Microscopic endometrial perivascular epithelioid cell nodules: a case report with the earliest presentation of a uterine perivascular epithelioid cell tumor

Neoplasms with perivascular epithelioid cell (PEC) differentiation (PEComas), as defined by the World Health Organization, are a family of mesenchymal tumors composed of histologically and immunohistochemically distinctive PECs, characteristically expressing both melanocytic and smooth muscle markers [1]. The concept of “PEC’ was first proposed by Bonetti et al. to identify a novel cell type, which exhibited an epithelioid appearance, clear-eosinophilic cytoplasm, and a perivascular distribution in renal angiomyolipoma (AML), clear-cell sugar tumor (CCST) of the lung, and lymphangioleiomyomatosis (LAM) [2]. Since that time, PEComas have been reported at a wide variety of anatomic sites [3‐5]. In the female genital tract, PEComas most frequently affect the uterus [6]. AML, LAM, sclerosing PEComa, and PEComa-not otherwise specified (NOS) were described in the uterus [6‐16]. We herein report a unique case of microscopic endometrial PEC nodules, which have never been reported in the literature, and they perhaps represent the earliest manifestation of uterine PEComas. We discuss their histologic and immunohistochemical features, the relationship between the microscopic PEC nodules and PEComas of the female genital tract, and their clinical significance.

Specimens of the first surgery included two myometrial fragments, one piece of pelvic soft tissue, and some ovarian fragments. Specimens of the second surgery included the uterine corpus, bilateral ovaries, and bilateral fallopian tubes. All specimens were fixed in 10 % formalin and processed for a histologic examination by conventional methods. Hematoxylin and eosin (H&E)-stained sections were analyzed. Immunohistochemistry was performed using a Benchmark XT automated slide stainer (Ventana Medical Systems, Tucson, AZ, USA). The primary antibodies are listed in Table 1.

Since HMB-45 immunoreactivity was described in renal AML and pulmonary CCST in 1991, the family of PEComas has expanded to include AML, LAM, CCST, clear-cell myomelanocytic tumor of the falciform ligament/ligamentum teres, and clear-cell tumors of other anatomic sites in the past 2 decades [1, 17‐21]. Although their distinctive morphologic, immunohistochemical, ultrastructural, and genetic features have been clearly demonstrated, their histogenesis and normal counterpart remain poorly understood [4]. Rare examples of early or precursor lesions of PEComas were reported in the literature [11, 12, 22‐25]. In the current case, some microscopic PEC nodules, which were entirely confined within the endometrium with no attachment to the myometrium or the PEComa despite serial sections, were seen. These findings indicate that the nodules were not a consequence of an infiltration from the PEComa in the myometrium, supporting such nodules representing an example of early or precursor PEC lesions. We speculate that a uterine PEComa may evolve from these nodules. Proliferation of PECs initially begins around small arteries in the endometrium and produces microscopic PEC nodules. Further proliferation of PECs expands the nodule size and subsequently extends to the myometrium forming a better-developed PEComa. Only three other cases with similar early or precursor PEC lesions were reported in female genital organs, with two cases being associated with uterine PEComas and the other regarded as the earliest LAM [11, 12, 22]. The authors used the designation ”PEComatosis” to describe such lesions [11, 12]. The clinical and pathologic features of these cases, including ours, are summarized in Table 2. The size of the early or precursor PEC lesions was usually < 2 mm. All cases, except the case reported by Liang et al., showed a similar immunohistochemical profile with diffuse positivity to melanocytic markers and focal positivity to smooth muscle markers. The case described by Liang et al. also revealed positive immunoexpressions for HMB-45 and SMA, but the extent of the staining was not described. Involvement of early PEC lesions in multiple organs was found in three patients, all of whom were complicated with tuberous sclerosis. The phenomenon may simply reflect the underlying genetic alterations associated with tuberous sclerosis, in which simultaneous or non-simultaneous development of multiple PEComas in many organs is characteristic [5]. Although patients with tuberous sclerosis are prone to having multiple PEComas and/or early PEC lesions in the female genital tract, the multicentricity of PEComas does not seem to be associated with adverse clinical outcomes [11, 16]. Our patient remained free of local recurrence and metastasis 168 months after excision of the uterine corpus, ovaries, and fallopian tubes.

AMLs, the most common member of PEComas, were reported to have similar early PEC lesions, which were designated as microharmatomas, small mesenchymal nodules, or intraglomerular lesions [4, 23‐25]. Such lesions were confined within glomeruli or located outside glomeruli. They contained only epithelioid cells or a mixture of epithelioid cells and adipocytes. Immunoreactivity with HMB-45 was consistently noted in epithelioid cells. Chowdhury et al., analyzing the histologic relationship between small mesenchymal nodules and AML, found that small-sized mesenchymal nodules tended to only be comprised of epithelioid cells without adipocytes or blood vessels, compared to large-sized ones which contained epithelioid cells, adipocytes, and blood vessels [24]. These findings support small mesenchymal nodules being the buds of AMLs. Precursor PEC lesions were also described in association with a urachal PEComa [26]. The precursor lesion, remote from the main urachal PEComa, was composed of capillaries lined by HMB-45-positive clear cells. A gradual transition from the precursor lesions to invasive PEComa nests was observed.

The present case is unique for the early PEC lesions being present in the endometrium, in contrast to other early PEC lesions, which were reported in the myometrium [11, 12, 22] (Table 2). The endometrial distribution of early PEC lesions may account for rare uterine PEComas, which present as polypoid endometrial lesions with minimal myometrial involvement [9, 10]. This also strengthens the supposition that uterine PEComas are distinctive from uterine smooth muscle tumors. Uterine PEComas being a distinct clinicopathologic entity can be challenged by frequent HMB-45 expression in epithelioid and conventional leiomyosarcomas of the uterus [27‐29]. The question arises as to whether uterine PEComas are a distinctive entity or whether they represent a subgroup of HMB-45-positive smooth muscle tumors [30]. Since smooth muscle tumors typically originate in the myometrium, the endometrial origin of early PEC lesions supports uterine PEComas being a distinctive tumor. The immunophenotype of early PEC lesions, with diffuse HMB-45 and rare SMA expressions, also argues against their origin from smooth muscle cells.

Another immunophenotypic feature requiring further discussion is CD10 expression. In the present case, CD10 expression was restricted to microscopic endometrial PEC nodules, whereas the well-formed PEComa, LAM, and sclerosing PEComa in the myometrium showed negative CD10 reactivity. CD10 expression was reported in 25 % of uterine PEComas which underwent CD10 staining [6]. Although CD10 reactivity is not low in uterine PEComas, the significance of CD10 expression remains to be elucidated. Herein, we try to clarify the role of CD10. First, CD10 is commonly used to distinguish endometrial stromal nodules/sarcomas from uterine smooth muscle tumors [31]. CD10, although occasionally positive in smooth muscle tumors, is relatively specific for endometrial stromal cells. It would be interesting to determine whether or not the development of uterine PEComas is related to endometrial stromal cells. In CD10-positive uterine PEComas, tumor locations were not limited to the endometrium [10, 32, 33]. CD10 expression was not always detected in uterine PEComas which extensively involved the endometrium [13]. On the basis of the aforementioned findings, the hypothesis that uterine PEComas are derived from endometrial stromal cells is not supported. Second, CD10 is used as a cell surface marker of mesenchymal stem cells and tumor stem cells [34]. It is of paramount importance to understand whether CD10 is a surface marker of tumor stem cells of uterine PEComas. Unfortunately, CD10 is not routinely used in diagnosing PEComas. CD10 staining was not done in the other uterine PEComas with early or precursor PEC lesions [11, 12]. More cases with early or precursor PEC lesions stained with CD10 are required to answer this question.

Moreover, since uterine PEComas may present as endometrial lesions with CD10 expression, PECs may be misinterpreted as endometrial stromal cells, particularly in an endometrial curettage specimen. Plump cells with more abundant clear to light eosinophilic cytoplasm and a radial arrangement around vessels should raise suspicion of the existence of PECs. A misdiagnosis can be avoided by awareness of the existence of endometrial PEComas and assistance with immunostaining using melanocytic markers.

The present case showed microscopic endometrial PEC nodules widely distributed in the endometrium of adenomyosis, pelvic endometriosis, an ovarian endometriotic cyst, and the endometrium of the uterine cavity in a tuberous sclerosis patient. The microscopic endometrial PEC nodules were clearly demonstrated to have progressed into a uterine PEComa. Further studies focusing on early HMB-45-positive PEC nodules may be helpful in clarifying the pathogenesis of PEComas. The role of CD10 in PEComas also needs to be further evaluated.