Prospective longitudinal study of ultrasound screening for endometrial abnormalities in women with breast cancer receiving tamoxifen

Abstract

Objective The goal of this work was to study the role of transvaginal ultrasonography (TVUS) together with colorflow Doppler imaging (CFDI) in the detection of significant endometrial abnormalities induced by tamoxifen. Methods Over a 6-year period, 304 women on tamoxifen as adjuvant therapy for breast cancer were recruited into the current study. Standard demographic data as well as duration of tamoxifen use were collected. Patients were assessed at study entry and at yearly intervals with TVUS together with CFDI. All patients had an endometrial biopsy at the time of study entry, and repeat endometrial evaluations were done subsequently only if there were abnormal ultrasound findings or the presence of irregular vaginal bleeding. All ultrasonic characteristics and Doppler flow measurements were recorded. Descriptive statistics were used to describe the study group. Logistic regression was used to identify significant treatment- and ultrasound-related factors associated with the presence of significant uterine pathology. Results One thousand and sixty-one ultrasound assessments were performed on 304 patients over a 6-year period. The mean age was 52.33 (range, 29–79). Seventy-two percent of the patients were postmenopausal at the time of breast cancer diagnosis. The median concentrations of estrogen and progesterone receptor were 75 and 73 fmol/L, respectively. Fifty-eight percent of the patients had received cytotoxic chemotherapy. The mean duration of tamoxifen use was 48.2 months. Thirty-two percent of the ultrasound examinations had associated significant uterine pathology defined as conditions that required further medical or surgical investigation and treatment. However, 80% of the abnormalities represented benign polyps. Six cases of primary endometrial cancer were detected. All cases presented with irregular bleeding. No recurrence of disease was detected at a median follow-up of 48 months. One case of metastatic breast cancer to the uterus was encountered. By setting the endometrial thickness cutoff at more than 9 mm to represent significant abnormality in this patient population, the sensitivity was 63.3%, specificity was 60.4%, positive predictive value was 43.3%, and negative predictive value was 77.5%. To detect endometrial cancer, the endometrial thickness cutoff at 9 mm had a positive predictive value of only 1.4%. Logistic regression analysis showed only endometrial thickness greater than 9 mm (OR 3.99, CI = 1.26–12.65, P = 0.018) and spiral artery pulsatility index measurement (OR 4.18, CI = 1.25–13.92, P = 0.02) to be associated with significant uterine abnormalities. Conclusions Routine sequential ultrasound surveillance in asymptomatic women on tamoxifen is not useful because of its low specificity and positive predictive value. A significant portion of screened asymptomatic women would need to undergo needless surgical evaluations of their endometrium if widespread use of ultrasound is implemented in this patient population.