Back to the list
Congress: ECR25
Poster Number: C-11988
Type: Poster: EPOS Radiologist (scientific)
Authorblock: R. Niemantsverdriet1, F. Hartmann1, M. P. A. Starmans1, M. Ronot2, R. L. Miclea3, V. Vilgrain2, M. Thomeer1, S. Klein1, LAI Consortium1; 1Rotterdam/NL, 2Paris/FR, 3Maastricht/NL
Disclosures:
Ruben Niemantsverdriet: Nothing to disclose
Frederik Hartmann: Nothing to disclose
Martijn Pieter Anton Starmans: Nothing to disclose
Maxime Ronot: Nothing to disclose
Razvan Lucian Miclea: Nothing to disclose
Valérie Vilgrain: Nothing to disclose
Maarten Thomeer: Nothing to disclose
Stefan Klein: Nothing to disclose
LAI Consortium: Nothing to disclose
Keywords: Liver, MR, Diagnostic procedure, Cancer
Purpose Accurate diagnosis of solid-appearing liver lesions on MRI scans is of utmost importance for subsequent treatment and prognosis. For example, the decision to perform a biopsy, surgery or chemotherapy is based on the findings of the radiologist. However, this diagnosis can be challenging, subjective, and time-consuming. This is partly due to the wide variety of lesions, differences in scan protocols and the rarity of certain subtypes. Artificial intelligence (AI) based on machine learning for automated liver lesion phenotyping has potential...
Read more Methods and materials Twelve hospitals from three continents are retrospectively collecting routine MRI scans of pathologically confirmed solid-appearing liver lesions. With no restrictions on age, sex, or underlying liver diseases, the project aims to gather anonymised clinical and imaging data from over 3000 patients. The dataset will include common and rare types of malignant and benign lesions. With multiple MRI sequences per patient, this will be the first and largest publicly available MRI dataset focused on solid-appearing liver lesions. This dataset will be...
Read more Results A consortium agreement describing data inclusion and data management procedures, governance mechanisms, intellectual property strategy, and publication policy has been approved by all partners. As a starting point for the LAI dataset, the dataset described by Starmans et al. [1] is used. This dataset consists of 486 patients with over 2500 scans, including 156 patients with hepatocellular carcinoma, 65 with intrahepatic cholangiocarcinoma, 139 with hepatocellular adenoma, and 126 with focal nodular hyperplasia. For each patient, at least a T2-weighted MRI...
Read more Conclusion The LAI dataset will offer a valuable resource for developing and benchmarking AI algorithms aimed at automating the MRI-based diagnosis of solid-appearing liver lesions. The dataset and resulting algorithms hold the potential to enhance radiologists' decision-making and assist in patient referrals to specialised centres. 
Read more References [1] Starmans MP, Miclea RL, Vilgrain V, Ronot M, Purcell Y, Verbeek J, Niessen WJ, Ijzermans JN, De Man RA, Doukas M, Klein S, and Thomeer MG (2023). Automated assessment of T2-Weighted MRI to differentiate malignant and benign primary solid liver lesions in noncirrhotic livers using radiomics. Academic Radiology, 31(3), 870–879. https://doi.org/10.1016/j.acra.2023.07.024  
Read more
GALLERY