Günter Köhler, Katja Evert, Marek Zygmunt and Matthias Evert
3Endometrial stromal tumors – endometrial stromal nodule, endometrial stromal tumor with sex cord-like elements (ESTSCLE), uterine tumor resembling ovarian sex-cord tumor (UTROSCT) and similar tumors
The current WHO Classification (36) divides EST and similar neoplasms into ESN, LG-ESS, HG-ESS and UUS. UTROSCT also belong to the EST family of tumors. ESN are benign stromal tumors. There are special forms of EST with different types of differentiation, like the endometrial stromal tumor with endometrioid glands, mixed endometrial stromal tumors and smooth muscle tumors as well as endometrial stromal tumors with sex cord-like elements (ESTSCLE), and uterine tumors resembling ovarian sex-cord tumor (UTROSCT). Uterine stromal sarcomas (LG-ESS, HG-ESS, UUS) belong to the family of pure homologous sarcomas and are discussed in separate chapters (Chapters 4 and 5).
General, pathogenesis, pathological-anatomical features
ESN is a benign ESTwith histological features unambiguously reminiscent of endometrial stroma in the proliferation phase. Overall, ESN are only very rarely diagnosed, accounting for only 3% of all EST in the DKSM database (25). ESN have been observed to arise in women aged 23–86. The majority of patients are premenopausal, with a median and mean age of 47 resp. 53 years. It can already be deduced from this age distribution that ESN share an epidemiologic relationship with LG- ESS,which, according to the DKSM data, arise in patients aged 18 to 80 years. Median age of patients with LG-ESS is 46 years, suggesting that, like ESN, the majority of such tumors arise in premenopausal women (25). These two types of neoplasm in fact share clear pathogenetic commonalities. On the one hand, the differences between the two in terms of their microscopic features are more gradual than abrupt. Both LG-ESS and ESN exhibit t(7;17) translocations and consecutive fusions of the JAZF1 and SUZ12 (previously: JJAZ1) genes as well as t(6;7) translocations (JAZF1/SUZ12 gene fusion) (26). The t(7;17) (p15;q21) translocation is regarded as an early genetic aberration in the development of ESN and LG-ESS. Interestingly, this translocation can also be observed in ESN that exhibit focal smooth muscle differentiation, which allows the conclusion that such tumors originate from a progenitor cell (35), and could also serve to explain why ESN and LG-ESS so frequently arise within the myometrium. LG-ESS, ESN and CLM express common immunohistochemical markers (CD10, SMA,WT-1), another factor that supports the notion that stromal and myogenic tumors might derive from the same precursor cells (48). What is rather conspicuous from an epidemiologic perspective is that ESN account for only a very small proportion of all EST. Subtle analyses have revealed that,when ESN and LG-ESS are summed together, the latter account for a substantially larger share (84%) than the former (16%) (45). Given the fact that microscopically differentiating between LG-ESS and ESN is very difficult, this can essentially only be explained by the fact that a non-negligible number of ESN (for example such with limited infiltration or myogenic differentiation) are probably classified or diagnosed as LG-ESS just to be “on the safe side”. That could also help to explain why there is so much variation in terms of the propensity of such tumors to recur. The situation might actually be very similar to that of STUMP and LMS.
Endometrial stromal nodules are rare benign endometrial stromal tumors that predominantly arise in premenopausal women. They are regarded as a precursor of low-grade endometrial stromal sarcoma.
Similar to LM, ESN are usually situated intramurally. Their intramural growth serves to visibly compress the surrounding myometrium, and potentially the endometrium as well. The tumor is thus well-circumscribed. ESN have a yellowish, tan, occasionally almost white color on the predominantly smooth, fleshy cut surface. Sometimes cystic structures, necrosis and focal hemorrhages can be visible in the cut-open specimen. Nodular or polypoid growth into the cavum uteri is not uncommon. ESN are usually considerably softer than LM. Their size can range from only a few mm to 22 cm, with a mean diameter of 5–7 cm. Multiple ESN can develop simultaneously.
Microscopy also very frequently reveals that ESN are situated entirely within the myometrium. Myometrial compression caused by the tumor’s strong displacing growth is usually easy to see, and can appear to resemble an actual capsule in some cases. The densely packed, diffusely growing small oval to spindle cells strongly resemble the endometrial stromal cells of the proliferative endometrium. The cells contain scant cytoplasm and usually have unclear margins. The oval to spindle, relatively uniform nuclei are practically void of pleomorphism, have unsuspicious nucleoli and exhibit a prominent coloration that often makes them appear deep blue in histological specimens.
The cells often focally whorl around the vessels (1). Virtually all ESN contain ample uniform vessels (15) reminiscent of endometrial spiral arterioles. The vessels can occasionally be thick-walled and hyalinized, but are usually situated within areas of smooth muscle differentiation in such cases.
Densely cellular tumors are predominant (roughly 80%), while 20% are hypocellular with fibrous, hyalinized, myxoid or edematous sections (15). ESN with an abundance of epithelioid cells have also been described in the literature (33). Occasional areas consisting of foam cells within the tumor are another noticeable feature (31). There are no entirely cytologic criteria that can be drawn on to differentiate ESN from LG-ESS. Hyalinized areas, cystic degeneration and necroses occur in both of these types of tumors, and are thus not suitable specific DD criteria (1).
In microscopy, the tumor is usually sharply delineated from the myometrium due to the absence of infiltration (Fig. 3.1.1 (A)).
Fig. 3.1.1: Histological findings of an endometrial stromal nodule and diagnostic differentiation from cellular leiomyoma; (A) a clear, sharp margin between the stromal nodule (top left) and the myometrium (bottom right) constitutes an important characteristic feature, the stromal cells with scant cytoplasm morphologically correspond to those of low-grade endometrial stromal sarcoma; (B) cellular leiomyoma can have a very similar appearance and needs to be ruled out via immunohistochemical testing.
The tumor margins can also be slightly irregular or poorly defined as a result of tongue- or finger-like protrusions of the main tumor up to 3mm deep into the bordering myometrium. Even the presence of small individual nodules located no further than 3mm away from the tumor does not serve to exclude such cases from being diagnosed as ESN. Tumors in such cases are referred to as ESN or EST with limited infiltration (15, 49, 51) or LG-ESS with limited infiltration (1). VI and LVI are absent by definition. Tumors in which the infiltrations or satellite masses extend beyond 3mm or in which there is VI will be LG-ESS. In some cases, tumors with focal margin irregularities that extend up to 9mm beyond the main border of the tumor still do not have to be classified as LG-ESS (15). Such infiltrations usually only occur focally and should be limited to three localizations in order to fulfill the ESN criteria (1). These criteria must indeed appear vague even to non-pathologists, and constitute a substantial reason why ESN and LG-ESS share such overlap in a diagnostic “gray area”.
Smooth muscle differentiations are far from the exception in ESN, and are present in up to 50% of cases (15). At times, the smooth muscle cells can also appear to resemble epithelioid cells. Tumors in such cases are referred to as ESN with smooth muscle differentiation, regardless of the extent of the smooth muscle component (32). In the past, tumors in which the smooth muscle tissue accounted for > 30% were termed stromal myoma. Today, these tumors constitute a special subgroup of EST called mixed endometrial stromal smooth muscle tumors (see below). Larger sections of smooth muscle differentiation can be “interlocked” with stromal cells in a cog- or phalanx-like fashion, a situation that can be misinterpreted as myometrial infiltration by stromal cells and thus result in a misdiagnosis as LG-ESS. Larger areas of smooth muscle differentiation can thus cause differential diagnostic problems, especially in curettage specimens (see below). Besides fibrous elements, 24% of tumors also contain focal visible degrees of sex cord-like differentiation arranged in cords/bundles in a glandular epithelioid structure (15). These tumors are further subdivided into two subtypes along the lines of the extent of sex cord-like differentiation in the tumor – ESTSCLE and UTROSCT (in which sex cord-like differentiation is predominant) (see below).
The majority of ESN exhibit no mitotic activity or MI < 3M/10HPF. MI of up to 15 M/10 HPF have been described in the literature (49), but only 10% of ESN have > 5 M/10 HPF (15). MI is not relevant for the diagnostic context and has no impact on prognosis. Cysts resulting from cystic degeneration as well as infarct-type necroses are observed in 24 and 68% of ESN, respectively (15, 17).
Regarding IHC, ESN virtually always express ER and PGR, vimentin and CD10. Pure ESN without further differentiation are usually negative for desmin and hcaldesmon (34). The corresponding lead markers can be found in cases in which tumors exhibit additional morphologic differentiations, for instance: desmin and h-caldesmon when there is smooth muscle differentiation; inhibin, calretinin, CD99 and O-13 when there is sex cord-like differentiation; myogenin when there is skeletal muscle differentiation. CD10 constitutes the most reliable and most characteristic marker for endometrial stromal cells, and thus for ESN and LG-ESS. CD10 is typically diffusely expressed and strongly positive in both ESN and LG-ESS (see also the table on p. XXVI). In ESN, areas with smooth muscle differentiation also stain with CD10. ESN usually show strong and diffuse positive immunoreactive staining for the WT-1 antibody (48).
Endometrial stromal nodules can barely be macroscopically discerned from leiomyoma or lowgrade endometrial stromal sarcoma, also in terms of their localization. Endometrial stromal nodules are noticeably soft and of a conspicuously yellow color.
The only real histological difference between endometrial stromal nodules and low-grade endometrial stromal sarcomas is the absence of infiltration in cases of the former. It consists of very densely packed small cells with low levels of pleomorphism that are reminiscent of proliferative endometrial stroma. Endometrial stromal nodules typically express CD10. Smooth muscle and sex cord-like differentiations with corresponding immunophenomena are possible.
Like LG-ESS, ESN are often incidental findings made on the basis of specimens obtained via HE. Uterus enlargement, menorrhagia and possibly AUB in the form of additional bleeding are the central clinical symptoms. Patients occasionally also complain of abdominal pains when there is uterine enlargement. The noticeable softness of such tumors often becomes apparent either during gynecologic examination or intraoperatively. Overall, at the latest during surgery, ESN or LG-ESS should be considered as possible diagnoses when a tumor is soft and appears very smooth and yellowish on the cut surface. In principle, symptoms largely correspond to those of LM and can thus also be entirely absent. Accordingly, and analogous to LG-ESS, they are usually operated on under the diagnosis LM.
Diagnostics primarily involve sonography, and also include HSC in combination with targeted biopsy or curettage when there is AUB. The problems pertaining to curettage are discussed at greater length in the chapter on LG-ESS (Chapter 4). The aforementioned diagnostic steps are often not taken when LM is assumed. Similar to smooth muscle tumors and LG-ESS, the hysteroscopic picture usually only reveals a smooth bulge in the endometrium caused by a displacing tumor (Fig. 3.1.2).
Fig. 3.1.2: Hysteroscopic findings of an endometrial stromal nodule, differentiation from leiomyoma is barely possible on gross observation.
An irregular surface or polypoid structures are also possible features. The cut surface usually has a conspicuous yellowish color. The tumor appears softer than LM in biopsy. The diagnostic flowchart can be drawn on in ambiguous or uncertain cases (cf. Chapter 6).
Symptoms, findings and clinical diagnostics all largely correspond to those of leiomyoma or lowgrade endometrial stromal sarcoma. Hysteroscopy and curettage are indicated when there is abnormal uterine bleeding.
Given the rarity of ESN, only very few data are available that can give insight into their imaging characteristics. Sonography does not yield findings that are specific to ESN, but typically reveals a well-circumscribed heterogeneous tumor that is largely hypoechoic (38). While the presence of anechoic structures (cysts or necroses) is not a finding that is specific to ESN, they can nonetheless be suggestive of ESN when the tumor in which they are situated is hypoechoic. Large ESN also appear as wellcircumscribed heterogeneous hypoechoic masses with cystic/necrotic and solid parts (17) (Fig. 3.1.3 (A), (B)).
Fig. 3.1.3: (A) endometrial stromal nodules with predominantly cystic structures in sonography; (B) the tumor also appeared to be cystic in uterus-sparing surgery, the findings were later pathologico-anatomically characterized as pseudocysts within the tumor.
Since the only microscopic and macroscopic difference between ESN and LG-ESS is the absence of myometrial and VI in the former, it cannot be realistically expected that these two types of neoplasm will exhibit greatly differing imaging-related features. Assessing the margins of the tumor with a view to identifying infiltrations can, however, be helpful.
Cases of ESN that are isoechoic to the myometrium or to LM have been reported (8). Accordingly, as is also the case with LG-ESS, the majority of ESN are classified as LM in sonography. Findings similar to those for ESN and LG-ESS have also been described for LMS and DLM. Overall, the sonographic picture merely serves to suggestively indicate the presence of an unusual mesenchymal mass.
CT reveals unspecific tumor enlargement with occasionally discernible cysts. The tumor shows heterogeneous enhancement in CECT that is absent in the cystic sections. CT is not an appropriate tool for reaching a specific diagnosis (41).
T1W-MRI usually reveals a tumor with SI that is isointense to the myometrium (38). Administration of gadolinium shows homogeneous to inhomogeneous enhancement that can be lower than or identical to that of normal myometrium. Cystic structures remain hypointense (41). In one case (38), only the areas on the margins of the tumor showed enhancement, while the rest of the tumor did not. The non-enhancing areas apparently corresponded to hemorrhagic degeneration. T2W reveals a heterogeneous tumor that is intermediately intense to hyperintense. SI is slightly higher than that of the myometrium and skeletal musculature (18). Hypointense bands within the tumor, a typical feature of LG-ESS (cf. Chapter 4), are absent in T2W (38). These features serve to distinguish ESN from LM rather well. Cystic structures that appear hypointense in T1W are hyperintense in T2W. Analogous to LG-ESS that exhibit worm- or tongue-like invasion, T2W should also reveal hypointense bands in ESN with more pronounced smooth muscle differentiation. LG-ESS with little or no certain myometrial invasion are typically separated from the myometrium by a narrow hypointense rim (cf. Chapter 4). This rim might well serve as a criterion for differentiation from DLM and LG-ESS with diffuse infiltration (18).
It is generally assumed that ESN cannot be reliably differentiated from LG-ESS with limited or little infiltration in MRI (18).
Endometrial stromal nodules have no specific sonographic criteria or features. They typically present as well-circumscribed masses with heterogeneous hypoechoicity, sometimes with cystic structures. In T2W-MRI endometrial stromal nodules are heterogeneously intermediate to hyperintense. Signal intensity is slightly higher than that of normal myometrium, and tumors are usually surrounded by a hypointense rim. These characteristics and features are well suited for differentiating endometrial stromal nodules from leiomyoma. Differentiation from low-grade endometrial stromal sarcoma is often not possible.
From a clinical perspective, ESN and LG-ESS are often mistaken for LM, both preoperatively and intraoperatively. In fact, the majority of both of these types of neoplasm are subjected to surgery under the indication LM, and such tumors are thus often incidental findings (cf. Chapter 4). Intramurally localized ESN are only exceptionally accessible to curettage, as is also the case with other intramural mesenchymal tumors. ESN are, however, usually softer than LM on palpation and are usually solitary. Cut-open ESN have a yellowish-browncolor, a smooth cut surface and lack a clear pseudo-capsule,making them relatively easy to discern macroscopically from LM with their more whitish color and whorled structures. DD from CLM, which are also rather soft (Fig. 3.1.1 (A), (B)), is particularly challenging from a clinical and pathological- anatomical perspective. Discerning between CLM and ESN is practically impossible using purely clinical methods. CLM and STUMP can be similar to ESN in terms of color and consistency, and can also contain cystic structures and necroses. Polypoid protrusion or “bulging” into the cavum militates against a diagnosis of LM, CLM or STUMP.
LG-ESS constitutes the most important DD in terms of microscopic and macroscopic features. ESN cannot be discerned from LG-ESS solely on the basis of cytology. A tumor can only be definitively and reliably diagnosed as ESN when the entirety of the tumor’s margins is observable/assessable. A reliable diagnosis of ESN cannot be reached when the tumor borders are incomplete following curettage, morcellation or enucleation, but the clinician might nonetheless have to settle for that diagnosis, even though LG-ESS cannot be ruled out. Differentiating ESN with limited infiltration from LG-ESS is very difficult even when the margins of the tumor are all clearly assessable, but also has prognostic consequences. The same applies for intraoperative frozen sections, whose value is already limited as it is (see below).
In histological tissue specimens, the uniformity of the cells, dense cellularity, strong staining in the nucleoli and the fact that they are typically well-delineated from the myometrium are all factors collectively suggestive of CLM. The latter, however, are characterized by relatively large, thick-walled vessels and narrow clefts in the tumor. The clefts possibly correspond to compressed vessels (34). Thick-walled arterioles are occasionally also found in ESN, but are usually situated at the periphery/near the margins of the tumor in such cases (15). Performing frozen sections without IHC analysis (Fig. 3.1.1 (B)) is thus greatly prone to error for this group of neoplasms, and must be regarded critically by clinicians. In contrast to CLM, ESN exhibit strong and diffuse CD10 expression. With the exception of ESN with smooth muscle differentiation, and in contrast to CLM, ESN do not stain positively for desmin. H-caldesmon appears to be the most important differential diagnostic marker as it virtually never shows staining in ESN (48). Since CD10 can also be traceable in 20–55% of CLM, the latter can best be discerned from ESN by drawing on a combination of desmin, h-caldesmon and CD10 (cf. Tab. 4.1.1) (10, 34). However, it should be noted that LMS can also be CD10 positive and h-caldesmon negative (4, 42). Assessing CD10 is generally not enough for safely and reliably distinguishing between ESN and smooth muscle neoplasms.
Extensive smooth muscle differentiations with interposed stromal elements can serve to mimic invasion and subsequently result in overdiagnosis as LG-ESS. Related problems tend to arise in cases in which extensive morcellation has been performed leaving many small pieces of tissue. ESN with epithelioid cells with ample eosinophilic cytoplasm (33) can be confused with PEComa in microscopy. Effort must be devoted to differentiating ESN from ESTSCLE (see below) and UTROSCT (see below) when sex cord-like elements are found to be present.
Stromal endometriosis constitutes a further differential diagnostic challenge, a nomenclature used to describe endometriosis with sparse or no glandular differentiation, sometimes also referred to as stromatosis. Stromal endometriosis is sometimes observed to arise after long-lasting or highly-dosed progestin therapy (29). Stromal endometriosis usually arises on the surface of the peritoneum in the vicinity of other areas of endometriosis and does not develop actual tumors as such (7). Cyclic changes or older hemorrhages with hemosiderin can be recognizable in the stromal endometriosis. Stromal endometriosis occasionally arises in the cervix (12), where it can also be confused with ESN or LG-ESS. An AS can be overlooked in specimens obtained via biopsy or curettage when no glands are visible in the small amount of tissue available.
ESN with focal endometrioid glandular differentiation are usually endometrial stromal tumors with endometrioid glands. There is also an LG-ESS equivalent to such neoplasms. AS must also be considered from a DD perspective whenever glands are present. However, cuff-like accumulations of stromal cells around the vessels make AS generally easy to recognize as such (cf. Chapter 6, Vol. 2).
There is no particular screening that we are aware of. Nor is such a screening necessary due to the rarity of such tumors and the fact that they are benign. The low rate of cellular atypia also serves to rule out incidental detection of ESN in the context of cytological screenings for cervical carcinoma.
Low-grade endometrial stromal sarcoma and common leiomyoma are the most important clinical differential diagnoses. Microscopically differentiating endometrial stromal nodules from cellular leiomyoma can be very difficult without applying additional immunohistochemical methods, especially in frozen sections. The same applies for rarer differential diagnoses, like PEComa, ESTSCLE and UTROSCT.
There is no specific screening for endometrial stromal nodules, nor is such screening necessary given the rarity and benignity of such tumors.
Course, prognosis, primary surgery, adjuvant and additive radio-, chemo-and hormone therapy
ESN are benign tumors that do not recur or metastasize (49). Accordingly, prognosis is good so long as the tumor margins are well-circumscribed and open to assessment. Such assessment is largely based on tumors that have been removed via THE. No reliable prognostic assessments can be made concerning the application of conservative surgery to ESN since case numbers are so low. Among the few cases that have been recounted in the literature, none have recurred (15, 43, 49). From a pathological-anatomical perspective, it is assumed that limited myometrial invasion of up to 3mm depth is unlikely to impact unfavorably on prognosis. This notwithstanding, a case is recounted in the literature in which a uterine ESN with limited invasion exhibited extrauterine dissemination, though the lesion in question was probably in fact very early LG-ESS (24). A case has been reported in which an ESN with more than three permeations not exceeding 3mm metastasized (47). Due to a lack of data, the prognostic impact of the presence of tumor margin irregularities of 3 to 10mmdepth remains uncertain (15, 52). For practical purposes, it is therefore recommended that such tumors also be classified as LG-ESS, and that it is pointed out independently that the degree of infiltration at hand is far lower than is usually observed in LG-ESS (1). In cases in which tumors are subjected to morcellation or RX resection, or in which the tumor borders cannot be assessed in their entirety, pathologists must unequivocally point out that the tumor in question may also be LG-ESS.
Endometrial stromal nodules are benign tumors that neither recur nor metastasize. This statement only applies when all tumor margins are reliably observable or assessable.