As medical imaging is central to diagnostic processes, automating the generation of radiology reports has become increasingly relevant to assist radiologists with their heavy workloads. Most current methods rely solely on global image features, failing to capture fine-grained organ relationships crucial for accurate reporting. To this end, we propose CT-GRAPH, a hierarchical graph attention network that explicitly models radiological knowledge by structuring anatomical regions into a graph, linking fine-grained organ features to coarser anatomical systems and a global patient context. Our method leverages pretrained 3D medical feature encoders to obtain global and organ-level features by utilizing anatomical masks. These features are further refined within the graph and then integrated into a large language model to generate detailed medical reports. We evaluate our approach for the task of report generation on the large-scale chest CT dataset CT-RATE. We provide an in-depth analysis of pretrained feature encoders for CT report generation and show that our method achieves a substantial improvement of absolute 7.9\% in F1 score over current state-of-the-art methods. The code is publicly available at https://github.com/hakal104/CT-GRAPH.

Medical image understanding plays a crucial role in enabling automated diagnosis and data-driven clinical decision support. However, its progress is impeded by two primary challenges: the limited availability of high-quality annotated medical data and an overreliance on global image features, which often miss subtle but clinically significant pathological regions. To address these issues, we introduce RegionMed-CLIP, a region-aware multimodal contrastive learning framework that explicitly incorporates localized pathological signals along with holistic semantic representations. The core of our method is an innovative region-of-interest (ROI) processor that adaptively integrates fine-grained regional features with the global context, supported by a progressive training strategy that enhances hierarchical multimodal alignment. To enable large-scale region-level representation learning, we construct MedRegion-500k, a comprehensive medical image-text corpus that features extensive regional annotations and multilevel clinical descriptions. Extensive experiments on image-text retrieval, zero-shot classification, and visual question answering tasks demonstrate that RegionMed-CLIP consistently exceeds state-of-the-art vision language models by a wide margin. Our results highlight the critical importance of region-aware contrastive pre-training and position RegionMed-CLIP as a robust foundation for advancing multimodal medical image understanding.

We developed a pipeline for registering pre-surgery Magnetic Resonance (MR) images and post-resection Ultrasound (US) images. Our approach leverages unpaired style transfer using 3D CycleGAN to generate synthetic T1 images, thereby enhancing registration performance. Additionally, our registration process employs both affine and local deformable transformations for a coarse-to-fine registration. The results demonstrate that our approach improves the consistency between MR and US image pairs in most cases.

Medical image segmentation plays a crucial role in AI-assisted diagnostics, surgical planning, and treatment monitoring. Accurate and robust segmentation models are essential for enabling reliable, data-driven clinical decision making across diverse imaging modalities. Given the inherent variability in image characteristics across modalities, developing a unified model capable of generalizing effectively to multiple modalities would be highly beneficial. This model could streamline clinical workflows and reduce the need for modality-specific training. However, real-world deployment faces major challenges, including data scarcity, domain shift between modalities (e.g., CT vs. MRI), and privacy restrictions that prevent data sharing. To address these issues, we propose FedGIN, a Federated Learning (FL) framework that enables multimodal organ segmentation without sharing raw patient data. Our method integrates a lightweight Global Intensity Non-linear (GIN) augmentation module that harmonizes modality-specific intensity distributions during local training. We evaluated FedGIN using two types of datasets: an imputed dataset and a complete dataset. In the limited dataset scenario, the model was initially trained using only MRI data, and CT data was added to assess its performance improvements. In the complete dataset scenario, both MRI and CT data were fully utilized for training on all clients. In the limited-data scenario, FedGIN achieved a 12 to 18% improvement in 3D Dice scores on MRI test cases compared to FL without GIN and consistently outperformed local baselines. In the complete dataset scenario, FedGIN demonstrated near-centralized performance, with a 30% Dice score improvement over the MRI-only baseline and a 10% improvement over the CT-only baseline, highlighting its strong cross-modality generalization under privacy constraints.

Spinal radiology reports and physician-completed questionnaires serve as crucial resources for medical decision-making for patients experiencing low back and neck pain. However, due to the time-consuming nature of this process, individuals with severe conditions may experience a deterioration in their health before receiving professional care. In this work, we propose an ensemble framework built on top of pre-trained BERT-based models which can classify patients on their need for surgery given their different data modalities including radiology reports and questionnaires. Our results demonstrate that our approach exceeds previous studies, effectively integrating information from multiple data modalities and serving as a valuable tool to assist physicians in decision making.

This paper reviews the current research on Alzheimer's disease and the use of deep learning, particularly 3D-convolutional neural networks (3D-CNN), in analyzing brain images. It presents a predictive model based on MRI and clinical data from the ADNI dataset, showing that deep learning can improve diagnosis accuracy and sensitivity. We also discuss potential applications in biomarker discovery, disease progression prediction, and personalised treatment planning, highlighting the ability to identify sensitive features for early diagnosis.

This study aims to enhance the accuracy of fetal ultrasound image classification using convolutional neural networks, specifically EfficientNet. The research focuses on data collection, preprocessing, model training, and evaluation at different pregnancy stages: early, midterm, and newborn. EfficientNet showed the best performance, particularly in the newborn stage, demonstrating deep learning's potential to improve classification performance and support clinical workflows.

Routine duplex ultrasound surveillance is recommended after femoral-popliteal and femoral-tibial-pedal vein bypass grafts at various post-operative intervals. Currently, there is no systematic method for bypass graft surveillance using a set of peak systolic velocities (PSVs) collected during these exams. This research aims to explore the use of recurrent neural networks to predict the next set of PSVs, which can then indicate occlusion status. Recurrent neural network models were developed to predict occlusion and stenosis based on one to three prior sets of PSVs, with a sequence-to-sequence model utilized to forecast future PSVs within the stent graft and nearby arteries. The study employed 5-fold cross-validation for model performance comparison, revealing that the BiGRU model outperformed BiLSTM when two or more sets of PSVs were included, demonstrating that increasing duplex ultrasound exams improve prediction accuracy and reduces error rates. This work establishes a basis for integrating comprehensive clinical data, including demographics, comorbidities, symptoms, and other risk factors, with PSVs to enhance lower extremity bypass graft surveillance predictions.

This study proposes RRG-LLM, a model designed to enhance RRG by effectively learning medical domain with minimal computational resources. Initially, LLM is finetuned by LoRA, enabling efficient adaptation to the medical domain. Subsequently, only the linear projection layer that project the image into text is finetuned to extract important information from the radiology image and project it onto the text dimension. Proposed model demonstrated notable improvements in report generation. The performance of ROUGE-L was improved by 0.096 (51.7%) and METEOR by 0.046 (42.85%) compared to the baseline model.

The diagnosis of craniosynostosis, a condition involving the premature fusion of cranial sutures in infants, is essential for ensuring timely treatment and optimal surgical outcomes. Current diagnostic approaches often require CT scans, which expose children to significant radiation risks. To address this, we present a novel deep learning-based model utilizing multi-view X-ray images for craniosynostosis detection. The proposed model integrates advanced multi-view fusion (MVF) and cross-attention mechanisms, effectively combining features from three X-ray views (AP, lateral right, lateral left) and patient metadata (age, sex). By leveraging these techniques, the model captures comprehensive semantic and structural information for high diagnostic accuracy while minimizing radiation exposure. Tested on a dataset of 882 X-ray images from 294 pediatric patients, the model achieved an AUROC of 0.975, an F1-score of 0.882, a sensitivity of 0.878, and a specificity of 0.937. Grad-CAM visualizations further validated its ability to localize disease-relevant regions using only classification annotations. The model demonstrates the potential to revolutionize pediatric care by providing a safer, cost-effective alternative to CT scans.