Performance of the WID-qEC test versus sonography to detect uterine cancers in women with abnormal uterine bleeding (EPI-SURE): a prospective, consecutive observational cohort study in the UK

Abstract

Background: To detect uterine cancer, simpler and more specific index tests are needed to triage women with abnormal uterine bleeding to a reference histology test. We aimed to compare the performance of conventional index imaging tests with the novel WID-qEC DNA methylation test in terms of detecting the presence or absence of uterine cancers in women with abnormal uterine bleeding.

Methods: EPI-SURE was a prospective, observational study that invited all women aged 45 years and older with abnormal uterine bleeding attending a tertiary gynaecological diagnostic referral centre at University College London Hospital (London, UK) to participate. Women meeting these inclusion criteria who consented to participate were included. Pregnant women and those with previous hysterectomy were excluded. A cervicovaginal sample for the WID-qEC test was obtained before standard assessment using index imaging tests (ie, ultrasound) and, where applicable, reference histology (ie, biopsy, hysteroscopy, or both) was performed. Technicians performing the WID-qEC test were masked to the final clinical outcome. The result of the WID-qEC test is defined as the sum of the percentage of fully methylated reference (ΣPMR) of the ZSCAN12 and GYPC regions. Patients were followed until diagnostic resolution or until June 12, 2023. The primary outcome was to assess the real-world performance of the WID-qEC test in comparison with ultrasound with regard to the area under the receiver-operating-characteristic curve (AUC), sensitivity, specificity, and positive and negative predictive values. EPI-SURE is registered with ISRCTN (16815568).

Findings: From June 1, 2022, to Nov 24, 2022, 474 women were deemed eligible to participate. 74 did not accept the invitation to participate, and one woman withdrew after providing consent. 399 women were included in the primary analysis cohort. Based on 603 index imaging tests, 186 (47%) women were recommended for a reference histology test (ie, biopsy, hysteroscopy, or both). 12 women were diagnosed with cancer, 375 were not diagnosed with cancer, and 12 had inconclusive clinical outcomes and were considered study dropouts. 198 reference histology test procedures detected nine cases of cancer and missed two; one further cancer was directly diagnosed at hysterectomy without a previous reference test. The AUC for detection of uterine cancer based on endometrial thickness in mm was 87·2% (95% CI 71·1-100·0) versus 94·3% (84·7-100·0) based on WID-qEC (p=0·48). Endometrial thickness assessment on ultrasound scan was possible in 379 (95%) of the 399 women and a prespecified cut-off of 4·5 mm or more showed a sensitivity of 90·9% (95% CI 62·3-98·4), a specificity of 79·1% (74·5-82·9), a positive predictive value of 11·8% (6·5-20·3), and a negative predictive value of 99·6% (98·0-99·9). The WID-qEC test was possible in 390 (98%) of the 399 patients with a sensitivity of 90·9% (95% CI 62·3-98·4), a specificity of 92·1% (88·9-94·4), a positive predictive value of 25·6% (14·6-41·1), and a negative predictive value of 99·7% (98·3-99·9), when the prespecified threshold of 0·03 ΣPMR or more was applied. When a higher threshold (≥0·3 ΣPMR) was applied the specificity increased to 97·3% (95% CI 95·1-98·5) without a change in sensitivity.

Interpretation: The WID-qEC test delivers fast results and shows improved performance compared with a combination of imaging index tests. Triage of women with abnormal uterine bleeding using the WID-qEC test could reduce the number of women requiring histological assessments for identification of potential malignancy and specifically reduce the false positive rate.

Funding: The Eve Appeal, Land Tirol, and the European Research Council under the European Union's Horizon 2020 Research and Innovation Programme.