Physics-informed autoencoder for prostate tissue microstructure profiling with hybrid multidimensional MRI

Batuhan Gundogdu, Aritrick Chatterjee, Milica Medved, Ulas Bagci, Gregory S Karczmar, Aytekin Oto

Radiology AI - 20252025

Abstract

The Physics-Informed Autoencoder model provided accurate and explainable measurements of prostate tissue microstructure using hybrid multidimensional MRI and was more robust to image noise compared with the conventional least-squares algorithm.

Abstract source

Cite this work

Batuhan Gundogdu, Aritrick Chatterjee, Milica Medved, Ulas Bagci, Gregory S Karczmar, Aytekin Oto (2025). Physics-informed autoencoder for prostate tissue microstructure profiling with hybrid multidimensional MRI. Radiology AI - 2025 https://doi.org/10.1148/ryai.240167

BibTeX

Abstract

Cite this work

Related publications

Hybrid multi-dimensional MRI prostate cancer detection via Hadamard network-based bias correction and residual networks

Align then refine: text-guided 3D prostate lesion segmentation

Clinical and reader associated correlates of clinically significant magnetic resonance imaging–invisible prostate cancer based on negative magnetic resonance imaging

Cancer extent and cellularity but not inflammatory density are associated with MRI-invisible prostate cancer on AI-based digital pathology deep learning models

Artificial intelligence for prostate segmentation on MRI in global populations