Ultrasound detection of endometrial cancer in women with postmenopausal bleeding: Systematic review and meta-analysis

Abstract

Objective: To assess the performance of endometrial thickness (ET) cut-offs for detecting endometrial cancer (EC) in women with postmenopausal bleeding (PMB) and evaluate the clinical utility of additional ultrasound measures such as endometrial volume (EV), vascular flow index (VFI), vascularization index (VI), and uterine artery flow index (FI).

Methods: Clinicaltrials.gov and MEDLINE database via PubMed were queried for studies published between 1/1990 and 3/2016 using specific MeSH terms. Original, peer-reviewed cohort studies reporting EC outcomes and specific ultrasound findings by PMB status were included.

Results: Study design, country, clinical setting inclusion/exclusion criteria, aggregate study-level demographic and clinical data were extracted from 44 studies including 17,339 women with PMB and 1341 cases of EC (7.7%). In women with PMB and EC (n = 417), pooled mean ET was 16.4 mm (95% CI, 14.8-18.1 mm). In women with PMB without EC, pooled mean ET was 4.1 mm. 31 studies reported outcomes using different ET cut-off values ranging from 3 to 20 mm. Compared to ≥3 or 4 mm, a cutoff of ≥5 mm had similar sensitivity (96.2, 95%CI 92.3, 98.1) with improved specificity for EC (51.5, 95%CI 42.3-60.7), allowing to reduce the rate of invasive workup for PMB by 17%. EV, VI, VFI, and FI were significantly correlated with EC, but performance of specific cut-offs was not analyzed due to limited data.

Conclusion: Among women with PMB mean ET is substantially higher in women with EC compared to those without EC. An ET cutoff of ≥5 mm shows an acceptable tradeoff between sensitivity and specificity for diagnosis of EC.

Keywords: Endometrial cancer; Postmenopausal bleeding; Ultrasound.