Let distortion-guided restoration: a physics-informed learning framework to correct prostate diffusion MRI artifacts.
AI-corrected prostate MRI distortion scans identified all confirmed lesions
Let distortion-guided restoration: a physics-informed learning framework to correct prostate diffusion MRI artifacts.
Conventional correction strategies typically require additional acquisitions and may be limited in severe artifact settings.
To develop a physics-informed deep learning framework, distortion-guided restoration (distortion-guided restoration), for acquisition-free correction of ssEPI distortions in prostate diffusion-weighted imaging.
Model performance was evaluated in a retrospective study on 34 clinical scans with severe susceptibility artifacts drawn from the same clinical datasets, using quantitative image metrics and blinded radiologist scoring.
AI correction found every lesion pathology later confirmed, missing none
On synthetic data, the proposed distortion-guided restoration model achieved the highest peak signal-to-noise ratio (peak signal-to-noise ratio) and lowest normalized mean squared error (normalized mean squared error) across low-b diffusion-weighted imaging (0.089; 95% CI, 0.072-0.105) and apparent diffusion coefficient maps (0.062; 95% CI, 0.053-0.072), outperforming FSL TOPUP and FUGUE (all P < .001).
This proof-of-concept study suggests that a physics-informed hybrid convolutional neural network-diffusion framework offers a practical and acquisition-free solution for correcting severe prostate diffusion-weighted imaging distortions.
It warrants further development for clinical utility and implementation.