Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET

[2406.12632] Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Electrical Engineering and Systems Science > Image and Video Processing arXiv:2406.12632 (eess) [Submitted on 18 Jun 2024 ( v1 ), last revised 15 Apr 2026 (this version, v3)] Title: Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET Authors: Junho Moon , Symac Kim , Haejun Chung , Ikbeom Jang View a PDF of the paper titled Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET, by Junho Moon and 3 other authors View PDF HTML (experimental) Abstract: Positron emission tomography (PET) provides molecular biomarkers for Alzheimer's disease and related dementias (ADRD) and is increasingly used for diagnosis, staging, and clinical trial enrichment. However, its use is limited by cost, regulatory restrictions, and the invasiveness of radiotracer injection. Although current frameworks emphasize multimodal biomarker assessment, including the amyloid/tau/neurodegeneration (A/T/N) scheme, these barriers constrain access to PET imaging. Cross-modal image synthesis may help address this gap by reconstructing unavailable modalities from routine scans. Because PET is clinically valuable for regional uptake patterns rather than exact voxel-wise intensities, perceptual losses that capture higher-level semantic features are well suited to PET synthesis. Existing 2D, 3D, and 2.5D perceptual losses for 3D synthesis each have limitations, including restricted volumetric context, scarcity of pretrained 3D models, and difficulty balancing optimization across anatomical planes. In this study, we synthesize tau PET from structural MRI by generating 3D pseudo-[18F]flortaucipir standardized uptake value ratio (SUVR) maps from 3D T1-weighted MR images. We propose a cyclic 2.5D perceptual loss that alternates optimization across axial, coronal, and sagittal planes during training to improve volumetric consistency. We also standardize PET SUVRs by scanner manufacturer, reducing inter-manufacturer variability and better preserving high-uptake regions. Using cohorts spanning the ADRD spectrum from the ADNI and the SCAN cohort, we show that the method generalizes across U-Net, UNETR, SwinUNETR, CycleGAN, and Pix2Pix, with strong performance. Notably, it improves agreement between synthesized SUVRs and measured PET in brain regions relevant to Alzheimer-type tau pathology. Code is publicly available at this https URL . Comments: Published in Human Brain Mapping, available at this https URL Subjects: Image and Video Processing (eess.IV) ; Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2406.12632 [eess.IV] (or arXiv:2406.12632v3 [eess.IV] for this version) https://doi.org/10.48550/arXiv.2406.12632 Focus to learn more arXiv-issued DOI via DataCite Related DOI : https://doi.org/10.1002/hbm.70508 Focus to learn more DOI(s) linking to related resources Submission history From: Junho Moon [ view email ] [v1] Tue, 18 Jun 2024 13:59:10 UTC (37,413 KB) [v2] Thu, 15 May 2025 05:17:41 UTC (36,588 KB) [v3] Wed, 15 Apr 2026 10:38:58 UTC (29,316 KB) Full-text links: Access Paper: View a PDF of the paper titled Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET, by Junho Moon and 3 other authors View PDF HTML (experimental) TeX Source view license Current browse context: eess.IV < prev | next > new | recent | 2024-06 Change to browse by: cs cs.CV eess References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET.#8955
• Fast Cross-Strength Multi-Contrast Brain MRI Translation using Latent Bridge Matching#1034105
• Multimodal imaging and AI in brain tumor surgery: current tools and emerging integration.#479276
• A Multimodal Feature Distillation with Mamba-Transformer Network for Brain Tumor Segmentation with Incomplete Modalities#1117198
• Three-color high-definition whole-brain imaging: simultaneous analysis of multiple anatomical structures with colocalized cytoarchitectural information.#465185