Combined PET/CT imaging provides critical insights into both anatomic and molecular processes, yet traditional single-tracer approaches limit multidimensional disease phenotyping; to address this, we developed the PET Unified Multitracer Alignment (PUMA) framework-an open-source, postprocessing tool that multiplexes serial PET/CT scans for comprehensive voxelwise tissue characterization. <b>Methods:</b> PUMA utilizes artificial intelligence-based CT segmentation from multiorgan objective segmentation to generate multilabel maps of 24 body regions, guiding a 2-step registration: affine alignment followed by symmetric diffeomorphic registration. Tracer images are then normalized and assigned to red-green-blue channels for simultaneous visualization of up to 3 tracers. The framework was evaluated on longitudinal PET/CT scans from 114 subjects across multiple centers and vendors. Rigid, affine, and deformable registration methods were compared for optimal coregistration. Performance was assessed using the Dice similarity coefficient for organ alignment and absolute percentage differences in organ intensity and tumor SUV_mean <b>Results:</b> Deformable registration consistently achieved superior alignment, with Dice similarity coefficient values exceeding 0.90 in 60% of organs while maintaining organ intensity differences below 3%; similarly, SUV_mean differences for tumors were minimal at 1.6% ± 0.9%, confirming that PUMA preserves quantitative PET data while enabling robust spatial multiplexing. <b>Conclusion:</b> PUMA provides a vendor-independent solution for postacquisition multiplexing of serial PET/CT images, integrating complementary tracer data voxelwise into a composite image without modifying clinical protocols. This enhances multidimensional disease phenotyping and supports better diagnostic and therapeutic decisions using serial multitracer PET/CT imaging.

High-resolution magnetic resonance imaging (HR MRI) can provide accurate and rich information for doctors to better detect subtle lesions, delineate tumor boundaries, evaluate small anatomical structures, and assess early-stage pathological changes that might be obscured in lower resolution images. However, the acquisition of HR MRI images often requires prolonged scanning time, which causes the patient's physical and mental discomfort. The patient's slight movement may produce the motion artifacts and make the obtained MRI image become blurry, affecting the accuracy of clinical diagnosis. To tackle these problems, we propose a novel method, Mamba-enhanced Diffusion Model (MDM) for perception-aware blind super-resolution of Magnetic Resonance Imaging, which includes two important components: kernel noise estimator and SR reconstructor. Specifically, we propose a Perception-aware Blur Kernel Noise estimator (PBKN estimator), which takes advantage of the diffusion model to estimate the blur kernel from lowresolution images. Meanwhile, we construct a novel progressive feature reconstructor, which takes the estimated blur kernel and the content information of LR images as prior knowledge to reconstruct more accurate SR MRI images by using diffusion model. Moreover, we design a novel Semantic Information Fusion Mamba (SIF-Mamba) module for the SR reconstruction task. SIF-Mamba is specifically designed in the progressive feature reconstructor to capture the global context of MRI images and improve the feature reconstruction. The extensive experiments demonstrate that our proposed MDM achieves better SR reconstruction results than several outstanding methods. Our codes are available at https://github.com/YXDBright/MDM.

Metastatic castration-resistant prostate cancer has a high rate of mortality with a limited number of effective treatments after hormone therapy. Radiopharmaceutical therapy with [¹⁷⁷Lu]Lu-prostate-specific membrane antigen-617 (LuPSMA) is one treatment option; however, response varies and is partly predicted by PSMA expression and metabolic activity, assessed on [⁶⁸Ga]PSMA-11 or [¹⁸F]DCFPyL and [¹⁸F]FDG PET, respectively. Automated methods to measure these on PET imaging have previously yielded modest accuracy. Refining computational workflows and standardizing approaches may improve patient selection and prognostication for LuPSMA therapy. <b>Methods:</b> PET/CT and quantitative SPECT/CT images from an institutional cohort of patients staged for LuPSMA therapy were annotated for total disease burden. In total, 676 [⁶⁸Ga]PSMA-11 or [¹⁸F]DCFPyL PET, 390 [¹⁸F]FDG PET, and 477 LuPSMA SPECT images were used for development of automated workflow and tested on 56 cases with externally referred PET/CT staging. A segmentation framework, the Global Threshold Regional Consensus Network, was developed based on nnU-Net, with processing refinements to improve boundary definition and overall label accuracy. <b>Results:</b> Using the model to contour disease extent, the mean volumetric Dice similarity coefficient for [⁶⁸Ga]PSMA-11 or [¹⁸F]DCFPyL PET was 0.94, for [¹⁸F]FDG PET was 0.84, and for LuPSMA SPECT was 0.97. On external test cases, Dice accuracy was 0.95 and 0.84 on PSMA and FDG PET, respectively. The refined models yielded consistent improvements compared with nnU-Net, with an increase of 3%-5% in Dice accuracy and 10%-17% in surface agreement. Quantitative biomarkers were compared with a human-defined ground truth using the Pearson coefficient, with scores for [⁶⁸Ga]PSMA-11 or [¹⁸F]DCFPyL, [¹⁸F]FDG, and LuPSMA, respectively, of 0.98, 0.94, and 0.99 for disease volume; 0.98, 0.88, and 0.99 for SUV_mean; 0.96, 0.91, and 0.99 for SUV_max; and 0.97, 0.96, and 0.99 for volume intensity product. <b>Conclusion:</b> Delineation of disease extent and tracer avidity can be performed with a high degree of accuracy using automated deep learning methods. By incorporating threshold-based postprocessing, the tools can closely match the output of manual workflows. Pretrained models and scripts to adapt to institutional data are provided for open use.

Accurate brain tumor segmentation is essential for preoperative evaluation and personalized treatment. Multi-modal MRI is widely used due to its ability to capture complementary tumor features across different sequences. However, in clinical practice, missing modalities are common, limiting the robustness and generalizability of existing deep learning methods that rely on complete inputs, especially under non-dominant modality combinations. To address this, we propose AdaMM, a multi-modal brain tumor segmentation framework tailored for missing-modality scenarios, centered on knowledge distillation and composed of three synergistic modules. The Graph-guided Adaptive Refinement Module explicitly models semantic associations between generalizable and modality-specific features, enhancing adaptability to modality absence. The Bi-Bottleneck Distillation Module transfers structural and textural knowledge from teacher to student models via global style matching and adversarial feature alignment. The Lesion-Presence-Guided Reliability Module predicts prior probabilities of lesion types through an auxiliary classification task, effectively suppressing false positives under incomplete inputs. Extensive experiments on the BraTS 2018 and 2024 datasets demonstrate that AdaMM consistently outperforms existing methods, exhibiting superior segmentation accuracy and robustness, particularly in single-modality and weak-modality configurations. In addition, we conduct a systematic evaluation of six categories of missing-modality strategies, confirming the superiority of knowledge distillation and offering practical guidance for method selection and future research. Our source code is available at https://github.com/Quanato607/AdaMM.

Deep learning has significantly advanced automatic medical diagnostics, releasing human resources from clinical pressure, yet the persistent challenge of data scarcity in this area hampers its further improvements and applications. To address this gap, we introduce a novel ensemble framework called 'Efficient Transfer and Self-supervised Learning based Ensemble Framework' (ETSEF). ETSEF leverages features from multiple pre-trained deep learning models to efficiently learn powerful representations from a limited number of data samples. To the best of our knowledge, ETSEF is the first strategy that combines two pre-training methodologies (Transfer Learning and Self-supervised Learning) with ensemble learning approaches. Various data enhancement techniques, including data augmentation, feature fusion, feature selection, and decision fusion, have also been deployed to maximise the efficiency and robustness of the ETSEF model. Five independent medical imaging tasks, including endoscopy, breast cancer detection, monkeypox detection, brain tumour detection, and glaucoma detection, were tested to demonstrate ETSEF's effectiveness and robustness. Facing limited sample numbers and challenging medical tasks, ETSEF has demonstrated its effectiveness by improving diagnostic accuracy by up to 13.3% compared to strong ensemble baseline models and up to 14.4% compared with recent state-of-the-art methods. Moreover, we emphasise the robustness and trustworthiness of the ETSEF method through various vision-explainable artificial intelligence techniques, including Grad-CAM, SHAP, and t-SNE. Compared to large-scale deep learning models, ETSEF can be flexibly deployed and maintain superior performance for challenging medical imaging tasks, demonstrating potential for application in areas lacking training data. The code is available at Github ETSEF.

Many recent approaches in representation learning implicitly assume that uncorrelated views of a data point are sufficient to learn meaningful representations for various downstream tasks. In this work, we challenge this assumption and demonstrate that meaningful structure in the latent space does not emerge naturally. Instead, it must be explicitly induced. We propose a method that aligns representations from different views of the data to align complementary information without inducing false positives. Our experiments show that our proposed self-supervised learning method, Consistent View Alignment, improves performance for downstream tasks, highlighting the critical role of structured view alignment in learning effective representations. Our method achieved first and second place in the MICCAI 2025 SSL3D challenge when using a Primus vision transformer and ResEnc convolutional neural network, respectively. The code and pretrained model weights are released at https://github.com/Tenbatsu24/LatentCampus.

Image registration is a fundamental task in medical image analysis. Deformations are often closely related to the morphological characteristics of tissues, making accurate feature extraction crucial. Recent weakly supervised methods improve registration by incorporating anatomical priors such as segmentation masks or landmarks, either as inputs or in the loss function. However, such weak labels are often not readily available, limiting their practical use. Motivated by the strong representation learning ability of visual foundation models, this paper introduces SAMIR, an efficient medical image registration framework that utilizes the Segment Anything Model (SAM) to enhance feature extraction. SAM is pretrained on large-scale natural image datasets and can learn robust, general-purpose visual representations. Rather than using raw input images, we design a task-specific adaptation pipeline using SAM's image encoder to extract structure-aware feature embeddings, enabling more accurate modeling of anatomical consistency and deformation patterns. We further design a lightweight 3D head to refine features within the embedding space, adapting to local deformations in medical images. Additionally, we introduce a Hierarchical Feature Consistency Loss to guide coarse-to-fine feature matching and improve anatomical alignment. Extensive experiments demonstrate that SAMIR significantly outperforms state-of-the-art methods on benchmark datasets for both intra-subject cardiac image registration and inter-subject abdomen CT image registration, achieving performance improvements of 2.68% on ACDC and 6.44% on the abdomen dataset. The source code will be publicly available on GitHub following the acceptance of this paper.

Vision-language pretraining has advanced image-text alignment, yet progress in radiology remains constrained by the heterogeneity of clinical reports, including abbreviations, impression-only notes, and stylistic variability. Unlike general-domain settings where more data often leads to better performance, naively scaling to large collections of noisy reports can plateau or even degrade model learning. We ask whether large language model (LLM) encoders can provide robust clinical representations that transfer across diverse styles and better guide image-text alignment. We introduce LLM2VEC4CXR, a domain-adapted LLM encoder for chest X-ray reports, and LLM2CLIP4CXR, a dual-tower framework that couples this encoder with a vision backbone. LLM2VEC4CXR improves clinical text understanding over BERT-based baselines, handles abbreviations and style variation, and achieves strong clinical alignment on report-level metrics. LLM2CLIP4CXR leverages these embeddings to boost retrieval accuracy and clinically oriented scores, with stronger cross-dataset generalization than prior medical CLIP variants. Trained on 1.6M CXR studies from public and private sources with heterogeneous and noisy reports, our models demonstrate that robustness -- not scale alone -- is the key to effective multimodal learning. We release models to support further research in medical image-text representation learning.

<i>"Just Accepted" papers have undergone full peer review and have been accepted for publication in <i>Radiology: Artificial Intelligence</i>. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content</i>. Purpose To introduce an open-source deep learning brain segmentation model for fully automated brain MRI segmentation, enabling rapid segmentation and facilitating large-scale neuroimaging research. Materials and Methods In this retrospective study, a deep learning model was developed using a diverse training dataset of 1900 MRI scans (ages 24-93 with a mean of 65 years (SD: 11.5 years) and 1007 females and 893 males) with reference labels generated using a multiatlas segmentation method with human supervision. The final model was validated using 71391 scans from 14 studies. Segmentation quality was assessed using Dice similarity and Pearson correlation coefficients with reference segmentations. Downstream predictive performance for brain age and Alzheimer's disease was evaluated by fitting machine learning models. Statistical significance was assessed using Mann-Whittney U and McNemar's tests. Results The DLMUSE model achieved high correlation (r = 0.93-0.95) and agreement (median Dice scores = 0.84-0.89) with reference segmentations across the testing dataset. Prediction of brain age using DLMUSE features achieved a mean absolute error of 5.08 years, similar to that of the reference method (5.15 years, <i>P</i> = .56). Classification of Alzheimer's disease using DLMUSE features achieved an accuracy of 89% and F1-score of 0.80, which were comparable to values achieved by the reference method (89% and 0.79, respectively). DLMUSE segmentation speed was over 10000 times faster than that of the reference method (3.5 seconds vs 14 hours). Conclusion DLMUSE enabled rapid brain MRI segmentation, with performance comparable to that of state-of-theart methods across diverse datasets. The resulting open-source tools and user-friendly web interface can facilitate large-scale neuroimaging research and wide utilization of advanced segmentation methods. ©RSNA, 2025.

Brain network analysis plays a crucial role in identifying distinctive patterns associated with neurological disorders. Functional magnetic resonance imaging (fMRI) enables the construction of brain networks by analyzing correlations in blood-oxygen-level-dependent (BOLD) signals across different brain regions, known as regions of interest (ROIs). These networks are typically constructed using atlases that parcellate the brain based on various hypotheses of functional and anatomical divisions. However, there is no standard atlas for brain network classification, leading to limitations in detecting abnormalities in disorders. Recent methods leveraging multiple atlases fail to ensure consistency across atlases and lack effective ROI-level information exchange, limiting their efficacy. To address these challenges, we propose the Atlas-Integrated Distillation and Fusion network (AIDFusion), a novel framework designed to enhance brain network classification using fMRI data. AIDFusion introduces a disentangle Transformer to filter out inconsistent atlas-specific information and distill meaningful cross-atlas connections. Additionally, it enforces subject- and population-level consistency constraints to improve cross-atlas coherence. To further enhance feature integration, AIDFusion incorporates an inter-atlas message-passing mechanism that facilitates the fusion of complementary information across brain regions. We evaluate AIDFusion on four resting-state fMRI datasets encompassing different neurological disorders. Experimental results demonstrate its superior classification performance and computational efficiency compared to state-of-the-art methods. Furthermore, a case study highlights AIDFusion's ability to extract interpretable patterns that align with established neuroscience findings, reinforcing its potential as a robust tool for multi-atlas brain network analysis. The code is publicly available at https://github.com/AngusMonroe/AIDFusion.