A diagnostic medical foundation model (MedFM) is an artificial intelligence (AI) system engineered to accurately determine diagnoses across various medical imaging modalities and specialties. To train MedFM, we created the PubMed Central Medical Images Dataset (PMCMID), the largest annotated medical image dataset to date, comprising 16,126,659 images from 3,021,780 medical publications. Using AI- and ontology-based methods, we identified 4,482,237 medical images (e.g., clinical photos, X-rays, ultrasounds) and generated comprehensive annotations. To optimize MedFMs performance and assess biases, 13,266 images were manually annotated to establish a multimodal benchmark. MedFM achieved physician-level performance in diagnosis tasks spanning radiology, dermatology, and infectious diseases without requiring specific training. Additionally, we developed the Image2Paper app, allowing clinicians to upload medical images and retrieve relevant literature. The correct diagnoses appeared within the top ten results in 88.4% and at least one relevant differential diagnosis in 93.0%. MedFM and PMCMID were made publicly available. FundingResearch reported here was partially supported by the National Cancer Institute (NCI) (R01 CA260271), the Saudi Company for Artificial Intelligence (SCAI) Authority, and the German Federal Ministry for Economic Affairs and Climate Action (BMWK) under the project DAKI-FWS (01MK21009E). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Medical image reconstruction from measurement data is a vital but challenging inverse problem. Deep learning approaches have achieved promising results, but often requires paired measurement and high-quality images, which is typically simulated through a forward model, i.e., retrospective reconstruction. However, training on simulated pairs commonly leads to performance degradation on real prospective data due to the retrospective-to-prospective gap caused by incomplete imaging knowledge in simulation. To address this challenge, this paper introduces imaging Knowledge-Informed Dynamic Optimal Transport (KIDOT), a novel dynamic optimal transport framework with optimality in the sense of preserving consistency with imaging physics in transport, that conceptualizes reconstruction as finding a dynamic transport path. KIDOT learns from unpaired data by modeling reconstruction as a continuous evolution path from measurements to images, guided by an imaging knowledge-informed cost function and transport equation. This dynamic and knowledge-aware approach enhances robustness and better leverages unpaired data while respecting acquisition physics. Theoretically, we demonstrate that KIDOT naturally generalizes dynamic optimal transport, ensuring its mathematical rationale and solution existence. Extensive experiments on MRI and CT reconstruction demonstrate KIDOT's superior performance.

Recently, Ultra-High Field MRI (UHF-MRI) has become more available and one of the best tools to study the brain. One common step in quantitative neuroimaging is the brain segmentation. However, the differences between UHF-MRI and 1.5-3T images are such that the automatic segmentation techniques optimized at these field strengths usually produce unsatisfactory segmentation results for UHF images. It has been particularly challenging to perform quantitative analyses as typically done with 1.5-3T data, considerably limiting the potential of UHF-MRI. Hence, we propose a novel Deep Learning (DL)-based segmentation technique called GOUHFI: Generalized and Optimized segmentation tool for Ultra-High Field Images, designed to segment UHF images of various contrasts and resolutions. For training, we used a total of 206 label maps from four datasets acquired at 3T, 7T and 9.4T. In contrast to most DL strategies, we used a previously proposed domain randomization approach, where synthetic images generated from the label maps were used for training a 3D U-Net. GOUHFI was tested on seven different datasets and compared to techniques like FastSurferVINN and CEREBRUM-7T. GOUHFI was able to the segment six contrasts and seven resolutions tested at 3T, 7T and 9.4T. Average Dice-Sorensen Similarity Coefficient (DSC) scores of 0.87, 0.84, 0.91 were computed against the ground truth segmentations at 3T, 7T and 9.4T. Moreover, GOUHFI demonstrated impressive resistance to the typical inhomogeneities observed at UHF-MRI, making it a new powerful segmentation tool that allows to apply the usual quantitative analysis pipelines also at UHF. Ultimately, GOUHFI is a promising new segmentation tool, being the first of its kind proposing a contrast- and resolution-agnostic alternative for UHF-MRI, making it the forthcoming alternative for neuroscientists working with UHF-MRI or even lower field strengths.

Background and Objective: Precise preoperative planning and effective physician training for coronary interventions are increasingly important. Despite advances in medical imaging technologies, transforming static or limited dynamic imaging data into comprehensive dynamic cardiac models remains challenging. Existing training systems lack accurate simulation of cardiac physiological dynamics. This study develops a comprehensive dynamic cardiac model research framework based on 4D-CTA, integrating digital twin technology, computer vision, and physical model manufacturing to provide precise, personalized tools for interventional cardiology. Methods: Using 4D-CTA data from a 60-year-old female with three-vessel coronary stenosis, we segmented cardiac chambers and coronary arteries, constructed dynamic models, and implemented skeletal skinning weight computation to simulate vessel deformation across 20 cardiac phases. Transparent vascular physical models were manufactured using medical-grade silicone. We developed cardiac output analysis and virtual angiography systems, implemented guidewire 3D reconstruction using binocular stereo vision, and evaluated the system through angiography validation and CABG training applications. Results: Morphological consistency between virtual and real angiography reached 80.9%. Dice similarity coefficients for guidewire motion ranged from 0.741-0.812, with mean trajectory errors below 1.1 mm. The transparent model demonstrated advantages in CABG training, allowing direct visualization while simulating beating heart challenges. Conclusion: Our patient-specific digital-physical twin approach effectively reproduces both anatomical structures and dynamic characteristics of coronary vasculature, offering a dynamic environment with visual and tactile feedback valuable for education and clinical planning.

While the acute phase of the COVID-19 pandemic has subsided, its long-term effects persist through Post-Acute Sequelae of COVID-19 (PASC), commonly known as Long COVID. There remains substantial uncertainty regarding both its duration and optimal management strategies. PASC manifests as a diverse array of persistent or newly emerging symptoms--ranging from fatigue, dyspnea, and neurologic impairments (e.g., brain fog), to cardiovascular, pulmonary, and musculoskeletal abnormalities--that extend beyond the acute infection phase. This heterogeneous presentation poses substantial challenges for clinical assessment, diagnosis, and treatment planning. In this paper, we focus on imaging findings that may suggest fibrotic damage in the lungs, a critical manifestation characterized by scarring of lung tissue, which can potentially affect long-term respiratory function in patients with PASC. This study introduces a novel multi-center chest CT analysis framework that combines deep learning and radiomics for fibrosis prediction. Our approach leverages convolutional neural networks (CNNs) and interpretable feature extraction, achieving 82.2% accuracy and 85.5% AUC in classification tasks. We demonstrate the effectiveness of Grad-CAM visualization and radiomics-based feature analysis in providing clinically relevant insights for PASC-related lung fibrosis prediction. Our findings highlight the potential of deep learning-driven computational methods for early detection and risk assessment of PASC-related lung fibrosis--presented for the first time in the literature.

Ultrasound localization microscopy (ULM) has revolutionized microvascular imaging by breaking the acoustic diffraction limit. However, different ULM workflows depend heavily on distinct prior knowledge, such as the impulse response and empirical selection of parameters (e.g., the number of microbubbles (MBs) per frame M), or the consistency of training-test dataset in deep learning (DL)-based studies. We hereby propose a general ULM pipeline that reduces priors. Our approach leverages a DL model that simultaneously distills microbubble signals and reduces speckle from every frame without estimating the impulse response and M. Our method features an efficient channel attention vision transformer (ViT) and a progressive learning strategy, enabling it to learn global information through training on progressively increasing patch sizes. Ample synthetic data were generated using the k-Wave toolbox to simulate various MB patterns, thus overcoming the deficiency of labeled data. The ViT output was further processed by a standard radial symmetry method for sub-pixel localization. Our method performed well on model-unseen public datasets: one in silico dataset with ground truth and four in vivo datasets of mouse tumor, rat brain, rat brain bolus, and rat kidney. Our pipeline outperformed conventional ULM, achieving higher positive predictive values (precision in DL, 0.88-0.41 vs. 0.83-0.16) and improved accuracy (root-mean-square errors: 0.25-0.14 λ vs. 0.31-0.13 λ) across a range of signal-to-noise ratios from 60 dB to 10 dB. Our model could detect more vessels in diverse in vivo datasets while achieving comparable resolutions to the standard method. The proposed ViT-based model, seamlessly integrated with state-of-the-art downstream ULM steps, improved the overall ULM performance with no priors.

Osteoarthritis (OA) is a major cause of disability worldwide, with symptoms like joint pain, limited functionality, and decreased quality of life, potentially leading to deformity and irreversible damage. Chemical changes in joint tissues precede imaging alterations, making early diagnosis challenging for conventional methods like X-rays. Although Raman imaging provides detailed chemical information, it is time-consuming. This paper aims to achieve rapid osteoarthritis diagnosis and grading using a self-developed Raman imaging system combined with deep learning denoising and acceleration algorithms. Our self-developed aberration-corrected line-scanning confocal Raman imaging device acquires a line of Raman spectra (hundreds of points) per scan using a galvanometer or displacement stage, achieving spatial and spectral resolutions of 2 μm and 0.2 nm, respectively. Deep learning algorithms enhance the imaging speed by over 4 times through effective spectrum denoising and signal-to-noise ratio (SNR) improvement. By leveraging the denoising capabilities of deep learning, we are able to acquire high-quality Raman spectral data with a reduced integration time, thereby accelerating the imaging process. Experiments on the tibial plateau of osteoarthritis patients compared three excitation wavelengths (532, 671, and 785 nm), with 671 nm chosen for optimal SNR and minimal fluorescence. Machine learning algorithms achieved a 98 % accuracy in distinguishing articular from calcified cartilage and a 97 % accuracy in differentiating osteoarthritis grades I to IV. Our fast Raman imaging system, combining an aberration-corrected line-scanning confocal Raman imager with deep learning denoising, offers improved imaging speed and enhanced spectral and spatial resolutions. It enables rapid, label-free detection of osteoarthritis severity and can identify early compositional changes before clinical imaging, allowing precise grading and tailored treatment, thus advancing orthopedic diagnostics and improving patient outcomes.

Incoherent k-space under-sampling and deep learning-based reconstruction methods have shown great success in accelerating MRI. However, the performance of most previous methods will degrade dramatically under high acceleration factors, e.g., 8$\times$ or higher. Recently, denoising diffusion models (DM) have demonstrated promising results in solving this issue; however, one major drawback of the DM methods is the long inference time due to a dramatic number of iterative reverse posterior sampling steps. In this work, a Single Step Diffusion Model-based reconstruction framework, namely SSDM-MRI, is proposed for restoring MRI images from highly undersampled k-space. The proposed method achieves one-step reconstruction by first training a conditional DM and then iteratively distilling this model. Comprehensive experiments were conducted on both publicly available fastMRI images and an in-house multi-echo GRE (QSM) subject. Overall, the results showed that SSDM-MRI outperformed other methods in terms of numerical metrics (PSNR and SSIM), qualitative error maps, image fine details, and latent susceptibility information hidden in MRI phase images. In addition, the reconstruction time for a 320*320 brain slice of SSDM-MRI is only 0.45 second, which is only comparable to that of a simple U-net, making it a highly effective solution for MRI reconstruction tasks.

Deep learning reconstruction (DLR) provides an elegant solution for MR acceleration while preserving image quality. This advancement is crucial for body imaging, which is frequently marred by the increased likelihood of motion-related artifacts. Multiple vendor-specific models focusing on T2, T1, and diffusion-weighted imaging have been developed for the abdomen, pelvis, and chest, with the liver and prostate being the most well-studied organ systems. Variational networks with supervised DL models, including data consistency layers and regularizers, are the most common DLR methods. The common theme for all single-center studies on this subject has been noninferior or superior image quality metrics and lesion conspicuity to conventional sequences despite significant acquisition time reduction. DLR also provides a potential for denoising, artifact reduction, increased resolution, and increased signal-noise ratio (SNR) and contrast-to-noise ratio (CNR) that can be balanced with acceleration benefits depending on the imaged organ system. Some specific challenges faced by DLR include slightly reduced lesion detection, cardiac motion-related signal loss, regional SNR variations, and variabilities in ADC measurements as reported in different organ systems. Continued investigations with large-scale multicenter prospective clinical validation of DLR to document generalizability and demonstrate noninferior diagnostic accuracy with histopathologic correlation are the need of the hour. The creation of vendor-neutral solutions, open data sharing, and diversifying training data sets are also critical to strengthening model robustness.

We introduce a new type of foundational model for parsing human anatomy in medical images that works for different modalities. It supports supervised or unsupervised training and can perform matching, registration, classification, or segmentation with or without user interaction. We achieve this by training a neural network estimator that maps query locations to atlas coordinates via regression. Efficiency is improved by sparsely sampling the input, enabling response times of less than 1 ms without additional accelerator hardware. We demonstrate the utility of the algorithm in both CT and MRI modalities.