Tagged MRI enables tracking internal tissue motion non-invasively. It encodes motion by modulating anatomy with periodic tags, which deforms along with tissue. However, the entanglement between anatomy, tags and motion poses significant challenges on post processing. Existence of tags and imaging blur hinders downstream tasks such as segmenting anatomy. Tag fading, due to T1-relaxation, disrupts brightness constancy assumption for motion tracking. For decades, these challenges are handled in isolation and sub-optimally. In contrast, we introduce a blind and nonlinear inverse framework for tagged MRI that, for the first time, unifies these tasks: anatomical image recovery, high-resolution cine image synthesis, and motion estimation. At its core, the synergy of MR physics and generative priors enables us to blindly estimate the unknown forward imaging models, high-resolution underlying anatomy, while simultaneously tracking 3D diffeomorphic Lagrangian motion over time. Experiments on tagged brain MRI demonstrate that our approach yields high-resolution anatomy images, cine images, and more accurate motion than specialized methods.

Medical image challenges have played a transformative role in advancing the field, catalyzing innovation and establishing new performance benchmarks. Image registration, a foundational task in neuroimaging, has similarly advanced through the Learn2Reg initiative. Building on this, we introduce the Large-scale Unsupervised Brain MRI Image Registration (LUMIR) challenge, a next-generation benchmark for unsupervised brain MRI registration. Previous challenges relied upon anatomical label maps, however LUMIR provides 4,014 unlabeled T1-weighted MRIs for training, encouraging biologically plausible deformation modeling through self-supervision. Evaluation includes 590 in-domain test subjects and extensive zero-shot tasks across disease populations, imaging protocols, and species. Deep learning methods consistently achieved state-of-the-art performance and produced anatomically plausible, diffeomorphic deformation fields. They outperformed several leading optimization-based methods and remained robust to most domain shifts. These findings highlight the growing maturity of deep learning in neuroimaging registration and its potential to serve as a foundation model for general-purpose medical image registration.

Image harmonization (IH) makes images from different domains consistent, enabling multi-domain comparison with reliable quantitative measurements. Optical coherence tomography (OCT) could benefit from IH, as variations in OCT systems cause significant discrepancies in image quality due to speckle and noise. However, IH approaches typically rely on paired images during the training phase, and such images are not available in OCT data due to sparse image acquisition. A Schrödinger bridge (SB) finds an optimal coupling between arbitrary probability spaces with respect to a reference path measure and can be applied to unpaired IH. SBs are the core of many generative diffusion models, when one of the probability spaces is easy to sample from (i.e., Gaussian). SBs have been applied to natural images for image-to-image translation; however, they have not been applied to OCT IH because a SB does not guarantee anatomical consistency. In this paper, we use a dual diffusion implicit bridge (DDIB) for OCT IH, which finds independent SBs between arbitrary domains and a common Gaussian probability space. We generalize the assumption of DDIB and use segmentation masks as a condition to improve anatomical consistency. We conduct DDIB experiments with and without segmentation masks as a condition and analyze performance in terms of anatomical consistency and harmonization quality. The method does not require any paired training data and can in principle be quickly adapted to new domains.

Tagged magnetic resonance imaging (tMRI) is a valuable tool for visualizing and quantifying tissue deformation in vivo. Its use is often hampered, however, by tag fading, long computation times, and the challenge of ensuring diffeomorphic, incompressible motion fields. In this paper, we describe a novel integration of the harmonic phase (HARP) approach to tMRI analysis with an unsupervised deep learning-based registration framework to estimate 2D and 3D motion fields that are diffeomorphic and nearly incompressible. The resulting method, called deep sinusoidally transformed HARP, or DSHARP, enables end-to-end network training by implementing a transformation of the harmonic phase to remove phase-wrapping discontinuities. It produces diffeomorphic motion by estimating a stationary velocity field from which motion is computed using the scaling and squaring technique. Finally, it encourages incompressibility using a novel Jacobian determinant loss term during network training. We evaluated DSHARP on 2D and 3D phantom data with simulated incompressible motions, real 3D human tongue data acquired during speech from both healthy and glossectomy subjects, and cardiac tagged MRI from the public STACOM 2011 benchmark. Our approach outperforms HARP, SinMod, SyN, PVIRA, VoxelMorph, and DeepTag in tracking accuracy, computation speed, and preservation of incompressibility.

Accurate assessment of Alzheimer's disease (AD) using neuroimaging is important for understanding disease progression and supporting clinical research. Structural MRI (sMRI) is widely used for this purpose, but most deep learning approaches primarily rely on intensity-based features, which may not fully capture subtle morphological variations. Jacobian determinant maps (JSM) provide complementary information by describing localized brain deformations; however, conventional multimodal fusion strategies such as early or late fusion may not fully capture the interactions between these modalities. To address this, we explore a cross-attention fusion approach that explicitly models the relationship between sMRI intensity and JSM-derived deformation features for AD classification. Using data from the Alzheimer's Disease Neuroimaging Initiative (ADNI), we compare cross-attention with pairwise self-attention and four baseline fusion methods. The proposed approach achieved mean ROC-AUC scores of 0.903 (0.033) for AD vs. cognitively normal (CN) and 0.692 (0.061) for mild cognitive impairment (MCI) vs. CN under five-fold cross-validation. The proposed approach achieved the best performance among fusion strategies and achieved higher or comparable performance with only 1.56 million parameters. These findings suggest that cross-attention fusion may serve as a potentially useful approach for integrating structural and deformation-based MRI features in AD classification tasks.

Medical image harmonization aims to reduce the differences in appearance caused by scanner hardware variations to allow for consistent and reliable comparisons across devices. Harmonization based on paired images from different devices has limited applicability in real-world clinical settings. On the other hand, unpaired harmonization typically does not guarantee anatomy consistency, which is problematic because anatomical information preservation is paramount. The Schrödinger bridge framework has achieved state-of-the-art style transfer performance with natural images by matching distributions of unpaired images, but this approach can also introduce anatomy changes when applied to medical images. We show that such changes occur because the Schrödinger bridge uses the square of the Euclidean distance between images as the transport cost in an entropy-regularized optimal transport problem. Such a transport cost is not appropriate for measuring anatomical distances, as medical images with the same anatomy need not have a small Euclidean distance between them. In this paper, we propose a latent metric Schrödinger bridge (LMSB) framework to improve the anatomical consistency for the harmonization of medical images. We develop an invertible network that maps medical images into a latent Euclidean metric space where the distances among images with the same anatomy are minimized using the pullback latent metric. Within this latent space, we train a Schrödinger bridge to match distributions. We show that the proposed LMSB is superior to the direct application of a Schrödinger bridge to harmonize optical coherence tomography (OCT) images.

Spatially varying regularization accommodates the deformation variations that may be necessary for different anatomical regions during deformable image registration. Historically, optimization-based registration models have harnessed spatially varying regularization to address anatomical subtleties. However, most modern deep learning-based models tend to gravitate towards spatially invariant regularization, wherein a homogenous regularization strength is applied across the entire image, potentially disregarding localized variations. In this paper, we propose a hierarchical probabilistic model that integrates a prior distribution on the deformation regularization strength, enabling the end-to-end learning of a spatially varying deformation regularizer directly from the data. The proposed method is straightforward to implement and easily integrates with various registration network architectures. Additionally, automatic tuning of hyperparameters is achieved through Bayesian optimization, allowing efficient identification of optimal hyperparameters for any given registration task. Comprehensive evaluations on publicly available datasets demonstrate that the proposed method significantly improves registration performance and enhances the interpretability of deep learning-based registration, all while maintaining smooth deformations. Our code is freely available at http://bit.ly/3BrXGxz.

Recent advances in deep learning-based medical image registration have shown that training deep neural networks~(DNNs) does not necessarily require medical images. Previous work showed that DNNs trained on randomly generated images with carefully designed noise and contrast properties can still generalize well to unseen medical data. Building on this insight, we propose using registration between random images as a proxy task for pretraining a foundation model for image registration. Empirical results show that our pretraining strategy improves registration accuracy, reduces the amount of domain-specific data needed to achieve competitive performance, and accelerates convergence during downstream training, thereby enhancing computational efficiency.

Medical image registration is critical for clinical applications, and fair benchmarking of different methods is essential for monitoring ongoing progress in the field. To date, the Learn2Reg 2020-2023 challenges have released several complementary datasets and established metrics for evaluations. Building on this foundation, the 2024 edition expands the challenge’s scope to cover a wider range of registration scenarios, particularly in terms of modality diversity and task complexity, by introducing three new tasks, including large-scale multi-modal registration and unsupervised inter-subject brain registration, as well as the first microscopy-focused benchmark within Learn2Reg. The new datasets also inspired new method developments, including invertibility constraints, pyramid features, keypoints alignment and instance optimisation.

In recent years, unsupervised learning for deformable image registration has been a major research focus. This approach involves training a registration network using pairs of moving and fixed images, along with a loss function that combines an image similarity measure and deformation regularization. For multi-modal image registration tasks, the correlation ratio has been a widely-used image similarity measure historically, yet it has been underexplored in current deep learning methods. Here, we propose a differentiable correlation ratio to use as a loss function for learning-based multi-modal deformable image registration. This approach extends the traditionally non-differentiable implementation of the correlation ratio by using the Parzen windowing approximation, enabling backpropagation with deep neural networks. We validated the proposed correlation ratio on a multi-modal neuroimaging dataset. In addition, we established a Bayesian training framework to study how the trade-off between the deformation regularizer and similarity measures, including mutual information and our proposed correlation ratio, affects the registration performance.

Dynamic positron emission tomography (PET) imaging combined with radiotracer kinetic modeling is a powerful technique for visualizing biological processes in the brain, offering valuable insights into brain functions and neurological disorders such as Alzheimer's and Parkinson's diseases. Accurate kinetic modeling relies heavily on the use of a metabolite-corrected arterial input function (AIF), which typically requires invasive and labor-intensive arterial blood sampling. While alternative non-invasive approaches have been proposed, they often compromise accuracy or still necessitate at least one invasive blood sampling. In this study, we present the deep learning-derived arterial input function (DLIF), a deep learning framework capable of estimating a metabolite-corrected AIF directly from dynamic PET image sequences without any blood sampling. We validated DLIF using existing dynamic PET patient data. We compared DLIF and resulting parametric maps against ground truth measurements. Our evaluation shows that DLIF achieves accurate and robust AIF estimation. By leveraging deep learning's ability to capture complex temporal dynamics and incorporating prior knowledge of typical AIF shapes through basis functions, DLIF provides a rapid, accurate, and entirely non-invasive alternative to traditional AIF measurement methods.

BACKGROUND AND PURPOSE:Prostate cancer remains a significant clinical challenge, particularly in characterizing lesions and predicting patient outcomes. With the growing availability of advanced imaging techniques like [18F]DCFPyL PET/CT, there is an urgent need for intelligent tools that can facilitate clinical decision-making. This study aimed to develop artificial intelligence (AI) models for lesion characterization and outcome prediction in prostate cancer (PCa) patients. MATERIALS AND METHODS:PCa patients who underwent [18F]DCFPyL PET/CT imaging were divided into training and internal test sets (n = 238) and a prospective test set (n = 36). Lesions were scored using the PSMA-Reporting and Data System (RADS) and assessed for malignancy, treatment response, and survival outcomes. Single- and multi-modality deep learning models were trained for four tasks: PSMA-RADS scoring, malignancy classification, treatment response prediction, and survival prediction. RESULTS:The input concatenation model, which combined PET and CT modalities, demonstrated superior performance across all tasks. For the internal test set, the area under the receiver operating characteristic curves (AUROCs) were 0.81 (95 % CI: 0.80-0.81) for PSMA-RADS scoring, 0.79 (95 % CI: 0.78-0.80) for malignancy classification, and 0.74 (95 % CI: 0.73-0.77) for treatment response prediction. In the prospective test set, the AUROCs were 0.72 (95 % CI: 0.69-0.75) for PSMA-RADS scoring, 0.70 (95 % CI: 0.68-0.71) for malignancy classification, and 0.70 (95 % CI: 0.67-0.72) for treatment response prediction. The C-indices for survival predictions were 0.58 (95 % CI: 0.57-0.59) and 0.60 (95 % CI: 0.60-0.63) for the internal and prospective test sets, respectively. CONCLUSION:Our study highlights the potential of AI to improve lesion characterization and identify patients at high risk of disease progression.

Optical coherence tomography (OCT) images are often acquired as highly anisotropic volumes, where the scanning step is dense along the fast axis but sparse along the slow axis. This affects image analysis, such as image registration for longitudinal alignment. To create more isotropic volumes, bicubic interpolation can be used along the slow axis, but it generally produces blurry features. Registration-based interpolation can reduce blurriness, but often fails to generate realistic OCT images. Deep generative models can sample realistic images, but lack the structural consistency constraints required for interpolation. In this paper, we propose an unsupervised image interpolation method that combines registration-based interpolation with a deep generative model to overcome their individual limitations and improve the structural accuracy and realism of interpolated OCT images. We compare the proposed method with both bicubic and registration-based interpolation on real OCT datasets, and show that it achieves the best interpolation performance.

Objective: Deep learning-based deformable image registration has achieved strong accuracy, but remains sensitive to variations in input image characteristics such as artifacts, field-of-view mismatch, or modality difference. We aim to develop a general training paradigm that improves the robustness and generalizability of registration networks. Methods: We introduce surrogate supervision, which decouples the input domain from the supervision domain by applying estimated spatial transformations to surrogate images. This allows training on heterogeneous inputs while ensuring supervision is computed in domains where similarity is well defined. We evaluate the framework through three representative applications: artifact-robust brain MR registration, mask-agnostic lung CT registration, and multi-modal MR registration. Results: Across tasks, surrogate supervision demonstrated strong resilience to input variations including inhomogeneity field, inconsistent field-of-view, and modality differences, while maintaining high performance on well-curated data. Conclusions: Surrogate supervision provides a principled framework for training robust and generalizable deep learning-based registration models without increasing complexity. Significance: Surrogate supervision offers a practical pathway to more robust and generalizable medical image registration, enabling broader applicability in diverse biomedical imaging scenarios.

Accurate intraoperative image guidance is critical for achieving maximal safe resection in brain tumor surgery, yet neuronavigation systems based on preoperative MRI lose accuracy during the procedure due to brain shift. Aligning post-resection intraoperative ultrasound (iUS) with preoperative MRI can restore spatial accuracy by estimating brain shift deformations, but it remains a challenging problem given the large anatomical and topological changes and substantial modality intensity gap. The ReMIND2Reg 2025 Challenge provides the largest public benchmark for this task, built upon the ReMIND dataset. It offers 99 training cases, 5 validation cases, and 10 private test cases comprising paired 3D ceT1 MRI, T2 MRI, and post-resection 3D iUS volumes. Data are provided without annotations for training, while validation and test performance are evaluated on manually annotated anatomical landmarks. Metrics include target registration error (TRE), robustness to worst-case landmark misalignment (TRE30), and runtime. By establishing a standardized evaluation framework for this clinically critical and technically complex problem, ReMIND2Reg aims to accelerate the development of robust, generalizable, and clinically deployable multimodal registration algorithms for image-guided neurosurgery.

Producing spatial transformations that are diffeomorphic is a key goal in deformable image registration. As a diffeomorphic transformation should have positive Jacobian determinant |J| everywhere, the number of voxels with |J|<0 has been used to test for diffeomorphism and also to measure the irregularity of the transformation. For digital transformations, |J| is commonly approximated using a central difference, but this strategy can yield positive |J|'s for transformations that are clearly not diffeomorphic -- even at the voxel resolution level. To show this, we first investigate the geometric meaning of different finite difference approximations of |J|. We show that to determine if a deformation is diffeomorphic for digital images, the use of any individual finite difference approximation of |J| is insufficient. We further demonstrate that for a 2D transformation, four unique finite difference approximations of |J|'s must be positive to ensure that the entire domain is invertible and free of folding at the pixel level. For a 3D transformation, ten unique finite differences approximations of |J|'s are required to be positive. Our proposed digital diffeomorphism criteria solves several errors inherent in the central difference approximation of |J| and accurately detects non-diffeomorphic digital transformations. The source code of this work is available at https://github.com/yihao6/digital_diffeomorphism.

Tagged magnetic resonance imaging (tMRI) has been employed for decades to measure the motion of tissue undergoing deformation. However, registration-based motion estimation from tMRI is difficult due to the periodic patterns in these images, particularly when the motion is large. With a larger motion the registration approach gets trapped in a local optima, leading to motion estimation errors. We introduce a novel "momenta, shooting, and correction" framework for Lagrangian motion estimation in the presence of repetitive patterns and large motion. This framework, grounded in Lie algebra and Lie group principles, accumulates momenta in the tangent vector space and employs exponential mapping in the diffeomorphic space for rapid approximation towards true optima, circumventing local optima. A subsequent correction step ensures convergence to true optima. The results on a 2D synthetic dataset and a real 3D tMRI dataset demonstrate our method's efficiency in estimating accurate, dense, and diffeomorphic 2D/3D motion fields amidst large motion and repetitive patterns.

Deep learning technologies have dramatically reshaped the field of medical image registration over the past decade. The initial developments, such as regression-based and U-Net-based networks, established the foundation for deep learning in image registration. Subsequent progress has been made in various aspects of deep learning-based registration, including similarity measures, deformation regularizations, network architectures, and uncertainty estimation. These advancements have not only enriched the field of image registration but have also facilitated its application in a wide range of tasks, including atlas construction, multi-atlas segmentation, motion estimation, and 2D-3D registration. In this paper, we present a comprehensive overview of the most recent advancements in deep learning-based image registration. We begin with a concise introduction to the core concepts of deep learning-based image registration. Then, we delve into innovative network architectures, loss functions specific to registration, and methods for estimating registration uncertainty. Additionally, this paper explores appropriate evaluation metrics for assessing the performance of deep learning models in registration tasks. Finally, we highlight the practical applications of these novel techniques in medical imaging and discuss the future prospects of deep learning-based image registration.

Purpose:Deformable image registration establishes non-linear spatial correspondences between fixed and moving images. Deep learning-based deformable registration methods have been widely studied in recent years due to their speed advantage over traditional algorithms as well as their better accuracy. Most existing deep learning-based methods require neural networks to encode location information in their feature maps and predict displacement or deformation fields through convolutional or fully connected layers from these high-dimensional feature maps. We present vector field attention (VFA), a novel framework that enhances the efficiency of the existing network design by enabling direct retrieval of location correspondences. Approach:VFA uses neural networks to extract multi-resolution feature maps from the fixed and moving images and then retrieves pixel-level correspondences based on feature similarity. The retrieval is achieved with a novel attention module without the need for learnable parameters. VFA is trained end-to-end in either a supervised or unsupervised manner. Results:We evaluated VFA for intra- and inter-modality registration and unsupervised and semi-supervised registration using public datasets as well as the Learn2Reg challenge. VFA demonstrated comparable or superior registration accuracy compared with several state-of-the-art methods. Conclusions:VFA offers a novel approach to deformable image registration by directly retrieving spatial correspondences from feature maps, leading to improved performance in registration tasks. It holds potential for broader applications.

Background This study aimed to develop deep learning (DL) models for lesion characterization and outcome prediction in prostate cancer (PCa) patients using Prostate-Specific Membrane Antigen (PSMA) PET/CT imaging. Methods The study included 358 confirmed PCa patients who underwent [18F]DCFPyL PET/CT imaging. Patients were divided into training and internal test sets (n = 275), prospective test set (n = 64), and external test set (n = 19). Lesions were evaluated using PSMA-Reporting and Data System (RADS) scores, malignancy classification, treatment response and survival prediction, followed by DL models trained for each of these tasks. The performance of multi-modality (PET + CT) models was compared to single-modality models, with the best models from the internal and prospective test sets applied to the external test set. Results The input concatenation model, incorporating both PET and CT data, demonstrated the highest performance across all tasks. For PSMA-RADS scoring, the area under the receiver operating characteristic curve (AUROC) was 0.81 (95% CI: 0.80–0.81) for the internal test set, 0.72 (95% CI: 0.69–0.75) for the prospective test set, and 0.68 (95% CI: 0.68–0.69) for the external test set. For malignancy classification, the model achieved AUROCs of 0.79 (95% CI: 0.78–0.80), 0.70 (95% CI: 0.68–0.71), and 0.62 (95% CI: 0.61–0.63) in the internal, prospective, and external test sets, respectively. The AUROC for treatment response prediction was 0.74 (95% CI: 0.73–0.77) for the internal test set, 0.70 (95% CI: 0.67–0.72) for the prospective test set, and 0.72 (95% CI: 0.70–0.73) for the external dataset. The C-index for survival was 0.58 (95% CI: 0.57–0.59), 0.60 (95% CI: 0.60–0.63) and 0.59 (95% CI: 0.57–0.62) in the internal, prospective, and external test sets, respectively. Conclusions The DL model utilizing input concatenation of PET and CT data outperformed single-modality models in PSMA-RADS scoring, malignancy classification, treatment response assessment, and survival prediction, highlighting its potential as a clinical tool.