
IDN
IHC results
FISH results
S-1001, Mart-1 , Melan-A , MITF , HMB45 , p16 , Ki67 \5% S-100 , Mart-1 , Melan-A , MITF , HMB451, p161, Ki67 10% S-1001, Mart-11, HMB45 , Ki67 10%
N/A
-
-
-
6p25 Gain
-
-
-
8q24 Gain
-
-
-
6p25 Gain, 8q24 gain
-
-
-
N/A
-
-
-
11q13 Gain
-
-
-
9p21 Deletion 6q23 Loss N/A 6p25 Gain, 11q13 gain N/A N/A 6p25 Gain, 11q13 gain 6p25 Gain 6p25 Gain, 6q23 loss 9p21 Deletion N/A
-
-
1
N/A
-
-
-
6p25 Gain
-
-
-
N/A
-
-
-
N/A 11q13 Gain, 9p21 deletion N/A 9p21 Deletion
-
-
1 -
8q24 Gain
-
-
-
6p25 Gain, 6q23 loss, 11q13 gain 8q24 Gain N/A 6p25 Gain, 11q13 gain
-
1
-
1
-
-
28
Blue nevus S-1001, Melan-A1, HMB451, Ki67 10% Dark macule, rule out lentigo NOS Melan-A1, HMB451, Ki67 \10% Blue pigmented lesion; rule Blue nevus out blue nevus Spitzoid IDN BCC vs SCC IDN Ki67 10% Spitzoid IDN BCC Spitzoid S-1001, Melan-A1, HMB45e Spitzoid Melan-A1, Ki67 5% Nevus IDN Melan-A1, HMB451, p16 Spitzoid HMB451, p161, Ki67 \5% IDN Mart-11, Ki67 5% Two-toned nodule IDN 3mm; BCC vs SCC Spitzoid Dome-shaped firm NOS hemorrhagic papule; sclerosing hemangioma DF Changing nodule X years; IDN pigmented BCC vs malignant melanoma 5-mm Pink/white firm papule; IDN irritated nevus vs DN vs recurrent nevus vs scar 5 mm; Atypical nevus vs NOS malignant melanoma Changing pigmented lesion NOS S-1001, Mart-11 Spitzoid IDN IDN P16 2- to 3-wk History of 4 dark blue-black papules grouped on medial forehead; traumatic vs melanocytic Lesion present for 9 mo, NOS S-1001, Melan-A1, Ki67 10% enlarging; rule out NF vs DF vs other Melanocytic lesion Blue nevus Ki67 10%
29 30 31
NOS NOS IDN
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
20
21
22 23 24 25 26
27
Atypical nevus
S-1001, Melan-A1
AT PNI AI
Continued
J AM ACAD DERMATOL
6 Sidiropoulos et al
Table II. Cont’d Case
32 33 34 35 36 37 38 39
Clinical presentation
Recurrent inflamed nodule; inflamed cyst 1.5-cm Cystic nodule; rule out cyst 4-mm Pink papule; rule out NMSC Solitary 1.2-cm red violaceous plaque History of melanoma; rule out melanoma 5-mm Red papule; DN
Histologic subtype
IDN NOS NOS NOS Spitzoid
Spitzoid NOS
Rule out DN Small nodular lesion
NOS NOS
42
Patient with a solitary mass; DDx includes a benign nevus with congenital features/scar, and an atypical Spitzoid tumor/ melanoma Single pigmented lesion
IDN
46 47 48 49
FISH results
S-1001, Mart-11, MITF1
6p25 Gain, 6q23 loss N/A
1
1
-
N/A
-
-
-
N/A
-
-
-
6p25 Gain, 9p21 deletion
1
-
-
6p25 Gain, 9p21 deletion
1
-
-
N/A
-
-
-
6p25 Gain N/A
-
-
-
6p25 Gain
-
1
-
6p25 Gain, 8q24 gain 6p25 Gain, 8q24 gain 1 6p25 Gain, 11q13 gain, 9p21 deletion 6p25 Gain, 8q24 gain 6p25 Gain, 11q13 gain -
-
-
-
-
S-1001, Mart-11, MITF1, HMB45 S-1001, Melan-A1, HMB451, p16 , Ki67 5% Ki67 5%
Spitzoid
40 41
43 44 45
IHC results
P16 S-1001, Melan-A1, MITF1, HMB451 P161 S-1001, Melan-A1, MITF1, Sox101
NOS NOS Spitzoid Spitzoid NOS NOS Spitzoid
S-1001, Melan-A1
AT PNI AI
AI, Angioinvasion; AT, angiotropism; BCC, basal cell carcinoma; DDx, differential diagnosis; DF, dermatofibroma; DN, dysplastic nevus; FISH, fluorescence in situ hybridization; IDN, intradermal nevus; IHC, immunohistochemistry; N/A, not available; NF, neurofibroma; NMSC, nonmelanoma skin cancer; NOS, not otherwise specified; PNI, perineural invasion; SCC, squamous cell carcinoma.
morphology, blue nevuselike morphology, an associated IDN, and NOS morphology had a statically significant relationship with recurrence.
DISCUSSION Numerous previous studies, mostly reported as single-center series and case reports, have identified patients with a solitary focus of melanoma confined to the dermis with no known separate primary melanoma and negative metastatic findings on workup, who largely remain disease free with a favorable prognosis.1-9 As the prognosis and treatment for a PDM differs greatly from that of a stage IIIC or stage IV M1a metastatic melanoma, the distinction is critical. At the histologic level this distinction is often impossible. Alternatively, at the clinical level the diagnosis may be obvious in many cases because of a known nearby or widely
metastatic separate primary melanoma. Hence, good clinical correlation may solve many of the cases. However in a considerable proportion of cases, some diagnostic uncertainty may remain even after thorough clinical and histologic evaluation. Importantly, patients with CMM are candidates for adjuvant therapy and regular imaging. Typically patients with PDM would not receive adjuvant therapy or require regular imaging. In comparing the clinical features of PDM to typical primary cutaneous melanomas there are some interesting distinctions. In 12 of 28 cases where a clinical differential diagnosis was available, a melanocytic neoplasm was not among the differential diagnoses mentioned on the requisition. In these 12 cases the primary differential diagnosis was nonmelanoma skin cancer, dermatofibroma, cyst, or scar. Hence the ABCD rule will often not apply to
J AM ACAD DERMATOL
Sidiropoulos et al 7
Table III. Treatment, metastatic staging workup, and outcome of primary dermal melanoma cases Case
WLE
SLNB results
Imaging
1 2
1 1
Negative (2007) ND
CTXR/CT1
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Negative (2008) Negative (2009) Negative (2010) Positive (2010) Negative (2011) Negative (2010) ND Negative (2011) ND ND Negative (2011) ND Negative (2011) Negative (2011) Negative (2011) Negative (2011) Negative (2011) Negative (2011) Negative (2011) Positive (2010) Negative (2011) Negative (2012) Negative (2012) ND ND Negative (2013) ND Negative (2012) ND Positive (2013) Negative (2012)
CTXRND N/A PETN/A N/A CT/PET/MRIN/A N/A ND N/A N/A N/A ND XRXRCT1 CT/PETCT/PET1 CT/PET/MRI1 XRXRN/A N/A CTCT/PETXRND N/A CT/PET-
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Negative (2012) Negative (2012) ND Positive (2013) Positive (2013) Positive (2013) ND Negative (2013) Negative (2013) ND ND ND ND Negative (2010) ND Negative (2005)
CTXRN/A CTN/A CT/PETND N/A N/A ND ND N/A N/A ND N/A N/A
Follow-up, mo
Region of concern
Outcome
Lungs
NED DM (lungs)
74 23
NED NED NED NED NED NED NED NED N/A NED NED LR (regional) NED NED NED NED NED NED NED DM (liver/lungs) DM (liver/lungs) NED NED LR (regional) NED NED LR (lymph node) NED NED NED LR (regional)
60 50 37 36 36 29 28 24 N/A 23 7 22 21 21 29 26 23 28 22 24 17 2 12 18 18 18 13 8 11 10 14
Alive DOD December 2009 Alive Alive Alive Alive Alive Alive Alive Alive N/A Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive
LR (regional) NED NED NED LR (regional) NED NED NED NED NED NED NED NED NED NED NED
10 7 6 5 5 5 6 5 5 3 3 99 86 43 108 81
Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive
Liver/lungs Liver/lungs
Enhancing soft-tissue mass suspicious for local regional metastasis, no visceral involvement
Current status
CT, Computerized tomography; DM, distant metastasis; MRI, magnetic resonance imaging; LR, local recurrence; N/A, not available; ND, not done; NED, no evidence of disease; PET, positron emission tomography; SLN, sentinel lymph node; SLNB, sentinel lymph node biopsy; WLE, wide local excision; XR, x-ray.
8 Sidiropoulos et al
J AM ACAD DERMATOL
Fig 1. A, Low-power magnification showing sheetlike proliferation of melanoma cells filling the dermis with no involvement of the epidermis. B, Low-power magnification revealing bulbous extensions of the melanoma extending into the subcutaneous tissue. C, Intermediatepower magnification showing sheetlike proliferation of atypical and pleomorphic melanocytes. D, High-power showing high-grade nuclear atypia, pleomorphism, and mitotic activity.
Fig 2. A, Low-power image of a large primary dermal melanoma, Spitzoid type. B, High-power view demonstrating high-grade nuclear atypia with multiple mitotic figures, including an atypical form. C, Fluorescence in situ hybridization (FISH) image showing homozygous 9p21 deletion within the tumor cells. D, FISH image showing intact centromeric component of chromosome 9 as control.
J AM ACAD DERMATOL
Sidiropoulos et al 9
Table IV. Comparison of morphologic and American Joint Committee on Cancer parameters between primary dermal melanoma and cutaneous metastatic melanoma cases Group Variable name
Sample size Regression present Angiotropism present Perineural invasion present Angioinvasion present SLNB Negative Positive Not done Associated nevus present Male Site Head and neck Trunk Extremities Acral Clark level (3 vs 4 vs 5) 3 4 5 Mean age, y (SE) Median age, y (range) Mean mitotic level (SE) Median mitotic level (range) Mean follow-up time, mo (SE) Median follow-up time, mo (range) GEP class I GEP class II
PDM
CMM
49 0 (0%) 5 (10%) 3 (6%)
15 0 (0%) 2 (13%) 2 (13%)
2 (4%)
3 (20%)
27 8 14 13
(55%) (16%) (29%) (27%)
0 0 0 0
(0%) (0%) (0%) (0%)
25 (51%)
9 (60%)
16 19 11 3
(33%) (39%) (22%) (6%)
6 2 7 0
2 43 4 51.2 55.0 4.2 2.0
(4%) (88%) (8%) (18.7) (8-83) (6.2) (1-32)
3 5 7 70.2 72.0 7.9 5.0
P value
* .66 .58 .08 *
.03 .57 .13
(40%) (13%) (47%) (0%) .0001 (20%) (33%) (47%) (11.1) \.0001 (51-64) .0004 (7.7) .06 (0-29) .02
26.3 (25.5) 21.0 (2-108) 11 2
6 9
.023
CMM, Cutaneous metastatic melanoma; GEP, gene expression profiling; PDM, primary dermal melanoma; SLNB, sentinel lymph node biopsy. *No statistical test because both percentages are 0%.
PDM and a melanocytic neoplasm may not even be suspected. Previous studies have reported 5-year survival of 80% to 100%.1-3,5-7 In all these studies, almost all patients had surgical resection alone as treatment. In our study only a single patient died of disease thus far and only 3 patients have developed visceral metastases at a mean follow-up of 20 months (range 2-74 months). This was despite a mean Breslow of depth of 3.01 mm and mean mitotic rate of 4.3/mm2. Interestingly, none of the parameters included in Table I or II had a statistically significant relationship
with recurrence. This is in stark contrast to conventional melanomas in which Breslow depth, mitotic rate, and ulceration have repeatedly been shown to have a statistically significant correlation with outcome. It is not surprising, however, that Breslow depth is less correlated to outcome because, as opposed to conventional melanomas that are thought to start at the dermoepidermal junction, PDM may originate at any level of the dermis and hence even early lesions may have an elevated Breslow depth. Another challenge in predicting behavior of PDM is that this group is very heterogeneous. As we learn more about intermediate-grade melanocytic tumors it may become apparent that some cases classified as PDM are in fact intermediate-grade melanocytic tumors with low risk for aggressive behavior. Hence, moving forward it may be of value to classify PDMs into distinct phenotypic or molecular categories. We noted 4 distinct morphologic patterns: Spitzoid type, blue nevuselike, PDM with an associated IDN, and NOS. An associated benign IDN was identified in 14 cases (14/49; 29%). We did not identify a statistically significant relationship between any of the specific morphologic patterns and recurrence of melanoma, which is likely in part related to the relatively small number of recurrences. However, being familiar with the fact that PDM may occur with these various patterns may help its clinician recognition when seeing one of these distinct patterns. The fact that an associated IDN component was found in 29% of the cases is an important finding because it strongly supports the diagnosis of a primary melanoma over a metastasis. In comparing the clinical and pathological features of PDM versus CMM, an associated nevus component, lack of subcutaneous involvement, and younger age all had a statistically significant relationship with PDM (P = .03, P = .0001, and P # .0001, respectively). In addition, angioinvasion was more characteristic of CMM and infrequently identified in PDM, a finding that approached statistical significance. There was a tendency for specific chromosomal copy number aberrations among the specific morphologic subtypes of PDM. Mainly homozygous deletion of 9p21 was particularly common in Spitzoid PDM, whereas in the PDMs associated with an IDN, 7 of 12 cases had copy number gains at 6p25 and within the NOS group, 5 of 8 had copy number gains in 6p25 and 5 of 5 had copy number gains in 8q24. The 5 of 5 positive cases with 8q24 is related to the fact that this subtype of PDM can be morphologically recognized as a pattern of dermal melanoma that is amelanotic, not associated with a nevus, and typically occurs in areas of
J AM ACAD DERMATOL
10 Sidiropoulos et al
nonchronically sun-damaged skin.16 Hence, based on the morphologic findings, this should help guide molecular testing when PDM is suspected. An important finding from this study are the results from the DecisionDx-Melanoma assay (Castle Biosciences Inc). This test classifies melanomas using messenger RNA expression profiling into a class I or class II type. The development of this assay was in uveal melanoma, and a modification of this assay has since been developed for cutaneous melanoma and is able to distinguish between high- and low-risk cutaneous melanomas.12,13,15,1719 These studies showed 6-year metastasis-free survival of 97% for class I and 19% for predicted class II of metastasis (P\.0001). As we hypothesized, a class I result is significantly more frequent among PDM compared with CMMs. This test may be an additional aid in those cases that cannot be solved by the histologic and clinical findings. We believe this study should help allow for better recognition of this subtype of melanoma and provide guidance for ancillary molecular testing both for establishing a diagnosis of melanoma by FISH or comparative genomic hybridization (CGH) testing and differentiating primary versus metastatic melanoma using the DecisionDx-Melanoma assay (Castle Biosciences Inc). REFERENCES 1. Bowen GM, Chang AE, Lowe L, Hamilton T, Patel R, Johnson TM, et al. Solitary melanoma confined to the dermal and/or subcutaneous tissue: evidence for revisiting the staging classification. Arch Dermatol 2000;136:1397-9. 2. Swetter SM, Ecker PM, Johnson DL, Harvell JD. Primary dermal melanoma: a distinct subtype of melanoma. Arch Dermatol 2004;140:99-103. 3. Cassarino DS, Cabral ES, Kartha RV, Swetter SM. Primary dermal melanoma: distinct immunohistochemical findings and clinical outcome compared with nodular and metastatic melanoma. Arch Dermatol 2008;144:49-56. 4. Lee CC, Faries MB, Ye X, Morton DL. Solitary dermal melanoma: beginning or end of the metastatic process? Ann Surg Oncol 2009;16:578-84. 5. Giuliano AE, Moseley HS, Morton DL. Clinical aspects of unknown primary melanoma. Ann Surg 1980;191:98-104. 6. Schlagenhauff B, Stroebel W, Ellwanger U, Meier F, Zimmermann C, Breuninger H, et al. Metastatic melanoma of unknown primary
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
origin shows prognostic similarities to regional metastatic melanoma: recommendations for initial staging examinations. Cancer 1997;80:60-5. Anbari KK, Schuchter LM, Bucky LP, Mick R, Synnestvedt M, Guerry D 4th, et al. Melanoma of unknown primary site: presentation, treatment, and prognosis; a single institution study; University of Pennsylvania pigmented lesion study group. Cancer 1997;79:1816-21. Katz KA, Jonasch E, Hodi FS, Soiffer R, Kwitkiwski K, Sober AJ, et al. Melanoma of unknown primary: experience at Massachusetts General Hospital and Dana-Farber Cancer Institute. Melanoma Res 2005;15:77-82. Reintgen DS, McCarty KS, Woodard B, Cox E, Seigler HF. Metastatic malignant melanoma with an unknown primary. Surg Gynecol Obstet 1983;156:335-40. American Joint Committee on Cancer. Melanoma of the skin. In: AJCC cancer staging manual. 7th ed. New York: Springer; 2010. pp. 325-44. Balch CM, Gershenwald JE, Soon SJ, Thompson JF, Atkins MB, Byrd DR, et al. Final version of 2009 AJCC melanoma staging and classification. J Clin Oncol 2009;27:6199-206. Dhillon N, Roger AR, Delman KA, Maetzold D, Oelschlager KM, Lyle S, et al. Gene expression profile signature (DecisionDx-Melanoma) to predict visceral metastatic risk in patients with stage I and stage II cutaneous melanoma. J Clin Oncol 2012; 30(suppl; abstr 8543). Lawson DH, Russell M, Wilkinson J, Amaria RN, Gonzalez R, Gerami P, et al. Gene expression profile of primary cutaneous melanomas to distinguish between low and high risk metastasis. J Clin Oncol 2013;31(suppl; abstr 9022). 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. Am J Surg Pathol 2009;33:1146-56. Onken MD, Worley LA, Tuscan MD, Harbour JW. An accurate, clinically feasible multi-gene expression assay for predicting metastasis in uveal melanoma. J Mol Diagn 2010;12:461-8. Pouryazdanparast P, Brenner A, Haghighat Z, Guitart J, Rademaker A, Gerami P, et al. The role of 8q24 copy number gains and c-MYC expression in amelanotic cutaneous melanoma. Mod Pathol 2012;25:1221-6. Onken MD, Worley LA, Char DH, Augsburger JJ, Correa ZM, Nudleman E, et al. Collaborative ocular oncology group report no. 1: prospective validation of a multi-gene prognostic assay in uveal melanoma. Ophthalmology 2012;119: 1596-603. Harbour JW, Chen R. The DecisionDx-UM gene expression profile test provides risk stratification and individualized patient care in uveal melanoma. PLoS Curr 2013;9:5. Onken MD, Worley LA, Ehlers JP, Harbour JW. Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death. Cancer Res 2004;64:7205-9.
ARTICLE
Primary dermal melanoma: A unique subtype of melanoma to be distinguished from cutaneous metastatic melanoma A clinical, histologic, and gene expressioneprofiling study Michael Sidiropoulos, MD, MS,a Roxana Obregon, BA,a Chelsea Cooper, BA,a Lauren Meldi Sholl, MS,a Joan Guitart, MD,a,b and Pedram Gerami, MDa,b Chicago, Illinois Background: Primary dermal melanoma (PDM) is a subtype of melanoma confined to the dermis that may be morphologically impossible to distinguish from cutaneous metastatic melanoma (CMM). Objective: We sought to better characterize PDM by describing the clinical, histologic, and molecular features of 49 cases of PDM and determine whether a gene expressioneprofiling test could help distinguish PDM from CMM. Methods: We describe 49 cases of PDM and determined whether any clinical or histopathologic features had a statistically significant relationship with outcome. Secondly, we performed a melanoma gene expressioneprofiling test on a subset of the PDM and CMM cases. Results: Overall recurrence was infrequent and seen in 9 of 49 cases. Six patients had locoregional recurrences and 3 patients had distant metastasis. None of the clinical or histologic parameters showed a statistically significant relationship with recurrence. There was a statistically significant association of a class I signature by DecisionDx-Melanoma assay (Castle Biosciences Inc, Friendswood, TX) for PDM whereas CMM were more frequently class II (P value = .023). Limitations: The mean follow-up time was 26 months. Conclusions: Most conventional staging parameters used for prognosis in cutaneous melanoma have limited applicability to PDM. The melanoma prognostic assay may be a useful tool for distinguishing PDM from CMM. ( J Am Acad Dermatol http://dx.doi.org/10.1016/j.jaad.2014.07.051.) Key words: American Joint Committee on Cancer; cutaneous metastatic melanoma; fluorescence in situ hybridization; gene expression profile; melanoma; primary dermal melanoma.
rimary dermal melanoma (PDM) is a variant of melanoma confined to the dermis and/or subcutaneous tissue without an epidermal component.1-3 The main histologic differential diagnosis for PDM is cutaneous metastatic melanoma
P
(CMM). Diagnosis of PDM requires excluding the possibility of metastasis from a separate primary melanoma. There are few case series in the literature characterizing PDM, but the existing literature suggests a remarkably good, long-term survival
From the Department of Dermatologya and Robert H. Lurie Cancer Center,b Feinberg School of Medicine, Northwestern University. Supported by the Irene D. Pritzker Foundation and partially supported by Castle Biosciences Inc. Disclosure: Dr Gerami has served as a consultant to Castle Biosciences Inc, Myriad Genetics, and DermTech Inc, receiving honoraria. Dr Guitart has served as a consultant to Castle Biosciences Inc, receiving honoraria. The other authors declared no conflicts of interest. Dr Sidiropoulos and Ms Obregon contributed equally to this article.
Accepted for publication July 29, 2014. Reprint requests: Pedram Gerami, MD, Department of Dermatology, Feinberg School of Medicine, Northwestern University, 676 N Saint Clair St, Suite 1765, Chicago, IL 60611. E-mail: [email protected]. Published online September 23, 2014. 0190-9622/$36.00 Ó 2014 by the American Academy of Dermatology, Inc. http://dx.doi.org/10.1016/j.jaad.2014.07.051
1
J AM ACAD DERMATOL
2 Sidiropoulos et al
compared with similarly staged conventional and 15 cases of CMM were identified in our dermatocutaneous melanomas.1-9 pathology database at Northwestern University. Differentiating PDM from CMM has significant All lesions were evaluated by 3 dermatopathologists staging and therapeutic implications. Although (J. G., P. G., and M. S.). Clinical histories and demost patients with PDMs including those with scriptions of lesions were obtained from medical higher Breslow depths have excellent long-term surrecords. Fourteen cases were diagnosed within our vival (80%-100%), the presence of CMM institution and 35 cases were external consultation indicates stage IIIC disease cases. In addition to a histowith estimated 10-year surpathologic diagnosis consisvival of 24%.10,11 Patients CAPSULE SUMMARY tent with PDM, inclusion with stage IIIC disease may criteria required a thorough Primary dermal melanoma has a far be candidates for adjuvant physical examination better prognosis than cutaneous therapy or more aggressive excluding the possibility of a metastatic melanoma, although these imaging protocols. Alterseparate primary melanoma, entities may in some cases be natively for patients with ulceration, or regression. indistinguishable. PDM, excision or excision The following clinical with sentinel lymph node bifeatures were assessed and Most conventional prognostic markers opsy (SLNB) and clinical documented for each PDM used for cutaneous melanoma have follow-up may be adequate case: patient’s age, sex, site of limited applicability to primary dermal therapy. At the histologic the lesion, and clinical appearmelanoma. level, this distinction is at times ance as initially described by The melanoma prognostic assay may impossible. The vast majority the submitting dermatologist help distinguish primary dermal of cases can be confidently (Tables I and II). A number of melanoma from cutaneous metastatic diagnosed based on the clinhistopathologic parameters melanoma. ical history and examination noted in Tables I and II were findings; however, there will recorded for each case. We remain a proportion of cases also subtyped the cases into 1 that, despite adequate clinical information, cannot be of 4 morphologic categories: Spitzoid; blue nevuselike resolved. The goal of the current study is to describe in morphology; associated with an intradermal nevus depth the clinical, histologic, immunohistochemical (IDN); and none of the above (not otherwise specified (IHC), and molecular characteristics of a cohort of [NOS]). PDMs and determine whether a novel molecular In 33 of 49 PDM cases and in 25 of 49 PDM cases prognostic test for melanoma can assist in differentifluorescence in situ hybridization (FISH) and IHC ating this entity from CMM. studies were performed, respectively. The results The prognostic test for melanoma, also known of the FISH findings and the IHC stains are reported as DecisionDx-Melanoma (Castle Biosciences Inc, in Table II. FISH was performed using probes Friendswood, TX), is a gene expressioneprofiling targeting chromosome 6p25 (RREB1), chromosome (GEP) test that looks at the messenger RNA 6q23 (MYB), chromosome 11q13 (CCND1), the expression profile of 28 genes. The data are centromeric portion of chromosomes 6 (CEP6), reported as either a class I (low risk) or class II chromosome 9p21 (CDKN2A) and 8q24 (MYC), as (high risk) signature. The assay has been shown previously described.14 to be an independent prognostic marker for In 13 PDM cases there was sufficient tissue cutaneous melanoma in multivariate analysis.12,13 available for messenger RNA GEP testing. Fifteen Because PDM has a good prognosis, we exCMM cases were evaluated as control specimens. pected a considerable proportion of PDMs to Thirteen PDM and 15 CMM cases with formalin-fixed have a class I signature whereas most CMMs paraffin embedded melanoma samples were would be class II. Hence, we studied a subset of macrodissected and analyzed blindly by the the PDMs that had adequate tissue for the GEP DecisionDx-Melanoma GEP signature assay (Castle testing and a control group of CMMs to evaluate Biosciences Inc).12,13,15 Fisher exact test was used to this hypothesis. compare the frequency of class I and class II results in the PDM and CMM cases. Clinical follow-up was available for 48 of the 49 METHODS reported PDM cases. All patients had a physical After obtaining approval from the Northwestern examination excluding a separate primary University Cancer Center, Chicago, IL, and internal melanoma. In 22 of 49 patients, imaging studies institutional review board, 49 cases of PDM d
d
d
J AM ACAD DERMATOL
Abbreviations used: CGH: CMM: FISH: GEP: IDN: IHC: NOS: PDM: SLNB:
comparative genomic hybridization cutaneous metastatic melanoma fluorescence in situ hybridization gene expression profiling intradermal nevus immunohistochemical not otherwise specified primary dermal melanoma sentinel lymph node biopsy
were performed to exclude metastatic disease. Clinical data were retrieved from medical records and patient interviews (Table III). None of the patients had clinically palpable lymph node disease at the initial presentation. Frequency tables were generated to summarize the categorical variables between PDM cases with and without recurrence. The Fisher exact test was used for comparison. Two-sample t test was used to compare continuous variables between the 2 groups. Comparisons between the groups were considered statistically significant at a P value less than .05.
RESULTS The clinical findings for the 49 PDM cases are summarized in Table I. The database review of 6000 patients resulted in 49 patients (0.82%) with a diagnosis of PDM. The 49 PDM included 25 (51%) male and 24 (49%) female patients, ranging in age from 8 to 83 years (mean 51.2 years). The most common anatomic site of involvement was the trunk (19/49; 39%), followed by the head and neck (16/49; 33%), extremities (11/49; 22%) and acral area (3/49; 6%). The histopathologic and IHC findings are summarized in Tables II and III. Histopathologically, all 49 cases (100%) of PDM were characterized by a dermalbased melanocytic neoplasm (Fig 1). There was no evidence of an overlying in situ component or regression in any of the cases analyzed. The mean Breslow depth was at least 3.01 mm (range 0.5 to $ 10.05 mm). All cases demonstrated a Clark level equal to or greater than III: Clark level III (2/49; 4%), Clark level IV (34/49; 69%), and Clark level V (5/49; 11%). In 8 cases the base of the lesion was transected and a definitive assessment of the Clark level could not be made (8/49; 16%). The mean mitotic rate was 4.3 mitoses/mm2 (range 1-32/mm2). An associated benign IDN was identified in 14 cases (14/49; 29%) (Fig 2). Angiotropism (5/49; 10%), angioinvasion (2/49; 4%), and perineural invasion (3/49; 6%) were uncommonly identified. Morphologically, the cases could be roughly categorized into 4 patterns: Spitzoid (13/49; 26%), blue nevuselike (3/49; 6%), PDM in
Sidiropoulos et al 3
association with an IDN (14/49; 29%), and NOS (19/ 49; 39%). Among the cases with a clinical description, the most common differential diagnosis was melanocytic nevus or melanoma (16/28; 57%), nonmelanoma skin cancer (6/28; 22%), dermatofibroma (2/ 28; 7%), cyst (2/28; 7%), and scar (2/28; 7%). Of 49 cases, 33 were tested by FISH and criteria for FISH were met in 28 of 33 cases analyzed (85%) (Table II). The following copy number aberrations were noted: gains at 6p25 (19/33; 58%), gains at 11q13 (8/33; 24%), deletions of 6q23 (4/33; 12%), homozygous deletions of 9p21 (7/27; 26%), and gains at 8q24 (7/11; 64%). Among the Spitzoid PDM, the most common copy number aberration was a homozygous deletion of 9p21 seen in 6 of 10 cases (60%). Among the blue nevuselike PDM, 2 of 3 cases (67%) showed copy number gains at 6p25 and 2 of 3 cases (67%) had copy number gains at 11q13. Among the PDM associated with an IDN, the most common copy number aberration was gain of 6p25 seen in 7 of 12 cases (58%). Among the NOS cases, 6 of 8 (75%) showed copy number gains at 6p25 and of the 5 tested for 8q24 change, 5 (100%) showed copy number gains at 8q24. Comparison of morphologic and American Joint Committee on Cancer parameters between PDM and CMM cases (Table IV) demonstrated statistical significance with the presence of an associated nevus in PDM (P = .03), Clark level (P = .0001), and age (P \.0001). GEP using the DecisionDx-Melanoma assay (Castle Biosciences Inc) in 13 PDM cases demonstrates class I (low risk) expression in 11 cases and class II (high risk) expression in only 2 cases (Table IV). Among the 15 cases of CMM, the assay demonstrated class I expression in 6 cases and class II expression in 9 cases. The findings are summarized in Table IV. The Fisher exact test gave a P value of .023 showing a statistically significant association of class I signature with PDM. At least 2 months of clinical follow-up was available in 48 of 49 patients (Table III). The average follow-up time was 26 months and it ranged from 2 to 108 months. In 22 of 48 patients (46%), imaging studies were performed and 3 of 22 cases (14%) showed evidence of distant metastasis. Overall, 9 of 48 (19%) patients developed recurrence despite re-excision with clear margins from the primary melanoma. Three patients (3 of 48; 6%) developed distant metastatic disease, 2 to the lung and liver and 1 with isolated lung metastasis. Locoregional recurrence of melanoma was present in 6 patients (6 of 48; 13%), 5 of whom developed in transit metastasis and 1 with regional lymph node disease. SLNB was performed in 34 of 48 (71%) patients and
J AM ACAD DERMATOL
4 Sidiropoulos et al
Table I. American Joint Committee on Cancer characteristics of primary dermal melanoma cases Case
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
Age, y
Sex
Site
Site code
60 75 70 17 55 71 30 54 83 41 32 29 57 48 45 80 52 53 51 59 28 51 75 39 72 68 8 63 33 66 74 26 63 51 63 58 24 62 47 35 55 72 58 58 38 14 62 17 66
F M M F M F F M M F F M M M F M F M F M F F M F F F M M M M M F M M F F F F M F F F F F M M M M M
Eye, right Forehead Ear, left Arm, left Nose Scalp, right Shoulder, left Buttock, right Sideburn, right Thigh, left Nose, tip Neck, posterior Back, right Ear, left Ear, right Back, mid Scalp, mid Ear, left Thigh, left upper Thigh, left Scalp, crown Abdomen, mid Breast, right Ankle, left Cheek, left Forehead, medial Back Scapula, left Calf, left Arm, left Ear, right Thigh, left Back, upper Back, left upper Back, right Thigh, right Foot, left Back, left Umbilicus Arm, right Back, left Great toe, right Arm, right Arm, right Chest, mid Back, right upper Back, mid Back Chest wall, posterior
1 1 1 3 1 1 2 2 1 3 1 1 2 1 1 2 1 1 3 3 1 2 2 4 1 1 2 2 3 3 1 3 2 2 2 3 4 2 2 3 2 4 3 3 2 2 2 2 2
Breslow depth, mm
At
At At At
At At At At At
At
At
At
At
At At
3.3 1.05 4.8 least 7.5 0.5 6.5 least 0.9 least 5.0 least 0.9 4.5 0.85 least 1.0 least 3.2 least 0.9 least 1.16 1.2 least 2.4 1.1 1.5 2.7 3.5 0.65 6.7 least 1.1 1.33 1.1 least 2.8 7.5 0.6 1.7 least 2.5 7 7.6 10.05 0.53 least 5.6 1.9 1.5 9.3 least 1.45 least 1.7 3.8 1.2 1.4 3.3 3.9 0.9 3.5 2.2
Clark level
Ulceration
Mitoses, /mm2
AJCC T stage
4 4 4 5 4 5 4 4 least 4 4 least least 4 least 4 4 4 4 4 4 4 4 least 4 4 4 4 3 4 4 4 4 5 3 least 4 4 5 least least 4 4 4 4 4 4 4 4
No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No
1 2 6 4 2 3 4 30 3 3 2 2 8 1 1 1 3 1 5 4 3 2 32 2 4 3 6 8 3 5 1 1 16 10 1 2 1 1 2 1 1 1 2 3 2 1 2 2 2
T3a T1b T4a T4a T1b T4a T1b T4a T1b T4a T1b T1b T3a T1b T2a T1b T3a T2a T2a T3a T3a T1b T4a T2a T2a T2a T3a T4a T1b T2a T3a T4a T4a T4a T1b T4a T2a T2a T4a T2a T2a T3a T2a T2a T3a T2a T1b T3a T3a
At
At At At
At
At
At At
4
4 4 4
4
4
4 4
AJCC, American Joint Committee on Cancer; F, female; M, male.
was positive in 6 (6 of 34; 18%) patients. A positive SLNB specimen did have a trend toward adverse prognosis according to Cox proportional hazards with a P value of .07.
None of the parameters studied including age, sex, anatomic site, Breslow depth, Clark level, mitotic rate, ulceration, regression, angioinvasion, angiotropism, perineural invasion, SLNB, Spitzoid
J AM ACAD DERMATOL
Sidiropoulos et al 5
Table II. Histopathologic characteristics of primary dermal melanoma cases Case
Clinical presentation
1
Histologic subtype
NOS
2
BCC vs SCC
NOS
3
8-mm Smooth, pink, pedunculated papule with crusted surface; fibrolipoma vs BCC vs SCC
IDN
IHC results
FISH results
S-1001, Mart-1 , Melan-A , MITF , HMB45 , p16 , Ki67 \5% S-100 , Mart-1 , Melan-A , MITF , HMB451, p161, Ki67 10% S-1001, Mart-11, HMB45 , Ki67 10%
N/A
-
-
-
6p25 Gain
-
-
-
8q24 Gain
-
-
-
6p25 Gain, 8q24 gain
-
-
-
N/A
-
-
-
11q13 Gain
-
-
-
9p21 Deletion 6q23 Loss N/A 6p25 Gain, 11q13 gain N/A N/A 6p25 Gain, 11q13 gain 6p25 Gain 6p25 Gain, 6q23 loss 9p21 Deletion N/A
-
-
1
N/A
-
-
-
6p25 Gain
-
-
-
N/A
-
-
-
N/A 11q13 Gain, 9p21 deletion N/A 9p21 Deletion
-
-
1 -
8q24 Gain
-
-
-
6p25 Gain, 6q23 loss, 11q13 gain 8q24 Gain N/A 6p25 Gain, 11q13 gain
-
1
-
1
-
-
28
Blue nevus S-1001, Melan-A1, HMB451, Ki67 10% Dark macule, rule out lentigo NOS Melan-A1, HMB451, Ki67 \10% Blue pigmented lesion; rule Blue nevus out blue nevus Spitzoid IDN BCC vs SCC IDN Ki67 10% Spitzoid IDN BCC Spitzoid S-1001, Melan-A1, HMB45e Spitzoid Melan-A1, Ki67 5% Nevus IDN Melan-A1, HMB451, p16 Spitzoid HMB451, p161, Ki67 \5% IDN Mart-11, Ki67 5% Two-toned nodule IDN 3mm; BCC vs SCC Spitzoid Dome-shaped firm NOS hemorrhagic papule; sclerosing hemangioma DF Changing nodule X years; IDN pigmented BCC vs malignant melanoma 5-mm Pink/white firm papule; IDN irritated nevus vs DN vs recurrent nevus vs scar 5 mm; Atypical nevus vs NOS malignant melanoma Changing pigmented lesion NOS S-1001, Mart-11 Spitzoid IDN IDN P16 2- to 3-wk History of 4 dark blue-black papules grouped on medial forehead; traumatic vs melanocytic Lesion present for 9 mo, NOS S-1001, Melan-A1, Ki67 10% enlarging; rule out NF vs DF vs other Melanocytic lesion Blue nevus Ki67 10%
29 30 31
NOS NOS IDN
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
20
21
22 23 24 25 26
27
Atypical nevus
S-1001, Melan-A1
AT PNI AI
Continued
J AM ACAD DERMATOL
6 Sidiropoulos et al
Table II. Cont’d Case
32 33 34 35 36 37 38 39
Clinical presentation
Recurrent inflamed nodule; inflamed cyst 1.5-cm Cystic nodule; rule out cyst 4-mm Pink papule; rule out NMSC Solitary 1.2-cm red violaceous plaque History of melanoma; rule out melanoma 5-mm Red papule; DN
Histologic subtype
IDN NOS NOS NOS Spitzoid
Spitzoid NOS
Rule out DN Small nodular lesion
NOS NOS
42
Patient with a solitary mass; DDx includes a benign nevus with congenital features/scar, and an atypical Spitzoid tumor/ melanoma Single pigmented lesion
IDN
46 47 48 49
FISH results
S-1001, Mart-11, MITF1
6p25 Gain, 6q23 loss N/A
1
1
-
N/A
-
-
-
N/A
-
-
-
6p25 Gain, 9p21 deletion
1
-
-
6p25 Gain, 9p21 deletion
1
-
-
N/A
-
-
-
6p25 Gain N/A
-
-
-
6p25 Gain
-
1
-
6p25 Gain, 8q24 gain 6p25 Gain, 8q24 gain 1 6p25 Gain, 11q13 gain, 9p21 deletion 6p25 Gain, 8q24 gain 6p25 Gain, 11q13 gain -
-
-
-
-
S-1001, Mart-11, MITF1, HMB45 S-1001, Melan-A1, HMB451, p16 , Ki67 5% Ki67 5%
Spitzoid
40 41
43 44 45
IHC results
P16 S-1001, Melan-A1, MITF1, HMB451 P161 S-1001, Melan-A1, MITF1, Sox101
NOS NOS Spitzoid Spitzoid NOS NOS Spitzoid
S-1001, Melan-A1
AT PNI AI
AI, Angioinvasion; AT, angiotropism; BCC, basal cell carcinoma; DDx, differential diagnosis; DF, dermatofibroma; DN, dysplastic nevus; FISH, fluorescence in situ hybridization; IDN, intradermal nevus; IHC, immunohistochemistry; N/A, not available; NF, neurofibroma; NMSC, nonmelanoma skin cancer; NOS, not otherwise specified; PNI, perineural invasion; SCC, squamous cell carcinoma.
morphology, blue nevuselike morphology, an associated IDN, and NOS morphology had a statically significant relationship with recurrence.
DISCUSSION Numerous previous studies, mostly reported as single-center series and case reports, have identified patients with a solitary focus of melanoma confined to the dermis with no known separate primary melanoma and negative metastatic findings on workup, who largely remain disease free with a favorable prognosis.1-9 As the prognosis and treatment for a PDM differs greatly from that of a stage IIIC or stage IV M1a metastatic melanoma, the distinction is critical. At the histologic level this distinction is often impossible. Alternatively, at the clinical level the diagnosis may be obvious in many cases because of a known nearby or widely
metastatic separate primary melanoma. Hence, good clinical correlation may solve many of the cases. However in a considerable proportion of cases, some diagnostic uncertainty may remain even after thorough clinical and histologic evaluation. Importantly, patients with CMM are candidates for adjuvant therapy and regular imaging. Typically patients with PDM would not receive adjuvant therapy or require regular imaging. In comparing the clinical features of PDM to typical primary cutaneous melanomas there are some interesting distinctions. In 12 of 28 cases where a clinical differential diagnosis was available, a melanocytic neoplasm was not among the differential diagnoses mentioned on the requisition. In these 12 cases the primary differential diagnosis was nonmelanoma skin cancer, dermatofibroma, cyst, or scar. Hence the ABCD rule will often not apply to
J AM ACAD DERMATOL
Sidiropoulos et al 7
Table III. Treatment, metastatic staging workup, and outcome of primary dermal melanoma cases Case
WLE
SLNB results
Imaging
1 2
1 1
Negative (2007) ND
CTXR/CT1
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Negative (2008) Negative (2009) Negative (2010) Positive (2010) Negative (2011) Negative (2010) ND Negative (2011) ND ND Negative (2011) ND Negative (2011) Negative (2011) Negative (2011) Negative (2011) Negative (2011) Negative (2011) Negative (2011) Positive (2010) Negative (2011) Negative (2012) Negative (2012) ND ND Negative (2013) ND Negative (2012) ND Positive (2013) Negative (2012)
CTXRND N/A PETN/A N/A CT/PET/MRIN/A N/A ND N/A N/A N/A ND XRXRCT1 CT/PETCT/PET1 CT/PET/MRI1 XRXRN/A N/A CTCT/PETXRND N/A CT/PET-
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Negative (2012) Negative (2012) ND Positive (2013) Positive (2013) Positive (2013) ND Negative (2013) Negative (2013) ND ND ND ND Negative (2010) ND Negative (2005)
CTXRN/A CTN/A CT/PETND N/A N/A ND ND N/A N/A ND N/A N/A
Follow-up, mo
Region of concern
Outcome
Lungs
NED DM (lungs)
74 23
NED NED NED NED NED NED NED NED N/A NED NED LR (regional) NED NED NED NED NED NED NED DM (liver/lungs) DM (liver/lungs) NED NED LR (regional) NED NED LR (lymph node) NED NED NED LR (regional)
60 50 37 36 36 29 28 24 N/A 23 7 22 21 21 29 26 23 28 22 24 17 2 12 18 18 18 13 8 11 10 14
Alive DOD December 2009 Alive Alive Alive Alive Alive Alive Alive Alive N/A Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive
LR (regional) NED NED NED LR (regional) NED NED NED NED NED NED NED NED NED NED NED
10 7 6 5 5 5 6 5 5 3 3 99 86 43 108 81
Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive Alive
Liver/lungs Liver/lungs
Enhancing soft-tissue mass suspicious for local regional metastasis, no visceral involvement
Current status
CT, Computerized tomography; DM, distant metastasis; MRI, magnetic resonance imaging; LR, local recurrence; N/A, not available; ND, not done; NED, no evidence of disease; PET, positron emission tomography; SLN, sentinel lymph node; SLNB, sentinel lymph node biopsy; WLE, wide local excision; XR, x-ray.
8 Sidiropoulos et al
J AM ACAD DERMATOL
Fig 1. A, Low-power magnification showing sheetlike proliferation of melanoma cells filling the dermis with no involvement of the epidermis. B, Low-power magnification revealing bulbous extensions of the melanoma extending into the subcutaneous tissue. C, Intermediatepower magnification showing sheetlike proliferation of atypical and pleomorphic melanocytes. D, High-power showing high-grade nuclear atypia, pleomorphism, and mitotic activity.
Fig 2. A, Low-power image of a large primary dermal melanoma, Spitzoid type. B, High-power view demonstrating high-grade nuclear atypia with multiple mitotic figures, including an atypical form. C, Fluorescence in situ hybridization (FISH) image showing homozygous 9p21 deletion within the tumor cells. D, FISH image showing intact centromeric component of chromosome 9 as control.
J AM ACAD DERMATOL
Sidiropoulos et al 9
Table IV. Comparison of morphologic and American Joint Committee on Cancer parameters between primary dermal melanoma and cutaneous metastatic melanoma cases Group Variable name
Sample size Regression present Angiotropism present Perineural invasion present Angioinvasion present SLNB Negative Positive Not done Associated nevus present Male Site Head and neck Trunk Extremities Acral Clark level (3 vs 4 vs 5) 3 4 5 Mean age, y (SE) Median age, y (range) Mean mitotic level (SE) Median mitotic level (range) Mean follow-up time, mo (SE) Median follow-up time, mo (range) GEP class I GEP class II
PDM
CMM
49 0 (0%) 5 (10%) 3 (6%)
15 0 (0%) 2 (13%) 2 (13%)
2 (4%)
3 (20%)
27 8 14 13
(55%) (16%) (29%) (27%)
0 0 0 0
(0%) (0%) (0%) (0%)
25 (51%)
9 (60%)
16 19 11 3
(33%) (39%) (22%) (6%)
6 2 7 0
2 43 4 51.2 55.0 4.2 2.0
(4%) (88%) (8%) (18.7) (8-83) (6.2) (1-32)
3 5 7 70.2 72.0 7.9 5.0
P value
* .66 .58 .08 *
.03 .57 .13
(40%) (13%) (47%) (0%) .0001 (20%) (33%) (47%) (11.1) \.0001 (51-64) .0004 (7.7) .06 (0-29) .02
26.3 (25.5) 21.0 (2-108) 11 2
6 9
.023
CMM, Cutaneous metastatic melanoma; GEP, gene expression profiling; PDM, primary dermal melanoma; SLNB, sentinel lymph node biopsy. *No statistical test because both percentages are 0%.
PDM and a melanocytic neoplasm may not even be suspected. Previous studies have reported 5-year survival of 80% to 100%.1-3,5-7 In all these studies, almost all patients had surgical resection alone as treatment. In our study only a single patient died of disease thus far and only 3 patients have developed visceral metastases at a mean follow-up of 20 months (range 2-74 months). This was despite a mean Breslow of depth of 3.01 mm and mean mitotic rate of 4.3/mm2. Interestingly, none of the parameters included in Table I or II had a statistically significant relationship
with recurrence. This is in stark contrast to conventional melanomas in which Breslow depth, mitotic rate, and ulceration have repeatedly been shown to have a statistically significant correlation with outcome. It is not surprising, however, that Breslow depth is less correlated to outcome because, as opposed to conventional melanomas that are thought to start at the dermoepidermal junction, PDM may originate at any level of the dermis and hence even early lesions may have an elevated Breslow depth. Another challenge in predicting behavior of PDM is that this group is very heterogeneous. As we learn more about intermediate-grade melanocytic tumors it may become apparent that some cases classified as PDM are in fact intermediate-grade melanocytic tumors with low risk for aggressive behavior. Hence, moving forward it may be of value to classify PDMs into distinct phenotypic or molecular categories. We noted 4 distinct morphologic patterns: Spitzoid type, blue nevuselike, PDM with an associated IDN, and NOS. An associated benign IDN was identified in 14 cases (14/49; 29%). We did not identify a statistically significant relationship between any of the specific morphologic patterns and recurrence of melanoma, which is likely in part related to the relatively small number of recurrences. However, being familiar with the fact that PDM may occur with these various patterns may help its clinician recognition when seeing one of these distinct patterns. The fact that an associated IDN component was found in 29% of the cases is an important finding because it strongly supports the diagnosis of a primary melanoma over a metastasis. In comparing the clinical and pathological features of PDM versus CMM, an associated nevus component, lack of subcutaneous involvement, and younger age all had a statistically significant relationship with PDM (P = .03, P = .0001, and P # .0001, respectively). In addition, angioinvasion was more characteristic of CMM and infrequently identified in PDM, a finding that approached statistical significance. There was a tendency for specific chromosomal copy number aberrations among the specific morphologic subtypes of PDM. Mainly homozygous deletion of 9p21 was particularly common in Spitzoid PDM, whereas in the PDMs associated with an IDN, 7 of 12 cases had copy number gains at 6p25 and within the NOS group, 5 of 8 had copy number gains in 6p25 and 5 of 5 had copy number gains in 8q24. The 5 of 5 positive cases with 8q24 is related to the fact that this subtype of PDM can be morphologically recognized as a pattern of dermal melanoma that is amelanotic, not associated with a nevus, and typically occurs in areas of
J AM ACAD DERMATOL
10 Sidiropoulos et al
nonchronically sun-damaged skin.16 Hence, based on the morphologic findings, this should help guide molecular testing when PDM is suspected. An important finding from this study are the results from the DecisionDx-Melanoma assay (Castle Biosciences Inc). This test classifies melanomas using messenger RNA expression profiling into a class I or class II type. The development of this assay was in uveal melanoma, and a modification of this assay has since been developed for cutaneous melanoma and is able to distinguish between high- and low-risk cutaneous melanomas.12,13,15,1719 These studies showed 6-year metastasis-free survival of 97% for class I and 19% for predicted class II of metastasis (P\.0001). As we hypothesized, a class I result is significantly more frequent among PDM compared with CMMs. This test may be an additional aid in those cases that cannot be solved by the histologic and clinical findings. We believe this study should help allow for better recognition of this subtype of melanoma and provide guidance for ancillary molecular testing both for establishing a diagnosis of melanoma by FISH or comparative genomic hybridization (CGH) testing and differentiating primary versus metastatic melanoma using the DecisionDx-Melanoma assay (Castle Biosciences Inc). REFERENCES 1. Bowen GM, Chang AE, Lowe L, Hamilton T, Patel R, Johnson TM, et al. Solitary melanoma confined to the dermal and/or subcutaneous tissue: evidence for revisiting the staging classification. Arch Dermatol 2000;136:1397-9. 2. Swetter SM, Ecker PM, Johnson DL, Harvell JD. Primary dermal melanoma: a distinct subtype of melanoma. Arch Dermatol 2004;140:99-103. 3. Cassarino DS, Cabral ES, Kartha RV, Swetter SM. Primary dermal melanoma: distinct immunohistochemical findings and clinical outcome compared with nodular and metastatic melanoma. Arch Dermatol 2008;144:49-56. 4. Lee CC, Faries MB, Ye X, Morton DL. Solitary dermal melanoma: beginning or end of the metastatic process? Ann Surg Oncol 2009;16:578-84. 5. Giuliano AE, Moseley HS, Morton DL. Clinical aspects of unknown primary melanoma. Ann Surg 1980;191:98-104. 6. Schlagenhauff B, Stroebel W, Ellwanger U, Meier F, Zimmermann C, Breuninger H, et al. Metastatic melanoma of unknown primary
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
origin shows prognostic similarities to regional metastatic melanoma: recommendations for initial staging examinations. Cancer 1997;80:60-5. Anbari KK, Schuchter LM, Bucky LP, Mick R, Synnestvedt M, Guerry D 4th, et al. Melanoma of unknown primary site: presentation, treatment, and prognosis; a single institution study; University of Pennsylvania pigmented lesion study group. Cancer 1997;79:1816-21. Katz KA, Jonasch E, Hodi FS, Soiffer R, Kwitkiwski K, Sober AJ, et al. Melanoma of unknown primary: experience at Massachusetts General Hospital and Dana-Farber Cancer Institute. Melanoma Res 2005;15:77-82. Reintgen DS, McCarty KS, Woodard B, Cox E, Seigler HF. Metastatic malignant melanoma with an unknown primary. Surg Gynecol Obstet 1983;156:335-40. American Joint Committee on Cancer. Melanoma of the skin. In: AJCC cancer staging manual. 7th ed. New York: Springer; 2010. pp. 325-44. Balch CM, Gershenwald JE, Soon SJ, Thompson JF, Atkins MB, Byrd DR, et al. Final version of 2009 AJCC melanoma staging and classification. J Clin Oncol 2009;27:6199-206. Dhillon N, Roger AR, Delman KA, Maetzold D, Oelschlager KM, Lyle S, et al. Gene expression profile signature (DecisionDx-Melanoma) to predict visceral metastatic risk in patients with stage I and stage II cutaneous melanoma. J Clin Oncol 2012; 30(suppl; abstr 8543). Lawson DH, Russell M, Wilkinson J, Amaria RN, Gonzalez R, Gerami P, et al. Gene expression profile of primary cutaneous melanomas to distinguish between low and high risk metastasis. J Clin Oncol 2013;31(suppl; abstr 9022). 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. Am J Surg Pathol 2009;33:1146-56. Onken MD, Worley LA, Tuscan MD, Harbour JW. An accurate, clinically feasible multi-gene expression assay for predicting metastasis in uveal melanoma. J Mol Diagn 2010;12:461-8. Pouryazdanparast P, Brenner A, Haghighat Z, Guitart J, Rademaker A, Gerami P, et al. The role of 8q24 copy number gains and c-MYC expression in amelanotic cutaneous melanoma. Mod Pathol 2012;25:1221-6. Onken MD, Worley LA, Char DH, Augsburger JJ, Correa ZM, Nudleman E, et al. Collaborative ocular oncology group report no. 1: prospective validation of a multi-gene prognostic assay in uveal melanoma. Ophthalmology 2012;119: 1596-603. Harbour JW, Chen R. The DecisionDx-UM gene expression profile test provides risk stratification and individualized patient care in uveal melanoma. PLoS Curr 2013;9:5. Onken MD, Worley LA, Ehlers JP, Harbour JW. Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death. Cancer Res 2004;64:7205-9.
