The depth of myometrial invasion (DMI) is one of the most important prognostic indicators and determinants of therapy in endometrial cancer. There are well-documented problems in recognizing DMI. We examined 100 previously diagnosed endometrioid endometrial carcinomas in hysterectomy specimens, reassessed DMI, and explored morphological features that complicated appraisal of myometrial invasion. The DMI was different from the original measurement in 29% of cases. Twelve percent of all cases (40% of cases with measurement discrepancies) involved differences in the assignment of invasion categories (noninvasive, < or =50% myometrial invasion, and >50% myometrial invasion). Nearly all endometrial cancers originally diagnosed as invasive were considered noninvasive on review. We examined whether the distribution of stromal metaplasia, noninvasive patterns (exophytic tumors, irregular endomyometrial junctions, and adenomyosis), and myometrial invasion patterns were different in cases with and without measurement discrepancies. Irregular endomyometrial junctions, exophytic tumors, and adenomyosis tended to coexist and were more common in the cases with DMI discrepancies. Although there seemed to be a relationship between smooth muscle metaplasia and exophytic tumors, it did not appear that smooth muscle metaplasia was significantly more common in cases with measurement difficulties. However, cases with extensive smooth muscle metaplasia posed problems with assessment of myometrial invasion. Patterns of myometrial invasion other than the conventional destructive pattern were sufficiently uncommon as to not impact on DMI measurement in large numbers of cases. Measuring the DMI is usually uncomplicated, but additional scrutiny should be paid to cases involving exophytic tumors, irregular endomyometrial junctions, adenomyosis, and extensive stromal smooth muscle metaplasia.