TY - GEN
T1 - Assessing the impact of a pre-processing pipeline on a tumor localization model in prostate MRI within a large multi-institutional dataset (NRGGU005)
AU - Alley, Stephanie
AU - Tonneau, Marion
AU - Olivié, Damien
AU - Tempany-Afdhal, Clare M.
AU - Choyke, Peter L.
AU - Turkbey, I. Baris
AU - van der Heide, Uulke A.
AU - Ellis, Rodney J.
AU - Boike, Thomas P.
AU - Pennington, J. Daniel
AU - Frazier, Arthur
AU - Lawton, Colleen A.F.
AU - Leong, Nelson
AU - Mihai, Alina M.
AU - Morgan, Scott C.
AU - Solanki, Abhishek A.
AU - Michalski, Jeff M.
AU - Feng, Felix Y.
AU - Sandler, Howard
AU - Menard, Cynthia
AU - Kadoury, Samuel
N1 - Publisher Copyright:
© 2025 SPIE.
PY - 2025
Y1 - 2025
N2 - Multi-parametric magnetic resonance imaging (mpMRI) is increasingly recognized as a valuable tool for characterizing prostate cancer, integrating T2-weighted (T2w), diffusion-weighted (DWI), and dynamic contrast-enhanced (DCE) imaging. Despite its high sensitivity in localizing tumors, the specificity of mpMRI was shown to be hindered by benign conditions that mimic cancerous tissue features. The aim of this study is to investigate the effect of a pre-processing pipeline integrating state-of-the-art registration, bias field correction and normalization tools. We validated this pre-processing pipeline on a large multi-site dataset of 468 patients from 109 institutions. After performing all pre-processing, tumor localization was determined using a model-based tumor localization approach that takes both multi-parametric MRI and prior clinical knowledge features as input. Results show deformable image registration yielded a significant improvement in tumor localization accuracy, both for the diameter analysis as well as the area under the curve comparison for the subset of patients with ground truth tumor delineations.
AB - Multi-parametric magnetic resonance imaging (mpMRI) is increasingly recognized as a valuable tool for characterizing prostate cancer, integrating T2-weighted (T2w), diffusion-weighted (DWI), and dynamic contrast-enhanced (DCE) imaging. Despite its high sensitivity in localizing tumors, the specificity of mpMRI was shown to be hindered by benign conditions that mimic cancerous tissue features. The aim of this study is to investigate the effect of a pre-processing pipeline integrating state-of-the-art registration, bias field correction and normalization tools. We validated this pre-processing pipeline on a large multi-site dataset of 468 patients from 109 institutions. After performing all pre-processing, tumor localization was determined using a model-based tumor localization approach that takes both multi-parametric MRI and prior clinical knowledge features as input. Results show deformable image registration yielded a significant improvement in tumor localization accuracy, both for the diameter analysis as well as the area under the curve comparison for the subset of patients with ground truth tumor delineations.
UR - https://www.scopus.com/pages/publications/105004556500
U2 - 10.1117/12.3043835
DO - 10.1117/12.3043835
M3 - Conference contribution
AN - SCOPUS:105004556500
T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE
BT - Medical Imaging 2025
A2 - Gimi, Barjor S.
A2 - Krol, Andrzej
PB - SPIE
T2 - Medical Imaging 2025: Clinical and Biomedical Imaging
Y2 - 18 February 2025 through 21 February 2025
ER -