And Diagnostic Reproducibility

The lexicon for endometrial cancer precursors is large and confusing, reflecting repeated reclassifications of these lesions based on relatively scant, and usually retrospectively collected data (52-59). Not surprisingly, the same terms sometimes have been

Fig. 5. Serous carcinoma. Although some serous carcinomas have a distinctive papillary configuration with highly atypical neoplastic cells covering fibrovascular cores, much of the tumor might be made up of glands with either a scalloped apical cell border or intraglandular papillae.

applied to lesions of differing histological appearance in different classification schemes. For example, although some have suggested that all hyperplasias by definition are at least mildly atypical, others have confined atypical hyperplasia to only the lesions, which most closely approximate carcinoma histologically. However, the upper end of this spectrum has also been designated marked adenomatous hyperplasia or carcinoma in situ in other classification schemes. Much confusion has also resulted from the difficulty in determining the natural history and biological behavior of these lesions because they are usually sampled only following the onset of symptoms, and they are often found to be highly heterogeneous when the endometrium is examined thoroughly in the hysterectomy specimen.

The current classification, adopted by both the International Society of Gynecological Pathologists and the WHO, was proposed by Kurman, Kaminski, and Norris in 1985 (57). They divided endometrial hyperplasia into four groups according to the presence or absence of cytological atypia, and the degree of architectural complexity and crowding, as follows: simple hyperplasia, complex hyperplasia, simple atypical hyperplasia, and complex atypical hyperplasia. Simple hyperplasia is defined as an increase in the number of endometrial glands, which may be dilated with little crowding or have an irregular outline and exhibit crowding (Fig. 7). Complex hyperplasia is characterized by glands with irregular outlines, marked structural complexity, and back-to-back crowding (Fig. 8). Atypical hyperplasia is used to designate a proliferation of glands exhibiting

Fig. 6. Clear cell adenocarcinoma. This form of adenocarcinoma is characterized by cells with either abundant optically clear cytoplasm or a hobnail arrangement of cells growing as cystically dilated glands, tubules, papillae, or solid masses.

cytological atypia, recognized as nuclear enlargement, the presence of nucleoli, or a change from an elongated to more ovoid or round nucleus (Figs. 9 and 10). The chro-matin might be either evenly or irregularly dispersed. This classification was justified on the basis of differing outcomes for the various groups, with progression to carcinoma in 1% of patients with simple hyperplasia, 3% of those with complex hyperplasia, 8% of those with simple atypical hyperplasia, and 29% of those with complex atypical hyper-plasia. In that retrospective study, about two-thirds of the women received some surgical or hormonal modulation during the interval between initial diagnosis and hysterectomy, which varied from 1 to 27 years (mean of 15 years). As noted by Kendall et al. (60), the definitions of architectural complexity and nuclear atypia potentially rest on a multitude of criteria, and some but not all criteria may be fully developed in any given case.

Concern about the reliability of the pathologists to apply multiple criteria to distinguish among these forms of hyperplasia led to three articles and several abstracts assessing the reproducibility of the diagnoses. Skov et al. (61) compared the reproducibility of the WHO classifications of 1975 and 1994 of endometrial hyperplasia, circulating slides of 128 cases of hyperplasia to 6 experienced gynecological pathologists in Denmark. They found intraobserver reproducibility to be moderate for both systems, and interobserver reproducibility to be slight to moderate for various subtypes. Kendall et al. (60) examined the reproducibility of the diagnosis of endometrial hyperplasia and well-differentiated adenocarcinoma among five pathologists of varying experience at

Fig. 7. Simple hyperplasia. Hyperplasia is generally characterized by a preferential proliferation of endometrial glands, resulting in an increase in the ratio of glands to stroma. The glands may be cystically dilated, or have branching or bifurcation. In simple hyperplasia, easily identifiable stroma separates the abnormal appearing glands.

Fig. 7. Simple hyperplasia. Hyperplasia is generally characterized by a preferential proliferation of endometrial glands, resulting in an increase in the ratio of glands to stroma. The glands may be cystically dilated, or have branching or bifurcation. In simple hyperplasia, easily identifiable stroma separates the abnormal appearing glands.

Johns Hopkins Hospital. Intraobserver agreement was substantial. Interobserver agreement was highly variable, with almost perfect agreement for the diagnosis of prolifera-tive endometrium or well-differentiated adenocarcinoma, substantial reproducibility for simple hyperplasia, moderate reproducibility for complex nonatypical and atypical hyperplasia, but only slight for simple atypical hyperplasia. Bergeron et al. (62) conducted a European multicenter study of the reproducibility of the WHO classification of endometrial hyperplasia. The intraobserver agreement was moderate, and the interobserver reproducibility was fair to moderate. They suggested condensing atypical hyperplasia and well-differentiated adenocarcinoma into a category of endometrioid neoplasia, and designating nonatypical, simple, and complex hyperplasia as hyperplasia. The gynecological oncology group recently completed a study of the reproducibility of the community-based diagnosis of atypical hyperplasia in about 300 women, the results of which have been reported in abstract form. A panel of the three gynecological pathol-ogists each reviewed the slides, blinded to the interpretations of the other panelists. A panel diagnosis was defined as agreement of assignment of a case to one of the five categories by two of three or all three-panel pathologists. The panel agreed with the diagnosis of atypical hyperplasia in only 39% of cases, interpreting about one quarter of cases as a less significant process and about one quarter of cases as adenocarcinoma. Interobserver reproducibility among the panel pathologists was the lowest for the

Fig. 8. Complex hyperplasia. In addition to higher glandular complexity, little stroma separates the glands from one another.

diagnosis of atypical hyperplasia, with a k value of about 0.27. Hysterectomy subsequently performed within 12 weeks of initial diagnosis revealed the presence of coexisting adenocarcinoma in the uterus in about 43% of cases. These disturbing results raise serious questions about the ability of pathologists to recognize and distinguish endometrial precancers from carcinoma in biopsies or curettings.

Factors contributing to low reproducibility include:

1. The fragmentary nature of biopsies or curettings;

2. The presence of borderline lesions;

3. Uncertainty about the significance of focal hyperplasia;

4. Artifacts relating to poor fixation, cryotomy, and staining;

5. The inadequacy of published descriptions and understanding of terms used to define architectural or cytological atypia; and

6. The difficulty associated with the translation of verbal descriptions into light microscopic interobserver reproducibility for images.

The difficulty in identifying distinctive subjective histological characteristics associated with differing biological behavior encouraged investigations using morphometric analysis. Colgan et al. (63), and Baak et al. (64-67) investigated features of gland cell nuclei. Colgan et al. (63) found that a linear discriminant function selected the mean and standard deviation of the maximal nuclear diameter as the most useful predictor of eventual progression of atypical endometrial hyperplasia to adenocarcinoma. However,

Fig. 9. Atypical hyperplasia. Hyperplasia is also divided according to the absence or presence of cyto-logical atypia. This atypia is usually defined as the presence of nuclear enlargement, abnormal chro-matin distribution, and especially the presence of nucleoli. Because fixation and staining of specimens are variable, comparison of features with adjacent nonhyperplastic gland nuclei may be helpful.

Fig. 9. Atypical hyperplasia. Hyperplasia is also divided according to the absence or presence of cyto-logical atypia. This atypia is usually defined as the presence of nuclear enlargement, abnormal chro-matin distribution, and especially the presence of nucleoli. Because fixation and staining of specimens are variable, comparison of features with adjacent nonhyperplastic gland nuclei may be helpful.

a significant number of false-negatives and false-positives resulted from this classification rule, particularly when the study set included other types of endometrial proliferations (64,68,69). Norris et al. (70) examined DNA light absorbance and added epithelial cellularity to the features examined. Although DNA content was not helpful, the addition of epithelial cellularity to the nuclear features improved prognostication.

Baak et al. (66,69) have performed exhaustive studies, noting that a combination of architectural and nuclear features improved the ability to discriminate lesions, which progressed from those that failed to progress to carcinoma. Using a relatively comprehensive list of 22 architectural and nuclear features, linear stepwise regression, and discriminant analyses, they concluded that volume percent stroma, standard deviation of the shortest nuclear axis, and the outer surface density of the glands are the most important discriminant factors (D-score). In a follow-up study (71), using another set of 55 biopsies and curettings with follow-up hysterectomy averaging 3 months after initial sampling, this morphometric assessment was both relatively sensitive and specific.

During the last 10 years, Mutter, Baak, and colleagues (72) have attempted to determine whether the morphometric data distinguish clonal lesions from those that are polyclonal. They initially determined that endometrial hyperplasia included both poly-clonal and clonal proliferations using a polymerase chain reaction assay for nonrandom X chromosome inactivation (73). Subsequent studies demonstrated that the clonal lesions were highly correlated with a morphometrically identifiable subset of lesions

Fig. 10. Typical hyperplasia. In this hyperplasia, the nuclei retain a stratified arrangement, with dense chromatin and inconspicuous or absent nucleoli.

with a low D-score, and that these could be recognized subjectively by pathologists by having a volume percent stroma of less than 55% (74). They suggest classifying such lesions as endometrial intraepithelial neoplasia (EIN) (75) in contrast to those poly-clonal lesions with different morphology and morphometry, which they would consider to represent endometrial hyperplasia. This subject is discussed in more detail in Chapter 6. To date, the reproducibility of this distinction made subjectively by pathologists has not been formally evaluated.

It is evident that in spite of efforts to provide rigorous histological criteria that repro-ducibly would distinguish various categories of endometrial cancer precursors, such criteria still elude pathologists. If a morphological classification scheme is to fulfill these criteria, it almost certainly will have to be based on correlations with clonality, morphometry, gene expression or protein expression profiles as well as outcome data. Further, it is evident that the molecular alterations and pathways involved in endometrial carcinogenesis need to be correlated with histological features.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook


Post a comment