This study aims to integrate CT imaging with occupational health surveillance data to construct a multimodal model for preclinical CWP identification and individualized risk evaluation. CT images and occupational health surveillance data were retrospectively collected from 874 coal workers, including 228 Stage I and 4 Stage II pneumoconiosis patients, along with 600 healthy and 42 subcategory 0/1 coal workers. First, the YOLOX was employed for automated 3D lung extraction to extract radiomics features. Second, two feature selection algorithms were applied to select critical features from both CT radiomics and occupational health data. Third, three distinct feature sets were constructed for model training: CT radiomics features, occupational health data, and their multimodal integration. Finally, five machine learning models were implemented to predict the preclinical stage of CWP. The model's performance was evaluated using the receiver operating characteristic curve (ROC), accuracy, sensitivity, and specificity. SHapley Additive exPlanation (SHAP) values were calculated to determine the prediction role of each feature in the model with the highest predictive performance. The YOLOX-based lung extraction demonstrated robust performance, achieving an Average Precision (AP) of 0.98. 8 CT radiomic features and 4 occupational health surveillance data were selected for the multimodal model. The optimal occupational health surveillance feature subset comprised the Length of service. Among 5 machine learning algorithms evaluated, the Decision Tree-based multimodal model showed superior predictive capacity on the test set of 142 samples, with an AUC of 0.94 (95% CI 0.88-0.99), accuracy 0.95, specificity 1.00, and Youden's index 0.83. SHAP analysis indicated that Total Protein Results, original shape Flatness, diagnostics Image original Mean were the most influential contributors. Our study demonstrated that the multimodal model demonstrated strong predictive capability for the preclinical stage of CWP by integrating CT radiomic features with occupational health data.

Accurate brain parcellation in diffusion MRI (dMRI) space is essential for advanced neuroimaging analyses. However, most existing approaches rely on anatomical MRI for segmentation and inter-modality registration, a process that can introduce errors and limit the versatility of the technique. In this study, we present a novel deep learning-based framework for direct parcellation based on the Desikan-Killiany (DK) atlas using only diffusion MRI data. Our method utilizes a hierarchical, two-stage segmentation network: the first stage performs coarse parcellation into broad brain regions, and the second stage refines the segmentation to delineate more detailed subregions within each coarse category. We conduct an extensive ablation study to evaluate various diffusion-derived parameter maps, identifying an optimal combination of fractional anisotropy, trace, sphericity, and maximum eigenvalue that enhances parellation accuracy. When evaluated on the Human Connectome Project and Consortium for Neuropsychiatric Phenomics datasets, our approach achieves superior Dice Similarity Coefficients compared to existing state-of-the-art models. Additionally, our method demonstrates robust generalization across different image resolutions and acquisition protocols, producing more homogeneous parcellations as measured by the relative standard deviation within regions. This work represents a significant advancement in dMRI-based brain segmentation, providing a precise, reliable, and registration-free solution that is critical for improved structural connectivity and microstructural analyses in both research and clinical applications. The implementation of our method is publicly available on github.com/xmindflow/DKParcellationdMRI.

Generative artificial intelligence models facilitate open-data sharing by proposing synthetic data as surrogates of real patient data. Despite the promise for healthcare, some of these models are susceptible to patient data memorization, where models generate patient data copies instead of novel synthetic samples, resulting in patient re-identification. Here we assess memorization in unconditional latent diffusion models by training them on a variety of datasets for synthetic data generation and detecting memorization with a self-supervised copy detection approach. We show a high degree of patient data memorization across all datasets, with approximately 37.2% of patient data detected as memorized and 68.7% of synthetic samples identified as patient data copies. Latent diffusion models are more susceptible to memorization than autoencoders and generative adversarial networks, and they outperform non-diffusion models in synthesis quality. Augmentation strategies during training, small architecture size and increasing datasets can reduce memorization, while overtraining the models can enhance it. These results emphasize the importance of carefully training generative models on private medical imaging datasets and examining the synthetic data to ensure patient privacy.

The core role of medical images in disease diagnosis makes their quality directly affect the accuracy of clinical judgment. However, due to factors such as low-dose scanning, equipment limitations and imaging artifacts, medical images are often accompanied by non-uniform noise interference, which seriously affects structure recognition and lesion detection. This paper proposes a medical image adaptive denoising model (MI-ND) that integrates multi-scale convolutional and Transformer architecture, introduces a noise level estimator (NLE) and a noise adaptive attention module (NAAB), and realizes channel-spatial attention regulation and cross-modal feature fusion driven by noise perception. Systematic testing is carried out on multimodal public datasets. Experiments show that this method significantly outperforms the comparative methods in image quality indicators such as PSNR, SSIM, and LPIPS, and improves the F1 score and ROC-AUC in downstream diagnostic tasks, showing strong prac-tical value and promotional potential. The model has outstanding benefits in structural recovery, diagnostic sensitivity, and cross-modal robustness, and provides an effective solution for medical image enhancement and AI-assisted diagnosis and treatment.

The composition of urinary calculi serves as a critical determinant for personalized surgical strategies; however, such compositional data are often unavailable preoperatively. This study aims to develop a machine learning-based preoperative prediction model for stone composition and evaluate its clinical utility. A retrospective cohort study design was employed to include patients with urinary calculi admitted to the Department of Urology at the Second Affiliated Hospital of Zhengzhou University from 2019 to 2024. Feature selection was performed using least absolute shrinkage and selection operator (LASSO) regression combined with multivariate logistic regression, and a binary prediction model for urinary calculi was subsequently constructed. Model validation was conducted using metrics such as the area under the curve (AUC), while Shapley Additive Explanations(SHAP) values were applied to interpret the predictive outcomes. Among 708 eligible patients, distinct prediction models were established for four stone types: calcium oxalate stones: Logistic regression achieved optimal performance (AUC = 0.845), with maximum stone CT value, 24-hour urinary oxalate, and stone size as top predictors (SHAP-ranked); infection stones: Logistic regression (AUC = 0.864) prioritized stone size, urinary pH, and recurrence history; uric acid stones: LASSO-ridge-elastic net model demonstrated exceptional accuracy (AUC = 0.961), driven by maximum CT value, 24-hour oxalate, and urinary calcium; calcium-containing stones: Logistic regression attained better prediction (AUC = 0.953), relying on CT value, 24-hour calcium, and stone size. This study developed a machine learning prediction model based on multi-algorithm integration, achieving accurate preoperative discrimination of urinary stone composition. The integration of key imaging features with metabolic indicators enhanced the model's predictive performance.

Achieving clinical level performance and widespread deployment for generating radiology impressions encounters a giant challenge for conventional artificial intelligence models tailored to specific diseases and organs. Concurrent with the increasing accessibility of radiology reports and advancements in modern general AI techniques, the emergence and potential of deployable radiology AI exploration have been bolstered. Here, we present ChatRadio-Valuer, the first general radiology diagnosis large language model for localized deployment within hospitals and being close to clinical use for multi-institution and multi-system diseases. ChatRadio-Valuer achieved 15 state-of-the-art results across five human systems and six institutions in clinical-level events (n=332,673) through rigorous and full-spectrum assessment, including engineering metrics, clinical validation, and efficiency evaluation. Notably, it exceeded OpenAI's GPT-3.5 and GPT-4 models, achieving superior performance in comprehensive disease diagnosis compared to the average level of radiology experts. Besides, ChatRadio-Valuer supports zero-shot transfer learning, greatly boosting its effectiveness as a radiology assistant, while ensuring adherence to privacy standards and being readily utilized for large-scale patient populations. Our expeditions suggest the development of localized LLMs would become an imperative avenue in hospital applications.

Accelerating Magnetic Resonance Imaging (MRI) reduces scan time but often degrades image quality. While Implicit Neural Representations (INRs) show promise for MRI reconstruction, they struggle at high acceleration factors due to weak prior constraints, leading to structural loss and aliasing artefacts. To address this, we propose PrIINeR, an INR-based MRI reconstruction method that integrates prior knowledge from pre-trained deep learning models into the INR framework. By combining population-level knowledge with instance-based optimization and enforcing dual data consistency, PrIINeR aligns both with the acquired k-space data and the prior-informed reconstruction. Evaluated on the NYU fastMRI dataset, our method not only outperforms state-of-the-art INR-based approaches but also improves upon several learning-based state-of-the-art methods, significantly improving structural preservation and fidelity while effectively removing aliasing artefacts.PrIINeR bridges deep learning and INR-based techniques, offering a more reliable solution for high-quality, accelerated MRI reconstruction. The code is publicly available on https://github.com/multimodallearning/PrIINeR.

Epilepsy affects over 50 million people worldwide, with anti-seizure medications (ASMs) as the primary treatment for seizure control. However, ASM selection remains a "trial and error" process due to the lack of reliable predictors of effectiveness and tolerability. While machine learning approaches have been explored, existing models are limited to predicting outcomes only for ASMs encountered during training and have not leveraged recent biomedical foundation models for this task. This work investigates ASM outcome prediction using only patient MRI scans and reports. Specifically, we leverage biomedical vision-language foundation models and introduce a novel contextualized instruction-tuning framework that integrates expert-built knowledge trees of MRI entities to enhance their performance. Additionally, by training only on the four most commonly prescribed ASMs, our framework enables generalization to predicting outcomes and effectiveness for unseen ASMs not present during training. We evaluate our instruction-tuning framework on two retrospective epilepsy patient datasets, achieving an average AUC of 71.39 and 63.03 in predicting outcomes for four primary ASMs and three completely unseen ASMs, respectively. Our approach improves the AUC by 5.53 and 3.51 compared to standard report-based instruction tuning for seen and unseen ASMs, respectively. Our code, MRI knowledge tree, prompting templates, and TREE-TUNE generated instruction-answer tuning dataset are available at the link.

We propose a deep learning-based approach that integrates MRI sequence parameters to improve the accuracy and generalizability of quantitative image synthesis from clinical weighted MRI. Our physics-driven neural network embeds MRI sequence parameters -- repetition time (TR), echo time (TE), and inversion time (TI) -- directly into the model via parameter embedding, enabling the network to learn the underlying physical principles of MRI signal formation. The model takes conventional T1-weighted, T2-weighted, and T2-FLAIR images as input and synthesizes T1, T2, and proton density (PD) quantitative maps. Trained on healthy brain MR images, it was evaluated on both internal and external test datasets. The proposed method achieved high performance with PSNR values exceeding 34 dB and SSIM values above 0.92 for all synthesized parameter maps. It outperformed conventional deep learning models in accuracy and robustness, including data with previously unseen brain structures and lesions. Notably, our model accurately synthesized quantitative maps for these unseen pathological regions, highlighting its superior generalization capability. Incorporating MRI sequence parameters via parameter embedding allows the neural network to better learn the physical characteristics of MR signals, significantly enhancing the performance and reliability of quantitative MRI synthesis. This method shows great potential for accelerating qMRI and improving its clinical utility.

Retrospective cross-sectional study. To develop a multi-view fusion framework that effectively identifies suspect keratoconus cases and facilitates the possibility of early clinical intervention. A total of 573 corneal topography maps representing eyes classified as normal, suspect, or keratoconus. We designed the Corneal Multi-View Fusion Transformer (CMVFT), which integrates features from seven standard corneal topography maps. A pretrained ResNet-50 extracts single-view representations that are further refined by a custom-designed Multi-Scale Attention Module (MSAM). This integrated design specifically compensates for the representation gap commonly encountered when applying Transformers to small-sample corneal topography datasets by dynamically bridging local convolution-based feature extraction with global self-attention mechanisms. A subsequent fusion Transformer then models long-range dependencies across views for comprehensive multi-view feature integration. The primary measure was the framework's ability to differentiate suspect cases from normal and keratoconus cases, thereby creating a pathway for early clinical intervention. Experimental evaluation demonstrated that CMVFT effectively distinguishes suspect cases within a feature space characterized by overlapping attributes. Ablation studies confirmed that both the MSAM and the fusion Transformer are essential for robust multi-view feature integration, successfully compensating for potential representation shortcomings in small datasets. This study is the first to apply a Transformer-driven multi-view fusion approach in corneal topography analysis. By compensating for the representation gap inherent in small-sample settings, CMVFT shows promise in enabling the identification of suspect keratoconus cases and supporting early intervention strategies, with prospective implications for early clinical intervention.