80 papers
This study demonstrates that dynamic MRI using intravenous gadolinium-based contrast agents can successfully track fluid and solute transport along the posterior lymphatic pathway in the ocular glymphatic system of healthy humans, revealing distinct clearance patterns between the vitreous body and the intraorbital optic nerve over a four-hour period.
This paper demonstrates that a self-supervised Video Diffusion Encoder can effectively detect Myofascial Pain Syndrome in the upper trapezius muscle using B-mode ultrasound videos from a small prospective cohort, offering a data-efficient alternative to conventional deep learning methods that require large annotated datasets.
The paper introduces ReMIND, a vision-language framework trained on over 73,000 patient visits to enable robust, generalizable, and clinically coherent automated interpretation of multi-sequence brain MRI through large-scale instruction tuning and specialized report generation strategies.
This study demonstrates that accelerated proteomic aging, particularly in late life, is significantly associated with a higher burden of cerebral small vessel disease markers, including white matter hyperintensities and various types of brain infarcts, across two large US community-based cohorts.
This study demonstrates that the Sampled Iterative Local Approximation (SILA) algorithm can accurately model longitudinal tau PET trajectories and estimate individual tau positivity onset ages in the meta-temporal region across two cohorts, though its accuracy in the entorhinal cortex is limited in individuals with dementia.
This cross-sectional study of 705 youth reveals that even mild elevations in hemoglobin A1c within the non-diabetic range are associated with reduced cortical thickness and gray matter volume, particularly in frontal, cingulate, and occipital regions, with these negative structural effects being significantly amplified in neighborhoods characterized by greater socioeconomic deprivation.
This study demonstrates that achieving clinically significant weight loss (>5% of body weight) over two years significantly slows the progression of knee cartilage degeneration, as measured by the sensitive Cartilage Thickness Score (CTh-Score), in individuals with and without radiographic osteoarthritis, whereas conventional structural metrics failed to detect this beneficial effect.
This study utilizes 3D mesh-based shape analysis of gluteus maximus MRI data from the UK Biobank to reveal that spatially localized muscle remodelling, rather than global volume or fat fraction alone, provides sex-specific biomarkers for metabolic dysfunction and type 2 diabetes risk.
This study presents and evaluates a deployable, explainable deep learning system based on DenseNet121 and Grad-CAM that achieves accurate tuberculosis detection from chest X-rays on both desktop and mobile platforms, demonstrating its potential as an offline decision support tool for resource-constrained healthcare settings.
This study demonstrates that the LAQUA system, a multimodal large language model-based tool, achieves substantial agreement with expert radiation oncologists in evaluating auto-contouring quality, suggesting its feasibility as an efficient primary screening tool to reduce clinical workload despite a potential risk of overestimating contour acceptability.
The study demonstrates that QTMnet, a deep neural network trained on synthetic data to perform AIF-independent perfusion estimation, significantly outperforms traditional 2CXM methods in classifying low-grade versus high-grade gliomas.
This study demonstrates that chest radiograph-derived age acceleration is independently associated with an increased risk of future lung cancer, although adding this metric to established risk prediction models provides minimal improvement in discrimination.
This matched case-control study demonstrates that Fourier-based analysis of bilateral breast asymmetry in raw full-field digital mammography (FFDM) images provides the strongest short-term breast cancer risk prediction, outperforming both clinical FFDM and synthetic digital breast tomosynthesis (DBT) images despite the latter's higher spatial resolution.
This paper introduces TWISTARE, a novel dual-interleaved steady-state 3D-GRE technique that enables fast, whole-brain quantitative T2 mapping with simultaneous RF field estimation at 7T, effectively overcoming ultra-high field challenges like SAR limitations and B1 inhomogeneities while achieving precision comparable to gold-standard methods.
This study analyzes FDA-authorized radiology AI devices to demonstrate how low disease prevalence creates a false positive paradox that undermines positive predictive value, arguing for the mandatory disclosure of false discovery and omission rates to guide clinically and ethically appropriate AI selection.
This study demonstrates that peripheral blood measures of ATP-linked mitochondrial respiration are significantly associated with altered alpha and theta cortical rhythms and aperiodic signaling in individuals with Alzheimer's disease, suggesting that peripheral mitochondrial function can serve as a biomarker for the neurophysiological energetic changes underlying the disease.
Although French radiologists recognize the potential of AI to enhance precision and personalization in their field and overwhelmingly express a willingness to adopt it, its current implementation is hindered primarily by high costs and insufficient equipment.
This study demonstrates that simulated MRI artifacts significantly and differentially degrade the accuracy of deep learning-based brain age prediction across three algorithms (Pyment, MIDI, and MCCQR), highlighting the critical need for artifact-aware evaluation and mitigation strategies in clinical research.
This multicenter study in the Souss Massa region of Morocco evaluated patient radiation doses for conventional lumbar spine radiography across four hospitals, finding that while significant variations existed between facilities, all measured entrance surface doses and dose area products remained below established diagnostic reference levels.
This study demonstrates that the cross-scanner reliability of brain MRI foundation model embeddings varies significantly based on pretraining strategy, with models incorporating biological metadata achieving scanner-robust performance comparable to traditional morphometric baselines, while purely self-supervised models exhibit substantial scanner-induced variance.