MASC: Metal-Aware Sampling and Correction via Reinforcement Learning for Accelerated MRI

Zhengyi Lu, Ming Lu, Chongyu Qu, Junchao Zhu, Junlin Guo, Marilyn Lionts, Yanfan Zhu, Yuechen Yang, Tianyuan Yao, Jayasai Rajagopal, Bennett Allan Landman, Xiao Wang, Xinqiang Yan, Yuankai Huo
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:2602-2620, 2026.

Abstract

Metal implants in MRI cause severe artifacts that degrade image quality and hinder clinical diagnosis. Traditional approaches address metal artifact reduction (MAR) and accelerated MRI acquisition as separate problems. We propose MASC, a unified reinforcement learning framework that jointly optimizes metal-aware k-space sampling and artifact correction for accelerated MRI. To enable supervised training, we construct a paired MRI dataset using physics-based simulation, generating k-space data and reconstructions for phantoms with and without metal implants. This paired dataset provides simulated 3D MRI scans with and without metal implants, where each metal-corrupted sample has an exactly matched clean reference, enabling direct supervision for both artifact reduction and acquisition policy learning. We formulate active MRI acquisition as a sequential decision-making problem, where an artifact-aware Proximal Policy Optimization (PPO) agent learns to select k-space phase-encoding lines under a limited acquisition budget. The agent operates on undersampled reconstructions processed through a U-Net-based MAR network, learning patterns that maximize reconstruction quality. We further propose an end-to-end training scheme where the acquisition policy learns to select k-space lines that best support artifact removal while the MAR network simultaneously adapts to the resulting undersampling patterns. Experiments demonstrate that MASC’s learned policies outperform conventional sampling strategies, and end-to-end training improves performance compared to using a frozen pre-trained MAR network, validating the benefit of joint optimization. Cross-dataset experiments on FastMRI with physics-based artifact simulation further confirm generalization to realistic clinical MRI data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-lu26a, title = {MASC: Metal-Aware Sampling and Correction via Reinforcement Learning for Accelerated MRI}, author = {Lu, Zhengyi and Lu, Ming and Qu, Chongyu and Zhu, Junchao and Guo, Junlin and Lionts, Marilyn and Zhu, Yanfan and Yang, Yuechen and Yao, Tianyuan and Rajagopal, Jayasai and Landman, Bennett Allan and Wang, Xiao and Yan, Xinqiang and Huo, Yuankai}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {2602--2620}, year = {2026}, editor = {Huo, Yuankai and Gao, Mingchen and Kuo, Chang-Fu and Jin, Yueming and Deng, Ruining}, volume = {315}, series = {Proceedings of Machine Learning Research}, month = {08--10 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v315/main/assets/lu26a/lu26a.pdf}, url = {https://proceedings.mlr.press/v315/lu26a.html}, abstract = {Metal implants in MRI cause severe artifacts that degrade image quality and hinder clinical diagnosis. Traditional approaches address metal artifact reduction (MAR) and accelerated MRI acquisition as separate problems. We propose MASC, a unified reinforcement learning framework that jointly optimizes metal-aware k-space sampling and artifact correction for accelerated MRI. To enable supervised training, we construct a paired MRI dataset using physics-based simulation, generating k-space data and reconstructions for phantoms with and without metal implants. This paired dataset provides simulated 3D MRI scans with and without metal implants, where each metal-corrupted sample has an exactly matched clean reference, enabling direct supervision for both artifact reduction and acquisition policy learning. We formulate active MRI acquisition as a sequential decision-making problem, where an artifact-aware Proximal Policy Optimization (PPO) agent learns to select k-space phase-encoding lines under a limited acquisition budget. The agent operates on undersampled reconstructions processed through a U-Net-based MAR network, learning patterns that maximize reconstruction quality. We further propose an end-to-end training scheme where the acquisition policy learns to select k-space lines that best support artifact removal while the MAR network simultaneously adapts to the resulting undersampling patterns. Experiments demonstrate that MASC’s learned policies outperform conventional sampling strategies, and end-to-end training improves performance compared to using a frozen pre-trained MAR network, validating the benefit of joint optimization. Cross-dataset experiments on FastMRI with physics-based artifact simulation further confirm generalization to realistic clinical MRI data.} }
Endnote
%0 Conference Paper %T MASC: Metal-Aware Sampling and Correction via Reinforcement Learning for Accelerated MRI %A Zhengyi Lu %A Ming Lu %A Chongyu Qu %A Junchao Zhu %A Junlin Guo %A Marilyn Lionts %A Yanfan Zhu %A Yuechen Yang %A Tianyuan Yao %A Jayasai Rajagopal %A Bennett Allan Landman %A Xiao Wang %A Xinqiang Yan %A Yuankai Huo %B Proceedings of The 9th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Yuankai Huo %E Mingchen Gao %E Chang-Fu Kuo %E Yueming Jin %E Ruining Deng %F pmlr-v315-lu26a %I PMLR %P 2602--2620 %U https://proceedings.mlr.press/v315/lu26a.html %V 315 %X Metal implants in MRI cause severe artifacts that degrade image quality and hinder clinical diagnosis. Traditional approaches address metal artifact reduction (MAR) and accelerated MRI acquisition as separate problems. We propose MASC, a unified reinforcement learning framework that jointly optimizes metal-aware k-space sampling and artifact correction for accelerated MRI. To enable supervised training, we construct a paired MRI dataset using physics-based simulation, generating k-space data and reconstructions for phantoms with and without metal implants. This paired dataset provides simulated 3D MRI scans with and without metal implants, where each metal-corrupted sample has an exactly matched clean reference, enabling direct supervision for both artifact reduction and acquisition policy learning. We formulate active MRI acquisition as a sequential decision-making problem, where an artifact-aware Proximal Policy Optimization (PPO) agent learns to select k-space phase-encoding lines under a limited acquisition budget. The agent operates on undersampled reconstructions processed through a U-Net-based MAR network, learning patterns that maximize reconstruction quality. We further propose an end-to-end training scheme where the acquisition policy learns to select k-space lines that best support artifact removal while the MAR network simultaneously adapts to the resulting undersampling patterns. Experiments demonstrate that MASC’s learned policies outperform conventional sampling strategies, and end-to-end training improves performance compared to using a frozen pre-trained MAR network, validating the benefit of joint optimization. Cross-dataset experiments on FastMRI with physics-based artifact simulation further confirm generalization to realistic clinical MRI data.
APA
Lu, Z., Lu, M., Qu, C., Zhu, J., Guo, J., Lionts, M., Zhu, Y., Yang, Y., Yao, T., Rajagopal, J., Landman, B.A., Wang, X., Yan, X. & Huo, Y.. (2026). MASC: Metal-Aware Sampling and Correction via Reinforcement Learning for Accelerated MRI. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:2602-2620 Available from https://proceedings.mlr.press/v315/lu26a.html.

Related Material