Transformers have revolutionized medical image restoration, but the quadratic complexity still poses limitations for their application to high-resolution medical images. The recent advent of the Receptance Weighted Key Value (RWKV) model in the natural language processing field has attracted much attention due to its ability to process long sequences efficiently. To leverage its advanced design, we propose Restore-RWKV, the first RWKV-based model for medical image restoration. Since the original RWKV model is designed for 1D sequences, we make two necessary modifications for modeling spatial relations in 2D medical images. First, we present a recurrent WKV (Re-WKV) attention mechanism that captures global dependencies with linear computational complexity. Re-WKV incorporates bidirectional attention as basic for a global 16 receptive field and recurrent attention to effectively model 2D dependencies from various scan directions. Second, we develop an omnidirectional token shift (Omni-Shift) layer that enhances local dependencies by shifting tokens from all directions and across a wide context range. These adaptations make the proposed Restore-RWKV an efficient and effective model for medical image restoration. Even a lightweight variant of Restore-RWKV, with only 1.16 million parameters, achieves comparable or even superior results compared to existing state-of-the-art (SOTA) methods. Extensive experiments demonstrate that the resulting Restore-RWKV achieves SOTA performance across a range of medical image restoration tasks, including PET image synthesis, CT image denoising, MRI image superresolution, and all-in-one medical image restoration. Code is available at: https://github.com/Yaziwel/Restore-RWKV.

The escalating adoption of high-resolution, large-field-of-view imagery amplifies the need for efficient compression methodologies. Conventional techniques frequently fail to preserve critical image details, while data-driven approaches exhibit limited generalizability. Implicit Neural Representations (INRs) present a promising alternative by learning continuous mappings from spatial coordinates to pixel intensities for individual images, thereby storing network weights rather than raw pixels and avoiding the generalization problem. However, INR-based compression of large images faces challenges including slow compression speed and suboptimal compression ratios. To address these limitations, we introduce COLI (Compressor for Large Images), a novel framework leveraging Neural Representations for Videos (NeRV). First, recognizing that INR-based compression constitutes a training process, we accelerate its convergence through a pretraining-finetuning paradigm, mixed-precision training, and reformulation of the sequential loss into a parallelizable objective. Second, capitalizing on INRs' transformation of image storage constraints into weight storage, we implement Hyper-Compression, a novel post-training technique to substantially enhance compression ratios while maintaining minimal output distortion. Evaluations across two medical imaging datasets demonstrate that COLI consistently achieves competitive or superior PSNR and SSIM metrics at significantly reduced bits per pixel (bpp), while accelerating NeRV training by up to 4 times.

Deformable image registration (DIR) is critical for accurate clinical diagnosis and effective treatment planning. However, patient movement, significant intensity differences, and large breathing deformations hinder accurate anatomical alignment in multi-modal image registration. These factors exacerbate the entanglement of anatomical and modality-specific style information, thereby severely limiting the performance of multi-modal registration. To address this, we propose a novel LAtent Diffusion-based Domain-Adaptive feature disentangling (LADDA) framework for unsupervised multi-modal medical image registration, which explicitly addresses the representation disentanglement. First, LADDA extracts reliable anatomical priors from the Latent Diffusion Model (LDM), facilitating downstream content-style disentangled learning. A Domain-Adaptive Feature Disentangling (DAFD) module is proposed to promote anatomical structure alignment further. This module disentangles image features into content and style information, boosting the network to focus on cross-modal content information. Next, a Neighborhood-Preserving Hashing (NPH) is constructed to further perceive and integrate hierarchical content information through local neighbourhood encoding, thereby maintaining cross-modal structural consistency. Furthermore, a Unilateral-Query-Frozen Attention (UQFA) module is proposed to enhance the coupling between upstream prior and downstream content information. The feature interaction within intra-domain consistent structures improves the fine recovery of detailed textures. The proposed framework is extensively evaluated on large-scale multi-center datasets, demonstrating superior performance across diverse clinical scenarios and strong generalization on out-of-distribution (OOD) data.

Semantic segmentation of medical images is pivotal in applications like disease diagnosis and treatment planning. While deep learning automates this task effectively, it struggles in ultra low-data regimes for the scarcity of annotated segmentation masks. To address this, we propose a generative deep learning framework that produces high-quality image-mask pairs as auxiliary training data. Unlike traditional generative models that separate data generation from model training, ours uses multi-level optimization for end-to-end data generation. This allows segmentation performance to guide the generation process, producing data tailored to improve segmentation outcomes. Our method demonstrates strong generalization across 11 medical image segmentation tasks and 19 datasets, covering various diseases, organs, and modalities. It improves performance by 10-20% (absolute) in both same- and out-of-domain settings and requires 8-20 times less training data than existing approaches. This greatly enhances the feasibility and cost-effectiveness of deep learning in data-limited medical imaging scenarios.

The objective and reliable quantification of adenoid size is pivotal for precise clinical diagnosis and the formulation of effective treatment strategies. Conventional manual measurement techniques, however, are often labor-intensive and time-consuming. To develop and validate a fully automated system for measuring adenoid size using deep learning (DL) on lateral head and neck radiographs. In this retrospective study, we analyzed 711 lateral head and neck radiographs collected from two centers between February and July 2023. A DL-based adenoid size measurement system was developed, utilizing Fujioka's method. The system employed the RTMDet network and RTMPose networks for accurate landmark detection, and mathematical formulas were applied to determine adenoid size. To evaluate consistency and reliability of the system, we employed the intra-class correlation coefficient (ICC), mean absolute difference (MAD), and Bland-Altman plots as key assessment metrics. The DL-based system exhibited high reliability in the prediction of adenoid, nasopharynx, and adenoid-nasopharyngeal ratio measurements, showcasing strong agreement with the reference standard. The results indicated an ICC for adenoid measurements of 0.902 [95%CI, 0.872-0.925], with a MAD of 1.189 and a root mean square (RMS) of 1.974. For nasopharynx measurements, the ICC was 0.868 [95%CI, 0.828-0.899], with a MAD of 1.671 and an RMS of 1.916. Additionally, the adenoid-nasopharyngeal ratio measurements yielded an ICC of 0.911 [95%CI, 0.883-0.932], a MAD of 0.054, and an RMS of 0.076. The developed DL-based system effectively automates the measurement of the adenoid-nasopharyngeal ratio, adenoid, and nasopharynx on lateral neck or head radiographs, showcasing high reliability.

Lymphomas are a diverse group of disorders characterized by the clonal proliferation of lymphocytes. While definitive diagnosis of lymphoma relies on histopathology, immune-phenotyping and additional molecular analyses, imaging modalities such as PET/CT, CT, and MRI play a central role in the diagnostic process and management, from assessing disease extent, to evaluation of response to therapy and detecting recurrence. Artificial intelligence (AI), particularly deep learning models like convolutional neural networks (CNNs), is transforming lymphoma imaging by enabling automated detection, segmentation, and classification. This review elaborates on recent advancements in deep learning for lymphoma imaging and its integration into clinical practice. Challenges include obtaining high-quality, annotated datasets, addressing biases in training data, and ensuring consistent model performance. Ongoing efforts are focused on enhancing model interpretability, incorporating diverse patient populations to improve generalizability, and ensuring safe and effective integration of AI into clinical workflows, with the goal of improving patient outcomes.

The success of deep learning in 3D medical image segmentation hinges on training with a large dataset of fully annotated 3D volumes, which are difficult and time-consuming to acquire. Although recent foundation models (e.g., segment anything model, SAM) can utilize sparse annotations to reduce annotation costs, segmentation tasks involving organs and tissues with blurred boundaries remain challenging. To address this issue, we propose a region uncertainty estimation framework for Computed Tomography (CT) image segmentation using noisy labels. Specifically, we propose a sample-stratified training strategy that stratifies samples according to their varying quality labels, prioritizing confident and fine-grained information at each training stage. This sample-to-voxel level processing enables more reliable supervision information to propagate to noisy label data, thus effectively mitigating the impact of noisy annotations. Moreover, we further design a boundary-guided regional uncertainty estimation module that adapts sample hierarchical training to assist in evaluating sample confidence. Experiments conducted across multiple CT datasets demonstrate the superiority of our proposed method over several competitive approaches under various noise conditions. Our proposed reliable label propagation strategy not only significantly reduces the cost of medical image annotation and robust model training but also improves the segmentation performance in scenarios with imperfect annotations, thus paving the way towards the application of medical segmentation foundation models under low-resource and remote scenarios. Code will be available at https://github.com/KHan-UJS/NoisyLabel.

Radiology reports can change during workflows, especially when residents draft preliminary versions that attending physicians finalize. We explored how large language models (LLMs) and embedding techniques can categorize these changes into textual, semantic, or clinically actionable types. We evaluated 400 adult CT reports drafted by residents against finalized versions by attending physicians. Changes were rated on a five-point scale from no changes to critical ones. We examined open-source LLMs alongside traditional metrics like normalized word differences, Levenshtein and Jaccard similarity, and text embedding similarity. Model performance was assessed using quadratic weighted Cohen's kappa (κ), (balanced) accuracy, F₁, precision, and recall. Inter-rater reliability among evaluators was excellent (κ = 0.990). Of the reports analyzed, 1.3 % contained critical changes. The tested methods showed significant performance differences (P < 0.001). The Qwen3-235B-A22B model using a zero-shot prompt, most closely aligned with human assessments of changes in clinical reports, achieving a κ of 0.822 (SD 0.031). The best conventional metric, word difference, had a κ of 0.732 (SD 0.048), the difference between the two showed statistical significance in unadjusted post-hoc tests (P = 0.038) but lost significance after adjusting for multiple testing (P = 0.064). Embedding models underperformed compared to LLMs and classical methods, showing statistical significance in most cases. Large language models like Qwen3-235B-A22B demonstrated moderate to strong alignment with expert evaluations of the clinical significance of changes in radiology reports. LLMs outperformed embedding methods and traditional string and word approaches, achieving statistical significance in most instances. This demonstrates their potential as tools to support peer review.

Artificial intelligence (AI)-empowered data processing tools improve our ability to assess, measure, and enhance medical interventions. AI-based tools automate the extraction of data from histories, test results, imaging, prescriptions, and treatment outcomes, and transform them into unified, accessible records. They are powerful in converting unstructured data such as clinical notes, magnetic resonance images, and electroencephalograms into structured, actionable formats. For example, in the extraction and classification of diseases, symptoms, medications, treatments, and dates from even incomplete and fragmented clinical notes, pathology reports, images, and histological markers. Especially because the demographics within correctional facilities greatly diverge from the general population, the adoption of electronic health records and AI-enabled data processing will play a crucial role in improving disease detection, treatment management, and the overall efficiency of health care within prison systems.

This review paper comprehensively summarizes healthcare provider (HCP) evaluation of explanations produced by explainable artificial intelligence methods to support point-of-care, patient-specific, clinical decision-making (CDM) within medical settings. It highlights the critical need to incorporate human-centered (HCP) evaluation approaches based on their CDM needs, processes, and goals. The review was conducted in Ovid Medline and Scopus databases, following the Institute of Medicine's methodological standards and PRISMA guidelines. An individual study appraisal was conducted using design-specific appraisal tools. MaxQDA software was used for data extraction and evidence table procedures. Of the 2673 unique records retrieved, 25 records were included in the final sample. Studies were excluded if they did not meet this review's definitions of HCP evaluation (1156), healthcare use (995), explainable AI (211), and primary research (285), and if they were not available in English (1). The sample focused primarily on physicians and diagnostic imaging use cases and revealed wide-ranging evaluation measures. The synthesis of sampled studies suggests a potential common measure of clinical explainability with 3 indicators of interpretability, fidelity, and clinical value. There is an opportunity to extend the current model-centered evaluation approaches to incorporate human-centered metrics, supporting the transition into practice. Future research should aim to clarify and expand key concepts in HCP evaluation, propose a comprehensive evaluation model positioned in current theoretical knowledge, and develop a valid instrument to support comparisons.