Endometrial surveillance in tamoxifen users: role, timing and accuracy of hysteroscopic investigation: observational longitudinal cohort study

Aims To determine the role of endometrial hysteroscopy in relation to the ultrasound, clinical and histological features of the endometrium during tamoxifen(TAM) use. Method We performed an observational longitudinal cohort study (2007–2012) that investigated the endometrium of 151 TAM users with hysteroscopy and histology. For all patients, ultrasound endometrial thickness(ET), gynecological history, years of adjuvant treatment and indications for hysteroscopy were recorded. Results Hysteroscopic findings showed that 100% of patients referred for simple follow-up had no evidence of endometrial disease. A strong correlation was found between history of abnormal uterine bleeding (with or without endometrial thickening) and hysteroscopic suspicion of endometrial atypia, confirmed by histology. Hysteroscopy had 83.3% sensitivity, 99% specificity, 83.3% positive-predictive-value(PPV) and 99% negative-predictive-value(NPV) in detecting endometrial atypia. No correlation was found between ET>5mm without bleeding and histological atypia. Similarly, the duration of treatment was not related to ET and histological atypia. Endometrial stromal hyperplasia was detected by histology in 70.5% of patients with ET between 5 and 10mm. In contrast, no atypia was detected when ET was <5mm. Ultrasound performed using a 5mm cut-off threshold for ET showed 100% sensitivity, 15% specificity, 4% PPV and 100% NPV in detecting endometrial atypia, while a 10mm cut-off threshold resulted in 84% sensitivity, 69% specificity, 10% PPV and 99% NPV. Conclusion Low-risk TAM users do not require different endometrial surveillance than general population. Hysteroscopy could play a fundamental role in determining the endometrial status of patients before the initiation of TAM treatment and in assessing the endometrial status of patients when bleeding occurs.