97 papers
Radiology and imaging serve as the eyes of modern medicine, allowing doctors to peer inside the human body without making a single incision. This rapidly evolving field uses technologies like X-rays, MRI scans, and ultrasound to detect diseases, guide treatments, and monitor patient recovery. As new research emerges, these visual tools become increasingly sophisticated, offering deeper insights into conditions ranging from broken bones to complex neurological disorders.
At Gist.Science, we bridge the gap between raw scientific data and public understanding by processing every new preprint in this category from medRxiv. Our team translates these complex studies into both plain-language overviews and detailed technical summaries, ensuring that breakthroughs in medical imaging are accessible to everyone, from students to specialists. Below are the latest papers in radiology and imaging, ready for you to explore.
This retrospective study demonstrates that various large language models, particularly GPT-5, can achieve substantial to almost perfect agreement with radiologists in labeling specific incidental findings in lung cancer screening CT reports, though no single model-prompt combination proved optimal across all finding types.
This longitudinal study demonstrates that early, progressive demyelination in normal-appearing and lesioned white matter, measured by MTsat within the first year of a multiple sclerosis diagnosis, is independently associated with subsequent disability progression, suggesting a critical early therapeutic window that is not captured by conventional inflammatory MRI markers.
Although multiple deep learning architectures achieve similarly high baseline discrimination for intraoral ultrasonography, they exhibit drastically different and often unreliable behaviors under synthetic occlusion, with EfficientNet-B0 showing severe performance collapse, overconfidence, and poor reproducibility compared to other models.
This study evaluates the fMRI foundation models BrainLM and BrainHarmonix on the Prevent-AD dataset and finds that, unlike in language and vision domains, they currently fail to outperform simple baselines for clinical prediction tasks while requiring substantial code adaptation that hinders their practical adoption by independent researchers.
This study demonstrates that VoxelMap, a deep learning framework, can generate accurate, real-time 3D volumetric images of respiratory motion from 2D fluoroscopic projections on a clinical C-arm system using deformable phantoms, with model uncertainty serving as a reliable indicator of prediction accuracy.
This paper introduces GLIODISSECT, a novel voxel-wise framework for decomposing hypoxia-powered BOLD MRI responses into distinct temporal and amplitude components, demonstrating its feasibility and superior descriptive capability over single-amplitude summaries in healthy controls and glioblastoma patients.
This study demonstrates that deep learning models, particularly the EVIT U-Net, can effectively and rapidly automatically segment key abdominal organs in pediatric trauma ultrasound with high accuracy and expert-validated acceptability, thereby overcoming manual segmentation bottlenecks to accelerate large-scale dataset development.
This study demonstrates that combining optimized acquisition strategies (strong background suppression, water excitation) with a novel regularized motion-compensated reconstruction enables robust, high-resolution (2.0-mm isotropic) 3D-GRASE pCASL perfusion imaging at 3T that effectively mitigates motion-related degradation.
This study demonstrates that the signal loss caused by cardiac pulsation artifacts in left-lobe diffusion-weighted imaging is inversely correlated with liver stiffness, establishing it as a simple, intrinsic marker capable of accurately differentiating cirrhotic from non-cirrhotic livers.
This study utilized 7 T short-echo-time 1H-MRS and OPLS-DA classification to characterize the neurochemical profiles of the Alzheimer's disease continuum, successfully distinguishing Alzheimer's disease from healthy controls with high accuracy by identifying key metabolites like myo-inositol and ascorbate alongside specific macromolecular components.