82 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 prospective randomised crossover study demonstrates that vendor-integrated AI software significantly improves operational efficiency and reduces mental demand and physical effort for sonographers performing abdominal ultrasound, although the magnitude of time savings varies between operators and the technology reshapes rather than eliminates the workflow.
This study demonstrates that using CT images with EEG electrodes and an extended Hounsfield unit range provides the most accurate attenuation correction for simultaneous PET/EEG imaging, introducing only a small, systematic bias compared to scans without electrodes, while metal artifact reduction techniques offer minimal additional benefit.
This cross-dataset analysis reveals that while standard automated metrics like BLEU and ROUGE are highly sensitive to textual changes, they fail to distinguish clinically meaningful errors, with CheXbert emerging as the most aligned metric for assessing radiology report quality despite only moderate overall performance.
This retrospective study demonstrates that applying Low-Rank Adaptation (LoRA) to a frozen Rad-DINO ViT-B/14 foundation model improves multi-label classification performance on the NIH ChestX-ray14 dataset, while explicitly clarifying that the reported results are descriptive rather than confirmatory due to prior exposure to test labels and highlighting the sensitivity of metrics to counterfactual label-flip analyses.
The author presents CT ECV Mapper, an open-source 3D Slicer application that enables both interactive and unattended batch processing of voxelwise extracellular volume mapping for the liver and hepatic tumors using conventional single-energy multiphase CT, demonstrating high feasibility and robust quality control on a large public dataset.
This multicenter prospective study validates a report-guided semi-supervised learning framework with a lesion-only teacher model (RG-SSL-LOC) for biparametric MRI, demonstrating superior lesion segmentation and robust case-level detection of clinically significant prostate cancer that significantly enhances multimodal diagnosis and reduces unnecessary biopsies compared to existing methods.
This paper introduces RCC-AID, an openly available dataset comprising 129 expertly annotated contrast-enhanced CT scans from 91 renal cell carcinoma patients, designed to address the scarcity of public lesion annotations and foster reproducible AI research in medical imaging.
BioDeformUNet is a deep learning model that achieves near real-time, biomechanically informed liver image registration with accuracy comparable to traditional biomechanical algorithms while offering a 34-fold speedup in inference time to facilitate efficient intra-procedural evaluation of minimal ablative margins.
This exploratory neuroimaging study of 122 participants suggests that severe Long COVID symptoms, including tremor-like sensations and neuromuscular fatigue, may be associated with structural and microstructural abnormalities at the mammillary body-fornix-hypothalamic interface and connected brainstem-cerebellar pathways, though the findings require prospective replication to establish causal mechanisms.
This study demonstrates that the quantitative accuracy of long axial field-of-view PET systems is primarily determined by the spatial distribution of defective detector blocks rather than just their number, revealing that sparse defects are tolerable in higher numbers while clustered defects cause significant localized biases, thereby supporting a re-evaluation of quality control thresholds to balance system uptime with diagnostic reliability.