Parasitism as a potential contributor to massive clam mortality at the Blake Ridge Diapir methane-hydrate seep

Transcription

1 See discussions, stats, and author profiles for this publication at: Parasitism as a potential contributor to massive clam mortality at the Blake Ridge Diapir methane-hydrate seep Article in Journal of the Marine Biological Association of the UK December 2005 DOI: /S CITATIONS 8 READS 98 4 authors, including: Anne M Mills University of Virginia 71 PUBLICATIONS 458 CITATIONS SEE PROFILE Cindy Lee Van Dover Duke University 195 PUBLICATIONS 6,677 CITATIONS SEE PROFILE Some of the authors of this publication are also working on these related projects: Deep-Sea Mining View project Managing Impacts of Deep-seA resource exploitation (MIDAS) View project All content following this page was uploaded by Anne M Mills on 01 June The user has requested enhancement of the downloaded file.

2 Seminars in Diagnostic Pathology (2010) 27, Endometrial hyperplasia Anne M. Mills, MD, Teri A. Longacre, MD From the Department of Pathology, Stanford University, Stanford, California. KEYWORDS Endometrial hyperplasia; Atypical endometrial hyperplasia; Endometrial intraepithelial neoplasia Endometrial hyperplasia is a heterogeneous set of pathologic lesions that range from mild, reversible glandular proliferations to direct cancer precursors. These lesions comprise a continuum of morphologic appearances, with the earliest proliferation represented by crowded glands with simple tubular architecture lined by cells resembling proliferative endometrium, whereas advanced proliferations in this continuum are characterized by crowded glands with complex architecture, often containing cells with nuclear atypia resembling low-grade endometrioid adenocarcinoma. The former early proliferations may be isolated to an endometrial polyp, but advanced proliferations are generally more diffusely present throughout the endometrium. There are at least three major classification systems for endometrial carcinoma precursor lesions, each of which trend toward overlap at the complex end of the spectrum. Although some classifications are based on a series of molecular genetic alterations (which may or may not translate into biologically or clinically relevant risk lesions), each classification scheme ultimately uses a series of histologic features, usually a combination of architecture and cytology, to establish a diagnosis of hyperplasia. Because different pathologists may apply different histologic criteria for endometrial hyperplasia depending on the classification system used, this article will provide an overview of the classifications used in current daily practice, present the histologic criteria and relative merits of each classification system, and discuss common and not so common causes of misclassification Elsevier Inc. All rights reserved. Introduction Endometrial hyperplasia typically occurs in response to estrogen stimulation unchecked by progesterone. Given its association with hormonal imbalance, it is not surprising that endometrial hyperplasia is most often encountered around or following menopause, although premenopausal women with sclerocystic ovaries or obesity are also at increased risk. In the perimenopausal setting, estrogen excess may be caused by either exogenous hormone replacement therapy or endogenous factors, such as anovulation. Obesity contributes to estrogen excess through conversion Address reprint requests and correspondence: Teri A. Longacre, MD, Stanford University School of Medicine, Room L235, 300 Pasteur Dr, Stanford, CA address: longacre@stanford.edu. of androgens into estrogen within adipose tissue. Tamoxifen therapy may also contribute to increased risk of hyperplasia. Most women with endometrial hyperplasia present with abnormal uterine bleeding. Although endometrial hyperplasia represents a risk factor for the development of adenocarcinoma, the degree of this risk is subject to controversy because it varies considerably across the various categories of hyperplasia. Treatment is informed both by morphologic classification and clinical scenario. In postmenopausal women, hysterectomy is often appropriate even for low-risk lesions, whereas in younger women who wish to maintain fertility, less invasive therapies are often initially employed, even when the diagnostic considerations also include low-grade endometrioid adenocarcinoma. It is important to emphasize that endometrial hyperplasia is not the precursor for all histologic subtypes of endometrial carcinoma; high-grade /$ -see front matter 2010 Elsevier Inc. All rights reserved. doi: /j.semdp

3 200 Seminars in Diagnostic Pathology, Vol 27, No 4, November 2010 Figure 1 Endometrial hyperplasia. There is gland crowding with a resultant gland-to-stroma ratio of 3:1. This feature is best detected on low magnification. serous carcinoma appears to be largely estrogen independent. Gross features Endometrial hyperplasia typically translates as a thickened endometrial stripe on transvaginal ultrasound and increased volume of endometrial tissue on hysteroscopy or curettage, although it may not be associated with any gross abnormalities. Although hyperplasia may be associated with abundant endometrial tissue, the amount is typically no more impressive than that observed in normal secretory-phase endometrium with the appearance being that of diffuse, somewhat polypoid, tan, velvety tissue. In some instances, localized hyperplasia may mimic a polyp, arise in a background of a polyp, or involve adenomyosis, including deep foci (the outer half of the myometrium). glandular hyperplasia), the 3:1 threshold decreases the likelihood of overdiagnosis of hyperplasia, with minimal if any risk of underdiagnosing a lesion that is at risk for progression. We do not include disordered proliferative endometrium in the classification of endometrial hyperplasia because there is no evidence that it poses a risk for progression to carcinoma. 2 Luminal spaces and villoglandular structures are both included in the glandular component for this calculation. If the gland-to-stroma ratio fails to meet the required cut-off, a lesion is probably best classified as disordered proliferative endometrium rather than hyperplasia. In most instances, the constituent cells are larger than those of normal endometrium and exhibit proliferative features, including nuclear pseudostratification and increased mitotic rate (Figure 2); in the absence of the latter features, a metaplastic endometrial process should be considered. Although hyperplastic endometrium might have cytologic atypia, this criterion is neither required nor sufficient for this diagnosis. The assessment of cytologic atypia is poorly reproducible, but the various criteria include nuclear rounding and enlargement, nuclear pleomorphism, loss of polarity, increased nuclear-to-cytoplasm ratio, prominent nucleoli, irregular nuclear borders, vesicular chromatin, and clumped chromatin. Atypical cells may show tufting and focal stratification, although extensive nuclear stratification merits a diagnosis of malignancy. Hyperplastic endometria with cytologic atypia pose a more significant risk for endometrial adenocarcinoma than hyperplastic endometria without atypia. 3,4 Although the morphologic features of hyperplasia usually occur in concert with increased overall volume when compared with the background endometrial tissue, in rare instances the architectural criteria for hyperplasia are satis- Microscopic features Endometrial hyperplasia is morphologically defined as proliferating endometrium with architectural abnormalities. These architectural changes range from cystic dilatation to more complex glandular configurations, including budding and branching, papillary infoldings, villous and villoglandular growths, and cribriform structures. In addition to abnormal architecture, a diagnosis of hyperplasia usually requires increased glandular density with a gland-to-stroma ratio of 3:1 (Figure 1). 1 Some authors have proposed a lower gland-to-stroma ratio ( 2:1), but because a variety of nonhyperplastic processes might exhibit a gland-to-stroma ratio that can approach this level (eg, focal artifactual compression of glands in curettage specimens, benign endometrial polyps, disordered proliferative endometrium, and occasionally pronounced morular metaplasia in the absence of Figure 2 Endometrial hyperplasia. Cells are enlarged (relative to the uninvolved normal endometrium, if present) with enlarged nuclei and nuclear pseudostratification, and mitotic figures are typically present (although not depicted in this image). Benign mimics (eg, florid secretory or gestational endometrium) can be excluded by evaluating suspected hyperplasic proliferations on higher magnification.

4 Mills and Longacre Endometrial Hyperplasia 201 Figure 3 Squamous metaplasia. (A) A spindled morular proliferation within the lumen of the endometrial glands is the more common type of squamous metaplasia and can be seen (rarely) in atrophic endometrium, disordered proliferative endometrium, endometrial hyperplasia, and carcinoma. It can be confused with squamous proliferations of the uterine cervix. (B) P16 positivity is seen. This finding may further pose differential diagnostic problems with cervical proliferations. fied in the absence of increased volume. Isolated foci of hyperplasia may also be identified in a background of otherwise unremarkable endometrium ( focal hyperplasia ); their significance is uncertain. Hyperplasic endometria are often associated with a variety of metaplasias. The latter signify the presence of epithelial cell differentiation not ordinarily identified in endometrial cells or only in a minority of them (eg, ciliated cells). Squamous metaplasia (Figure 3) and ciliated cell change (Figure 4) are the most common, although eosinophilic (Figure 5), mucinous (Figure 6), papillary (Figure 7), and clear cell or hobnail cell changes also occur. Mixed patterns of metaplasia are common (Figure 8). Squamous metaplasia can be extensive, filling and expanding hyperplastic glands (Figure 9), further architecturally complicating the crowded appearance of the lesion. Importantly, the presence and extent of these metaplasias have no bearing on prognosis and should be ignored when assessing for associated carcinoma. Other histologic features frequently seen in association with hyperplasia include ectatic blood vessels and stromal breakdown. World Health Organization classification system The most commonly used classification system for endometrial hyperplasia was outlined by the World Health Organization (WHO) in 1994 (Table 1). The classification criteria used in this system derived from the 1985 study by Kurman and colleagues, which correlated morphologic features with clinical outcome. 4 This study showed a 23% risk of progression to carcinoma for atypical endometrial hyperplasia but only 2% for nonatypical hyperplasia, informing the rationale for separating atypical and nonatypical hyperplasia. In this system, hyperplasias are further subdivided into simple and complex based on architectural features, although the subdivision has a less dramatic bearing on Figure 4 Ciliated cell metaplasia. Ciliated cells are present in normal endometrium, but increase under estrogenic stimulation. The terms ciliated cell change or ciliated cell prominence have been proposed because their presence does not strictly qualify as a metaplastic process. Figure 5 Eosinophilic cell metaplasia. Note slightly enlarged nuclei and small nucleoli.

5 202 Seminars in Diagnostic Pathology, Vol 27, No 4, November 2010 Figure 6 (A) Mucinous metaplasia closely simulates endocervical mucinous epithelial cells and can be mistaken for an endocervical process (and vice versa). (B) The cells have tall cytoplasm and compressed basally located nuclei. A key that the lesion may be endometrial is the presence of other types of metaplasia, typically eosinophilic cell metaplasia. prognosis than the presence or absence of cytologic atypia. The latter has been linked to a 20-30% progression rate to carcinoma in subsequent reports. 5 The four categories of endometrial hyperplasia recognized by the WHO classification schema are (1) simple hyperplasia without atypia, (2) complex hyperplasia without atypia, (3) simple atypical hyperplasia, and (4) complex atypical hyperplasia. 1. Simple hyperplasia without atypia Simple hyperplasia is characterized by densely packed, cystically dilated, variable sized glands separated by normal intervening stroma (Figure 10). Although the glandular patterns in simple hyperplasia may be similar to what is observed in disordered proliferative endometria, the later fails to meet the 3:1 gland-to-stroma ratio. Ciliated cells are often quite prominent and squamous morular metaplasia might occur. Simple hyperplasia can be distinguished from cystic atrophy and normally occurring patches of inactive glands based on cytologic features. Hyperplastic glands are composed of columnar cells with pseudostratification and include mitotic figures, whereas atrophic/inactive glands bear a single layer of mitotically inactive flattened or cuboidal epithelium. 2. Complex hyperplasia without atypia Complex hyperplasia without atypia is defined as glands with abnormal, irregular architecture set in a background of scant intervening stroma that is generally less prominent than in simple hyperplasia. Some stroma must be present, however, and even in cases where glands are apparently back to back, close examination reveals basement membrane lining individual glands and a rim of intervening endometrial-type stroma between them. In addition to backto-back and cribriform-like arrangements, other glandular architectural abnormalities warranting designation of complex hyperplasia include outpouchings, infoldings, and budding (Figure 11). Squamous or morular metaplasia as well as eosinophilic and ciliated cell changes are common. Com- Figure 7 Prominent papillary change. The cytologic features are bland and the cells often have abundant eosinophilic cytoplasm. Figure 8 Mixed patterns of metaplasia. In this example, ciliated cell and eosinophilic cell changes are present within glands that exhibit complex architecture, but there is no overall increase of the gland-to-stroma ratio.

6 Mills and Longacre Endometrial Hyperplasia 203 Figure 9 Extensive morular metaplasia. The squamous cells fill the glands to the extent that their outer contours are essentially obliterated, imparting the appearance of infiltration into the adjacent stroma. plex hyperplasia may be seen next to areas showing simple hyperplasia. Cytologically, the epithelial cells are identical to those seen in simple hyperplasia and are characterized by pseudostratification (one to four layers of cells including focal tufting), smooth oval nuclei with evenly dispersed chromatin, inconspicuous nucleoli, and a variable number of mitoses (Figure 12). These features are reminiscent of those seen in normal proliferative endometrial cells except for the nuclei that are often larger in hyperplasia. 3. Simple atypical hyperplasia The finding of endometrial glands showing cytologic atypia in the absence of architectural complexity is rare. Criteria for nuclear atypia in this setting mirror those required for complex atypical hyperplasia and are described below. 4. Complex atypical hyperplasia Most cases of atypical endometrial hyperplasia fall into this category, which is differentiated from complex nonatypical hyperplasia solely on the basis of cytologic abnormalities (Figure 13). Unlike the smooth, oval nuclei seen in nonatypical hyperplasia, atypical nuclei are rounder and may have irregular membranes. Chromatin is often unevenly dispersed and clumpy, imparting a vesicular appearance, and enlarged nucleoli are seen (perhaps the most reproducible criterion for atypical endometrial hyperplasia). True stratification with loss of polarity in relation to the basement membrane is common and differs from the pseudostratification seen with nonatypical hyperplasia. Apoptotic bodies may be seen. Neither the degree nor the extent of atypia needed to establish the diagnosis of atypical hyperplasia is well specified, but in general, atypia should involve a significant proportion of the hyperplastic proliferation rather than an isolated focus to warrant this diagnosis (Figure 14). However, occasional curettage specimens may contain glandular tissue with marked complexity, but only focal cytologic atypia. In these instances, we classify the architecture (eg, complex hyperplasia, borderline, or carcinoma see below) and note the degree and extent of atypia. The surrounding stroma is typically compressed, but may appear fibrous or exhibit features of stromal breakdown (the latter may artifactually increase the apparent gland-to-stroma ratio). Stromal foam cells may also be seen and help to localize the proliferation to the endometrium (vs cervix) in cases in which the morphologic features are ambiguous. 1 Although the criteria appear straightforward, the application of this system can be difficult in clinical practice. The determination of cytologic atypia is particularly troublesome and accounts for a considerable lack of reproducibility in the classification of endometrial hyperplasia (interobserver variability in the classification of simple vs complex hyperplasia also exists, but this distinction does not appear to carry the same clinical import). 3 An additional confounding factor in the practical application of these criteria is the observation that occasional low-grade endometrial adenocarcinomas (and presumably their precursor lesions) exhibit minimal cytologic atypia; this is particularly a problem with complex mucinous endometrial proliferations. Historical terms Prior to the establishment of formal classification systems by the International Society of Gynecological Pathologists and subsequently the WHO, the language describing endometrial hyperplasia was variably employed. The term atypical hyperplasia, for instance, was historically used to connote a variety of morphologies. The term was initially introduced by Novak and Rutledge and was intended to designate proliferative endometria with increased glands and scant intervening stroma with no mention of cytologic atypia. 6 Campbell and Barter subsequently subdivided atypical hyperplasia into grades based on how closely the glandular patterns resembled carcinoma, again without reference to cytologic atypia. 7 Subsequently, Vellios applied the term to endometria with cytologic atypia irrespective of architectural crowding. 8 These different definitions have contributed to persistent confusion around the term. This Table 1 World Health Organization classification of endometrial hyperplasia Simple hyperplasia without atypia Complex hyperplasia without atypia Simple hyperplasia with atypia Complex hyperplasia with atypia

7 204 Seminars in Diagnostic Pathology, Vol 27, No 4, November 2010 Figure 10 Simple hyperplasia. Minimally branched, closely packed glands with a gland-to-stroma ratio of 3:1 are seen. confusion is best resolved by assigning the full WHO designation (eg, complex atypical hyperplasia ) and indicating that this system is being employed. Other historical terms that might generate confusion include cystic hyperplasia, which roughly parallels the current WHO category of simple hyperplasia without atypia, but also encompasses some cases of cystic atrophy (see Differential Diagnosis ). The terms adenomatous hyperplasia and carcinoma in situ have also been employed, although they have fallen out of favor because of limited reproducibility and thus are not recommended Endometrial intraepithelial neoplasia (EIN) classification system The WHO classification system has proven clinically useful but remains an imperfect identifier of precancerous lesions and, as mentioned earlier, suffers from interobserver variability. 13 The deficiencies of the WHO system have spurred the development of alternative schemes that attempt to better recognize precancerous endometrial lesions, the most notable being the EIN system. 14,15 The molecular foundation of the EIN system rests on a series of meticulous studies characterizing the expression of phosphatase and tensin homolog (PTEN) in endometrial glands. 16,17 PTEN expression initially appeared to be closely correlated with development of endometrioid adenocarcinoma; however, subsequent studies have shown a less direct correlation with progression to malignancy. 18,19 In practice, criteria for the diagnosis of EIN are not entirely dissimilar from those used in the WHO system and are centered primarily around the relative proportions of tissue occupied by glands and stroma (Table 2). In the EIN system this is quantified as a volume percentage stroma (VPS) score and a cut-off of 55% is used for the diagnosis of hyperplasia or precancer (foci must measure 1 mm). 14,15 However, this system acknowledges that benign conditions, such as polyps, normal basalis, and secretory endometrium, may also have VPS scores below this cut-point; therefore, careful morphologic assessment is still required. Although specific cytologic criteria are not provided, the cytology of hyperplastic endometria differs from the surrounding endo- Figure 11 Complex hyperplasia. Increased gland-to-stroma ratio ( 3:1) and gland complexity caused by branching, outward budding, internal papillary infoldings, or internal bridges are depicted. Figure 12 Hyperplastic glands (either simple or complex) are composed of enlarged cells (relative to normal endometrial cells) with enlarged nuclei, nuclear pseudostratification, and mitotic figures. However, the nuclear membranes are smooth and the chromatin is evenly dispersed. Nucleoli may be present, but they are small and rather indistinct.

8 Mills and Longacre Endometrial Hyperplasia 205 Table 2 Classification of endometrial hyperplasia and endometrial intraepithelial neoplasia Hyperplasia Endometrial intraepithelal neoplasia* Gland stroma (VPS 55%) Size 1 mm Cytology differs from surrounding endometrium *Note that loss of PTEN expression is not a criterion for EIN. Figure 13 Atypical hyperplasia. Distinct nucleoli, nuclear pleomorphism, irregular nuclear membranes, and dispersed or clumpy chromatin define this type of hyperplasia. metrium. Loss of PTEN has been proposed as a useful marker for EIN, but loss of PTEN is not specific for EIN and might be seen in histologically normal secretory as well as in disordered proliferative glands (Figure 15). 18 Although the WHO and EIN systems both recognize the importance of glandular density in predicting progression, the EIN system gives less credence than its predecessor to the importance of nuclear atypia. This leads to a slight dissonance between the sets of cases labeled as high risk by each classification system. For example, approximately 4% of glandular proliferations classified by the WHO as simple hyperplasia without atypia and 44% of proliferations classified as complex hyperplasia without atypia are classified as EIN, whereas 78% of atypical hyperplasias are classified as EIN, with most of the remaining atypical hyperplasias being diagnosed as carcinomas using this system (Figure 16). 14,15,20 The initial hope was that the EIN system would improve on the ability to predict the development of endometrial carcinoma; however, subsequent studies have failed to demonstrate that EIN carries a greater risk of progression than simple or complex atypical hyperplasia. 3,21 Moreover, the diagnostic reproducibility is similar among the two classification systems. Thus, although the PTEN hypothesis has shed considerable light on the molecular pathway leading to endometrial endometrioid adenocarcinoma, selecting which hyperplastic lesions will progress to carcinoma with sufficient tempo to warrant surgical intervention remains difficult. Endometrial neoplasia classification system Given the poor reproducibility in differentiating atypical hyperplasia from nonatypical hyperplasia on the one hand and from well-differentiated adenocarcinoma on the other, Bergeron and colleagues have proposed a simplified histologic classification based on a combined category for simple and complex hyperplasia, designated hyperplasia, and a combined category for atypical hyperplasia and well-differentiated adenocarcinoma, designated endometrioid neoplasia. 22 Using this scheme, glandular crowding is the best criterion to diagnose hyperplasia, whereas nuclear pleomorphism is the chief criterion for endometrioid neoplasia; nuclear enlargement, vesicular chromatin, and nucleoli are additional features. Preliminary studies have suggested that the diagnoses of hyperplasia and endometrioid neoplasia are highly reproducible when using this classification scheme, but this system has not been widely adopted (Table 3). Does focal hyperplasia exist? Figure 14 Atypical hyperplasia. There is extensive atypia in this proliferation. The significance of small foci of atypia set in a hyperplasic process that is otherwise banal is uncertain. Although endometrial hyperplasia generally is a diffuse process, in some instances hyperplasia may only focally involve the endometrium. This phenomenon is best identified in hysterectomy specimens, but can occasionally be encountered in an endometrial curetting. In the latter setting, the sampling consists of a dimorphic pattern of hyperplastic endometrium with separate fragments of normal cycling endometrium in premenopausal patients and inactive or weakly proliferative endometrium in postmenopausal patients. Apparent isolated involvement of a polyp might be seen. When the hyperplasic endometrial glands are atypical, there are few data relating to risk of progression, but a

9 206 Seminars in Diagnostic Pathology, Vol 27, No 4, November 2010 Figure 15 (A) Disordered proliferative endometrium shows (B) loss of PTEN expression. Although loss of PTEN is associated with endometrial neoplasia, it is not specific for it and is not recommended for use in routine diagnosis or in risk stratification. curettage followed by a trial of progestin therapy might be warranted in this situation. A practical approach to diagnosing endometrial hyperplasia In practice, the key distinctions to be made in endometrial hyperplasia are (1) the presence or absence of atypia, (2) the exclusion of benign mimics that pose no risk for progression to endometrial carcinoma, and (3) the exclusion of lowgrade endometrial endometrioid carcinoma. Most cases of atypical endometrial hyperplasia exhibit complex glandular architecture and diffuse atypia (ie, enlarged nucleoli may be focally present in some hyperplastic endometria, but the significance of focal cytologic atypia is uncertain). If the architecture is sufficiently complex, we diagnose these lesions as complex endometrial hyperplasia with focal cytologic atypia and specify the degree and extent in a comment section of the report. If the hyperplastic process is architecturally noncomplex and the sampling appears to be representative, we diagnose these lesions as hyperplasia and indicate the presence of focal atypia in the comment section, with the note that the degree and extent of atypia are insufficient to warrant a definitive diagnosis of atypical hyperplasia. At the other end of the spectrum are markedly complex endometrial glandular proliferations that appear to fall short of that typically seen in low-grade endometrioid adenocarcinoma, but are nevertheless worrisome for welldifferentiated adenocarcinoma (Figure 16). Because the endometrial sampling is essentially a screening procedure, these cases are best diagnosed as borderline or at least Figure 16 Endometrial hyperplasia spectrum. The degree of relatedness between the various classifications for endometrial hyperplasia and carcinoma is depicted. Note that the classification schemes are not equivalent, particularly with respect to the WHO and EIN. The degree of overlap (and nonoverlap) between WHO and EIN classification schemes is an approximation based on extrapolation of the published data.

10 Mills and Longacre Endometrial Hyperplasia 207 Table 3 European classification of endometrial hyperplasia 22 Endometrial hyperplasia Simple hyperplasia without atypia Complex hyperplasia without atypia Endometrial neoplasia Atypical hyperplasia Endometrial carcinoma complex atypical hyperplasia, cannot exclude well-differentiated adenocarcinoma (Table 4). 1,23 The exclusion of benign mimics that pose no risk for progression is discussed below. Differential diagnosis 1. Simple and complex endometrial hyperplasia without atypia Artifactual changes are among the chief mimickers of nonatypical hyperplasia. For instance, stromal endometrial collapse can lead to an apparent increase of the gland-tostroma ratio. Telescoping of glands, a particularly common phenomenon in fragmented curettage specimens, can impart a low-power appearance of glandular complexity (Figure 17). Surface epithelium can become on occasion so tightly coiled in a curettage specimen that it might closely simulate hyperplasia with a papillary architecture. Poor orientation can also lead to misinterpretation of an endometrial curettage, particularly when the basalis is overrepresented. In these cases, identification of the surface (or lack thereof) epithelium is crucial to avoid the diagnosis of hyperplasia. Disordered proliferative endometrium can closely resemble simple and, on occasion, complex hyperplasia (Figure 18). This occurs most commonly during the perimenarchal Figure 17 Telescoping of endometrial glands. This phenomenon can be misinterpreted as disordered proliferative endometrium or hyperplasia on low magnification. Higher magnification typically reveals bland cells without the usual features of hyperplasia (eg, nuclear enlargement, increased mitotic figures, and nuclear pseudostratification). and perimenopausal years in association with anovulatory cycles and unopposed estrogen. Disordered proliferative endometrial glands often exhibit irregular contours, cystic expansion, and focal crowding. However, architectural changes are limited and the gland-to-stroma ratio does not reach 3:1 when compared with endometrial hyperplasia. This is essentially a clinically benign process and should not be overdiagnosed as hyperplasia. Although progression to hyperplasia may occur, progression of disordered proliferative endometrium to carcinoma is exceedingly rare, if it occurs at all. 5 Endometrial polyps should also be in the differential of nonatypical hyperplasia, particularly on limited samples (Figure 19). Polyps may show architecturally complex glands and significant ciliated or squamous metaplasia, both common features of endometrial hyperplasia. Although en- Table 4 Classification systems for endometrial hyperplasia* WHO EIN European WHO (modified) Simple hyperplasia without atypia Hyperplasia Hyperplasia Hyperplasia without atypia Complex hyperplasia without atypia Simple hyperplasia with atypia EIN Endometrial neoplasia Hyperplasia with atypia Complex hyperplasia with atypia Borderline Carcinoma Carcinoma Carcinoma Abbreviations: EIN, endometrial intraepithelial neoplasia; WHO, World Health Organization. *Note that the classification schemes are not equivalent, particularly with respect to the WHO and EIN. The degree of overlap (and nonoverlap) between WHO and EIN classification schemes is based on extrapolation of the published data. Approximately 4% of glandular proliferations classified as simple hyperplasia without atypia and 44% of glandular proliferations classified as complex hyperplasia without atypia are classified as EIN, whereas 78% of atypical hyperplasias are classified as EIN, with most of the remaining atypical hyperplasias classified as carcinoma using the EIN system. Because the critical distinction in the WHO classification is nonatypical versus atypical hyperplasia, collapsing the four WHO categories for endometrial hyperplasia into two categories should improve interobserver reproducibility. 43 The introduction of the borderline category captures both the uncertainty and the poor interobserver reproducibility of diagnoses for glandular proliferations that lie within the morphologic transition zone of the hyperplasia carcinoma spectrum. 23 Simple hyperplasia with atypia is very uncommon. Most atypical hyperplasias are complex.

11 208 Seminars in Diagnostic Pathology, Vol 27, No 4, November 2010 Figure 18 Disordered proliferative endometrium. (A) Focal gland crowding is noted, but the overall gland-to-stroma ratio is less than 3:1 (often less than 2:1). (B) The constituent cells may exhibit various degrees of metaplasia, but cytologic atypia is absent. dometrial polyps may represent focal hyperplasia of the basalis, they do not appear to be estrogen related and do not carry the risks associated with real endometrial hyperplasia. The presence of dense stroma, thick-walled vessels, and normal adjacent endometrium helps in establishing a diagnosis of endometrial polyp and avoiding an overdiagnosis of hyperplasia. However, foci of hyperplasia or even carcinoma (both low and high grade) might be seen in this setting; thus, the glands contained within endometrial polyps should be carefully evaluated. Cystic atrophy is typically encountered in endometria from postmenopausal women. It is distinguished from cystic hyperplasia by the presence of attenuated and nonstratified epithelium lacking mitoses, often with prominent ciliated metaplasia, set in an eosinophilic, collagenous stroma (Figure 20). Because the underlying process reflects a low estrogenic state, the volume of involved endometrial tissue is less than that typically encountered in hyperplasia. 2. Simple and complex endometrial hyperplasia with atypia A misdiagnosis of atypical hyperplasia has a higher penalty than that of nonatypical hyperplasia because the former carries a greater risk of progression to cancer and is therefore more likely to trigger definitive clinical action. As with nonatypical hyperplasia, however, multiple artifactual or benign changes can complicate an accurate diagnosis of atypical hyperplasia. Unsettling artifacts include nuclear changes, such as nuclear pseudostratification caused by the tangential sectioning of hyperplastic epithelium, hobnaillike morphology resulting from hormonal therapy (see Hormonal Effects ), nuclear clearing caused by inadequate fixation, and reactive epithelial changes in endometritis. Similarly, normal proliferative endometrium can have rounded cells with coarse chromatin reminiscent of those in atypical hyperplasia. In these cases, the absence of other cytologic features of atypia helps to exclude the diagnosis of simple atypical hyperplasia. Metaplasias can be particularly problematic when assessing atypia in an endometrial proliferation, in part because these changes are more commonly seen under estrogenic stimulation and often occur in concert with hyperplasia and in part because the metaplastic epithelium may exhibit cytologic changes that begin to overlap those observed in atypical hyperplasia. However, in isolation, metaplasias have no neoplastic potential and their presence alone does not impart an increased risk for malignancy. Papillary syncytial change (also referred to as surface syncytial change and papillary syncytial metaplasia ) is among the most Figure 19 Endometrial polyp. (A) The glands in endometrial polyps are often crowded and (B) irregular in contour; however, recognition of collagenous stroma, thick-walled blood vessels, and overall polypoid configuration will prevent overdiagnosis of hyperplasia. Because hyperplasia and carcinoma may occur in polyps, these diagnoses should always be considered before rendering a diagnosis of benign endometrial polyp.

12 Mills and Longacre Endometrial Hyperplasia 209 Figure 20 Cystic atrophy. There are numerous dilated glands that are lined by atrophic endometrial cells. eye-catching of these alterations and is thought to represent a degenerative process rather than a true metaplasia (Figure 21). 24 Papillary syncytial change is characterized by a sheetlike or pseudopapillary proliferation of eosinophilic cells with indistinct cell borders and crowded nuclei. In some cases, it may be associated with prominent reactive cytologic atypia and even scattered mitoses. However, typically it is observed in the endometrial surface and superficial glands in the setting of stromal breakdown. Another form of eosinophilic metaplasia consists of cuboidal to columnar cells with pale to dark pink cytoplasm, resembling tubal metaplasia, but without cilia (Figure 22). The nuclei may exhibit hyperchromasia, posing concern for a cytologically atypical process, but the chromatin is smudged. The presence of enlarged nuclei and nucleoli in ciliated cells may lead to confusion with atypical hyperplasia. However, the glands lack nuclear pleomorphism as well as mitoses and peg cells and cilia are typically seen. Figure 22 Papillary syncytial change. There is nuclear atypia consisting of slightly enlarged nuclei with small nucleoli or smudged chromatin. Mitotic figures might be present but are typically sparse. Squamous metaplasia in endometrial glands may also cause concern for hyperplasia, particularly when extensive. Squamous metaplastic cells are typically nonkeratinizing and contain moderate amounts of eosinophilic cytoplasm. Nuclei are oval and uniform and mitoses are rare, but small nucleoli may be seen (Figure 23). Cells can form sheets and morules that fill the glandular lumens (Figure 24), creating the false impression of a significant solid component. Central necrosis, frequently encountered within the morules, should not be misinterpreted as evidence of atypia or malignancy (Figure 25). It is not uncommon for extensive morular metaplasia to be misdiagnosed as complex atypical hyperplasia or even carcinoma because of the sheet-like growth pattern and the presence of small nucleoli in the squamous metaplastic cells. 25 Classification should be based on an assessment of the architecture and cytology of the glandular components. Occasionally, metaplastic squa- Figure 21 Papillary syncytial change. A sheet-like proliferation of eosinophilic cells with indistinct cell borders is seen on the surface endometrium. Figure 23 Squamous metaplasia is usually cytologically banal, but some cases may show nucleoli.

13 210 Seminars in Diagnostic Pathology, Vol 27, No 4, November 2010 Figure 24 Squamous morules. Immature-appearing squamous cells extensively fill the gland lumens. mous epithelium exhibits prominent keratin production (Figure 26) and diffusely involves the uterus (ichthyosis uteri), a finding that can be associated with hyperplasia or carcinoma. 26 Mucinous metaplasia (Figure 27) might also present significant problems in the distinction between benign mucinous metaplasia, atypical mucinous hyperplasia, and welldifferentiated mucinous carcinoma. 1 In these instances, careful assessment of the glandular architecture is critical. Because mucinous carcinomas may be deceptively bland, whenever the glandular configuration of a mucinous proliferation is indeterminate between hyperplasia and carcinoma the term complex mucinous proliferation should be employed to convey this uncertainty. Evaluation for a possible cervical source is also prudent in these cases because distinction among cervical microglandular hyperplasia, endometrial atypical mucinous hyperplasia with microglandular Figure 25 Squamous or morular metaplasia. Foci of punctuate necrosis are present in the center of the morules. This finding has no bearing on risk for progression and should not be misinterpreted as evidence of atypia or malignancy. Figure 26 Squamous metaplasia. Prominent keratin production may be encountered. Distinction from a cervical process can be problematic when the curettage specimen consists entirely of keratinizing squamous epithelium. A differential curettage is often required to localize the lesion. features, and endometrial mucinous carcinoma with microglandular features can be problematic. 27,28 The architectural and nuclear changes seen in florid secretory or gestational endometria, when interpreted without the relevant clinical history, can also raise the possibility of atypical hyperplasia. Secretory changes that can lead to concern include complex infoldings, abundant vacuolated cytoplasm, enlarged hyperchromatic nuclei, and nuclear hobnailing. 29 These changes may also occur focally or diffusely within hyperplasic glands; when widespread, hyperplasia may be classified as secretory hyperplasia (Figure 28). Identification of bona fide cytologic atypia can be difficult in this setting. If real atypia cannot be excluded, recommendation for possible rebiopsy is warranted. Endometritis can elicit cytologic atypia, including nuclear enlargement and stratification as well as architectural glandular changes; however, the presence of abundant spindled, reactive stroma and plasma cells provides a clue to avoid the diagnosis of atypical hyperplasia. In some cases, examination of the sample or the clinical history may offer clues to the etiology behind the process. Fragmented benign endometrial polyps are also common mimickers of atypical hyperplasia, particularly in curetting specimens. Many polyps contain irregular, basalis-type glands that can bear some nuclear atypia. Although a diagnosis of polyp is easily rendered when their polypoid shape, dense stroma, and thick-walled blood vessels are readily apparent, in fragmented samples these features are not always obvious. In such cases additional sampling may be necessary to secure a diagnosis. Atypical polypoid adenomyomas (APAs) often contain complex, irregular glands with abnormal cytology suggestive of atypical hyperplasia or even carcinoma (Figure 29). 30 These localized, polypoid proliferations typically occur in young women (often 40 years of age) that are often nulliparous and not uncommonly have a history of infertil-

14 Mills and Longacre Endometrial Hyperplasia 211 Figure 27 Mucinous metaplasia. This type of metaplasia may be associated with (A) papillary metaplasia or other types of metaplasia and (B) closely mimics endocervical mucinous epithelium. ity. They are characterized by a biphasic proliferation of glandular elements and fibromuscular stroma. Curettage specimens may contain, admixed with the APA fragments, additional fragments of normal proliferative or secretory pattern endometrium, signifying a focal process When the glandular component of APAs has marked architectural complexity and cytologic atypia, the presence of smooth muscle and fibrous tissue intercalating between the abnormal glands and endometrial stroma helps exclude atypical hyperplasia. Although myoinvasive adenocarcinoma should also be considered whenever atypical glands are immediately juxtaposed to muscle, this is an unusual finding in an endometrial sampling and most such cases harbor fragments of adenocarcinoma without associated smooth muscle elsewhere in the sample. 31,34 Exuberant morular squamous metaplasia can also provide guarded reassurance of a diagnosis of APA because it is quite common in (although by no means exclusive to) this entity. 3. Atypical hyperplasia versus well-differentiated adenocarcinoma Perhaps the most important and problematic entity in the differential diagnosis of atypical hyperplasia (or EIN) is well-differentiated endometrioid carcinoma. This distinction remains a focus of debate and nidus of confusion in gynecologic surgical pathology Uncertainty about this morphologic definition persists in part because of clinical considerations. Among these is the fact that pathologists, surgeons, and patients are all understandably loath to allow potentially precancerous lesions to persist in women who are beyond childbearing years. The labile nature of endometrial precursor lesions also impedes their characterization: they may be unceremoniously shed and can resolve with hormonal changes. Despite the difficulties that arise in their application, pathologists can generally agree on three criteria for the diagnosis well-differentiated endometrioid carcinoma. They are: (1) the presence of large macroglands with complex cribriform or papillary internal structure; (2) a confluent, haphazard pattern of branching glands with interconnecting gland lumina without intervening stroma; and (3) extensive or confluent papillary formations. 23,38 The presence of any one of these three features merits a diagnosis of welldifferentiated endometrioid adenocarcinoma. However, it is often not possible to securely place a proliferation in the malignant category based on small, fragmented samples. When myometrial invasion is readily identified the distinction is easy; however, this is seldom sampled on curetting or biopsy specimens. Occasionally, there may be associated desmoplastic stroma, but this is also uncommon. Cases in which the distinction between the two processes cannot be made with confidence are best diagnosed as borderline or at least complex atypical hyperplasia, cannot exclude well Figure 28 Secretory hyperplasia. (A) Hyperplastic endometrial glands with branching show prominent subnuclear cytoplasmic vacuoles, (B) similar to secretory endometrium.

15 212 Seminars in Diagnostic Pathology, Vol 27, No 4, November 2010 Figure 29 Atypical polypoid adenomyoma. (A) Irregular endometrial glands with associated morular metaplasia are set in a prominent fibromuscular stroma. (B) The glands exhibit some degree of cytologic atypia. differentiated adenocarcinoma (Figure 30). 1,23 It is important to emphasize that some endometrial adenocarcinomas may exhibit less cytologic atypia than complex atypical hyperplasia; this is especially true with some mucinous adenocarcinomas. Other difficulties that may be encountered in endometrial sampling specimens include extensive fragmentation, extensive squamous or morular metaplasia, and detached fragments of spindled cell stroma in extensive breakdown that may mimic a desmoplastic stromal response; in these instances, a definitive diagnosis should be deferred and further evaluation, including consideration for repeat sampling, should be recommended. 1,23 Evaluation of progestin treatment in endometrial hyperplasia Hormonal administration alters baseline endometrial histology, and interpretation of endometrial samplings should always be informed by the patient s therapeutic regimen. Most hormone replacement therapies include both estrogen and progesterone and result in a muted or incomplete secretory pattern. Progesterone-only programs are usually associated with an expanded, pseudodecidualized stroma bearing atrophic or weakly secretory glands (Figure 31). Hormonal therapy is also often linked to mucinous changes. Tamoxifen is widely employed in women with breast cancer for its antiestrogenic properties at this site. In the uterus, however, it appears to have a paradoxically proliferative effect. Assessment of proliferative changes in women on tamoxifen may be complicated by the fact that this drug also leads to endometrial stromal fibrosis, resulting in scanty biopsy samples. Appropriate assessment in these samples is crucial, however, because of the drug s association with the development of endometrial hyperplasia and carcinoma. In cases where biopsy material is scanty and the presence of a fragmented polyp is excluded, consideration for resampling should be recommended if abnormal imaging or clinical findings persist. Patients with endometrial hyperplasia are often treated with progestin to induce a hormonal ablation of the hyperplastic process Response rates vary, but have been Figure 30 Complex glandular proliferation with atypia bordering on low-grade endometrioid adenocarcinoma (right). Because of the small volume of the complex proliferation and in the absence of more definitive features, a diagnosis of borderline or at least complex atypical hyperplasia, cannot exclude adenocarcinoma is adequate. Figure 31 Progestin effect. The endometrial glands appear inactive with associated stromal pseudodecidualization.

16 Mills and Longacre Endometrial Hyperplasia 213 Figure 32 (A and B) Mucinous metaplasia associated with progestin therapy. Because hyperplasia and carcinoma might also feature prominent mucinous metaplasia, the prior, pretreatment curettage should be evaluated in conjunction with the posttreatment sampling to determine whether a persistent lesion exists. reported in % of treated patients. Treatment tends to be more successful for nonatypical hyperplasia, but occasionally atypical hyperplasia and even well-differentiated endometrioid carcinoma respond to progestin therapy. Evaluation of the follow-up biopsy or curettage to assess overall response is best accomplished by comparison with the prior, pretreatment sampling. Common changes secondary to the progestin therapy include decreased gland confluence and complexity, increased metaplasias (squamous, eosinophilic, and mucinous), and decreased nuclear size (nuclear-tocytoplasmic ratio) and atypia (Figure 32). 42 Because persistence of significant cytologic atypia or complex architecture might be present only focally in the posttreated endometrium, these samplings should be carefully screened to alert the treating physician to the possibility of persistent and/or possibly recurrent atypical hyperplasia. Although the clinical significance of focal hyperplasia is generally not well defined (see Focal Hyperplasia ), the presence of focal hyperplasia in this setting implies either inadequate therapy (too low dose or too short interval or both) or incomplete response to medical management. References 1. Longacre TA, Atkins KA, Kempson RL, et al: The uterine corpus, in Mills S (ed): Sternberg s Diagnostic Surgical Pathology, vol 2 (ed 5). New York, NY, Lippincott Williams & Wilkins, 2009, pp Mazur MT: Endometrial hyperplasia/adenocarcinoma. A conventional approach. Ann Diagn Pathol 9: , Lacey JV Jr, Mutter GL, Nucci MR, et al: Risk of subsequent endometrial carcinoma associated with endometrial intraepithelial neoplasia classification of endometrial biopsies. Cancer 113: , Kurman R, Kaminski P, Norris H: The behavior of endometrial hyperplasia. A long-term study of untreated hyperplasia in 170 patients. Cancer 56: , Huang S, Amparo E, Fu Y: Endometrial hyperplasia: histologic classification and behavior. Surg Pathol 1: , Novak E, Rutledge F: Atypical endometrial hyperplasia simulating adenocarcinoma. Am J Obstet Gynecol 55:46-63, Campbell PE, Barter RA: The significance of a typical endometrial hyperplasia. J Obstet Gynaecol Br Commonw 68: , Vellios F: Endometrial hyperplasia and carcinoma in situ. Gynecol Oncol 2: , Welch W, Scully R: Precancerous lesions of the endometrium. Hum Pathol 8: , Hertig AT, Sommers SC, Bengloff H: Genesis of endometrial carcinoma; carcinoma in situ. Cancer 2: , 1949 [PubMed], illust 11. Buehl IA, Vellios F, Carter JE, Huber CP: Carcinoma in situ of the endometrium. Am J Clin Pathol 42: , Gusberg SB, Kaplan AL: Precursors of corpus cancer. IV. Adenomatous hyperplasia as stage O carcinoma of the endometrium. Am J Obstet Gynecol 87: , Kendall BS, Ronnett BM, Isacson C, et al: Reproducibility of the diagnosis of endometrial hyperplasia, atypical hyperplasia, and welldifferentiated carcinoma. Am J Surg Pathol 22: , Mutter GL: Histopathology of genetically defined endometrial precancers. Int J Gynecol Pathol 19: , Mutter GL, Baak JP, Crum CP, et al: Endometrial precancer diagnosis by histopathology, clonal analysis, and computerized morphometry. J Pathol 190: , Mutter GL, Ince TA, Baak JP, et al: Molecular identification of latent precancers in histologically normal endometrium. Cancer Res 61: , Mutter GL, Lin MC, Fitzgerald JT, et al: Altered PTEN expression as a diagnostic marker for the earliest endometrial precancers. J Natl Cancer Inst 92: , Mutter GL, Lin MC, Fitzgerald JT, et al: Changes in endometrial PTEN expression throughout the human menstrual cycle. J Clin Endocrinol Metab 85: , Lacey JV, Jr, Mutter GL, Ronnett BM, et al: PTEN expression in endometrial biopsies as a marker of progression to endometrial carcinoma. Cancer Res 68: , Mutter GL: Endometrial intraepithelial neoplasia (EIN): will it bring order to chaos? The Endometrial Collaborative Group. Gynecol Oncol 76: , Baak JP, Mutter GL, Robboy S, et al: The molecular genetics and morphometry-based endometrial intraepithelial neoplasia classification system predicts disease progression in endometrial hyperplasia more accurately than the 1994 World Health Organization classification system. Cancer 103: , Bergeron C, Nogales FF, Masseroli M, et al: A multicentric European study testing the reproducibility of the WHO classification of endometrial hyperplasia with a proposal of a simplified working classification for biopsy and curettage specimens. Am J Surg Pathol 23: , Longacre TA, Chung MH, Jensen DN, et al: Proposed criteria for the diagnosis of well-differentiated endometrial carcinoma. A diagnostic test for myoinvasion. Am J Surg Pathol 19: , 1995