AI-based CT quantification in SSc-ILD now moves beyond visual scoring to outcome-oriented biomarkers; deep-learning UIP probability stratifies FVC decline and survival, supporting human-in-the-loop decision support.
An integrated model combining carotid Plaque-RADS, degree of stenosis, and ancillary features (plaque burden, remodeling index, enhancement ratio, perivascular fat density) predicted symptomatic plaques with AUC 0.876 (95% CI: 0.828–0.924), and 0.919 in moderate stenosis subgrou…
Systematic review of 101 studies finds LLM-generated medical reports show moderate acceptance but persistent clinically significant errors and mixed workflow effects; evidence too heterogeneous to support autonomous use, requiring clinician supervision.
In a multicenter study of NSCLC, a PET/CT radiomics model predicted lymph node metastasis with AUC 0.845 (95% CI 0.716–0.973) and PD-L1 expression with AUC 0.776 (95% CI 0.641–0.911) from a single preoperative scan.
In an external validation cohort of 69 brain metastasis patients, a T1 postcontrast MRI radiomics model achieved a C-index of 0.574 (95% CI 0.491-0.658) for overall survival prediction, suggesting limited standalone discrimination.
Deep learning denoising of accelerated 3D T1-weighted brain MRI with transfer-learning fine-tuning or self-supervised Self2Self-AM achieved structural similarity index 0.943, cortical thickness discrepancy 0.16 mm, Dice 0.986, beating conventional denoising.
In a phantom, metal artefact reduction (MAR) restored analysability of coronary artery calcium near pacemaker leads, and deep learning reconstruction (DLR) yielded lowest artefact and CT offset. MAR had minimal effect without metal present.
MRI in three children with juvenile Huntington disease revealed severe bilateral caudate and lentiform atrophy with T2/FLAIR hyperintensity; SWI showed globus pallidus mineral deposition in two, and one had mild thoracic spinal cord thinning. Findings highlight consistent basal…
In 26 patients with periventricular hemorrhagic moyamoya disease, three microangioarchitectural features of target collateral vessels were present in 92.3% of hemorrhagic hemispheres vs 25.0% of contralateral non-hemorrhagic hemispheres (P=0.80). These patterns require longitudi…
In MS patients, intra-soma signal fraction (fis) on high-gradient diffusion MRI was decreased in atrophy-prone gray matter, and correlated with disability (EDSS, ρ=-0.44, p<0.05), suggesting network-based neurodegeneration.
A scoping review of 36 studies found no interventional radiology (IR) curriculum for medical students addressed all six steps of Kern's framework; they focused on problem identification and teaching strategies but lacked needs assessment and evaluation.
A machine reasoning framework with rejection option reduced false positives and anatomically implausible errors in CT kidney segmentation, while preserving acceptance of valid outputs. Robustness persisted under noise.
A deep learning model combining YOLOv10 region-of-interest detection and transformer fusion classified osteoporosis, osteopenia, and normal on knee radiographs with 98.87% accuracy and 98.23% AUC.
A self-supervised physics-guided deep learning framework transformed clinical T1w, T2w, FLAIR MRI into quantitative T1, T2, PD maps in >600 test sessions, showing scanner invariance (inter-group CV ≤1.1%) and high voxel-wise reproducibility (T1/T2 correlation >0.82).
A multistage deep learning model detected pelvic active bleeding on CT with an AUC of 0.901 and accuracy of 77.5%, matching specialists and outperforming residents (inference time 1.93 s).
Deep-learning system using pre/post neoadjuvant chemotherapy DCE-MRI and clinical data predicted residual tumor burden with AUCs 0.8885 internal, 0.8628 external, 0.8695 I-SPY2; NAC response AUCs 0.8400/0.8299; subtype AUCs 0.8066/0.7748/0.7661 in 1,515 patients.
FDA grants 510(k) clearance to Gleamer's ChestView US 1.1, a lung ultrasound device (product code MYN, Radiology committee).
The U.S. FDA has posted a 510(k) clearance for ReVISION 2 from Argus Cognitive, Inc., listing it as a radiology device under product code QIH.
An interpretable ensemble learning model using intratumoral & peritumoral MR-radiomics predicted high-grade cervical cancer with LVSI (HGVI), AUC 0.858 (95% CI 0.762-0.932) in the test set. SHAP analysis showed ADC_peritumoral_3mm as a top contributor.
CT attenuation-defined regional radiomics differentiated well-differentiated from dedifferentiated retroperitoneal liposarcoma with out-of-fold AUC 0.864 (95% CI 0.778-0.940) but did not significantly outperform quantitative CT (AUC 0.849; adjusted P=0.389).
In LVO stroke, a clinical-CT model predicted 90-day outcome (AUC 0.939 in CV), but adding CT imaging score to clinical variables (age, NIHSS, onset-to-CT time) did not significantly improve discrimination in testing (ΔAUC 0.0078, 95% CI -0.0071–0.0227, p=0.305).
A LASSO model identified interstitial lung disease (ILD) in Japanese cancer patients: sensitivity 33.5% (external 19.8%), specificity 99.3% (99.4%), positive predictive value (PPV) 76.7% (65.5%). High PPV aids case confirmation despite low sensitivity.
On MIMIC-CXR, a Siamese network using paired chest radiographs outperformed a single-image model for detecting resolved vs absent findings (AUROC up to 0.87 vs chance 0.45-0.60), demonstrating the value of prior images for longitudinal change detection.
Molecular imaging (radioiodine, FDG-PET, SPECT/CT, PET/MRI) combined with genomic profiling (BRAF, RAS, RET, TERT, TP53, etc.) enhances diagnosis, risk stratification, and selection of targeted therapies in thyroid cancer, advancing precision oncology.
New prospective research reveals about (177Lu)Lu-PSMA-I&T for patients with mCRPC.
Ketamine offers better sedation option than fentanyl in IR
AI model OccuNet detected 94.7% of femoral neck fractures initially missed or deemed indeterminate on radiographs, compared with 86.2% for musculoskeletal radiologists and 68.8% for emergency physicians. Overall sensitivity 97.5%, specificity 98.8%.
Paget's Disease Mimicking Metastasis: A Diagnostic Pitfall
<![CDATA[The Reading Room Podcast: Assessing the Current Landscape of AI and Reimbursement in Radiology, Part 1]]>
RSNA Explores Role of APPs to Address Workforce Challenges
