Approximately 75% of endometrial cancer occurs in women older than 55 yr of age. Postmenopausal bleeding is often considered endometrial cancer until proven otherwise. One diagnostic challenge is that endometrial biopsy or curettage generally yields limited samples from elderly patients. There are no well-defined and unified diagnostic criteria for adequacy of endometrial samples. Pathologists who consider any sample including those lacking endometrial tissue as "adequate" run the risk of rendering false-negative reports; on the contrary, pathologists requiring ample endometrial glands along with stroma tend to designate a greater number of samples as "inadequate," leading to unnecessary follow-up. We undertook a quantitative study of 1768 endometrial samples from women aged 60 yr and older aiming to propose validated adequacy criteria for diagnosing or excluding malignancy. Using repeat-procedure outcomes as reference, we found that samples exceeding 10 endometrial strips demonstrated high negative predictive value close to 100%. Such samples can be scant, yet appear to be sufficient in excluding malignant conditions. When tissue diminished to <10 strips, negative predictive value dropped significantly to 81%. The risk of undersampled malignancy rose to 19%. Among 274 malignant cases, only 4 cases yielded limited tissue yet >10 strips. In conclusion, we propose 10 endometrial strips as the minimum for adequate samples from postmenopausal women. Applying such validated adequacy criteria will greatly reduce false-negative errors and avoid unnecessary procedures while ultimately improving diagnostic accuracy. Our criteria may serve as a reference point in unifying the pathology community on this important and challenging topic.