Abstract
Endometrial stromal sarcoma is a rare uterine tumor associated with favorable outcomes despite its ability to recur and metastasize to distant sites. Most recurrences are local, being limited to the abdomen/pelvis, but distant metastases can occur. Metastatic endometrial stromal sarcoma can occur many months to years after the original diagnosis or may present prior to the primary, potentially creating a diagnostic challenge. We report a bi-institutional review of 10 cases of endometrial stromal sarcoma with extrapelvic metastases without a prior history of endometriosis. The histologic, immunophenotypic, and molecular characteristics of these tumors are analyzed in the context of a relevant literature review.
Introduction
Endometrial stromal sarcomas (ESSs) are rare tumors, representing less than 1% of all uterine cancers and 15% of uterine sarcomas.1-4 ESSs usually occur in adult women with a mean age of 52 years, 4 with the highest incidence in perimenopausal women. 3 ESS is usually a slow-growing indolent disease, reflected by a 5-year disease-specific survival of greater than 90% for stages I and II, and 50% for stages III and IV.2,4
Pathologically, ESSs are classified as low-grade, high-grade, and undifferentiated sarcomas. Low-grade ESS is composed of irregularly sized and shaped islands resembling proliferative endometrial cells with an infiltrating growth pattern, a network of arterioles, and without an associated stromal response. Cells are usually small with scant cytoplasm, uniform, oval to fusiform nuclei, minimal atypia, and a low mitotic rate. However, high mitoses do not exclude the diagnosis. There may be smooth muscle differentiation, fibromyxoid change, sex cord–like differentiation, or endometrioid-type glands. Immunohistochemistry (IHC) typically reveals diffuse positive staining for CD10, ER (estrogen receptor), PR (progesterone receptor), WT-1, and occasionally for desmin and smooth muscle actin. Many cases of low-grade ESS harbor the characteristic translocation t(7;17) (p21;q15), which results in a fusion between JAZF1 and SUZ12 (JJAZ1). Other translocations described include t(6;7)(p21;p25), t(6;10;10)(p21;q22;p11), and t(1;6)(p34;p21), which result in PHF1-JAZF1, EPC1-PHF1, and MEAF6-PHF1 rearrangements. 4
High-grade ESS typically has an infiltrative growth pattern and vasculature similar to low-grade ESS, but with a more destructive growth pattern and deep invasion. Typically hypercellular with nest-like architecture, the cells have high-grade, round-cell morphology, sometimes associated with a low-grade spindle cell component. Cells have eosinophilic to granular cytoplasm, irregular nuclear contours, vesicular chromatin, and variably distinct nucleoli. There is typically a high mitotic rate, necrosis, and lymphovascular invasion. CD10, ER, and PR are usually negative but cyclin D1 is characteristically strongly and diffusely positive. If there is an associated low-grade component, CD10, ER, and PR can be positive. 4 Most tumors have a distinctive fusion between the YWHAE and NUTM2A/B (previously known as FAM22A/B) genes subsequent to the translocation t(10;17)(q22;p13),4,5 which is associated with an aggressive clinical behavior and poor prognosis.3,6 Compared with low-grade ESS, high-grade ESS has more frequent and earlier recurrences, and its overall prognosis is between low-grade and undifferentiated ESS. 4
Recurrent disease in ESS may occur with long tumor-free intervals, sometimes even 20 years after the initial diagnosis.7,8 Most recurrences are local, being limited to the abdomen/pelvis, but distant metastases can occur, most commonly to the lungs.9,10 In addition, metastases have been reported in the bladder, 2 heart,1,11 thoracic spine, 12 colon, 13 breast, 14 brain, 15 and liver.16,17
We report a bi-institutional review of extrapelvic ESS without a prior history of endometriosis. The clinical, histologic, immunophenotypic, and molecular characteristics of these tumors are analyzed for discussion.
Methods
The Laboratory Information System databases in Saskatoon and Regina, Saskatchewan, were searched from 1996 to 2017 (21-year review). A total of 32 ESS cases were identified, 10 of which had pathologically confirmed extrapelvic metastases. A central pathology review was completed along with IHC and genetic analysis. IHC stains used in the pathological analysis of these lesions included HepPar, CD10, Cyclin D1, ER, PR, Ki-67, p53, vimentin, CK7, CK20, CD45, S100, NSE (neuron specific enolase), desmin, actin, calponin, inhibin, and calretinin.
For genetic analysis, an RNA-based assay for detecting the presence of fusion transcript using NanoString Element–based technology was used. A detailed description of the methods and its reliability in detecting JAZF1-SUZ12, YWHAE-NUTM2A/B, and ZC3H7B-BCOR/BCOR-ZC3H7B fusions have been previously published. 18 Briefly, RNA was extracted from paraffin scrolls or cores obtained from formalin-fixed paraffin embedded tumor tissue with a High Pure FFPET RNA isolation kit (Roche, Laval, Quebec, Canada). A total of 100 to 300 ng of RNA was hybridized to the tagged probe set to produce hybridized RNA/probe mixture, which is then cleaned and bound to a streptavidin-coated cartridge in a NanoString Prep Station (NanoString Technologies, Seattle, WA). Following the prep station completion, the cartridge was sealed and transferred to a NanoString Digital Analyzer device (NanoString Technologies) for optical barcodes counting in a second generation NanoString Digital Analyzer. A sample ratio of 5.0 or above was considered positive for the corresponding gene fusion. A sample ratio below 5.0 was considered negative for that fusion.
A literature review of the English language was completed in MEDLINE and PubMed, with the search terms “endometrial stromal sarcoma” and “metastases/neoplasm metastases.” The resulting publications and their relevant references were reviewed.
Results
Epidemiology
There were 32 cases of ESS identified over the past 21 years. Five cases occurred prior to the beginning of the database (1996), resulting in an average incidence of 1.35 cases per year over the past 21 years (range of 0 to 3 cases identified per year). With a population of just over a million people in Saskatchewan, the incidence of ESS is roughly one in a million.
Of the 32 ESS cases, 10 had pathologically confirmed extrapelvic metastases. There were 4 additional cases that had extrauterine involvement, but not distant metastases. The possibility that the remaining 22 cases had distant metastases that were not discovered or removed cannot be excluded. Of the 10 cases with distant metastases, 2 also had local spread at presentation, and 2 separate cases had local recurrences.
Histopathology
The low-grade cases (Figure 1A) available for review all displayed small monomorphic oval cells with scanty cytoplasm and small round and uniform nuclei accompanied by well-defined vascular spaces. Atypia was minimal, mitotic rates were low, and necrosis was absent. The IHC was also as expected, with positive staining for ER, PR, and CD10. There were no significant changes in morphology on recurrence, with the metastatic tumors displaying the same features and staining profile as the primaries.

(A) Hematoxylin-eosin stained section showing low-grade endometrial stromal sarcoma (ESS) at medium power. A network of arterioles (arrows) can be seen with surrounding cells resembling proliferative endometrium. The cells are small with uniform nuclei and a low mitotic rate. (B) Hematoxylin-eosin stained section showing high-grade ESS at medium power. The nuclei have a higher grade morphology and increased mitotic rate (arrows).
The 2 high-grade cases (Figure 1B) had a similar appearance and also maintained their histology after metastasizing. The tumors had a permeative growth pattern consisting of neoplastic cells arranged in a concentric manner around vascular spaces. The tumors were vaguely nested with round to spindle cells containing scant cytoplasm and large vesicular nuclei with irregular contours. Numerous atypical mitotic figures were identified throughout the tumors (up to 20/10 high-power field) with a proliferative index of 15% to 20%. Lymphovascular invasion was readily apparent and tumor necrosis was present. There were also accompanying low-grade areas, which were very limited in one of the tumors. Tumor cells in the high-grade areas were negative for ER, PR, and CD10, while both being positive for Cyclin D1. The low-grade areas were variably positive for ER, PR, and CD10.
Extrapelvic Metastases
A total of 10 patients had extrapelvic metastases of ESS (Table 1). There were 8 cases of low-grade ESS and 2 cases of high-grade ESS. Two cases had adnexal involvement at presentation, both low-grade, and 2 patients had distal involvement at the time of diagnosis, also both low-grade. Six patients had multiple episodes of distant metastases, including 1 out of the 2 high-grade ESS cases, and 4 low-grade ESS cases. The most common distant site overall was the lung (Figure 2A-E), in 6 patients. Two cases involved the colon/rectum, 1 involved the small bowel (Figure 2F), 2 involved the peri–appendiceal tissue, and 2 involved the liver. 17 One case had suspected bone involvement within the humerus on imaging (Figure 3), but there was no histological confirmation.
Histological, Immunohistochemical, and Molecular Data.
Abbreviations: ER, estrogen receptor; PR, progesterone receptor; NA, not applicable; R/L, right/left.
Poor RNA quality.

(A) Hematoxylin-eosin stained section showing low-grade endometrial stromal sarcoma (ESS) (^) involving the lung (*) at medium power. (B) Hematoxylin-eosin stained section showing high-grade ESS (^) involving the lung (*) at medium power. (C) Estrogen receptor immunohistochemistry showing positive staining in low-grade ESS involving the lung at medium power. (D) Cyclin D1 immunohistochemistry showing positive staining in high-grade ESS involving the lung at medium power. (E) CD10 immunohistochemistry showing positive staining in low-grade ESS at medium power. (F) Progesterone receptor immunohistochemistry showing positive staining in low-grade ESS at medium power.

Computed tomography imaging shows a lytic destructive process involving the left proximal humerus extending into the neck and greater tuberosity with an associated soft tissue mass, worrisome for metastatic disease.
Molecular Characteristics
Five cases had confirmed translocations (Table 1), with 3 being low-grade with detection of JAZF1/SUZ12 fusions (Figure 4), and 2 high-grade with detection of YWHAE/NUTM2 fusions (Figure 5). Of the remaining 5 cases, 4 had poor quality of RNA precluding complete analysis. As all of the cases with failed analysis were diagnosed as low-grade ESS on histology, the possibility that they harbored one of the other reported translocations in low-grade ESS cannot be ruled out.

Molecular analysis showing the characteristic JAZF1-SUZ12 fusion seen in low-grade endometrial stromal sarcoma.

Molecular analysis showing the characteristic YWHAE/NUTM2 fusion seen in high-grade endometrial stromal sarcoma.
Clinical Features
The clinical characteristics are summarized in Table 2. The average age at diagnosis was 40.3 years (range 20-57 years), and the average time between primary and first recurrence was 122.4 months (range 0-420 months). Only 2 cases were confined to the uterus at initial presentation, with the remaining involving either the adnexa or distant sites. The most common initial surgery was a total abdominal hysterectomy and bilateral salpingo-oophorectomy. Four cases had lymph node dissections completed; 3 of the 4 were negative while 1 had 2 positive lymph nodes. Six patients are still alive, while 4 are deceased, with the average survival being 16.6 years and 25.3 years, respectively.
Clinical Characteristics.
Abbreviations: TAH/BSO, total abdominal hysterectomy and bilateral salpingo-oophorectomy; NA, not applicable.
Discussion
ESS is most common in perimenopausal women, with possible presentations including abnormal uterine bleeding, abdominal pain, an enlarged uterus, or a pelvic mass. Patients may also be asymptomatic, with metastases being the initial presentation. Overall, stage is the most important prognostic factor. 4 Ten percent of patients will have adnexal involvement at presentation, and lymph node involvement at presentation may be from 10% to 30%.4,19 Most recurrences are local, limited to the abdomen/pelvis, but distant metastasis may also happen. Late recurrences are known to occur, even in early stage disease. 20 However, low-grade ESS typically have an overall excellent prognosis. Conversely, high-grade ESS more frequently experience disease recurrence and more often die of disease. 19
Metastatic ESS can occur many months to years after the original diagnosis or may present prior to the primary, due to either a delayed or missed diagnosis. The median time to recurrence has been reported from 34 to 65 months.20-22 However, much longer intervals have been reported, including 18 years 23 and 29 years. 24 . In one series, only 2 of 5 patients had a definitive previous diagnosis of ESS, 2 while another series reported 6 cases of delayed diagnoses due to an initial misdiagnosis. 25 In one series, the pulmonary lesion was detected prior to the uterine mass in 3 cases, and both pulmonary and uterine lesions were detected simultaneously in 1 patient. 26 Therefore, pathologists entertaining the diagnosis of distant involvement by ESS cannot always rely on a prior clinical history of uterine ESS.2,26
Diagnostic difficulty may also arise due to small biopsy size, varied histological features, nonspecific imaging appearance, and various radiologic manifestations.26,27 For example, pulmonary metastases of low-grade ESS can manifest as various patterns on computed tomography imaging, including the presence of a solitary nodule, multiple nodules, multiple cysts, and reticulonodular infiltrates.7,9,28-30 IHC can aid in the diagnosis of ESS. 31 Low-grade ESSs are usually positive for ER, PR, CD10, and at least focally for actin, while usually negative for desmin and h-caldesmon. Cyclin D1 is usually positive in high-grade ESS. 5 Histologic features of recurrent or metastatic ESS generally parallel those of the initial tumors, resembling stromal cells in normal proliferative endometrium. 2 However, recurrent tumors may demonstrate more nuclear pleomorphism and atypia compared with the primary tumor or show areas of high-grade features or leiomyomatous differentiation. 32 Conversely, tumors may appear more benign in appearance. In one study, the characteristic histological features of ESS, including spiral arteriole-like vasculature, were not apparent in 26% of metastatic nodules and secondary histological changes were observed in 93% of the nodules. Nuclear pleomorphism or mitotic figures were not apparent in many cases, and their scarcity could be suggestive of a benign spindle cell lesion. Awareness of the various secondary histological changes and use of IHC may reduce errors in diagnoses. 26
The most common distant metastatic site is the lungs. Approximately 10% of cases of ESS confined to the uterus at the time of diagnosis will develop pulmonary metastasis. 28 One study reported intervals from hysterectomy to subsequent pulmonary metastasis ranging from 2.5 to 20 years. 9 The differential diagnosis with pulmonary metastases includes benign metastasizing leiomyoma, carcinoid tumors, sclerosing hemangioma, and metastasis of other neoplasms. 27 Recurrences in the gastrointestinal tract are less frequent, with ESS metastatic to the colon being infrequently reported.13,33-36 On imaging, a variety of submucosal tumors may be on the differential. Most of the mesenchymal neoplasms (leiomyoma, fibroma, and schwannoma) can be excluded based on histology. 35 However, a gastrointestinal stromal tumor (GIST) may be confused with low-grade ESS. The presence of short fascicles or sheets of monotonous plump spindle cells, prominent arterioles, and perivascular whorl arrangement of the tumor cells should argue against the diagnosis of GIST.34,35 Moreover, the IHC profile would aid in the distinction between GIST and ESS with the former usually characterized by strong positive expression of CD117 (C-Kit) and/or DOG-1.
When trying to differentiate metastatic ESS from primary extrauterine endometrial stromal sarcoma (EESS), a uterine primary needs to be excluded. A primary uterine tumor can only be excluded if a hysterectomy is done concurrently. Similarly, if the hysterectomy occurred previously, the slides can be reviewed if available. 37 If there is a uterine tumor and endometriosis, there is a possibility of synchronous tumors since EESS often associated with foci of endometriosis.37,38 The largest series of primary EESS to date showed that the tumor was associated with endometriosis in 30 of 50 cases, suggesting an origin from ectopic endometrial stroma. In the absence of a uterine primary, and when no endometriosis is identified, the tumor may have originated from unrecognized stromal endometriosis or overgrowth of the tumor may have obscured the underlying endometriosis. 37 Alternatively, it has also been postulated that EESS could arise de novo from coelomic or subcoelomic multipotential epithelium/stem cells.37,39
Due to low numbers, evidence-based guidelines for treatment are not available currently, 40 and the treatment modality for recurrent or distant metastatic ESS remains individualized on a case to case basis.10,41 It has been suggested that the indolent growth potential of this tumor makes cytoreductive surgery, including repeat surgeries, appropriate for late recurrent disease. 2 The value of adjuvant therapy is controversial, but nevertheless considered. 40 In one series of metastatic pulmonary ESS, treatment modalities included surgery, selective embolization, chemotherapy, hormonal therapy, and radiation. 10 Other reports have also advocated using progestin, alone or in combination with GnRH agonists as adjuvant therapy and for treatment of recurrent disease.41-43 However, the optimal dose, regimen, and duration of hormonal treatment are not well established. 41
Although it is recognized that high-grade ESS have worse outcomes, low-grade ESS is clearly capable of recurring and metastasizing to distant sites. Of our 10 cases, 8 were low-grade. As there are no strict histological criteria for calling low-grade versus high-grade, molecular confirmation will aid in consistent classification. In addition, characteristic molecular alterations will provide increased confidence in the diagnosis of ESS in diagnostically challenging cases. It is difficult to draw any conclusions regarding outcome as most of our patients have survived for many years or were relatively recently diagnosed. However, regardless of the grade, patients remain at risk for future recurrences.
Conclusion
ESS is usually an indolent neoplasm with favorable outcomes despite its ability to recur and potentially metastasize to distant sites. Since patients usually do well with a long disease-free interval, recurrences that occur predominantly extrapelvic in the lung, liver, bone, or brain many years after the initial presentation are usually a diagnostic challenge. Although pulmonary metastases are most common, other sites including the liver are possible, and must be kept in mind. Correctly identifying distant metastatic ESS can be challenging, especially when the original clinical history may be misleading or absent. Even with classic features, metastatic ESS in an uncommon location can be diagnostically challenging. A combination of detailed history, imaging, histology, and IHC will assist in the differential diagnosis of endometriosis-related ESS versus metastatic ESS. Furthermore, characteristic molecular alterations, if accessible, can facilitate accurate identification. Due to their rarity, a central registry for complete documentation may facilitate the study of larger cohorts of these uncommon tumors that can have unusual and unpredictable biological behavior.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
Not applicable, because this article does not contain any studies with human or animal subjects.
Informed Consent
Not applicable, because this article does not contain any studies with human or animal subjects.
Trial Registration
Not applicable, because this article does not contain any clinical trials.
