Johannes Hofmanninger

Head of Research & Development at contextflow

[contextflow - A 4DMedical Company @contextflow_rad]

🚨Product update news🚨

The latest version of @contextflow_rad ADVANCE Chest CT is out:

1) Incidental coronary artery calcification detection & quantification (iCAC)!

2) ILD pattern updates

3) Configurable PDF reports

For more details, talk to us https://t.co/ZMoSa8eX50

[Johannes Hofmanninger @joh_hof]

A truly great event! The venue was fantastic, and I especially appreciated the talks covering such a broad spectrum of topics. Looking forward to the next one! 🩻

[contextflow - A 4DMedical Company @contextflow_rad]

Scenes from June's Radiology Symposium in Mainz!

This year we focused on how AI integration, lung cancer screening and agentic AI.

Thank you to every speaker and attendee and (ofc) to @SectraNews for five wonderful years 🙏

Want to join the next edition👉 [email protected]

[contextflow - A 4DMedical Company @contextflow_rad]

June marked our 5th Radiology Symposium in Mainz, an annual tradition organized in collaboration with our friends at @Sectra DACH.

In case you missed it, Guido Gebhardt was kind enough to write this thoughts about the event for @Radiologie Magazin.

https://t.co/jkVOwtgWck

[Johannes Hofmanninger @joh_hof]

@karpathy The scope of radiology is far broader than what people commonly believe. Even if reduced to image recognition alone it spans an unimaginable broad spectrum of tasks which unfortunately often don't benefit from transfer learning. Not to forget the challenging labelling...

[Johannes Hofmanninger @joh_hof]

@jeremyphoward Hidden treasure, had no idea

[contextflow - A 4DMedical Company @contextflow_rad]

🚀 Better emphysema quantification with AI

Old way: Thresholding at −950 HU (simple but misses nuance)
New way: AI (more accurate, scanner-agnostic)

Result → Improved detection suggests better treatment 💡

Read about our ML method for #emphysema https://t.co/DttlgOxkzN

[Johannes Hofmanninger @joh_hof]

@DrDominicNg Great summary and discussion!

[Johannes Hofmanninger @joh_hof]

@HarryEJAnthony @UNSURE_Workshop @matthias_perk

[Johannes Hofmanninger @joh_hof]

@elonmusk Heaviest, not the largest flying object

[Johannes Hofmanninger @joh_hof]

@cir_lab @MedUni_Wien Herzliche Gratulation! 👏

[ESR Journals @ESRJournals]

🥇 The most cited 2023 #EuropeanRadiologyExperimental article compared four #DeepLearning approaches trained on various datasets related to lung segmentation, illustrating importance of data diversity. (@joh_hof et al., @georg, @contextflow_rad)

Kudos!

🔗https://t.co/okIkFu7ZIw

[contextflow - A 4DMedical Company @contextflow_rad]

New case posted on the @CMRadiology platform: can you spot the easy-to-miss nodule? 🔎
Sign up or login to leave us a comment! https://t.co/b5RgW3K43l #lcs #lungcancer #pulmonarynodule #AI

[contextflow - A 4DMedical Company @contextflow_rad]

Automatic lung segmentation in CT is a critical part of image analysis. We developed an algorithm that yields high accuracy despite the presence of disease patterns.

In this whitepaper, we compare our algorithm with the established state-of-the-art. https://t.co/B6o0dMRv2s

[Johannes Hofmanninger @joh_hof]

Super happy and proud that we have created something that has been useful for so many and facilitates so much downstream research 🙂

[ESR Journals @ESRJournals]

🥇 And finally, our sincere congratulations to the most cited #EuropeanRadiologyExperimental article in 2022, which highlights the importance of #data diversity when developing and training #DeepLearning models!

🎉 Well done to @joh_hof et al.!

🔗 https://t.co/okIkFu7ZIw

[Die Tagespresse @DieTagespresse]

Our job is done.

[CIR Lab @cir_lab]

77 mio years of cerebral cortex evolution reconstructed. New @cir_lab paper by E Schwartz, KH Nenning, @katjaQheuer, @NathanJeffery17 @OrnellaBertrand @R3RT0, G Kasprian, D Prayer, and @georg: https://t.co/Ez5EkN22mF in @NatureComms @MedUni_Wien @brainhackorg

[Johannes Hofmanninger @joh_hof]

@Suhail @wasimlorgat Yes, that's terrible 😔 at least it seems to be recognized as an issue now

[contextflow - A 4DMedical Company @contextflow_rad]

Wish you could easily track your #lungcancer patient's progress over time? Now you can, with TIMELINE views.

TIMELINE shows changes in lung nodules over time, regardless of the number of priors. We'd love to show you how. Schedule a chat with us at #ECR23 https://t.co/NKOKNhzAIb

[Johannes Hofmanninger @joh_hof]

@realGeorgeHotz This might be to search what visual programming was to coding. There is some niche but never be productive/flexible enough. Quite unsure though how this will turn out 🤷

[contextflow - A 4DMedical Company @contextflow_rad]

Congrats to several of our team members and advisors for the most downloaded article 2021 from @ESR_Journals! @joh_hof Florian Prayer,@jeanny_pan Sebastian Röhrich,@HProsch and @georg. Looking forward to what comes out of @cir_labLab at @MedUni_Wien next! #EuropeanRadiology