Brendan Kelly
brenkelly.bsky.social
Brendan Kelly
@brenkelly.bsky.social
93 followers 8 following 61 posts
Posts Media Videos Starter Packs
Reposted by Brendan Kelly
🔦 #EuraSpotlight

How do we make radiology more accessible, engaging, and impactful in the digital age?

📢 European Radiology now requires graphical abstracts for Original Articles.

Led by: Brendan S. Kelly

📖 link.springer.com/article/10.1...
Reposted by Brendan Kelly
Amplifying the impact of original research through visual media: the importance of graphical abstracts.

A new Editorial by our Social Media Editor, @brenkelly.bsky.social, looks at the decision to adopt graphical abstracts as part of the publication process at #EuropeanRadiology.

shorturl.at/eeoBH
Reposted by Brendan Kelly
A recent #EuropeanRadiologyExperimental study comparing #ChatGPT-4 and #Gemini in analyzing adult brain CT referrals showed limited #chatbot performance (κ=0.3) in justifying scans, though they were accurate for complete referrals.
Reposted by Brendan Kelly
How do medical students view #radiology as a viable career?

A recent survey of 1,184 medical students among 21 universities reveals that while 99% acknowledge radiology's clinical importance, 64% rule out pursuing it as a career option due to limited patient interaction.
MSK and Spinal #essentials in @europeanradiology.bsky.social this week!
This addition to #ESREssentials found that percutaneous bone consolidation, including vertebroplasty and advanced spinal implant techniques, effectively treats bone pain from benign and malignant conditions. (Le Corroller & Arrigoni)

#EuropeanRadiology

link.springer.com/article/10.1...
🚀 Could AI help streamline MRI workflows, optimise biopsy decision-making, and reduce variability in prostate cancer detection?

🤔 What are your thoughts on Radiology lead research and CTiR?

Let’s discuss below! 👇
Not yet... The next crucial step is prospective validation, ensuring AI improves patient outcomes rather than just performing well on retrospective datasets.
🤔 What does this mean for the future of prostate MRI?
This study provides compelling evidence that AI can match expert performance, but does that mean AI is ready to replace human oversight?
✅ Sensitivity vs. Specificity Trade-off – At the predefined risk threshold, AI achieved:
📌 Sensitivity: 95% (Rads: 99%)
📌 Specificity: 67% (Rads: 73%)
While radiologists still had the edge, AI was remarkably close—raising the question: could AI help in reducing radiologist workload in routine cases?
✅ Performance in Real-World Settings – AI demonstrated robust generalisability, performing consistently across different scanner models and field strengths, an essential step in ensuring AI’s clinical viability.
✅ AI vs. Radiologists – This study evaluated an AI system for detecting clinically significant prostate cancer (GG≥2) on MRI across six hospitals with multiple scanner vendors. Impressively, AI performed non-inferior to MDT-supported radiologists, achieving an AUC of 0.91 vs. 0.95.
A standout? Francesco Giganti et al.'s multi-centre, multi-vendor validation study on AI-powered prostate cancer detection.
🔗 Read the full study here: lnkd.in/eYdXXGhR

💡 Key Insights:
LinkedIn
This link will take you to a page that’s not on LinkedIn
lnkd.in
One of the most exciting parts of #ECR2025 has been the Clinical Trials in Radiology sessions (CTiR) —showcasing cutting-edge, radiology-led imaging research that is shaping the future of our field.
Reposted by Brendan Kelly
This study of #AI use in Dutch radiology departments (2020-2022) showed steady growth in AI adoption, with 68 total implementations by 2022. However, key challenges like cost and IT integration remain.
Head and neck imaging gets its turn under the @europeanradiology.bsky.social spotlight! Have a look at the #Essentials of tinnitus imaging!
#EuraSpotlight on Tinnitus Imaging
🎧 Does deciding when tinnitus patients get imaging give you a headache? The latest #ESREssentials article breaks it down!

🔗 Full article here: link.springer.com/article/10.1...

🧵⬇️
Let’s reimagine how radiology tells its story—one image at a time.
Thanks to all who've helped make this happen!
— Brendan
#BrendansBest #Radiology #VisualScience #ESR #AIinRadiology
Authors now submit a visual summary that captures their main finding—starting at revision.
It’s creative, effective, and built for how we actually read content in 2025.
🖼️ One image, big impact.
#GraphicalAbstract #EuropeanRadiology
Graphical abstracts:
✅ Summarize research at a glance
✅ Boost social media reach
✅ Improve accessibility for clinicians, patients & policymakers
✅ Help radiology speak beyond its own walls
#ScientificCommunication #Radiology
🎨 Radiology is visual—so why shouldn’t our science be too?
Proud to share a new editorial in European Radiology, where we’ve officially introduced graphical abstracts for Original Articles.
📖 link.springer.com/article/10.1...
#BrendansBest
Embracing graphical abstracts in European Radiology - European Radiology
European Radiology -
link.springer.com
We shine the light on the #Essentials of #Lymphoma imaging in @europeanradiology.bsky.social this week!
#EuraSpotlight on Lymphoma Imaging
🎯 How do we image lymphoma? The latest #ESREssentials article provides key recommendations on when to use PET/CT vs. CT, biopsy considerations, and how to apply the Deauville score.

🔗 Full article here: link.springer.com/article/10.1...
(Doris Leithner et al.)
🧵⬇️
I’m very excited for the next phase of the “From the experts” series at @europeanradiology.bsky.social with @pintodrad.bsky.social who is a true expert in the field of #AI #Informatics and more! He has curated a great list, do check them out!
👋 Hello!

I'm Daniel Pinto dos Santos ( @pintodrad.bsky.social ), Deputy Editor for AI, IT and radiomics at #EuropeanRadiology. I am also Managing Senior Radiologist at the University Medical Center Mainz, working mostly in CT and abdominal radiology.
Reposted by Brendan Kelly
This study by Alice M. Munari et al. shows that the #ORADS system, originally developed and validated on adults, can effectively be applied to #pediatric ovarian mass MRIs.
Reposted by Brendan Kelly
Standardization and resources can help improve reproducibility and usability.
Could be such a great use case for AI. Here’s hoping. Great work!