Abstract
Cross-modalitymedicalimagetranslationcanreducetheburden of multi-modal acquisitions, yet the field remains constrained by two coupled limitations: methods operate on 2D slices or 3D patches rather thanwholevolumes,andtrainaseparatemodelforeachtranslationtask. Bothstemfromasinglecause,theabsenceofasufficientlystrongvolumetric prior, which forces generative models to learn anatomical appearance and cross-modality mapping simultaneously, an ill-posed problem at the scaleofavailablepaireddatasets.Weproposetodecoupletheseobjectives. A large-scale pretrained 3D variational autoencoder provides a compact latent representation of volumetric appearance, reducing translation to a conditional flow-matching problem. This compression makes wholevolume processing tractable, while a resolution-aware sampling strategy preserves native anatomical scale. We train a single model jointly across inter-modality(MRI→CT,CBCT→CT)andintra-modality(MRI→MRI) tasks over three multi-center datasets. Across all tasks, whole-volume processing outperforms its patch-based counterpart, and the multi-task model matches task-specific baselines while replacing N networks with one. Crucially, joint training unlocks capabilities inaccessible to taskspecificapproaches:zero-shotgeneralizationtoanatomicalregionsunseen during training, within 0.15 SSIM of the fully supervised model, and compositional cross-dataset translation along paths never directly supervised. These results suggest that combining a strong volumetric prior with multitask training is a scalable route toward synthesis systems that generalize beyond their training distribution. Code is available at https://github.com/arco-group/Whole-Volume-Latent-FM.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/SASHIMI_006.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=4gGmt5hdbb
BibTex
@InProceedings{MolDan_Compositional_MICCAISAT2026,
author = { Molino, Daniele AND Zoboli, Alessio AND Caruso, Camillo Maria AND Guarrasi, Valerio AND Soda, Paolo},
title = { { Compositional Cross-Modality Translation via Whole-Volume Multitask Latent Flow Matching } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17258},
month = {pending},
page = {pending}
}