Bowen disease with matrical differentiation: Report of an exceptional histopathologic presentation.

Matrical differentiation is the distinctive feature of pilomatricoma and other purely matrical adnexal neoplasms; additionally, foci of matrical differentiation have been also described in hybrid cysts of Gardner syndrome, as well as in a wide variety of benign and malignant cutaneous tumors, including basal cell carcinoma. We report an exceptional case of Bowen disease exhibiting multiple foci of matrical differentiation, as confirmed by means of immunohistochemical studies. Several types of divergent, non-squamous differentiation have been exceptionally reported in cutaneous squamous cell carcinoma in situ (cSCCIS), including sebaceous, mucinous/glandular, poroid, tricholemmal, and neuroendocrine differentiation; matrical differentiation may be added to this list. Our findings further emphasize the undifferentiated nature of neoplastic cells in cSCCIS.

Mitotic rate in primary melanoma: interobserver and intraobserver reliability, analyzed using H&E sections and immunohistochemistry.

BACKGROUND: In 2009, the AJCC issued a revised melanoma staging system. In addition to tumor thickness and ulceration, the mitotic rate was introduced as the third major prognostic parameter for the classification of primary cutaneous melanoma. Given that, according to the 2009 AJCC classification, the detection of one or more dermal tumor mitoses leads to an upstaging - from stage Ia to Ib - of melanomas with a tumor thickness of ≤ 1.0 mm, we set out to investigate the reproducibility of this new parameter. METHODS: In order to assess interobserver reliability, 17 dermatopathologists und pathologists - all well versed in the diagnosis of cutaneous melanoma - analyzed the mitotic rate in 15 thin primary cutaneous melanomas (mean tumor thickness 0.91 mm) using identical slides. Mitotic rates were determined on H&E and phosphohistone H3 (Ser10)-stained samples. Without knowledge of their previous assessment, five of the aforementioned examiners reevaluated the samples after more than one year in order to ascertain intraobserver reliability. RESULTS: Interobserver reliability of the mitotic rate in thin primary melanomas is disappointing and independent of whether H&E or immunohistochemically stained samples are used (kappa value: 0.088 [H&E], 0.154 [IH], respectively). Kappa values improved to 0.345 (H&E) and 0.403 (IH) when using a cutoff of 0/1 vs. 2+ mitoses. Similarly unsatisfactory, kappa values for intraobserver reliability ranged from 0.18 and 0.348, depending on the individual examiner. DISCUSSION: Given the unsatisfactory reproducibility and large variations in assessing the mitotic rate, it remains a matter of debate whether this diagnostic parameter should play a role in therapeutic decisions.

Mitoserate beim primären Melanom: Interobserver- und Intraobserver-Reproduzierbarkeit am HE-Schnitt und in der Immunhistologie.

HINTERGRUND: Die Melanomklassifikation wurde 2009 durch die AJCC revidiert. Für die Klassifizierung primärer Melanome wurde als dritte Größe neben Tumordicke und Ulzeration die Angabe der Mitoserate neu eingeführt. Gemäß der AJCC-2009-Klassifikation des Melanoms führt der Nachweis nur einer oder mehrerer dermaler Tumormitosen bei Melanomen ≤ 1,0 mm Tumordicke zu einer Umgruppierung des Tumors von T1a nach T1b. Dies erklärt, wie wichtig die Frage nach der Reproduzierbarkeit dieses neuen Parameters ist. METHODEN: Zur Prüfung der Interobserver-Reproduzierbarkeit der Mitoserate haben 17 Dermatopathologen und Pathologen, die in der Befundung des kutanen Melanoms sehr erfahren sind, die Mitoserate in 15 dünnen Melanomen mit einer mittleren Tumordicke von 0,91 mm an demselben Tumorschnitt bestimmt. Die Mitoserate wurde am HE-Schnitt und immunhistologisch (IH) mittels des mitosespezifischen Antikörpers Phospho-Histon-H3 (Ser10) bestimmt. Fünf Befunder wiederholten die Bestimmung nach mehr als einem Jahr ohne Kenntnis ihres Vorbefundes zur Ermittlung der Intraobserver-Reproduzierbarkeit. ERGEBNISSE: Die Interobserver-Reproduzierbarkeit der Mitoserate bei dünnen Melanomen ist unbefriedigend und unabhängig davon, ob die Mitoserate am HE-Schnitt oder am immungefärbten Schnitt bestimmt wird (κ-Werte: 0,088 [HE] bzw. 0,154 [IH]). Bei einer Diskriminationsschwelle von 0/1 vs. 2+ Mitosen verbesserte sich der κ-Wert auf 0,345 (HE) bzw. 0,403 (IH). Die Intraobserver-Reproduzierbarkeit lag mit κ-Werten zwischen 0,18 und 0,348 je nach Befunder ebenfalls im unbefriedigenden Bereich. DISKUSSION: Wegen der unbefriedigenden Reproduzierbarkeit und der großen Variation der Befunde zur Mitoserate bleibt es zweifelhaft, ob dieser Befund als Grundlage für Therapieentscheidungen herangezogen werden kann.

Helical tomotherapy of the complete scalp and the ipsilateral lymph nodes in a case of scalp angiosarcoma.

Scalp angiosarcoma represents a therapeutic challenge to all disciplines. This case report demonstrates the potential usefulness of helical tomotherapy (HT) as a new radiotherapeutic treatment option. A 71-year-old woman presented with a superficial angiosarcoma of the scalp, forehead, and left pre- and postauricular areas, with several nodular ulcerating and bleeding lesions. Irradiation of the gross tumor was performed with a total dose of 70 Gy in 2-Gy fractions and of the left cervical lymph nodes with 56 Gy in 1.6-Gy fractions. Good target coverage was achieved without compromising organs at risk, notably the brain. Treatment was very fast (661 seconds per fraction) and was administered with minimal acute toxicity (National Cancer Institute Common Toxicity Criteria: grade 2 erythema and grade 2 dysphagia). During treatment, tumor nodules dissolved into hyperkeratosis. We conclude that with HT, irradiation of the scalp and cervical lymph nodes can be conducted with minimal acute toxicity and without junction problems.

Survival according to BRAF-V600 tumor mutations--an analysis of 437 patients with primary melanoma.

The prognostic impact of BRAF-V600 tumor mutations in stage I/II melanoma patients has not yet been analyzed in detail. We investigated primary tumors of 437 patients diagnosed between 1989 and 2006 by Sanger sequencing. Mutations were detected in 38.7% of patients and were associated with age, histological subtype as well as mitotic rate. The mutational rate was 36.7% in patients with disease-free course and 51.7% in those with subsequent distant metastasis (p = 0.031). No difference in overall survival (p = 0.119) but a trend for worse distant-metastasis-free survival (p = 0.061) was observed in BRAF mutant compared to BRAF wild-type patients. Independent prognostic factors for overall survival were tumor thickness, mitotic rate and ulceration. An interesting significant prognostic impact was observed in patients with tumor thickness of 1 mm or less, with the mutation present in 6 of 7 patients dying from melanoma. In conclusion, no significant survival differences were found according to BRAF-V600 tumor mutations in patients with primary melanoma but an increasing impact of the mutational status was observed in the subgroup of patients with tumor thickness of 1 mm or less. A potential role of the mutational status as a prognostic factor especially in this subgroup needs to be investigated in larger studies.

A benign cutaneous plexiform hybrid tumor of perineurioma and cellular neurothekeoma.

There are several recent reports describing hybrid peripheral nerve sheath tumors showing a biphasic component of neoplastic cells. These combinations include a mixture of neurofibroma and schwannoma, schwannoma and perineurioma, neurofibroma and perineurioma, and perineurioma and granular cell tumor. A case of a triphasic combination of neurofibroma, schwannoma, and perineurioma has also been described. We describe the clinicopathologic and immunohistochemical characteristics of 9 cases of a benign cutaneous plexiform nerve sheath tumor located on the lips and exhibiting hybrid features of perineurioma and cellular neurothekeoma. Clinically, lesions were solitary dome-shaped papules located on the lips. Histopathologically, the neoplasms consisted of well-circumscribed but uncapsulated dermal nodules with a plexiform pattern. They were composed of nests or rounded aggregations of neoplastic cells embedded in a slightly myxoid stroma. Within the aggregates, cells were distributed in a storiform and lamellar pattern. Immunohistochemically, most neoplastic cells expressed strong immunoreactivity for S100A6, MiTF, NKI/C3, PGP9.5, EMA, and NSE, whereas variable, focal, and weaker positivity for CD34, claudin-1, and Glut-1 was seen in some cases. On the basis of these findings, we believe that this neoplasm is a distinctive benign cutaneous plexiform nerve sheath tumor with histopathologic and immunohistochemical hybrid features of perineurioma and cellular neurothekeoma.

Malignant dermatofibroma: clinicopathological, immunohistochemical, and molecular analysis of seven cases.

Dermatofibroma (cutaneous fibrous histiocytoma) represents a common benign mesenchymal tumor, and numerous morphological variants have been described. Some variants of dermatofibroma are characterized by an increased risk of local recurrences, and there are a few reported metastasizing cases. Unfortunately, an aggressive behavior cannot be predicted reliably by morphology at the moment, and we evaluated the value of array-comparative genomic hybridization (CGH) in this setting. Seven cases of clinically aggressive dermatofibromas were identified, and pathological and molecular features were evaluated. The neoplasms occurred in four female and in three male patients (mean age was 33 years, range 2-65 years), and arose on the shoulder, buttock, temple, lateral neck, thigh, ankle, and cheek. The size of the neoplasms ranged from 1 to 9 cm (mean: 3 cm). An infiltration of the subcutis was seen in five cases. Two neoplasms were completely excised, whereas an incomplete or marginal excision was reported in the remaining cases. Local recurrences were seen in six cases (time to the first recurrence ranged from 8 months to 9 years). Metastases were noted between 3 months and 8 years after diagnosis in six patients. Two patients died of disease, and two patients are alive with disease. Histologically, the primary tumors showed features of cellular dermatofibroma (four cases), cellular/aneurysmal dermatofibroma (one case), atypical/cellular dermatofibroma (one case), and classical dermatofibroma (one case). Mitotic figures ranged from 3 to 25 per 10 high-power fields, and focal necrosis was present in five cases. Interestingly, malignant transformation from cellular dermatofibroma to an obvious spindle cell/pleomorphic sarcoma was seen in one primary and in one recurrent neoplasm. Five neoplasms showed chromosomal aberrations by array-CGH, suggesting that these changes may represent an additional diagnostic tool in the recognition of cases of dermatofibroma with a metastatic potential.

A proposal for improving multicolor FISH sensitivity in the diagnosis of malignant melanoma using new combined criteria.

Fluorescence in situ hybridization (FISH) for the diagnosis of melanoma makes use of specific fluorescent probes to detect selected chromosomal alterations on paraffin-embedded tissue samples. To date, interpretation of FISH data has been based on numerical values generated by 2 different computational algorithms that of Abbott and that of Gerami. To further evaluate the value of FISH in the diagnosis of malignant melanoma, we selected 163 clinically and histologically unequivocal cases of malignant melanoma in a cohort of 575 melanocytic tumors and analyzed FISH data using the criteria of Abbott, Gerami, and new combined criteria. Depending on the used criteria, FISH was positive in the unequivocal malignant melanoma in 69.3% (113/163) of cases using the Abbott criteria, 74.2% (121/163) of cases using the Gerami criteria, and 82.2% (134/163) of cases using the combined criteria of Abbott and Gerami. Although use of all 3 criteria was associated with 100% FISH negativity in a cohort of 30 unequivocal benign melanocytic nevi, use of the combined criteria revealed more FISH-positive cases in ambiguous benign melanocytic lesions than the criteria of Abbott or Gerami alone: Abbott, 125 of 367; Gerami, 146 of 367; combined, 161 of 367. Furthermore, we show that 66% (8/12) of FISH-negative cases of unequivocal melanoma are positive when analyzed by array comparative genomic hybridization (aCGH), demonstrating that false-negative results remain despite the usage of the combined criteria for evaluation of FISH data. In these 8 FISH-negative aCGH-positive cases, copy number alterations were often located on chromosomes 9p, a chromosomal locus that is not targeted by the FISH probes currently used. In conclusion, the existing criteria for the evaluation of multicolor melanocytic FISH are limited by a nonnegligeable rate of false negativity that can be reduced by using newly proposed combined criteria but at the cost of increased detection of FISH positivity in ambiguous benign melanocytic lesions.

Histological and genetic evidence for a variant of superficial spreading melanoma composed predominantly of large nests.

Cutaneous melanomas are characterized by a range of histological appearances, and several morphological variants have been described. In this study, we report a variant of superficial spreading melanoma that is characterized by large, irregular junctional melanocytic nests. The junctional nests varied in shape and size, showed focal tendency to confluence, and were often surrounded by a cuff of epidermal keratinocytes. The melanocytes comprising the nests showed variable cytological atypia. In most of the cases, scant intraepidermal or junctional single melanocytes were seen, and other well-documented diagnostic criteria for melanoma were lacking, and as a result, histological recognition of these tumors as melanoma was difficult. Some cases were associated with an invasive dermal component or showed evidence of sun damage. To provide supporting evidence for malignancy, we analyzed these tumors for genomic aberrations. Using array comparative genomic hybridization (aCGH), we identified multiple genomic aberrations in all analyzed cases. A similar pattern of genomic aberrations was seen in a control group of bona fide superficial spreading melanomas, suggesting that these 'melanomas composed exclusively or predominantly of large nests' are indeed variants of superficial spreading melanoma. Fluorescence in-situ hybridization (FISH) was positive in 40% of the cases. However, using aCGH, the FISH-negative cases showed multiple genomic aberrations in regions that are not covered by FISH. The low sensitivity of the FISH test can be explained by the fact that FISH only evaluates four genomic loci for aberrations, whereas aCGH surveys the entire genome. In summary, we present histological and molecular genetic evidence for a morphological variant of superficial spreading melanoma. Awareness of the histological features will aid in their correct diagnosis as melanoma, and in difficult cases, judicious application of ancillary tests such as aCGH (rather than FISH) will assist accurate diagnosis.

The Malassezia genus in skin and systemic diseases.

In the last 15 years, the genus Malassezia has been a topic of intense basic research on taxonomy, physiology, biochemistry, ecology, immunology, and metabolomics. Currently, the genus encompasses 14 species. The 1996 revision of the genus resulted in seven accepted taxa: M. furfur, M. pachydermatis, M. sympodialis, M. globosa, M. obtusa, M. restricta, and M. slooffiae. In the last decade, seven new taxa isolated from healthy and lesional human and animal skin have been accepted: M. dermatis, M. japonica, M. yamatoensis, M. nana, M. caprae, M. equina, and M. cuniculi. However, forthcoming multidisciplinary research is expected to show the etiopathological relationships between these new species and skin diseases. Hitherto, basic and clinical research has established etiological links between Malassezia yeasts, pityriasis versicolor, and sepsis of neonates and immunocompromised individuals. Their role in aggravating seborrheic dermatitis, dandruff, folliculitis, and onychomycosis, though often supported by histopathological evidence and favorable antifungal therapeutic outcomes, remains under investigation. A close association between skin and Malassezia IgE binding allergens in atopic eczema has been shown, while laboratory data support a role in psoriasis exacerbations. Finally, metabolomic research resulted in the proposal of a hypothesis on the contribution of Malassezia-synthesized aryl hydrocarbon receptor (AhR) ligands to basal cell carcinoma through UV radiation-induced carcinogenesis.

Cutaneous mixed tumor, eccrine variant: a clinicopathologic and immunohistochemical study of 50 cases, with emphasis on unusual histopathologic features.

Mixed tumor, eccrine type, is a rare cutaneous adnexal neoplasm, mostly reported as isolated case reports. A systematic analysis of its histopathologic and immunohistochemical features has not previously been performed on a large series. The purpose of our investigation was to study a large number of cutaneous eccrine mixed tumors so as to fully characterize the entire spectrum of changes in the epithelial and stromal components, with an emphasis on unusual histopathologic features that may represent a diagnostic pitfall. This article reports a light microscopic and immunohistochemical study of 50 cases of eccrine mixed tumor, complemented by a literature review. Our study identified some unusual histopathologic features, thus extending the morphologic spectrum of this neoplasm. These included prominent cribriform areas, clear cell change, pseudorosette structures, prominent osseous metaplasia, and physaliphorous-like cells. Most of these features have not been previously recorded in eccrine mixed tumors and may represent a potential diagnostic pitfall.

Histopathological diagnostics of malignant melanoma in accordance with the recent AJCC classification 2009: Review of the literature and recommendations for general practice.

BACKGROUND: TNM classifications are the basis for diagnostic and therapeutic procedures in oncology. Histopathological reports have to enable a proper indexing of tumor specific findings into recent classifications. METHODS: A systematic review of the literature was performed to identify reports dealing with the assessment of mitotic rate and the processing and evaluation of sentinel node biopsies in malignant melanoma. On the basis of this review an expert panel of dermatopathologists and general pathologists discussed and agreed recommendations for general practice. RESULTS: Following recommendations were agreed with a broad consensus (93-100 % agreement): The determination of the mitotic rate in primary melanoma is performed on HE slides. The evaluation of an area of 1 mm(2) is sufficient. Only dermal mitoses are considered. The counted number of mitoses is provided as an integer value. The mitotic rate shall be determined in primary melanomas of ≤1.00 mm vertical tumor thickness according to the hot-spot method and provided as an integer value in relation to an area of 1 mm(2) . The determination of the mitotic rate in the case of thicker primary melanomas is desirable. In general, for the evaluation of each sentinel lymph node, 4 slides should be prepared. For diagnostic purposes, immunohistochemistry (preferably with antibodies against S100ß, Melan A and HMB-45) should be performed in addition to HE staining. The pathology report should provide information about micro-metastases and their longest extension (one-tenth of a millimeter). CONCLUSIONS: These recommendations are suitable for standardizing the histopathological diagnosis of malignant melanoma and for providing a common basis for clinical decisions and scientific research.

Expression of apoptosis-inducing factor (AIF) in keratoacanthomas and squamous cell carcinomas of the skin.

The complex biological trait 'susceptibility to apoptosis' is a nosological feature distinguishing squamous cell carcinomas (SCC) from keratoacanthomas (KA). The purpose of this study was to compare the expression of apoptosis-inducing factor (AIF), a major effector of the caspase-independent apoptosis pathway, in formalin-fixed SCC (N = 23) and KA (N = 29) resection specimens. SCC express statistically significant more AIF than KA both as proportion of AIF+ cells by immunohistochemistry (median: 54% vs 33%; P < 0.01) and as total AIF protein content by western blot quantification (six-fold increased; P < 0.01). However, the contribution of AIF to apoptosis, measured as fraction of apoptotic nuclei with overt DNA fragmentation by the TUNEL method that co-express AIF translocated to nucleus, is significantly less prevalent among SCC (median: 19% vs 48% in KA; P < 0.01). These findings indicate to a distinctive involvement of AIF in the progression of certain epithelial skin tumors that might be exploited as a promising treatment target.

Plasmacytoid dendritic cells: an overview of their presence and distribution in different inflammatory skin diseases, with special emphasis on Jessner's lymphocytic infiltrate of the skin and cutaneous lupus erythematosus.

BACKGROUND: Plasmacytoid dendritic cells (PDC) play a pivotal role in the induction of autoimmune diseases and other skin diseases. The present study focuses on the distribution patterns of PDC in patients with cutaneous lupus erythematosus (LE) and Jessner's lymphocytic infiltrate (LI) of the skin and compares them with other skin diseases. The goal was to scrutinize the involvement of PDC in LI, and to show that PDC present a specific pattern of distribution in various cutaneous disorders. METHODS: 353 skin biopsies of LE (various subtypes), LI, and other inflammatory skin diseases as well as two halo melanocytic nevi and 10 epithelial tumors were immunohistochemically investigated for the presence of PDC by employing antibodies against CD123 and CD2AP. RESULTS: PDC were constantly detected as distinct perivascular and periadnexal clusters in LE and LI. In other forms of dermatitis, PDC could be found as single cells or scattered throughout the infiltrate or beneath the epidermis. CONCLUSIONS: Our data suggest that the distribution of PDC in tumid LE and LI is identical, and this observation suggests that both designations signify one disease. The distinct PDC arrangement in LE represents as useful diagnostic tool in the differential diagnosis with other forms of dermatitis.

Plaque-like CD34-positive dermal fibroma ("medallion-like dermal dendrocyte hamartoma"): clinicopathologic, immunohistochemical, and molecular analysis of 5 cases emphasizing its distinction from superficial, plaque-like dermatofibrosarcoma protuberans.

Medallion-like dermal dendrocyte hamartoma (DH) and superficial (plaque-like) dermatofibrosarcoma protuberans (DFSP) are CD34-positive dermal neoplasms with overlapping clinicopathologic features. We analyzed the clinical, histomorphologic, and molecular criteria of 5 DH and 7 DFSP to delineate diagnostically relevant differences between incipient dermal DFSP and its benign look-alike, DH. We expand the clinical and histologic spectrum of DH. As medallion-like dermal DH is neither of dermal dendrocyte lineage nor a genuine hamartoma, we propose instead the descriptive term of plaque-like CD34-positive dermal fibroma (PDF). Both PDF/DH and DFSP presented as slightly pigmented and indurated plaques on neck, trunk, and extremities. Histologically, DFSP was characterized either by horizontally oriented spindle cell fascicles or by diffusely arranged fibroblasts within a slightly myxoid stroma in the upper two-thirds of the dermis, whereas PDF/DH presented with a cellular band-like fibroblastic proliferation mostly in the papillary and adjacent upper reticular dermis. Only one congenital PDF/DH in a 9-year-old boy extended into the septa of the subcutaneous fat. Formalin-fixed paraffin-embedded archival tissue was used for detection of the COL1A1-PDGFB gene rearrangement by multiplex reverse transcription-polymerase chain reaction (RT-PCR) and by dual color fusion fluorescence in-situ hybridization (FISH). Archival blocs older than 4 years did not yield amplifiable RNA because of RNA degradation, whereas FISH analysis was feasible in all investigated cases. FISH analysis revealed COL1A1-PDGFB gene rearrangement in all DFSP cases (n=7), whereas RT-PCR could detect the COL1A1-PDGFB fusion transcript only in 1 DFSP. Two cases were negative. In 4 archival cases with storage between 4.5 and 12 years, RNA had been degraded making these cases unsuitable for RT-PCR. In PDF/DH, both RT-PCR and FISH analysis did not reveal any evidence of COL1A1-PDGFB gene rearrangement. We show that PDF/DH and superficial (plaque-like) DFSP, subtle clinicopathologic differences notwithstanding, are morphologic look-alikes that can be kept apart by molecular studies of the COL1A1-PDGFB gene fusion. For the detection of the COL1A1-PDGFB gene rearrangement in diagnostically difficult cases, RT-PCR and FISH analysis are reliable and helpful diagnostic tools. In archival formalin-fixed paraffin-embedded tissue, however, FISH analysis is more robust and exhibits a higher clinical sensitivity than RT-PCR.

Cutaneous clear cell sarcoma: a clinicopathologic, immunohistochemical, and molecular analysis of 12 cases emphasizing its distinction from dermal melanoma.

Clear cell sarcoma (CCS) of tendons and aponeuroses/malignant melanoma (MM) of soft parts is a rare tumor and in the majority of cases presents a characteristic reciprocal translocation t(12;22)(q13;q12) that results in fusion of the EWS and ATF1 genes. Although the melanocytic differentiation of CCS is indisputable, its precise lineage remains unclear. Typically, the slowly growing tumor affects the extremities of adolescents or young adults, especially around the ankle and foot. CCS is classically regarded as a deep soft tissue tumor associated with tendons or aponeuroses. This traditional view is put into perspective by the description of primary CCS of the gastrointestinal tract that may have a variant fusion gene EWSR1-CREB1. We describe 12 cases of cutaneous CCS and discuss the differential diagnoses. These 12 cases share an identical immunohistochemical profile with MM and thus can easily be confused with a dermal variant of spindle cell MM or metastasis of MM. The patients' ages ranged from 6 to 74 years (median: 25 y), and there was a female predominance (10 females, 2 males). Most tumors (n = 9) were located on the extremities, 2 tumors arose on the back, and 1 on the abdomen. The mean tumor size was 0.97 cm (range, 0.4 to 1.7 cm). Six cases showed invasion of the subcutis, the other 6 cases were entirely dermal. Tumor necrosis was evident in 2 cases, melanin pigment in 2 cases, and ulceration in 1 tumor. All cases showed uniform nests and fascicles of pale spindled or slightly epitheloid cells with finely granular eosinophilic or clear cytoplasm. There was fair pleomorphism with plump spindled nuclei and significantly prominent nucleoli. Multinucleated wreath-like tumor giant cells were observed in two-thirds of cases, but were usually present only focally. The dense cellular aggregates were encased by delicate fibrous septa. The stroma showed a sclerotic reticulated pattern. Partly, the nests of spindle cells bordered the epidermis, prima vista mimicking junctional nests of melanocytes. The specific translocation pattern was confirmed in all cases by fluorescence in situ hybridization. Local recurrences and metastases developed in 2 and 3 patients, respectively, and 1 patient died of the disease.