Organizations: Department of Radiology, Columbia University Irving Medical Center, NY, US · Atal Bihari Vajpayee Institute of Medical Sciences, New Delhi, India · Department of Biomedical Informatics, Stony Brook University, NY, US
Background: Incomplete multiparametric MR examinations limit glioma characterization and downstream AI tools. Generative models can synthesize missing MRI sequences, but anatomical structure and clinically relevant tumor features are challenging to preserve. We propose GliomaINK, an anatomy-guided conditional flow matching framework for cross-sequence 3D brain MRI synthesis. Methods: A 3D SwinUNETR predicted the conditional velocity field while an auxiliary topology-preserving loss encouraged structural consistency. The model was trained on 1,083 patients from the BraTS-AG dataset and independently tested on an in-house clinical cohort. Downstream tasks include image quality, tumor segmentation, IDH1 mutation-status prediction, and overall survival prediction. Two radiologists independently evaluated the anatomical fidelity, pathological representation, and overall image quality of the synthesized images. Results: GliomaINK exhibited the highest PSNR across all six MRI translation directions in both evaluation cohorts. In the BraTS-AG test set, the PSNR reached 25.64±0.55 dB for the T2W-to-T1C synthesis, compared with 13.78±0.32 dB for Pix2Pix and 18.90±0.53 dB for SPADE. For downstream tumor segmentation, the method showed strong performance across tumor subregions with Dice scores reaching 0.928±0.014 for edema, and 0.885±0.040 for enhancing tumor. For the IDH1 mutation-status prediction, GliomaINK achieved the highest AUROC score in four of the six sequence configurations, reaching 0.827±0.088. Conclusion: Anatomy-guided conditional flow matching enables high-fidelity cross-sequence glioma MRI synthesis demonstrating potential for preservation of clinically relevant information in synthesized MRI and improving downstream tasks.
Glioblastoma exhibits diverse, infiltrative, and patient-specific growth patterns that are only partially visible on routine MRI, making it difficult to reliably assess true tumor extent and personalize treatment planning and follow-up. We present a biophysically-conditioned generative framework that synthesizes biologically realistic 3D brain MRI volumes from estimated, spatially continuous tumor-concentration fields. Our approach combines a generative model with tumor-infiltration maps that can be propagated through time using a biophysical growth model, enabling fine-grained control over tumor shape and growth while preserving patient anatomy. This enables us to synthesize consistent tumor growth trajectories directly in the space of real patients, providing interpretable, controllable estimation of tumor infiltration and progression beyond what is explicitly observed in imaging. We evaluate the framework on longitudinal glioblastoma cases and demonstrate that it can generate temporally coherent sequences with realistic changes in tumor appearance and surrounding tissue response. These results suggest that integrating mechanistic tumor growth priors with modern generative modeling can provide a practical tool for patient-specific progression visualization and for generating controlled synthetic data to support downstream neuro-oncology workflows. In longitudinal extrapolation, we achieve a consistent 75% Dice overlap with the biophysical model while maintaining a constant PSNR of 25 in the surrounding tissue. Our code is available at: https://github.com/valentin-biller/lgm.git
Valentin Biller, Niklas Bubeck, Lucas Zimmer +6
Technical University of Munich (TUM), Germany · Munich Center for Machine Learning (MCML) · Chair for AI in Healthcare and Medicine, TUM and TUM University Hospital, Munich, Germany +3
Accurate classification of diffuse gliomas is often hindered by domain shifts across centers and a lack of large, annotated datasets. We propose the Anatomically-conditioned Latent Diffusion Model (ALDM), a novel framework for data-efficient, few-shot 3D volumetric MRI synthesis. ALDM utilizes a two-stage approach: a 3D variational autoencoder learns anatomical priors from a data-rich source domain, while a conditional latent diffusion model, guided by tumor masks via a ControlNet, generates structurally coherent volumes for a data-scarce target domain. Evaluated in an extreme few-shot setting with only 16 target images, ALDM outperformed GAN and hybrid baselines, achieving a superior Frechet Inception Distance (FID) of 85.40 and a downstream classification AUC of 0.987. Qualitative results confirm that the model preserves sharp pathology boundaries and cross-modal consistency, with visual fidelity improving progressively during training. By capturing essential diagnostic features, ALDM provides a robust tool for clinical data augmentation in low-resource settings. Our implementation is available at https://github.com/Analytics-Everywhere-Lab/anatomically-conditioned-LDM.
Salman Shaik, Truong Thanh Hung Nguyen, Hung Cao
Analytics Everywhere Lab, University of New Brunswick, Canada
Generative inpainting of brain MRI volumes is essential for synthesizing healthy tissue in pathological regions, improving the accuracy and reliability of automated downstream brain analysis applications such as image registration, brain extraction, and segmentation. However, standard 3D approaches are computationally prohibitive, while efficient 2D slice-wise methods suffer from severe inter-slice discontinuities. Furthermore, traditional models rely on conditional training, requiring task-specific learning of masked inputs. We propose a zero-shot brain MRI inpainting framework utilizing 2.5D unconditional flow priors to capture spatial context along the superior-inferior axis without the overhead of full 3D convolutions. During training, our flow matching model learns the joint distribution of adjacent axial slice triplets, modeling the manifold of healthy brain anatomy while explicitly excluding pathological regions from the loss function. At inference, the model processes the input triplets autoregressively along the depth axis. We employ the Restora-Flow solver to constrain the unconditional prior using the input mask, achieving accurate zero-shot inpainting. Evaluations show our 2.5D strategy resolves the structural discontinuities of 2D baselines, synthesizing plausible healthy tissue while maintaining volumetric consistency across the axial, sagittal, and coronal planes. As a final step, we generate and average an ensemble of multiple stochastic reconstructions to form the final prediction. Quantitative results benchmarked on the official BraTS 2026 Inpainting Challenge validation set demonstrate the effectiveness of our proposed approach, yielding an SSIM of 0.816 ± 0.112, MSE of 0.007 ± 0.005, and PSNR of 22.923 ± 4.343. Code is available at https://github.com/imigraz/brats2026-inpainting.
Arnela Hadzic, Franz Thaler, Simon Johannes Joham +1
Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Graz, Austria