Combining Unpaired Endometriosis Ultrasounds and Magnetic Resonance Imaging Using Artificial Intelligence, To Detect Pouch of Douglas Obliteration for Endometriosis Diagnosis: The IMAGENDO Study.
Monday, April 8, 2024
2:32pm – 2:39pm
Location: 412
Authors: Jodie Avery, University of Adelaide Yuan Zhang, The University of Adelaide Mathew Leonardi, McMaster University Hu Wang, The University of Adelaide Steven Knox, Benson Radiology/ Royal Adelaide Hospital Mary Louise Hull, University of Adelaide ,
Recent ESHRE guidelines for the diagnosis of endometriosis, have been updated, with clinicians now being encouraged to use imaging prior to traditional laparoscopy, and the best modalities found were transvaginal ultrasound or MRI. In clinical practice, it is difficult to access clinicians who can diagnose endometriosis with one of these modalities, not to mention those who are proficient in both modalities. IMAGENDO comprises of interdisciplinary researchers from the Robinson Research Institute and the Australian Institute of Machine Learning at the University of Adelaide, as well as other collaborators from Australia, the USA, Canada, the UK, Indonesia and New Zealand. Our model leverages the ultrasound POD obliteration detection to improve the automated detection accuracy from MRIs, using an unpaired training set containing scans from both modalities. To the best of our knowledge, this is the first detection method that distills knowledge from ultrasounds to MRI based algorithms using unpaired data, with the objective of improving the accuracy of diagnosing endometriosis with MRI; It is also the first machine learning method that can automatically detect POD obliteration from MRI data with the goal of diagnosing endometriosis.