Cem Bilgin

[Cem Bilgin @cembilgincb]

8/8 Takeaway 🧠💡

In selected hospitalized patients with non-acute SDH who do not require urgent decompression, first-line standalone MMAE may lead to better early discharge outcomes than CM, especially when performed early.

Read more: https://t.co/rUS1PH0tR1

[Cem Bilgin @cembilgincb]

1/8 Check out this new @Radiology_RSNA study by Chen et al. 🧠📄✨

First-line standalone Middle Meningeal Artery Embolization (MMAE) was evaluated for hospitalized patients with non-acute subdural hematoma.

#RadInTraining @RITEditor @RadiologyEditor @VChernyakMD @SuhnyAbbara

[Cem Bilgin @cembilgincb]

7/8 Early MMAE also had a more favorable inpatient profile 🏥✨

It was associated with:

⬇️ Respiratory complications
⬇️ Infections
⬇️ AKI
⬇️ Length of stay
⬇️ Accrued frailty

Important nuance: timing analyses were exploratory.

[Cem Bilgin @cembilgincb]

9/9
Take-home: structural MRI + machine learning may uncover hidden FRDA heterogeneity and improve stratification for research & personalized care. 🚀🧠
#RadInTraining #Tweetorial @RITEditor @Radiology_Editor @VChernyakMD @SuhnyAbbara
#MachineLearning #FriedreichAtaxia

[Cem Bilgin @cembilgincb]

1/9
🧠Friedreich ataxia (FRDA) is progressive, but not all patients may follow the same neurodegenerative pattern.
This @radiology_rsna study used machine learning to identify 3 structural MRI–based biological subtypes of FRDA. 🔍✨
🔗 https://t.co/MX7S6dLFlP

[Cem Bilgin @cembilgincb]

8/9
One especially interesting finding 🤔
The early cerebral subtype showed a weaker relationship between MRI stage and clinical duration/severity than the other subtypes.

[Cem Bilgin @cembilgincb]

8/8
Proposed next steps:
• Integrate RadSearch into clinical workflows
• Expand to imaging modalities beyond CT/MRI
• Prospective, real-world validation
🔗 Code: https://t.co/FNKhpw0mnH
@RITEditor @VchernyakMD @DrLindaMoy
#Radiology #ArtificialIntelligence #LLM #SemanticSearch

[Cem Bilgin @cembilgincb]

1/8
Ever looked for a perfect teaching case or cases for a research project? New in @radiology_rsna: RadSearch, a semantic search model trained on >16K radiology reports, designed to retrieve relevant reports more accurately than standard methods
https://t.co/ewjgxeBhAM

[Cem Bilgin @cembilgincb]

7/8
Potential applications of RadSearch include:
✅ Teaching file creation
✅ Comparison case retrieval
✅ Research dataset curation
✅ LLM-based clinical decision support