🔵 RESEARCH
Radiomics Model Differentiates Uterine Fibroids From Adenomyomas on Ultrasound
A study in the Journal of Ultrasound in Medicine developed machine learning models using ultrasound radiomics to distinguish fibroids from adenomyomas.
These lesions look similar on conventional ultrasound and require different treatments. No performance metrics, sample size, or external validation details are yet available — clinical utility remains unproven until prospective data and comparisons against experienced gynecologists emerge.
Journal of Ultrasound in Medicine · Fri, 10 Ap
Read more → (paywall)
🔵 RESEARCH
Vision-Language AI Model Tackles Lung Nodule Analysis and Risk Stratification
NPJ Digital Medicine published a graphicalized vision-language model for comprehensive lung nodule analysis and risk stratification.
Multimodal AI that combines imaging with language-based reasoning could shift lung screening competition from detection accuracy to full workflow integration. Pure-play nodule detection vendors face commoditization risk as these capabilities become table stakes. Prospective validation against Lung-RADS is needed.
NPJ Digital Medicine · 2026-04-11
Read more → (paywall)
🔵 RESEARCH
JACR Study Finds Linens, Not Equipment, Drive Ultrasound Carbon Emissions
A JACR study found disposable linens and supplies, not equipment energy use, account for the vast majority of ultrasound's annual carbon footprint.
Imaging departments chasing sustainability goals should redirect efforts toward reusable supplies rather than equipment efficiency alone. Health systems under ESG pressure may reshape ultrasound procurement contracts around supply chain sustainability — a potential opening for eco-friendly disposables vendors.
Medical Xpress — Health Research · Sun, 12 Ap
Read more →
🩺 EDITOR’S TAKE — FOR CLINICIANS
Two new AI models promise better differentiation — uterine masses on ultrasound, lung nodules on CT. Neither has prospective validation yet. Watch for head-to-head data against experienced specialists before changing workflows.
📊 EDITOR’S TAKE — FOR INVESTORS
Vision-language models signal the next competitive frontier in lung screening AI, threatening pure-play nodule detection vendors with commoditization. Companies that integrate multimodal analysis into clinical workflows will hold pricing power.
🏭 EDITOR’S TAKE — FOR INDUSTRY
The JACR sustainability finding hands procurement teams a concrete lever: swap disposable linens for reusables. OEMs marketing green imaging should look past equipment efficiency and address supply chain emissions instead.