Harmonizing MR Images Across 100+ Scanners: Multi-site Validation with Traveling Subjects and Real-world Protocols

Savannah P. Hays, Lianrui Zuo, Muhammad Faizyab Ali Chaudhary, Kathleen M. Bartz, Samuel W. Remedios, Jinwei Zhang, Jiachen Zhuo, Murat Bilgel, Shiv Saidha, Ellen M. Mowry, Scott D. Newsome, Jerry L. Prince, Blake E. Dewey, Aaron Carass
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:4703-4721, 2026.

Abstract

Reliable harmonization of heterogeneous magnetic resonance (MR) image datasets, especially those acquired in pragmatic clinical trials, is critical to advance multi-center neuroimaging studies and translational machine learning in healthcare. We present an enhanced and rigorously validated version of the HACA3 harmonization algorithm, which we refer to as HACA3$^+$, incorporating key methodological enhancements: (1) an improved artifact encoder to better isolate and mitigate image artifacts, (2) background and foreground-sensitive attention mechanisms to increase harmonization specificity, and (3) extensive training using data spanning 100+ scanners from 64 independent sites, providing a broader diversity of scanners than other harmonization methods. Our study focuses on four commonly acquired MR image contrasts (T1-weighted, T2-weighted, proton density, & fluid-attenuated inversion recovery), reflecting realistic clinical protocols. We perform inter-site harmonization experiments using traveling subjects to assess the generalization and robustness of the harmonization model. We compare the results of the publicly available version of HACA3 and our implementation, HACA3$^+$. Downstream relevance is further established through whole brain segmentation and image imputation. Finally, we justify each enhancement through an ablation experiment.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-hays26a, title = {Harmonizing MR Images Across 100+ Scanners: Multi-site Validation with Traveling Subjects and Real-world Protocols}, author = {Hays, Savannah P. and Zuo, Lianrui and Chaudhary, Muhammad Faizyab Ali and Bartz, Kathleen M. and Remedios, Samuel W. and Zhang, Jinwei and Zhuo, Jiachen and Bilgel, Murat and Saidha, Shiv and Mowry, Ellen M. and Newsome, Scott D. and Prince, Jerry L. and Dewey, Blake E. and Carass, Aaron}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {4703--4721}, 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/hays26a/hays26a.pdf}, url = {https://proceedings.mlr.press/v315/hays26a.html}, abstract = {Reliable harmonization of heterogeneous magnetic resonance (MR) image datasets, especially those acquired in pragmatic clinical trials, is critical to advance multi-center neuroimaging studies and translational machine learning in healthcare. We present an enhanced and rigorously validated version of the HACA3 harmonization algorithm, which we refer to as HACA3$^+$, incorporating key methodological enhancements: (1) an improved artifact encoder to better isolate and mitigate image artifacts, (2) background and foreground-sensitive attention mechanisms to increase harmonization specificity, and (3) extensive training using data spanning 100+ scanners from 64 independent sites, providing a broader diversity of scanners than other harmonization methods. Our study focuses on four commonly acquired MR image contrasts (T1-weighted, T2-weighted, proton density, & fluid-attenuated inversion recovery), reflecting realistic clinical protocols. We perform inter-site harmonization experiments using traveling subjects to assess the generalization and robustness of the harmonization model. We compare the results of the publicly available version of HACA3 and our implementation, HACA3$^+$. Downstream relevance is further established through whole brain segmentation and image imputation. Finally, we justify each enhancement through an ablation experiment.} }
Endnote
%0 Conference Paper %T Harmonizing MR Images Across 100+ Scanners: Multi-site Validation with Traveling Subjects and Real-world Protocols %A Savannah P. Hays %A Lianrui Zuo %A Muhammad Faizyab Ali Chaudhary %A Kathleen M. Bartz %A Samuel W. Remedios %A Jinwei Zhang %A Jiachen Zhuo %A Murat Bilgel %A Shiv Saidha %A Ellen M. Mowry %A Scott D. Newsome %A Jerry L. Prince %A Blake E. Dewey %A Aaron Carass %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-hays26a %I PMLR %P 4703--4721 %U https://proceedings.mlr.press/v315/hays26a.html %V 315 %X Reliable harmonization of heterogeneous magnetic resonance (MR) image datasets, especially those acquired in pragmatic clinical trials, is critical to advance multi-center neuroimaging studies and translational machine learning in healthcare. We present an enhanced and rigorously validated version of the HACA3 harmonization algorithm, which we refer to as HACA3$^+$, incorporating key methodological enhancements: (1) an improved artifact encoder to better isolate and mitigate image artifacts, (2) background and foreground-sensitive attention mechanisms to increase harmonization specificity, and (3) extensive training using data spanning 100+ scanners from 64 independent sites, providing a broader diversity of scanners than other harmonization methods. Our study focuses on four commonly acquired MR image contrasts (T1-weighted, T2-weighted, proton density, & fluid-attenuated inversion recovery), reflecting realistic clinical protocols. We perform inter-site harmonization experiments using traveling subjects to assess the generalization and robustness of the harmonization model. We compare the results of the publicly available version of HACA3 and our implementation, HACA3$^+$. Downstream relevance is further established through whole brain segmentation and image imputation. Finally, we justify each enhancement through an ablation experiment.
APA
Hays, S.P., Zuo, L., Chaudhary, M.F.A., Bartz, K.M., Remedios, S.W., Zhang, J., Zhuo, J., Bilgel, M., Saidha, S., Mowry, E.M., Newsome, S.D., Prince, J.L., Dewey, B.E. & Carass, A.. (2026). Harmonizing MR Images Across 100+ Scanners: Multi-site Validation with Traveling Subjects and Real-world Protocols. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:4703-4721 Available from https://proceedings.mlr.press/v315/hays26a.html.

Related Material