Cervical cancer is a significant global health issue, and high-dose-rate brachytherapy (HDR-BT) is crucial for its treatment. However, manually creating HDR-BT plans is time-consuming and heavily relies on the planner's expertise, making standardization difficult. This study introduces two advanced deep learning models to address this need: Bi-branch Cross-Attention UNet (BiCA-UNet) and Dendrite Cross-Attention UNet (DCA-UNet). BiCA-UNet enhances the correlation between the CT scan and segmentation maps of the clinical target volume (CTV), applicator, bladder, and rectum. It uses two branches: one processes the stacked input of CT scans and segmentations, and the other focuses on the CTV segmentation. A cross-attention mechanism integrates these branches, improving the model's understanding of the CTV region for accurate dose predictions. Building on BiCA-UNet, DCA-UNet further introduces a primary branch of stacked inputs and three secondary branches for CTV, bladder, and rectum segmentations forming a dendritic structure. Cross attention with bladder and rectum segmentation helps the model understand the regions of organs at risk (OAR), refining dose prediction. Evaluation of these models using multiple metrics indicates that both BiCA-UNet and DCA-UNet significantly improve HDR-BT dose prediction accuracy for various applicator types. The cross-attention mechanisms enhance the feature representation of critical anatomical regions, leading to precise and reliable treatment plans. This research highlights the potential of BiCA-UNet and DCA-UNet in advancing HDR-BT planning, contributing to the standardization of treatment plans, and offering promising directions for future research to improve patient outcomes in the source data.

Cervical cancer is a leading cause of death from malignant tumors in women, and accurate evaluation of occult lymph node metastasis (OLNM) is crucial for optimal treatment. This study aimed to develop several predictive models-including Clinical model, Radiomics models (RD), Deep Learning models (DL), Radiomics-Deep Learning fusion models (RD-DL), and a Clinical-RD-DL combined model-for assessing the risk of OLNM in cervical cancer patients.The study included 130 patients from Center 1 (training set) and 55 from Center 2 (test set). Clinical data and imaging sequences (T1, T2, and DWI) were used to extract features for model construction. Model performance was assessed using the DeLong test, and SHAP analysis was used to examine feature contributions. Results showed that both the RD-combined (AUC = 0.803) and DL-combined (AUC = 0.818) models outperformed single-sequence models as well as the standalone Clinical model (AUC = 0.702). The RD-DL model yielded the highest performance, achieving an AUC of 0.981 in the training set and 0.903 in the test set. Notably, integrating clinical variables did not further improve predictive performance; the Clinical-RD-DL model performed comparably to the RD-DL model. SHAP analysis showed that deep learning features had the greatest impact on model predictions. Both RD and DL models effectively predict OLNM, with the RD-DL model offering superior performance. These findings provide a rapid, non-invasive clinical prediction method.

Alzheimer's disease is a progressive neurodegenerative disorder marked by cognitive decline, memory loss, and behavioral changes. Early diagnosis, particularly identifying Early Mild Cognitive Impairment (EMCI), is vital for managing the disease and improving patient outcomes. Detecting EMCI is challenging due to the subtle structural changes in the brain, making precise slice selection from MRI scans essential for accurate diagnosis. In this context, the careful selection of specific MRI slices that provide distinct anatomical details significantly enhances the ability to identify these early changes. The chief novelty of the study is that instead of selecting all slices, an approach for identifying the important slices is developed. The ADNI-3 dataset was used as the dataset when running the models for early detection of Alzheimer's disease. Satisfactory results have been obtained by classifying with deep learning models, vision transformers (ViT) and by adding new structures to them, together with the model proposal. In the results obtained, while an accuracy of 99.45% was achieved with EfficientNetB2 + FPN in AD vs. LMCI classification from the slices selected with SSIM, an accuracy of 99.19% was achieved in AD vs. EMCI classification, in fact, the study significantly advances early detection by demonstrating improved diagnostic accuracy of the disease at the EMCI stage. The results obtained with these methods emphasize the importance of developing deep learning models with slice selection integrated with the Vision Transformers architecture. Focusing on accurate slice selection enables early detection of Alzheimer's at the EMCI stage, allowing for timely interventions and preventive measures before the disease progresses to more advanced stages. This approach not only facilitates early and accurate diagnosis, but also lays the groundwork for timely intervention and treatment, offering hope for better patient outcomes in Alzheimer's disease. The study is finally evaluated by a statistical significance test.

Pulmonary disease can severely impair respiratory function and be life-threatening. Accurately recognizing pulmonary diseases in chest X-ray images is challenging due to overlapping body structures and the complex anatomy of the chest. We propose an adaptive multiscale feature fusion model for recognizing Chest X-ray images of pneumonia, tuberculosis, and COVID-19, which are common pulmonary diseases. We introduce an Adaptive Multiscale Fusion Network (AMFNet) for pulmonary disease classification in chest X-ray images. AMFNet consists of a lightweight Multiscale Fusion Network (MFNet) and ResNet50 as the secondary feature extraction network. MFNet employs Fusion Blocks with self-calibrated convolution (SCConv) and Attention Feature Fusion (AFF) to capture multiscale semantic features, and integrates a custom activation function, MFReLU, which is employed to reduce the model's memory access time. A fusion module adaptively combines features from both networks. Experimental results show that AMFNet achieves 97.48% accuracy and an F1 score of 0.9781 on public datasets, outperforming models like ResNet50, DenseNet121, ConvNeXt-Tiny, and Vision Transformer while using fewer parameters.

Generative image models have achieved remarkable progress in both natural and medical imaging. In the medical context, these techniques offer a potential solution to data scarcity-especially for low-prevalence anomalies that impair the performance of AI-driven diagnostic and segmentation tools. However, questions remain regarding the fidelity and clinical utility of synthetic images, since poor generation quality can undermine model generalizability and trust. In this study, we evaluate the effectiveness of state-of-the-art generative models-Generative Adversarial Networks (GANs) and Diffusion Models (DMs)-for synthesizing chest X-rays conditioned on four abnormalities: Atelectasis (AT), Lung Opacity (LO), Pleural Effusion (PE), and Enlarged Cardiac Silhouette (ECS). Using a benchmark composed of real images from the MIMIC-CXR dataset and synthetic images from both GANs and DMs, we conducted a reader study with three radiologists of varied experience. Participants were asked to distinguish real from synthetic images and assess the consistency between visual features and the target abnormality. Our results show that while DMs generate more visually realistic images overall, GANs can report better accuracy for specific conditions, such as absence of ECS. We further identify visual cues radiologists use to detect synthetic images, offering insights into the perceptual gaps in current models. These findings underscore the complementary strengths of GANs and DMs and point to the need for further refinement to ensure generative models can reliably augment training datasets for AI diagnostic systems.

Label scarcity remains a major challenge in deep learning-based medical image segmentation. Recent studies use strong-weak pseudo supervision to leverage unlabeled data. However, performance is often hindered by inconsistencies between pseudo labels and their corresponding unlabeled images. In this work, we propose \textbf{SynMatch}, a novel framework that sidesteps the need for improving pseudo labels by synthesizing images to match them instead. Specifically, SynMatch synthesizes images using texture and shape features extracted from the same segmentation model that generates the corresponding pseudo labels for unlabeled images. This design enables the generation of highly consistent synthesized-image-pseudo-label pairs without requiring any training parameters for image synthesis. We extensively evaluate SynMatch across diverse medical image segmentation tasks under semi-supervised learning (SSL), weakly-supervised learning (WSL), and barely-supervised learning (BSL) settings with increasingly limited annotations. The results demonstrate that SynMatch achieves superior performance, especially in the most challenging BSL setting. For example, it outperforms the recent strong-weak pseudo supervision-based method by 29.71\% and 10.05\% on the polyp segmentation task with 5\% and 10\% scribble annotations, respectively. The code will be released at https://github.com/Senyh/SynMatch.

Multi-sequence Magnetic Resonance Imaging (MRI) offers remarkable versatility, enabling the distinct visualization of different tissue types. Nevertheless, the inherent heterogeneity among MRI sequences poses significant challenges to the generalization capability of deep learning models. These challenges undermine model performance when faced with varying acquisition parameters, thereby severely restricting their clinical utility. In this study, we present PRISM, a foundation model PRe-trained with large-scale multI-Sequence MRI. We collected a total of 64 datasets from both public and private sources, encompassing a wide range of whole-body anatomical structures, with scans spanning diverse MRI sequences. Among them, 336,476 volumetric MRI scans from 34 datasets (8 public and 26 private) were curated to construct the largest multi-organ multi-sequence MRI pretraining corpus to date. We propose a novel pretraining paradigm that disentangles anatomically invariant features from sequence-specific variations in MRI, while preserving high-level semantic representations. We established a benchmark comprising 44 downstream tasks, including disease diagnosis, image segmentation, registration, progression prediction, and report generation. These tasks were evaluated on 32 public datasets and 5 private cohorts. PRISM consistently outperformed both non-pretrained models and existing foundation models, achieving first-rank results in 39 out of 44 downstream benchmarks with statistical significance improvements. These results underscore its ability to learn robust and generalizable representations across unseen data acquired under diverse MRI protocols. PRISM provides a scalable framework for multi-sequence MRI analysis, thereby enhancing the translational potential of AI in radiology. It delivers consistent performance across diverse imaging protocols, reinforcing its clinical applicability.

Stress is a natural response of the body to perceived threats, and it can have both positive and negative effects on brain hemodynamics. Stress-induced changes in pupil and eyelid size/shape have been used as a biomarker in several fMRI studies. However, there were limited knowledges regarding changes in behavior of pupil and eyelid dynamics, particularly on animal models. In the present study, the pupil and eyelid dynamics were carefully investigated and characterized in a newly developed awake rodent fMRI protocol. Leveraging deep learning techniques, the mouse pupil and eyelid diameters were extracted and analyzed during different training and imaging phases in the present project. Our findings demonstrate a consistent downwards trend in pupil and eyelid dynamics under a meticulously designed training protocol, suggesting that the behaviors of the pupil and eyelid can be served as reliable indicators of stress levels and motion artifacts in awake fMRI studies. The current recording platform not only enables the facilitation of awake animal MRI studies but also highlights its potential applications to numerous other research areas, owing to the non-invasive nature and straightforward implementation.

To visualize and assess hematoma growth trends by generating follow-up CT images within 24 h based on baseline CT images of spontaneous intracerebral hemorrhage (sICH) using Transformer-integrated Generative Adversarial Networks (GAN). Patients with sICH were retrospectively recruited from two medical centers. The imaging data included baseline non-contrast CT scans taken after onset and follow-up imaging within 24 h. In the test set, the peak signal-to-noise ratio (PSNR) and the structural similarity index measure (SSIM) were utilized to quantitatively assess the quality of the predicted images. Pearson's correlation analysis was performed to assess the agreement of semantic features and geometric properties of hematomas between true follow-up CT images and the predicted images. The consistency of hematoma expansion prediction between true and generated images was further examined. The PSNR of the predicted images was 26.73 ± 1.11, and the SSIM was 91.23 ± 1.10. The Pearson correlation coefficients (r) with 95 % confidence intervals (CI) for irregularity, satellite sign number, intraventricular or subarachnoid hemorrhage, midline shift, edema expansion, mean CT value, maximum cross-sectional area, and hematoma volume between the predicted and true follow-up images were as follows: 0.94 (0.91, 0.96), 0.87 (0.81, 0.91), 0.86 (0.80, 0.91), 0.89 (0.84, 0.92), 0.91 (0.87, 0.94), 0.78(0.68, 0.84), 0.94(0.91, 0.96), and 0.94 (0.91, 0.96), respectively. The correlation coefficient (r) for predicting hematoma expansion between predicted and true follow-up images was 0.86 (95 % CI: 0.79, 0.90; P < 0.001). The model constructed using a GAN integrated with Transformer modules can accurately visualize early hematoma changes in sICH.

Skull stripping is a common preprocessing step in Magnetic Resonance Imaging (MRI) pipelines and is often performed manually. Automating this process is challenging for preclinical data due to variations in brain geometry, resolution, and tissue contrast. Existing methods for MRI skull stripping often struggle with the low resolution and varying slice sizes found in preclinical functional MRI (fMRI) data. This study proposes a novel method that integrates a Dense UNet-based architecture with a feature extractor based on the Smart Swin Transformer (SST), called SST-DUNet. The Smart Shifted Window Multi-Head Self-Attention (SSW-MSA) module in SST replaces the mask-based module in the Swin Transformer (ST), enabling the learning of distinct channel-wise features while focusing on relevant dependencies within brain structures. This modification allows the model to better handle the complexities of fMRI skull stripping, such as low resolution and variable slice sizes. To address class imbalance in preclinical data, a combined loss function using Focal and Dice loss is applied. The model was trained on rat fMRI images and evaluated across three in-house datasets, achieving Dice similarity scores of 98.65%, 97.86%, and 98.04%. We compared our method with conventional and deep learning-based approaches, demonstrating its superiority over state-of-the-art methods. The fMRI results using SST-DUNet closely align with those from manual skull stripping for both seed-based and independent component analyses, indicating that SST-DUNet can effectively substitute manual brain extraction in rat fMRI analysis.