A longitudinal evaluation of postmenopausal bleeding and transvaginal sonographic measurement of the endometrium as predictors of endometrial cancer

• Published in: Ultrasound in obstetrics & gynecology Vol. 18; no. s1; p. CEG‐06
• Main Authors: Gull, B. E., Karlsson, B., Milsom, I., Granberg, S.
• Format: Journal Article
• Language: English
• Published: Oxford, UK Blackwell Science, Ltd 01.10.2001
• ISSN: 0960-7692; 1469-0705
• DOI: 10.1046/j.1469-0705.2001.abs09-4.x

Objective:  To evaluate postmenopausal bleeding (PMB) and transvaginal sonographic (TVS) measurement of endometrial thickness (ET) as predictors of endometrial cancer (EC) and atypical hyperplasia (AH) in women during a >10‐year period following a PMB. Study design:  Women presented with a PMB from November 1987 to October 1990 were included in this study (n = 394). The women underwent TVS with measurement of ET and a dilation and curettage (D & C). It was possible to assess the medical records of 339 of the 394 women (86%) >10 years after referral for PMB. During the follow up period the recurrence of a PMB, the development of EC and mortality were assessed. Results:  After the primary investigation, 39 of the 339 women were diagnosed as having EC (11.5%) and 5 women had AH (1.5%). The relative risk (RR) of diagnosing an EC in women referred for a PMB was 63.9 (CI 46.0 Ð 88.8) and the corresponding RR for EC and AH together was 72.1 (CI 52.8 Ð 98.5) compared to women of the same age from the same region of Sweden. The reliability of PMB as a diagnostic test for EC was assessed: sensitivity 18%; specificity 100%; positive predictive value 12%; and negative predictive value 100%. None of the women with an ET of 2 × 4 mm were diagnosed as having EC. The reliability of ET (cut‐off 2 × 4 mm) as a diagnostic test for EC was assessed: sensitivity 100%; specificity 60%; positive predictive value 25% and negative predictive value 100%. The incidence of EC or AH in women with an intact uterus followed 3 × 10 years was 5.8%. The corresponding figure for women who had 3 × 1 recurrent bleeding during follow up was 22.7%. No EC was diagnosed in women with a recurrent PMB who had an ET of 2 × 4 mm at the initial scan. No EC was diagnosed in the absence of a recurrent bleeding. Conclusion:  PMB incurs a 64‐fold increase risk for EC. No EC was missed when ET measurement (cut‐off 2 × 4 mm) was used even if the women were followed 3 × 10 years. About 5.8% of the women referred for a PMB later developed EC or atypia during the 10‐year observation period. There was no increased risk of EC or AH in women who did not have recurrent bleedings whereas women with a recurrent bleeding were a high risk group. A thick endometrium, with histopathological diagnosis of atrophy or insufficient for diagnosis, should be taken seriously and a guided biopsy should be taken or saline instillation sonography performed with ultrasound‐guided biopsy.