PREDICT-EC: A risk prediction tool to streamline diagnostic pathways for suspected endometrial cancer

RESEARCH QUESTION Can a personalised risk score guide decisions about invasive testing for suspected endometrial cancer that is clinically effective, equitable, economical and acceptable to patients and clinicians? BACKGROUND In the UK, nearly 10,000 women are diagnosed with endometrial cancer/year. The red flag symptom for endometrial cancer is postmenopausal bleeding (PMB). PMB is extremely common and only 5% of sufferers have sinister underlying pathology. A woman experiencing PMB is referred urgently to gynaecology clinic on the 'suspected cancer' pathway, where she is offered a transvaginal scan (TVS) followed by invasive tests if the scan fails to reassure, specifically hysteroscopy and/or endometrial biopsy. The risks of these invasive tests include anxiety, repeated hospital visits, bleeding, pain, infection and rarely, uterine perforation. A tool to select women at high risk of endometrial cancer for invasive tests while safely reassuring low-risk women, would improve patient care. It could save thousands of women every year in the UK alone from the psychological and physical burden of invasive testing and ensure a more effective use of healthcare resources. We have developed a risk prediction tool called PREDICT-EC to select women with PMB for invasive tests for endometrial cancer. It incorporates age, BMI, recurrent PMB, ever use of hormonal contraceptives, TVS findings and microscopic haematuria. We showed that PREDICT-EC would identify 99/100 women with endometrial cancer and avoid invasive tests in an additional 24/100 women without endometrial cancer, compared to standard care. AIMS AND OBJECTIVES This 5-year research programme aims to externally validate PREDICT-EC for NHS implementation by: 1. Establishing its clinical performance in women with suspected endometrial cancer; 2. Comparing its accuracy across important patient and disease parameters; 3. Characterising current NHS resource use and costs of its use in clinical practice; 4. Exploring the experiences of women, patients and clinicians; 5. Developing a clinician-facing digital risk score calculator for NHS use. METHODS We will externally validate PREDICT-EC through prospective observation of 3,000 women referred to secondary care with suspected endometrial cancer at 10 hospital sites. We will track participants on the standard diagnostic pathway, recording demographics, clinic visits, test results and clinical outcomes. Alongside routine care, we will calculate PREDICT-EC risk scores whilst blinded to final diagnoses. We will assess the predictive performance of the tool and compare its accuracy across important patient (age, ethnicity, socio-economic status) and disease parameters (atypical hyperplasia, type-1 vs type-2 cancers, FIGO stage). We will look at current NHS resource use and the costs of embedding PREDICT-EC into clinical practice. Using qualitative research methods, we will garner the experiences of women from diverse socio-cultural backgrounds, and the views of healthcare professionals from primary/secondary care. Finally, we will design a clinician-facing digital calculator that generates risk scores for NHS implementation. ANTICIPATED IMPACT AND DISSEMINATION This research has the potential to deliver a new model of care for women with suspected endometrial cancer that could be adopted by the NHS and rolled out internationally. We will share results via patient groups and social media, and publish in scientific journals and conferences.