Meeting Banner
Abstract #0723

Development of MRI-based 3D Radiomics Signatures for Preoperative Risk Stratification of Patients with Histology-proven Endometrial Cancer

Thierry L. Lefebvre1,2, Yoshiko Ueno3,4, Anthony Dohan5,6, Avishek Chatterjee1,7, Martin Vallières1,8, Eric Winter-Reinhold9, Sameh Saif3, Ives R. Levesque1,10, Xing Ziggy Zeng11, Reza Forghani3,9, Jan Seuntjens1, Philippe Soyer5,6, Peter Savadjiev3,12, and Caroline Reinhold3,9
1Medical Physics Unit, McGill University, Montreal, QC, Canada, 2Cancer Research UK Cambridge Institute, University of Cambridge, Cambridge, United Kingdom, 3Department of Diagnostic Radiology, McGill University, Montreal, QC, Canada, 4Department of Radiology, Kobe University Graduate School of Medicine, Kobe, Japan, 5Department of Radiology, Cochin Hospital, AP-HP.Centre, Paris, France, 6Faculté de Médecine, Université de Paris, Paris, France, 7Department of Precision Medicine, GROW - School for Oncology and Developmental Biology, Maastricht University, Maastricht, Netherlands, 8Department of Computer Science, Université de Sherbrooke, Sherbrooke, QC, Canada, 9Augmented Intelligence & Precision Health Laboratory, Research Institute of McGill University Health Centre, Montreal, QC, Canada, 10Research Institute of McGill University Health Centre, Montreal, QC, Canada, 11Department of Obstetrics and Gynecology, McGill University Health Centre, Montreal, QC, Canada, 12School of Computer Science, McGill University, Montreal, QC, Canada


Radiomics analysis on standard MRI prior to surgery holds potential to help identifying high-risk histopathological features of endometrial carcinoma, including FIGO stage, deep myometrial invasion, lymphovascular space invasion and tumor grade, thus supporting preoperative risk stratification for optimal patient management. This dual-center retrospective study evaluated the role of radiomics to assess high-risk phenotypes of endometrial cancer in women who underwent 1.5-T MRI before hysterectomy. Radiomics-based machine learning models provided consistent clinically acceptable performance for differentiating early from advanced FIGO stage endometrial carcinoma and for differentiating low- from high-risk histopathological markers in two independent datasets from different institutions on preoperative MRI.

This abstract and the presentation materials are available to members only; a login is required.

Join Here