Desmoplastic Malignant Melanoma: A Study of Ten Cases and Status of BRAF MutationCoupelon S.a · Franck F.b · Jarrousse A.S.b · Déchelotte P.b · Souteyrand P.a · D’Incan M.a
Departments of aDermatology and bPathology, Université d’Auvergne, CHU, Clermont-Ferrand, France Corresponding Author
Background: Desmoplastic malignant melanoma (DM) is a rare variant of melanoma. BRAF gene mutations have been poorly explored in this entity. Objective: To detect BRAF gene mutation in a series of DM. Methods: This is a single-center retrospective study of ten patients with DM, with a biomolecular analysis of BRAF mutation. Results: The male:female ratio was 2.3:1, with a mean patient age of 66.5 years. Melanoma arose in the head and neck region in 3 cases. The mean tumor thickness was 7.97 mm, Clark level was IV or V in all cases. Six melanomas were of the pure DM variant. Three patients had at least one local recurrence, two had regional node metastases, and two experienced systemic metastases which they died of (average follow-up 34.1 months). A V600E BRAF mutation was detected in only one patient. Conclusion: BRAF mutation seems to be a rare event in DM contrary to other melanoma variants.
Copyright © 2012 S. Karger AG, Basel
Desmoplastic malignant melanoma (DM) is a rare histologic variant of spindle cell melanoma, characterized by an invasive proliferation of fusiform melanocytes in the deep dermis, separated by a dense fibrocollagenous stroma . It accounts for <4% of all cutaneous melanomas . The concept was later expanded with the notion of neurotropism, the so-called desmoplastic neurotropic melanoma, which indicates the neural invasion and/or neural differentiation of the tumor .
The degree of desmoplasia within the tumor allows to distinguish two subtypes of DM: ‘pure’ DM (pDM) with prominent desmoplasia throughout the entire tumor, and ‘composed’ – or mixed – DM (cDM) in which desmoplasia is combined with >10% of non-desmoplastic melanoma [4,5,6].
DM clinically and pathologically differs from conventional melanoma, and this constitutes a real diagnostic pitfall for the unwary , leading to delayed diagnosis and an advanced tumoral stage at the time of diagnosis . Furthermore, with a higher recurrence rate but a lower incidence of lymph node metastases, and a propensity to develop lung metastases, DM’s behavior seems to be closer to soft tissue sarcoma than to conventional melanoma [8,9].
Gene expression profiling of DM seems to differ from conventional melanoma . Recently, BRAF activating mutations became of importance for therapeutic choice in melanoma practice. Although occurring in about half of conventional melanomas, the only published study exploring BRAF mutation in DM failed to detect it . We reported a single-center retrospective study of ten cases of DM, designed to evaluate BRAF mutation status.
Materials and Methods
From the database of our institutional Department of Pathology, we identified tumors diagnosed as DM between 2004 and 2011. Mucosal melanoma and desmoplastic recurrence of conventional melanoma were excluded. After obtaining their acceptance, all still alive patients were clinically evaluated in the first 3 months of 2011. Data concerning dead patients were picked up from clinical charts. For each corresponding patient we collected data concerning age, sex, family history of melanoma, personal antecedents of melanoma or cutaneous carcinoma, disease duration before diagnosis, tumor location, clinical presentation, staging at diagnosis, outcome and time of follow-up. All tumor samples were examined by two pathologists (S.C., F.F.). Diagnosis of DM was confirmed on the basis of an invasive proliferation of fusiform melanocytes in the dermis associated with fascicles or bundles of collagen ensheathing tumoral cells, and tumors were classified as either pDM or cDM. The following additional criteria were collected: tumor depth (Breslow thickness and Clark level), mitotic rate, ulceration, regression, pigmentation, atypical intraepidermal melanocytic component, neurotropism, lymphoid aggregates, stromal myxoid change, vascular involvement and solar elastosis. Immunohistochemical staining with S-100 protein and gp 100 antigen (HMB-45 antibody) were recorded for all pathology slides.
All primitive tumors and, in some cases, metastases were analyzed for the presence of a thymine → adenine (T → A) missense mutation at nucleotide 1799 in the BRAF gene. In order to obtain a sufficient number of tumor cells and then get round the tumor heterogeneity, the maximum tumor-rich area was selected for each FFPE block after reviewing the corresponding H&E-stained slides. DNA was extracted using the QIAamp DNA FFPE Tissue kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions, after microdissection of a 3–5 mm3 sample from the previously selected tumor-rich area, and then diluted to 100 ng/µl before PCR amplification.
In some patients, BRAF mutation detection was achieved by qRT-PCR, using Taqman 7900 technology. In most patients, the BRAF exon 15 was amplified from genomic DNA by PCR and analyzed by Sanger bi-directional sequencing , using the Big Dye Terminator V1.1 Sequencing Cycle kit (Applied Biosystems, Carlsbad, Calif., USA) in combination with a 3130XL Genetic Analyzer (Applied Biosystems). Final analysis was performed with the Sequence Scanner V1.0 software (Applied Biosystems).
The PCR was performed in a total volume of 25 µl with a typical reaction mixture containing 100 ng of DNA matrix, 0.5 µmol/l of primers and the 1× KAPA2G Robust HotStart ReadyMix (KAPA Biosystems, Boston, Mass., USA) which contains KAPA2G Robust HotStart DNA polymerase in a proprietary reaction buffer, 0.2 mm of each dNTP and 2 mm of MgCl2. Amplification was performed on a Veriti Thermal Cycler (Applied Biosystems) with an initial denaturation at 95°C for 3 min, followed by 40 cycles of 95°C denaturation for 15 s, annealing at 55 or 52°C for amplification of the 223 and 184 bp products respectively for 15 s, and extension at 72°C for 15 s.
The primer sets used for PCR amplification were BRAF 223F 5′-ATG CTT GCT CTG ATA GGA AAA TG-3′ and BRAF 223R 5′-TCA GGG CCA AAA ATT TAA TC-3′ for patients 1, 5, 6 and 10 and yielded an amplification product of 223 bp. The primer sequences BRAF 184F 5′-GAT CTA CTG TTT TCC TTT AC-3′ and BRAF 184R 5′-TAA TCA GTG GAA AAA TAG CC-3′ were used for patients 3, 8 and 9 and yielded an amplification product of 184 bp. Primers were designed with the use of Primer 3 software .
PCR products were purified using the Illustra Microspin 5300HR columns (GE Healthcare), followed by Sanger bi-directional sequencing  using the Big Dye Terminator V1.1 Sequencing Cycle kit (Applied Biosystems) according to manufacturer’s instructions in combination with a 3130XL Genetic Analyzer (Applied Biosystems). Final analysis was performed with the Sequence Scanner V1.0 software (Applied Biosystems).
Ten patients (7 males, 3 females) were included with 4 cDM and 6 pDM. The mean follow-up period from time of diagnosis was 34.1 months (median 24.5 months, range 13–86 months). Mean age at diagnosis was 66.5 years (median 66.5 years, range 39–80 years). pDM occurred at a mean age of 69.5 years, cDM at 62 years. The average delay before diagnosis of DM was 47 months, with a median of 18 months (range 2–180 months). Three patients were initially misdiagnosed as malignant schwannoma, lentigo malignant melanoma and deep dermatofibroma, respectively. None of the patients had had another primary melanoma before diagnosis of DM, but 4 had a family history of primary cutaneous melanoma, and 3 had also had non-melanoma skin cancers, such as actinic keratosis and basal cell carcinomas.
DM arose on the trunk in 4 cases, in the head and neck region in 3 cases and on the extremities in 3 cases. The head and neck region was more concerned in pDM than cDM. Only 2 tumors were clinically pigmented. Clinical ulceration was observed in two cases. All melanomas were diagnosed at a locally advanced stage, with a mean tumor thickness of 7.97 mm (median 8.25 mm, range 2.8–15 mm), 8.75 mm for pDM and 6.8 mm for cDM. All DM tumors were deeply invasive, with a Clark level IV or V. Mitotic rate was <1/mm2 in 6 cases and >4/mm2 in only one case. Cytonuclear atypia was often moderate. Ulceration was histologically present in 4 cases. Neurotropism was present in 7 cases, sometimes prominent, concerning large nerve trunks, and was more frequently observed in pDM than in cDM. Syringotropism was found in one case. Pigmentation was observed in the fusiform component of the tumor in only one patient. A grenz zone between epidermal and dermal fusiform component was present in 2 cases, lymphoid aggregates and a myxoid stroma in 6 tumors. Solar elastosis was present in 5 cases. All tumors strongly expressed S-100 protein, in fusiform and epithelioid components, whereas only one melanoma was positive for HMB-45 staining in the fusiform component.
Sentinel lymph node biopsy was performed in 4 patients. A micrometastasis was detected in only one case, with a desmoplastic phenotype. Complete lymph node dissection was negative. Stages at diagnosis were IIB in 7 cases, IIC in 2 cases and IIIB in 1 case.
During the follow-up period, 3 patients developed local recurrences, 2 had regional lymph node metastases, and 2 experienced systemic metastases which they died of. One patient had metastases in several organs (liver, lung, bone), the second a single pulmonary metastasis followed by a perineural extension along the trigeminal nerve, and then an intracranial extension with temporal metastases. At the end of the follow-up, 2 patients had died of their melanoma, one had lymph node metastases, and 7 were in complete remission.
A T → A missense mutation at nucleotide 1799 of exon 15 in the BRAF gene was found in one case only, a pDM of the trunk.
DM is a rare variant of malignant melanoma, whose clinical behavior differs from that of other melanomas. We reported a retrospective series of ten cases of DM, which did not significantly differ from those published elsewhere . When comparing with non-desmoplastic melanoma, patients were older, time to diagnosis was higher and a significant number of patients had an initial erroneous diagnosis.
Two features may explain the important delay to diagnosis. First, as was the case in 8 of our 10 patients, DM is often a clinically non-pigmented lesion. Then, microscopic features may be confusing with sarcoma or other spindle cells tumors that do not express HMB-45, as was the case in all but one of our patients. The consequence is that DM is diagnosed at a locally advanced stage, with a mean tumor thickness at the time of diagnosis higher than in conventional melanomas. However, DM overall survival is similar to that of patients with conventional melanoma with comparable Breslow index , as was the case in our study.
One peculiarity of our patients was the high incidence of family history of melanoma, and the greater incidence of cutaneous carcinomas in comparison to conventional melanomas, which was consistent with literature data . Unlike most studies, our patients’ DM occurred preferentially on the trunk rather than in the head and neck region.
Considering microscopic features, our patients differed from the literature by the frequency of tumor ulceration and, in one case, presence of syringotropism of tumoral cells. On the opposite, as reported in other studies, mitotic index was low and neurotropism frequently observed.
Chemotherapies targeting the RAS-RAF-MEK-ERK pathway have brought new hope for the treatment of melanoma . Vemurafenib inhibits the BRAF protein when an activating mutation is present. The most frequent mutation is a T → A missense mutation at the nucleotide 1799 in exon 15, resulting in a valine-to-glutamic acid change at residue 600 (p.V600E mutation). Between 29 and 66% of cutaneous primitive melanomas express this mutation [17,18]. To our knowledge, only one paper reported BRAF analysis in DM . The authors, in a series of 12 DM, failed to identify the activating mutation whereas, in our hands, one tumor did express it. This melanoma was a pDM, so we cannot implicate the epithelioid component in the presence of the mutation. So, contrary to Davison and al.’s  conclusions, we believe that, even in the case of DM, this mutation is worth exploring. However, taken together, our two series suggest that BRAF mutation is a rarer event in DM than in conventional melanoma. This is in accordance with genomic studies which demonstrated numerous differences in gene expression profiling between DM and conventional melanoma .
The close relationship between BRAF mutation and sun exposure has been pointed out by Maldonado et al. . Mutation is preferentially observed in melanomas arising in intermittent sun-exposed skin, i.e. on the trunk and extremities, and is rare in those arising in non-exposed or chronically sun-exposed skin (i.e. the head and neck). In this respect, it is worth noting that the mutated melanoma of our series developed on the trunk, which is an intermittent sun-exposed area.
Department of Dermatology, Université d’Auvergne, CHU
1 place Lucie Aubrac
FR–63001 Clermont-Ferrand (France)
Received: April 24, 2012
Accepted after revision: August 13, 2012
Published online: October 24, 2012
Number of Print Pages : 4
Number of Figures : 0, Number of Tables : 0, Number of References : 19
Vol. 225, No. 2, Year 2012 (Cover Date: November 2012)
Journal Editor: Saurat J.-H. (Geneva)
ISSN: 1018-8665 (Print), eISSN: 1421-9832 (Online)
For additional information: http://www.karger.com/DRM