Radiology educators can use AI assistants such as Claude and ChatGPT to convert passive slide decks into interactive multimedia modules for resident self-directed learning; the article outlines a pedagogy-first workflow, demonstrated with a hip arthrocentesis module.
Pathological complete response after neoadjuvant chemo was predicted with AUROC 0.76, 8.7% false positive rate, 75% positive predictive value by a machine learning model using breast imaging and clinical features (n=388).
FDA clears Mindray's Resona I10, I10W, I10 Exp, I10S, XI, and I10 CV, plus Nuewa I10, I10W, I10 Exp, I10S, and XI diagnostic ultrasound systems via 510(k) for general imaging.
Oxos secures $9.6M federal contract for portable x-ray
Hologic survey finds 37% of women skipped or delayed mammograms
Why I Donate: Liisa L. Bergmann, MD, MBA
Advanced Prostate MRI: What Challenging Cases Reveal About the Role of AI
SNMMI: PSMA-PET/CT-guided biopsy improves prostate cancer detection
The FDA cleared NextMR, an AI cloud-based platform that creates 3D bone visualizations from routine 1.5T and 3T spine MRI in under 60 seconds, potentially eliminating the need for additional CT imaging and radiation for surgical planning.
Visa restrictions affect international radiology talent.
Great Leaders Need to be Great Storytellers. #RadiologyLeadership
Modified Clear Cell Likelihood Score (ccLS) with pseudocapsule assessment improved benign vs malignant renal mass discrimination (AUC 77% vs 65.4%) without losing ccRCC identification accuracy (AUC 88.4% vs 86.8%).
AI Reads Thyroid Nodules Like a Radiologist, Zone by Zone
The European Congress of Radiology (ECR) 2027 received a record 10,442 abstracts, a 7% increase, with submissions from 110 countries. Strong growth from Argentina (+84%), Ireland (+77%), Brazil (+72%), the US (+68%), and Türkiye (+60%) offset a 32% drop from China.
New research defines 'true imaging deserts': 18% of U.S. counties (6.4M people) have no local radiologists or scanners. 36% more have scanners but no local radiologists, relying on teleradiology.
Oxos Medical has been awarded a BARDA contract worth up to $24.4M to develop and seek FDA clearance for its next-gen ultra-portable X-ray system, Oxos X3, aiming for point-of-care and mass-casualty triage.
RadNet & WIN-NextGen launch workforce initiative for ages 12-24 in Maryland/DC/Virginia, offering digital content, mentorship, shadowing. Aims to address imaging workforce shortages. Plans for internships & scholarships.
AI Reads Breast Microcalcifications With New Precision, Cutting Unneeded Biopsies
In 46 patients with major depression, ECT significantly reduced the Hurst index (excitation-inhibition balance) in the middle cingulate cortex and decoupled it from posterior cingulate, with changes linked to dopamine and histamine maps and synaptic genes.
A systematic review and meta-analysis compares federated learning and centralized learning for brain tumor classification and segmentation, pooling classification accuracy across identified studies.
FAP-Directed PET in Nononcologic Disease: What Radiologists Need to Know.
ARIA, Artificial Intelligence, and the Burden of Being Helpful.
Could FAP PET Become the Next [18F]FDG PET for Nononcologic Disease?
Wrong Turns: What to Do When You Realize You Chose the Wrong Path.
For resectable HCC beyond Milan criteria, preoperative transarterial radioembolization (TARE) achieved overall survival comparable to primary resection (aHR 1.14, 95% CI 0.50-2.55), whereas transarterial chemoembolization (TACE) had worse OS (aHR 2.72, 95% CI 1.39-5.33).
Synthetic PCASL images generated via CycleGAN from phantoms improved deep learning renal segmentation. A model trained only on synthetic data segmented cortex with Dice score 0.659 in real kidneys.
Four texture radiomic features were reproducible across in vivo CT, ex vivo CT, and micro-CT in NSCLC after normalization and volume matching, and reliably distinguished adenocarcinoma from squamous cell carcinoma.
ICG lymphography spatial atlas in 289 limbs with breast cancer-related lymphedema showed dermal backflow most prevalent in wrist, forearm, olecranon. Adding backflow percentage improved ML prediction of MDACC stage (best: logistic regression, random forest).
Adding ARNAI to AI segmentation pipeline increased mean DSC from 0.814 to 0.870 and reduced L4-L5 Cobb angle error by ~70% (4.7° MAE) on internal test set, though external validation showed limited agreement.
In a dual-center cohort, an extreme trees model using CT body composition indices achieved an AUC of 0.772 (95% CI 0.714-0.822) for predicting gallstone occurrence, outperforming the best single predictor, neutrophil-to-lymphocyte ratio (NLR; AUC 0.627).
