Tamoxifen has been shown to decrease the recurrence rate of breast cancer. Evidence that tamoxifen use may be associated with an increased risk of endometrial cancer has caused investigators to recommend routine invasive sampling. We have assessed a minimally invasive alternative for endometrial surveillance of tamoxifen-treated patients utilizing transvaginal ultrasound and saline infusion sonohysterography. Asymptomatic women (n = 44) with breast cancer on postoperative tamoxifen treatment were referred to our gynecological ultrasound unit. Initially, the endometrial echo was measured with unenhanced transvaginal ultrasound. If a distinct echo measured < or = 5 mm, no further procedure was performed. For thickened or inadequately visualized echoes, sonohysterography was performed. If a thin echo was noted on sonohysterography, no further procedure was performed. If focal changes were detected, hysteroscopy/dilatation and curettage (D&C) was performed. For generalized symmetrically thickened echoes, a blind biopsy was done. If sonohysterography was unsuccessful, hysteroscopy/D&C was performed. Eleven (25%) patients had thin unenhanced echoes of < or = 5 mm. Twenty-five (57%) patients had thickened endometrial echoes. Three (7%) had naturally occurring endometrial fluid outlining a polyp. An endometrial echo could not be visualized in five (11%) patients. Sonohysterography was successfully performed in 21 out of 30 (70%) patients with either thickened or non-visualized unenhanced echoes. Of these, two patients had thin endometria with coexisting myomas; seven had thin endometria with typical tamoxifen-induced subendometrial changes: and seven had focal polypoid thickening confirmed by hysteroscopy/D&C. Histology revealed carcinoma associated with two, proliferation in one and four polyps. Five patients had thickened unenhanced echoes with symmetrically thickened single-layer measurements on sonohysterography. Histology revealed that three were proliferative, one was inactive and one was hyperplastic. In the nine patients with unsuccessful sonohysterography, hysteroscopy/D&C revealed inactive endometria in six, and three polyps. Our paradigm of evaluating the endometrial response to tamoxifen is concluded to overcome the shortcomings of either unenhanced transvaginal ultrasound or blind biopsy alone while it kept the number of invasive sampling procedures to 55% (24 out of 44).