The incidence of endometrial cancers diagnosed on biopsy that have no residual cancer identified at hysterectomy is not well studied. The aim of our study was to determine the incidence and long-term follow-up of this "vanishing cancer" phenomenon. All slides from the initial biopsy/curettage and hysterectomy specimens were reviewed and the diagnosis confirmed by a gynecologic pathologist. The entire endometrium was serially sectioned and submitted for histologic examination. Clinical and pathologic variables were analyzed, including patient demographics, tumor histologic type and grade, stage, biopsy method, adjuvant therapy, surgical procedure, recurrence, and disease-specific survival. We identified 23 biopsy-proven cases of endometrial cancer with no residual disease on hysterectomy specimen. Of the 23 patients, 15 (65.2%) were diagnosed as endometrioid, 6 (26%) serous, 1 clear cell (4.3%), and 1 (4.3%) serous intraepithelial carcinoma. Seventeen underwent dilatation and curettage, and 6 had endometrial biopsy as the primary procedure. The median follow-up was 8.8 years (range, 1.2 to 17 y). Only 2 cases with serous carcinoma underwent adjuvant chemotherapy, and none received radiation therapy. Only 1 patient died of disease after 27 months and was diagnosed as FIGO grade II endometrioid carcinoma on dilatation and curettage. The inability to identify cancer in a hysterectomy specimen for biopsy-confirmed carcinoma does not indicate technical failure. Although there is no specific standard treatment for patients with "vanishing endometrial cancer," the prognosis is excellent; however, close follow-up is suggested.