# DEMIST: Decoupled Multi-stream latent diffusion for Quantitative Myelin Map Synthesis

arXiv: https://arxiv.org/abs/2511.12396 · HTML: https://mlsanity.com/paper/2511.12396/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Medical imaging and segmentation, one of 14 directions in Diffusion models.
- Builds on 18 of its 21 references that are in the map.
- First posted on arXiv 16 Nov 2025.

## Placed in

- Diffusion models / [Medical imaging and segmentation](https://mlsanity.com/diffusion/medical-imaging-and-segmentation/)

## Builds on (most cited first, 18 of 18)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [Adding Conditional Control to Text-to-Image Diffusion Models](https://mlsanity.com/paper/2302.05543/) (2023)
- [Neural Discrete Representation Learning](https://mlsanity.com/paper/1711.00937/) (2017)
- [The Medical Segmentation Decathlon](https://mlsanity.com/paper/2106.05735/) (2021)
- [Medical Image Synthesis with Context-Aware Generative Adversarial Networks](https://mlsanity.com/paper/1612.05362/) (2016)
- [Conversion Between CT and MRI Images Using Diffusion and Score-Matching Models](https://mlsanity.com/paper/2209.12104/) (2022)
- [Anatomically-Controllable Medical Image Generation with Segmentation-Guided Diffusion Models](https://mlsanity.com/paper/2402.05210/) (2024)
- [Generative AI for Medical Imaging: extending the MONAI Framework](https://mlsanity.com/paper/2307.15208/) (2023)
- [Adaptive Latent Diffusion Model for 3D Medical Image to Image Translation: Multi-modal Magnetic Resonance Imaging Study](https://mlsanity.com/paper/2311.00265/) (2023)
- [The Brain Tumor Segmentation (BraTS) Challenge 2023: Brain MR Image Synthesis for Tumor Segmentation (BraSyn)](https://mlsanity.com/paper/2305.09011/) (2023)
- [Bridging the Gap Between f-GANs and Wasserstein GANs](https://mlsanity.com/paper/1910.09779/) (2019)
- [cWDM: Conditional Wavelet Diffusion Models for Cross-Modality 3D Medical Image Synthesis](https://mlsanity.com/paper/2411.17203/) (2024)
- [CoCoLIT: ControlNet-Conditioned Latent Image Translation for MRI to Amyloid PET Synthesis](https://mlsanity.com/paper/2508.01292/) (2025)
- [Self-Supervised CSF Inpainting with Synthetic Atrophy for Improved Accuracy Validation of Cortical Surface Analyses](https://mlsanity.com/paper/2303.05777/) (2023)
- [Adaptive Whole-Body PET Image Denoising Using 3D Diffusion Models with ControlNet](https://mlsanity.com/paper/2411.05302/) (2024)

## Abstract (arXiv)

Quantitative magnetization transfer (qMT) imaging provides myelin-sensitive biomarkers, such as the pool size ratio (PSR), which is valuable for multiple sclerosis (MS) assessment. However, qMT requires specialized 20-30 minute scans. We propose DEMIST to synthesize PSR maps from standard T1w and FLAIR images using a 3D latent diffusion model with three complementary conditioning mechanisms. Our approach has two stages: first, we train separate autoencoders for PSR and anatomical images to learn aligned latent representations. Second, we train a conditional diffusion model in this latent space on top of a frozen diffusion foundation backbone. Conditioning is decoupled into: (i) semantic tokens via cross-attention, (ii) spatial per-scale residual hints via a 3D ControlNet branch, and (iii) adaptive LoRA-modulated attention. We include edge-aware loss terms to preserve lesion boundaries and alignment losses to maintain quantitative consistency, while keeping the number of trainable parameters low and retaining the inductive bias of the pretrained model. We evaluate on 163 scans from 99 subjects using 5-fold cross-validation. Our method outperforms VAE, GAN and diffusion baselines on multiple metrics, producing sharper boundaries and better quantitative agreement with ground truth. Our code is publicly available at https://github.com/MedICL-VU/MS-Synthesis-3DcLDM.