(Module II)
EIN Architecture

horizontal rule

Distribution of EIN Lesions: Focal, then progress to diffuse

    Origination from a single cell explains why EIN lesions begin as regional events within the endometrial compartment.  In a representative tissue sample this is often apparent as a contrast in architecture and cytology between the EIN and non-EIN components.  EIN lesions grow over time, and some patients have no remaining or available sampled background endometrium in a biopsy sample.  These can still be recognized as EIN when the cytology is sufficiently altered that it is readily recognizable as abnormal.

Examples of focal EIN lesions with adjacent non-EIN endometrium.  
Two bottom rows show lesions as they appear by H&E and delimited by PTEN immunohistochemistry.

 EINfocal2.jpg (86782 bytes) EINfocal3.jpg (73985 bytes) 
59956x03.jpg (189442 bytes) 59956x01.jpg (152818 bytes) 
59102x03.jpg (171006 bytes) 59102x01.jpg (146418 bytes)

    Practical limitations of tissue sampling make this a "soft" or nonessential criterion for EIN diagnosis.  Some EIN lesions may be sufficiently large that unaffected tissue is poorly represented in the sample.   Many postmenopausal or perimenopausal women with a thin atrophic endometrium in the background will preferentially have areas of bulkier EIN represented in the curetting or biopsy.   Additionally, a single lesion may present in multiple independent tissue fragments during sampling, making it difficult to reconstruct their interrelationships.  Despite these limitations, the distribution of a lesion is useful in distinguishing between the diffuse, field-wide effects, of an abnormal hormonal environment (anovulation, or persistent estrogen effect, diagnosed as endometrial hyperplasia) and the more focal EIN. 


***Hot Tips***

PTEN Immunohistochemistry can delimit many EIN lesions, and is useful for educational purposes.  It is NOT, however,  appropriate for routine clinical use in patient diagnosis of EIN because of poor sensitivity and specificity.

Half of EIN lesions retain normal PTEN function (low sensitivity) and almost half of normal endometria have scattered glands with loss of PTEN function (low specificity).    Loss of PTEN function occurs so early in endometrial carcinogenesis that affected cells must evolve additional damage before they even look different by H&E.  Why is PTEN of interest at all? Visit PTEN Central.

To learn more about PTEN immunohistochemistry uses   Click Here

horizontal rule

Size: Must exceed 1mm

    Architectural criteria for EIN diagnosis require a contiguous field of glands sufficiently large to permit estimation of the Volume Percentage Stroma.  For this reason lesions shown by morphometry to predict future or coexistent carcinoma were all at least 1-2mm in maximum dimension.  Clonal analysis was possible only with tissue areas of approximately 2mm in diameter or more.   Thus, the supportive morphometry and clonal analysis discussed in this website as a foundation for clinically predictive EIN diagnosis applies to a lesion with a minimum diameter of 1-2 mm -  a scale which usually encompasses more than 5-10 glands.  

    Individual atypical glands or clusters of fewer than 5-10 atypical glands have an undefined natural history and are best diagnosed descriptively using non-EIN terminology.  Occasionally deeper levels will expose diagnostic fields, but the isolated atypical gland will always remain a problem.  Their indeterminate course can be managed with a recommendation for repeat evaluation in 3-6 months. The clinical significance of a minimal disease burden at these earliest phases of growth will have to be defined in the context of objective and reproducible detection methods with greater physical resolution than those reported previously.  It is predicted that as specific markers for premalignant endometrial epithelia become available, either through immunohistochemistry, or genetic analysis of finely microdissected tissues, that earlier and earlier stages of premalignant disease will be discernible and their clinical course can then be defined.  In the meantime, the clinical ambiguity of a very small lesion does not diminish the predictive value already established for EIN lesions exceeding a threshold size of 1-2 mm, or 5-10 glands. 

horizontal rule

Architectural Pattern: Area of glands exceeds that of stroma

    Although EIN is an epithelial disease, visual assessment of the glands themselves is complicated by frequent artifactual displacement from associated stroma, pale staining of most epithelia, and visual "shimmering" between gland epithelia and lumens.  These may all be avoided by focusing on the stromal compartment which has the significant advantages of a more uniform composition throughout the specimen, and superior staining qualities.  More importantly, by focusing on the stroma itself only intact fragments in which stroma has not been avulsed from glands will be evaluated.  The cardinal architectural feature of endometrial precancers is glandular crowding, with a threshold quantitative cutoff for EIN lesions of less than half of the tissue occupied by stroma (VPS<55%).     

    Careful review of the graphic and histologic examples of varying VPS will assist in training your eye to classify patient material as above or below a diagnostic threshold of 55% VPS.   Fortunately, there are very few examples with VPS in the borderline range of 50-60%, as EIN lesions cluster with a median VPS of about 40% and non-EIN (benign) lesions cluster at a median of  approximately 75% VPS.   These differences are sufficiently great that visual assessment by a trained eye can be informative.

horizontal rule

    Following is a series of calibrated tiles schematically representing glands (light grey circles include combined epithelium and lumen) against a field of dark grey stroma. 

VPS80lg.GIF (2946 bytes) VPS70lg.GIF (3639 bytes) VPS60lg.GIF (8259 bytes)VPS50lg.GIF (8272 bytes) VPS40lg.GIF (5641 bytes) VPS30lg.GIF (6172 bytes)

    You should now view the detailed pages with representative fields of  measured endometria from the Type Collection.  These are arranged by VPS value, and accompanied by matching tiles.

    Click the VPS ranges shown in ascending sequence on the left sidebar of this page, or begin with the first (benign) Set by clicking here.  You will begin with examples of benign high VPS (>60%) lesions, and work your way through the borderline zone of 50-60% VPS to EIN endometria with VPS of 50% or less. 


horizontal rule

Copyright 1998-2015 by George L. Mutter, MD.  All Rights Reserved