Clear Cell Myomelanocytic Tumor (PEComa) of the Duodenum in a Child With a History of Neuroblastoma

Clear cell myomelanocytic tumors (PEComas) are rare tumors initially identified in the lung as “sugar” tumors.1 They were later described in many extrapulmonary sites under various names. Their histologic appearance is distinctive, but their histogenesis remains debated. An origin from a putative perivascular epithelioid cell has been proposed; hence, the term PEComa was coined by investigators1 and used in recent World Health Organization classifications2 to designate this group of tumors, regardless of their location. The alternative term clear cell myomelanocytic tumor has been frequently used in the literature. This descriptive term refers to the distinctive immunophenotype of the tumor cells, which coexpress smooth muscle cell and melanocytic markers.3 We report herein a case of PEComa in an unusual location, its diagnosis by laparoscopic biopsy, and its occurrence in a child with a history of neuroblastoma.

A 12-year-old boy presented with anemia due to iron deficiency. He had been treated 8 years previously for a well-differentiated, nonsecreting cervical neuroblastoma, first by surgery, then by chemotherapy, followed by autologous bone marrow transplantation for pleural extension. A complete remission was achieved. Gastroduodenoscopy showed an ulcerated duodenal lesion. An initial series of endoscopic biopsies was not diagnostic and showed only necrosis and granulation tissue. Imaging studies disclosed a large hypervascular lesion located in the duodenal wall, extending into the pancreaticoduodenal groove and measuring 4 cm in diameter. There was no fixation on metaiodobenzylguanidine scintigraphy, urinary cathecholamine levels were within the normal range, and results of a bone marrow examination were normal. Because of the child's history of neuroblastoma, it was decided to perform a biopsy of the lesion under laparoscopy. Histologic examination showed that the lesion was formed by a proliferation of large epithelioid cells with clear, glycogen-laden cytoplasm. Tumor cells were associated with thin fibrous septa containing numerous capillary vessels. Immunohistochemical examination showed that the tumor cells coexpressed smooth muscle cell markers (including α smooth muscle actin) and melanocytic antigens (including HMB-45). Based on histologic and immunohistochemical findings, the diagnosis of PEComa was proposed.

Because of the large size of the tumor and its ill-defined margins, surgical resection was proposed. A pylorus-preserving pancreaticoduodenectomy was performed. Macroscopic examination showed that the tumor, partially encapsulated and measuring 3.5 cm, was restricted to the duodenal wall, without invasion of the adjacent pancreas. On histologic examination, the lesion was found to involve the mucosal and submucosal layers and had invaded the whole thickness of the muscularis propria. The tumor cells were monomorphic, manifesting as large epithelioid cells with well-defined abundant clear cytoplasm, and the nuclei were large and regular (Figure 1). The tumor cell phenotype was verified by immunohistochemistry. All tumor cells expressed smooth muscle (α smooth muscle actin) and melanocytic (HMB-45, PNL2, and Melan-A) markers (Figure 2). Epithelial (cytokeratins and epithelial membrane antigen) and endocrine (chromogranin A and synaptophysin) markers were undetectable, as were protein S100 and CD117 (c-Kit). CD99 was detectable in most tumor cells. There was no necrosis or cellular atypia. The mitotic index was low, and the proliferation index, evaluated by Ki-67 staining, was less than 2%. No metastasis was detected in the 21 lymph nodes examined. No further treatment was performed. Clinical examination and complete imaging studies, including head computed tomography, have not disclosed any lesion suggestive of a diagnosis of tuberous sclerosis. At this writing, the patient is alive, without recurrence or metastasis, after 2 years of follow-up.

The case of PEComa reported herein is typical in its histologic and immunohistochemical features,1 which rule out other digestive clear cell tumors (including rare variants of adenocarcinomas or endocrine carcinomas). Its distinctive features also rule out clear cell sarcoma, which may occur in the duodenum4 and is associated with a t(12; 22) translocation resulting in an EWS/ATF-1 fusion transcript. However, our case is unusual in its location and presentation.

To the best of our knowledge, this is the first reported case of PEComa in the duodenum, although this tumor has been reported in other digestive sites, including the jejunum,5,6 ileum,7 pancreas,8 liver,1 and falciform ligament.3 A second unusual feature of our case was its mode of diagnosis. In the absence of suggestive imaging features, PEComa is usually diagnosed only after surgical resection. In the present case, the diagnosis was made by histologic examination of a laparoscopic biopsy specimen, based on the distinctive morphologic and immunophenotypic characteristics of the tumor cells,2 which are diagnostic even with limited biopsy material.

The third unusual feature of the present case is its occurrence in a patient with a history of a rare cervical neuroblastoma. Cervical (or cervicothoracic) tumors represented fewer than 10% of the cases of neuroblastoma in a large nationwide multicentric study.9 As illustrated by our case, cervical neuroblastoma usually has a favorable prognosis, with prolonged disease-free survival and limited risk of recurrence.9,10 Similarly, PEComa is an unusual lesion, especially in the digestive tract. The association between these 2 rare lesions may not be fortuitous. The following 2 hypotheses may be considered: (1) PEComas, or PEComa-like tumors, may derive from neuroblastomatous foci, although this is unlikely based on current concepts about the histogenesis of these tumors. (2) PEComas, at least in some subsets of patients, and neuroblastomas may share some common molecular mechanisms. The latter hypothesis is supported by the fact that PEComa and neuroblastoma belong to the spectrum of tumors known to occur in patients with tuberous sclerosis. The association between tuberous sclerosis and neuroblastoma is well known and may be due to a common genetic defect in the TSC2 locus on chromosome 16 that is implicated in tuberous sclerosis.11 The association between PEComa and tuberous sclerosis has been documented in 2 cases of uterine tumors,6,12 but genetic defects were not investigated in those cases. Little information is available about the possible molecular alterations occurring in PEComas.2 In one report, the loss of chromosome X and a t(3;10) translocation were identified by conventional cytogenetic techniques.3 In another report, no gain or loss of chromosomal material was shown by comparative genomic hybridization.6 No cytogenetic or molecular analysis was performed in the case presented herein. However, a complete clinical and radiologic investigation was done to rule out tuberous sclerosis, which is recommended by others, especially in patients with abdominopelvic PEComas.6 

Our case report illustrates the primary clinical dilemma raised by PEComas, namely, the evaluation of their risk of malignancy.6,13 In our case, the malignant character of the lesion was suggested by results of imaging studies demonstrating its size and possible pancreatic invasion. Malignancy was not found on final examination, with no adverse histologic features6 or indications of metastasis. However, evidence of disseminated disease is sometimes much delayed,6,14 which justifies a prolonged follow-up.

In conclusion, our case report identifies another possible site of occurrence for PEComa. In addition, it demonstrates that distinctive morphologic and immunohistochemical features allow for identification of this tumor in biopsy material.