Can Staging Magnetic Resonance Imaging (MRI) Features Prognosticate Patients Presenting With Endometrial Cancer?

Endometrial Cancer, Endometrial Neoplasms

MRI, Radiomics, Texture features, Prognostication, Endometrial cancer

Aim: Assess the value of MRI features in predicting prognosis in patients with endometrial cancer This study will examine the MRI features of women with confirmed endometrial cancer to see if textural features can prognosticate patients.

Aim: Assess the value of MRI features in predicting prognosis in patients with endometrial cancer Background: Endometrial cancer is the commonest gynaecological cancer in the United Kingdom. MR Imaging is used to provide pre-treatment local staging in Endometrial Cancer based on the FIGO staging system. This MRI stage is used by clinicians, in combination with pre-operative histology and clinical assessment of the patient, to counsel patients about their prognosis. Currently the information obtained from the MRI study is largely anatomical and concerns the extent of local tumour spread at the time of diagnosis. Each stage under the FIGO system is associated with a prognosis based on data from the literature. Important prognostic indicators include the depth of tumour invasion in the uterine muscle (myometrium), invasion of the cervical stroma by tumour, spread to lymph node tissues and the histological sub-type of tumour present. However, with ever improving MRI technology there are now newer sequences employed as part of the staging study which may potentially yield more information. In other cancer types, e.g. prostate, breast and liver, textural analysis of the tumour on pre-treatment imaging (CT/PET-CT/MRI) is being analysed to try and identify features which could be used as prognostic markers for patient outcome. The Study: All patients with confirmed cancer on histology will receive an Ultrasound and MRI as part of routine care. The MRI scan will include T1, T2, DWI and contrast enhanced sequences. Patients referred to Hammersmith Hospital from other trusts who are diagnosed with endometrial cancer will be approached for consent and enrolment into latter part of study. These patients will have had an MRI but may not have had all the sequences needed and may therefore require a further MRI scan with additional sequences. These images will be saved and analysed. Pathological findings (e.g. type and grade of cancer) will be correlated with the radiological results. Definitive management of endometrial cancer includes a hysterectomy for most patients. Following surgery patients will continue with routine standard of care and follow up. Patients will be followed up for 5 years to provide 5 year survival data. Images from ultrasound and MRI pelvis will be analysed to see if they can predict the 5 year survival. This prospective study will also be used to validate a retrospective model using MRI textural features in endometrial cancer patients.

United Kingdom