AI Risk Scores Link to Low-Grade Malignancy Before Clinical Detection
Retrospective analysis shows AI-generated risk scores from prior-year mammograms correlate with tumor grade and invasive lobular pathology.
Population-Based Models Increase Lung Cancer Screening Coverage
A scoping review of 56 programs finds population-based strategies reach up to 40% of at-risk adults, exceeding referral-based models.
Coronary CT Angiography Misses Over One Third of Calcified Plaques
High-resolution non-contrast CT identifies small, low-density calcifications overlooked by standard angiography in nearly 9% of patients.
Large Language Models Reduce Radiology Reporting Time and Transcription Errors
A multicenter study finds LLMs faster than conventional speech recognition but notes inconsistent efficiency and new error types.
AI Model Predicts Axillary Response Using Baseline Mammography
A deep learning tool identifies nodal conversion after neoadjuvant therapy, potentially reducing unnecessary axillary dissections.
TNM-9 Staging Shows Limited Prognostic Accuracy in Nasopharyngeal Carcinoma
A retrospective study finds the ninth edition TNM system lacks strong survival discrimination when patients receive PET/CT imaging.
Advanced MRI Detects Occult Neuronal Loss in Asymptomatic MCA Occlusion
Soma and neurite density imaging reveals microstructural gray matter damage in stroke-free patients with chronic arterial occlusion.
Genicular Artery Embolization Shows No Benefit Over Sham for Knee Osteoarthritis
A randomized trial finds that 12-month pain reduction after genicular artery embolization is statistically equivalent to a sham procedure.