ASFNR clinical state-of-practice: For posterior circulation large vessel occlusion stroke, use noncontrast CT to exclude hemorrhage, CT angiography to identify occlusion, and MRI with DWI as reference standard. Imaging scores like PC-ASPECTS help select patients for thrombectomy.
The ACR Appropriateness Criteria 2026 update provides imaging guidelines for suspected or confirmed diffuse lung disease, covering initial evaluation, acute exacerbation, and surveillance.
New expert consensus framework for symptomatic B3 breast lesions: assess concordance, lower excision threshold for discordant/palpable lesions, conservative management for concordant non-atypical lesions, and risk-based surveillance for atypia.
ACR Appropriateness Criteria for indolent lung cancer: For patients with slow-growing subsolid nodules or atypical cysts where surveillance is chosen over intervention, imaging follow-up aims to detect changes that alter management, with modality selection guided by accuracy and…
You Make the Call! Freestyle Radiology Quiz #22
ACR Releases Early Summary of 2027 MPFS Proposed Rule
Systematic review of 95 studies on commercial AI for pulmonary nodule CT assessment shows research shift from technical efficacy to diagnostic thinking and therapeutic efficacy, but patient outcomes and cost-effectiveness remain poorly studied; all studies had high risk of bias…
FFRCT Analysis Rejections: Comparison of Frequency and Associated Factors Between Two CT Scanners.
In prostate cancer biochemical recurrence, guidelines say salvage radiation shouldn't be withheld despite negative PSMA PET/CT. Yet imaging is often pursued. AJR Expert Panel consensus: use PSA kinetics and tumor biology to select further imaging, such as repeat PSMA PET/CT or a…
Artificial Intelligence-Augmented Diagnostic Synthesis: Extending the Radiologist's Clinical Role.
Does Your Pet Need an Interventional Radiologist? Pioneering Minimally Invasive Interventions Between Veterinary and Human Radiologists.
In 117 patients, a 7-minute 5-sequence 3T knee MRI with deep learning reconstruction showed high diagnostic accuracy for arthroscopy-verified tears: ACL sensitivity 97%, specificity 98%, AUC 0.97; medial meniscus AUC 0.87. Interreader agreement was good.
Super-resolution deep-learning reconstruction (SR-DLR) improved in-stent restenosis detection on coronary CTA: per-stent accuracy 89.1% vs 79.1% for hybrid iterative reconstruction (HIR) in a reader, with 73 patients and 110 stents, along with better image quality.
Beyond faster reporting: risk-aware implementation of large language models in radiology.
Letter to the Editor: The safety and effectiveness of ethanol embolosclerotherapy in the treatment of arterio-venous malformations: a systematic review and meta-analysis.
A deep-learning denoising algorithm applied to simulated low-dose (90% reduced) 18F-FDG breast PET/MRI significantly improved image quality, with lesions judged equivalent or slightly inferior to full-dose images in 89.9% of reads. Quantitative noise decreased (p<0.0001) and tum…
For abdominal radiology reports, AI-generated impressions were preferred 38% of the time, radiologist impressions 45%, no preference 17%. Overall, AI was rated equivalent or superior (OR 1.34, 95% CI 1.004-1.788, p=0.02).
Auto-segmentation for cervical cancer CT achieved mean Dice scores of 0.92 (femur), 0.89 (bowel bag), 0.88 (bladder) and reduced contour time from 77 min to 5 sec (p<0.001) in a prospective study.
ViT-MultiRAGNet, a retrieval-augmented Vision Transformer for mammogram classification, achieved AUC 0.998 (±0.009) on the RTM dataset, outperforming unimodal benchmarks (p<0.01).
Percutaneous Extra-Anatomic Splenocaval Shunt Creation for Refractory Portal Hypertensive Bleeding.
A narrative review of AI in emergency radiology finds high detection accuracy for critical findings and worklist prioritization, but limited prospective data on patient outcomes.
Letter to the Editor: Diagnostic assessment of artificial intelligence reconstruction on accelerated prostate MRI: a retrospective, paired, multireader multi-case study.
On in-distribution test data, deep learning models achieved accuracies of 0.93, 0.92, and 0.91 for detecting parenchymal abnormalities, any abnormality, and small opacity grade, respectively, comparable to certified B-readers. Performance dropped on out-of-distribution cases.
Deep learning reconstruction (DnCNN) accelerated 3D knee bone MRI from 30 minutes to as low as 2.5 minutes in a feasibility study, with iterative reconstruction reducing denoising artifacts at 5-minute scans.
A survey of 79 Australian radiographers found communication difficulties, inadequate accommodations, and lack of autism-specific training were key barriers to quality imaging for autistic individuals. Effective communication and individualized care improved the experience.
Delayed pulmonary artery-bronchial fistula following embolization of mucormycosis related pulmonary artery pseudoaneurysms.
A deep learning model using serial CT scans predicted progressive disease at the next chemotherapy follow-up visit in advanced pancreatic cancer, achieving an area under the curve of 0.77 on internal, 0.76 on external, and 0.74 on prospective test sets.
A multicenter RCT is testing whether adding Yifei Sanjie Pill to CT surveillance reduces 2-year progression of high-risk subsolid pulmonary nodules (≤8 mm). 589 participants randomized; primary results expected 2028.
Multimodal deep learning fusion model using CT perfusion-defined infarct core, clinical variables, and dual-phase CTA scores predicted parenchymal hematoma type 2 after endovascular therapy with AUC 0.845 in external validation (0.906 in training), outperforming clinical-imaging…
FDA 510(k) clearance for SkeletonPlanner (UPlanner B.V. - Presurgeo), a radiological image processing system. Product code QIH; reviewed by the Radiology committee.
FDA grants 510(k) clearance to ThinkSono GmbH for ThinkSono Guidance, an AI‑powered ultrasound system to assist in detecting deep vein thrombosis (DVT).
FDA cleared Better Diagnostics Dental Assist (BDDA) Version 1.0 (Better Diagnostics AI Corp.), an AI-driven radiological image processing system for dental imaging, under 510(k).
AI trained on 127,496 CT reports (R-Super) detected 56% more malignant tumors than radiologists and improved sensitivity by over 11% beyond mask-only training, enabling multi-cancer detection without scarce tumor masks.
Logistic regression using radiomic features of epicardial adipose tissue on photon-counting CT achieved only moderate AUCs of 0.65 (cross-validation) and 0.59 (held-out test) for classifying arterial hypertension. Lower mean EAT attenuation was the most robust finding.
