Background. At present, no proven recommendations can be made for the surveillance of tamoxifen-treated women. The aim of the present study was to evaluate ultrasonography and hysteroscopy in this setting. Methods. Three hundred and ten postmenopausal patients using tamoxifen underwent vaginal ultrasonography, hysteroscopy, and endometrial biopsy; 274 were asymptomatic and 49 had abnormal bleeding. Ultrasonographic endometrial thickness and echotexture were recorded. Hysteroscopic endometrial appearance, presence of focal endometrial lesions and polyps were also recorded. General or selective endometrial biopsy was performed. Ultrasonographic and hysteroscopic follow up was provided. Results. At ultrasonography, mean endometrial thickness was 10.8mm. At hysteroscopy, cystic atrophy and suspect focal lesions were detected in 49.2% and 5.3% of women, respectively. Polyps were present in 44.8% of women; 38.9% of these polyps were missed at ultrasonography, whereas 11.4% were suspected but were not found at hysteroscopy. At biopsy, non-atypical hyperplasia and atypical changes were found in 4.8% and 1.3% of patients, respectively. Three carcinomas were found, all in asymptomatic women. Logistic regression analysis showed that only suspect focal lesions at hysteroscopy were significantly associated with abnormal histology. With a 6-mm cut-off value for endometrial thickness, negative and positive predictive values for ultrasonography in detecting hyperplastic or neoplastic changes were 96% and 8%, respectively; the corresponding values for hysteroscopy were 96% and 65%. No additional carcinoma was found at follow up. Conclusions. No single ultrasonographic feature (echotexture and borders) is significantly associated with the detection of endometrial hyperplasia or carcinoma; hysteroscopy, although not predictive unless revealing a focal lesion, is more accurate in detecting polyps and hyperplastic changes.
ASJC Scopus subject areas
- Obstetrics and Gynaecology