Evaluation of DNA Methylation Markers for Endometrial Cancer Risk-stratification Using Patient-collected Urine and Vaginal Samples and Clinician-collected Cervical Samples From Women With Postmenopausal Bleeding

BACKGROUND:

Endometrial cancer (EC) incidence is rising and is the fourth most common cancer type in women worldwide with about 417,000 new cases and 90,000 deaths annually. Moreover, EC is the most common gynaecological cancer in developed countries including Denmark with about 800 new cases and 90 deaths yearly. Early diagnosis is essential since the prognosis for women with an advanced-stage EC is poor. No screening exists for EC and the most commonly observed clinical symptom is post-menopausal bleeding (PMB), which is a common condition affecting about 11% of postmenopausal women. Because 90% of EC cases, are preceded by PMB, referral for specialized gynaecological care through cancer patient pathways is indicated for all Danish women with PMB, causing anxiety among women and high healthcare costs. However, only 5-10% of women presenting with PMB have EC. Worldwide, transvaginal ultrasound (TVUS) evaluation of the endometrium (cut-off: thickness ≥5mm and/or irregular lining of the uterine cavity) is used to identify women with PMB who have the highest cancer risk and therefore require an endometrial biopsy with or without hysteroscopy. While TVUS evaluation is highly sensitive (94.8%) it suffers from low specificity (51%) causing high numbers of unnecessary invasive procedures in healthy women. Moreover, the diagnostic program depends on skilled specialists for ultrasound evaluation, endometrial sampling as well as histopathological examination. This reinforces the clinical urgent and unmet need for a simpler, non-invasive, user-provided and more specific test to aid EC diagnosis.

This study aims to determine whether DNA methylation testing in patient-collected urine and vaginal samples is a clinically safe non-invasive alternative to determine EC risk. Testing for elevated DNA hypermethylation levels of cancer-specific genes has been demonstrated to provide a promising biomarker for detection of early-stage cancers, including cervical and endometrial cancer. DNA hypermethylation has been linked to silencing of tumour suppressor genes, thereby contributing to cancer development. Biologically, the ability to use vaginal and urine material for EC detection by DNA methylation testing is explained by local shedding of EC-derived DNA and cell fragments into the vaginal debris and urine as well as transrenal excretion of tumour-shed circulating DNA into the urine.

Mainly small proof-of-concept and case-control studies have shown feasibility for detecting EC using DNA methylation testing in plasma, self-collected vaginal samples, and clinician-collected cervical scrapes. However, the first paper reporting on the diagnostic value of DNA methylation analysis in vaginal samples and urine as a liquid biopsy for EC was just published by co-supervisor Prof. Steenbergen last year. Here, case-control data revealed excellent diagnostic performance of EC-specific DNA methylation markers in patient-collected urine and vaginal samples with sensitivities of 89-90% and specificities of 90-92% for EC detection using marker panels including: CDH13, GHSR, SST, CDO1, and ZIC1.

Only one prospective cohort study has compared the diagnostic performance of DNA methylation testing in clinician-collected cervicovaginal samples with the conventional TVUS evaluation in terms of detecting the presence or absence of EC in women aged 45+ with abnormal uterine bleeding. The results demonstrated the WID-qEC DNA methylation test (markers: ZSCAN12 and GYPC) to have equal sensitivity (90.9%) and superior specificity (92.1 vs 79.1%) as compared with TVUS evaluation.

However, this will be the first study to compare DNA methylation testing in patient-collected urine and vaginal samples with TVUS evaluation.

STUDY DESIGN: Comparative diagnostic accuracy study, comparing diagnostic test methods at the time of diagnosis as well as the long-term safety after two years.

MATERIALS AND METHODS:

The vaginal, urine and cervical samples will be stored in a biobank at the Dept. of Pathology, Randers Regional Hospital (RHR) until shipment for DNA methylation testing analysis using quantitative methylation-specific PCR (qMSP) in Prof. Steenbergen's laboratory at Amsterdam University Medical Centre, The Netherlands. All samples will be tested for methylation markers, three for vaginal samples (CDO1, GHSR, ZIC1), three for urine (GHSR, CDH13, SST) and three for cervical samples (CDH13 + CDO1 + ZIC1). Results on TVUS evaluation and endometrial biopsies will be retrieved from the nationwide Danish Pathology Databank and the electronic patient journal linked to the patient through the Civil Registration System in Denmark, where all residents have a unique 10-digit code (CPR number), to which all health data is linked.

AIMs:

1. To determine the diagnostic accuracy of DNA methylation testing in patient-collected full-void urine and vaginal samples, as well as clinician-collected cervical samples to detect EC among women with postmenopausal bleeding, and 2) to compare the diagnostic test accuracy of DNA methylation with the current diagnostic evaluation with TVUS.

HYPOTHESES:

1. DNA methylation testing in urine samples can detect EC with comparable sensitivity and superior specificity compared with TVUS and with higher accuracy than DNA methylation testing in vaginal and cervical samples.
2. DNA methylation analysis in urine samples will have a long-term negative EC risk after 2 years equivalent to TVUS, but better than vaginal and cervical samples.

RESEARCH PLAN AND FEASIBILITY The study will be conducted in collaboration between University Clinic for Cancer Screening, Department of Public Health Programmes, Department of Pathology, RHR and with gynaecological departments in Central Denmark (Randers and Aarhus) and Southern Denmark Region (Odense) working as inclusion sites. State-of-the-art laboratory facilities are in place at Prof. Steenbergen's laboratory at Amsterdam UMC, supporting the feasibility of the study.

PERSPECTIVES AND COMMUNICATION If successful, the research will have international impact and result in a novel diagnostic tool that will differentiate between women with PMB who do or do not need specialist referral through cancer patients' pathways to have invasive endometrial biopsies. This will facilitate timely diagnosis, lower the anxiety among patients, and reduce the pressure on the healthcare system. Results will be published in international peer-reviewed journals and presented at conferences and will be concluded with a PhD dissertation. The results will be distributed to relevant clinical fora and implemented in national as well as international guidelines.