Objective: The aim of this study was to determine whether an endometrial thickness less than 5 mm on transvaginal ultrasound (TVUS) is sufficient to exclude benign endometrial lesions in postmenopausal women with bleeding and to determine a cutoff value below which benign endometrial pathology could be ruled out.
Methods: Electronic medical records of consecutive postmenopausal women presenting with vaginal bleeding suspicious for benign pathology were reviewed between September 2002 and December 2007. All women underwent TVUS with endometrial stripe measurement followed by saline infusion sonography (SIS). Accuracy of endometrial echo thickness for detecting intracavitary masses was compared with the reference standard of SIS. A receiver operating characteristic curve was constructed to calculate whether other cutoff values would be more accurate than 5 mm in detecting benign endometrial masses.
Results: A total of 1,097 women were referred during the study period; 135 met the inclusion criteria and underwent TVUS followed by SIS. The endometrial echo was less than 5 mm in 43% and 5 mm or greater in 57%. The overall prevalence of polyps or fibroids was 50%. Using an endometrial echo cutoff less than 5 mm, sensitivity was 76% (95% CI, 65-85), specificity was 63% (95% CI, 51-73), positive predictive value was 67%, and negative predictive value was 72%. The area under the receiver operating characteristic curve for detection of benign masses was 0.79 (95% CI, 0.72-0.87). We were unable to determine a cutoff value below which benign endometrial pathology could be excluded.
Conclusions: With an endometrial thickness cutoff of 5 mm a considerable amount of benign endometrial pathology in postmenopausal women with bleeding is missed, and SIS or hysteroscopy may be warranted.