Can ultrasound replace dilation and curettage? A longitudinal evaluation of postmenopausal bleeding and transvaginal sonographic measurement of the endometrium as predictors of endometrial cancer

Abstract

Objective: The purpose of this study was to evaluate postmenopausal bleeding and transvaginal sonographic measurement of endometrial thickness as predictors of endometrial cancer and atypical hyperplasia in women whose cases were followed for ≥10 years after referral for postmenopausal bleeding. Study design: Women (n = 394) who had postmenopausal bleeding from November 1987 to October 1990 underwent transvaginal sonographic measurement of endometrial thickness and curettage. It was possible to assess the medical records (regarding recurrence of a postmenopausal bleeding, development of endometrial cancer, and death) in 339 of the 394 women (86%) ≥10 years after referral for postmenopausal bleeding. Results: Thirty-nine of the 339 women (11.5%) had endometrial cancer, and 5 women (1.5%) had atypical hyperplasia. The relative risk of endometrial cancer in women who were referred for postmenopausal bleeding was 63.9 (95% CI, 46.0-88.8); the corresponding relative risk for endometrial cancer and atypical hyperplasia together was 72.1 (95% CI, 52.8-98.5) compared with women of the same age from the general population of the same region of Sweden. No woman with an endometrial thickness of ≤4 mm was diagnosed as having endometrial cancer. The relative risk of the development of endometrial cancer in women with an endometrial thickness of >4 mm was 44.5 (95% CI, 6.5-320.1) compared with women with an endometrial thickness of ≤4 mm. The reliability of endometrial thickness (cutoff value, ≤4 mm) as a diagnostic test for endometrial cancer was assessed: Sensitivity, 100%; specificity, 60%; positive predictive value, 25%; and negative predictive value, 100%. The incidence of endometrial cancer or atypical hyperplasia in women with an intact uterus whose cases had been followed for ≥10 years was 5.8% (15/257 women) compared with 22.7% (15/66 women) in women who had ≤1 episode of recurrent bleeding. No endometrial cancer was diagnosed in women with a recurrent postmenopausal bleeding who had an endometrial thickness of ≤4 mm at the initial scan. Conclusion: Postmenopausal bleeding incurs a 64-fold increase risk for endometrial cancer. There was no increased risk of endometrial cancer or atypia in women who did not have recurrent bleeding, whereas women with recurrent bleeding were a high-risk group. No endometrial cancer was missed when endometrial thickness measurement (cutoff value, ≤4 mm) was used, even if the women were followed up for ≤10 years. We conclude that transvaginal sonographic scanning is an excellent tool for the determination of whether further investigation with curettage or some form of endometrial biopsy is necessary (Am J Obstet Gynecol 2003;188:401-8.)