Objective Biomarkers in Endometrioid-Type Endometrial Carcinogenesis

Endometrial hyperplasia (EH) is a common disease. An estimated 180,000–200,000 new cases occur annually in the Western world; approx 39,200 new cases and 6600 deaths were expected in the United States in 2002 (1). About 8–20% of all EH is associated with subsequent endometrial cancer of the endometrioid type. Major problems in treating EH include poor diagnostic reproducibility and inaccurate prediction of cancer progression. This has resulted in enormous overtreatment. The World Health Organization (WHO)’s 1994 histologic classification is widely used but is not reproducible, does not adequately predict the risk of cancer progression, and lacks a molecular and cell biology basis. Computerized morphometric analysis has identified a multivariate combination of architectural and nuclear features, called the DS, which is reproducible, fits with therapeutic options, and accurately predicts cancer outcome. Cases with DS ≥1 have a negligible progression risk (0.3%), contrasting with a 37% progression risk for those with DS < 1 found in a large multicenter study with more than 18 yr of follow-up. Moreover, molecular-genetic studies have found a {tly|464-1} strong correlation between clonality and DS; monoclonal cases nearly always have DS < 0; cases with DS ≥ 1 are mostly polyclonal. This has resulted in a new classification, endometrial intraepithelial neoplasia (EIN), which does not mimic subjective WHO hyperplasia diagnoses but rather uses new criteria for prognostic prediction of cancer endpoints. Moreover, high-risk EIN-DS lesions often show clonal inactivation of the phosphatase and tensin homologue (PTEN) tumor-supressor gene by immunohistochemistry. Based on EIN-DS, patients can now be reproducibly assigned with high accuracy to high- and low-risk categories, permitting appropriate management. EIN-DS application is easy, can be done on standard histologic sections, and has reasonable cost. EIN-DS thus can be assessed in any pathology laboratory, but sections can also be sent to reference laboratories for measurement if necessary.