I am thrilled to see @RSNA news featuring our AI multiple disease classifier for CT. This project was a combined effort of so many brilliant minds and sleepless nights :).

📦Dataset:

📄 Journal: Radiology: Artificial Intelligence.👉 🧾 arXiv: 💻 Code + Benchmarks:. 🔗 GitLab: 🔗 GitHub:

🚨 New Dataset & Publication!.Duke Lung Cancer Screening dataset is now published in @Radiology_AI🎉. 📌 2061 patients.📌 3168 3D boxes (benign & malignant).📌 Pretrained models + benchmarks.📌 Fully open-access!. @DukeRadiology @DukeEngineering @DukeCVIT @JosephLo16

RT @Radiology_AI: @DukeRadiology Lung Cancer Screening set contains 2061 lung cancer screening low-dose CT scans, acquired within the past….

Check our open-access Lung Cancer screening dataset :) @DukeCVIT @DukeRadiology.

RT @Radiology_AI: Large collection of lung cancer screening low-dose CT scans for lung nodule classification with annotations performed sem….

4/.🏛️ Supported by:.@Duke_CVIT, @DukeEngineering, @DukeRadiology, and @DukeU.💸 Funding: @NIH (P41-EB028744, R01EB001838, R01HL155293).

📊 Highlights:.1️⃣  Virtual trials can replicate certain aspects of clinical trials effectively. 2️⃣ CT > CXR in lung cancer detection. 3️⃣ Provides insights by lesion type, reinforcing imaging technology evaluation.4️⃣ Safe, cost-effective, scalable alternatives to clinical trials.

2/.💡 Why VLST matters:.We show how Virtual Imaging Trials (VITs) can emulate certain aspects of clinical trials--faster, safer, and more cost-effective. Results? CT outperformed CXR in detecting simulated lung cancers—mirroring NLST findings. ✅.

🚨 Excited to share our latest publication!. Virtual Lung Screening Trial (VLST) 🫁🌐.An in silico study inspired by the NLST, comparing CT and CXR for lung cancer detection using simulated patients and AI readers. 🔗 .@DukeCVIT @JosephLo16 @EhsanSamei

Visited my wife at her lab today, gotta admit, her experiments seem way more fun than mine!. I think it’s too late to change the field 😎

RT @DukeRadiology: Join Mobina Ghojogh Nejad, MD @RSNA #RSNA24 AUTOMATED EMPHYSEMA SEGMENTATION BY ACTIVE LEARNING: CLOSER CONCORDANCE WITH….

It’s a wrap! 🙌 Thank you, @RSNA 2024, for an incredible opportunity to present our work on Virtual Imaging Trials for Lung Health. Inspired by the feedback and discussions with fellow innovators in radiology. Couldn’t make it? check out:

Imagine a future where lung cancer detection is faster, safer, and more personalized ! . Join me tomorrow at Physics discussion session to connect and discuss this future. also read this fantastic #RSNA2024 bulletin on VITs featuring Prof. @EhsanSamei

RT @DukeRadiology: Don't miss @JRadMD at #RSNA24! presents: "Lung Cancer Screening: U.S. Perspective, Current Status, & Future Directions."….

RT @DukeRadiology: ". Virtual trials offer an unparalleled trustworthy and enabling alternative to guesswork." Join @EhsanSamei for an Edu….

It’s overwhelming to see the sheer number of presentations at #RSNA2024 !Scratching my head trying to decide which ones to attend—wish I had a time machine!. If you have any “AI can’t miss,” DM me—I’d love your recommendations. Excited for my first #RSNA2024 experience!🚀.

What if clinical trials could happen without real patients? 🤔 With Virtual Imaging Trials, we simulate the entire diagnostic process—using virtual patients, scanners, and AI readers. ✅ NLST Sensitivity: 94%.✅ NLST Specificity: 73%. Visit my poster at #RSNA24 to learn more!

@DukeEngineering @DukeRadiology @DukeCVIT @EhsanSamei @ehsanabadi1 @vancoillielie.