Zhongnan FANG

[Stanford AIDE @stanford_aide]

We are excited to invite you to participate in an online research study at @StanfordMed focused on improving patient-centered communication in healthcare 🩻 This study involves using a grading scheme designed to assess the quality of patient-friendly radiology reports 🗒️

[Zhongnan FANG @zhongnanf]

Excited to share our paper recently published in Radiology journal demonstrating that LLMs achieve performance comparable to experts in clinical history quality control, alongside an overview of the clinical history quality at our institution.

[Stanford AIDE @stanford_aide]

Can we use open-source LLMs to assess the quality of clinical histories? 🤔 Our latest results, published in @radiology_rsna show that we can!

[npj Digital Medicine @npjDigitalMed]

Large language models (#LLMs) can effectively summarize medical evidence, but proprietary LLMs often lack transparency and create vendor dependency.

This study showed that fine-tuning open-source LLMs significantly improves their performance, bringing some models, like LongT5, close to the capabilities of proprietary models like GPT-3.5, with smaller fine-tuned models sometimes outperforming larger, zero-shot models.

https://t.co/YoWMrytUjc

[Zhongnan FANG @zhongnanf]

Intriguing study on whether AI in healthcare truly enhances physician efficiency. A generalizable uncertainty measure we are working on could be a solution for this issue.

[Curt Langlotz @curtlanglotz]

A study validating that detection algorithms don't always save time: "With lengthened read times for falsely flagged examinations, system inefficiencies may outweigh the potential benefits of using the tool in a high volume, low prevalence environment." https://t.co/WMRlFNnKbF

[Zhongnan FANG @zhongnanf]

Exciting work by @loublanks on a novel 3D CT vision-language foundation model! Can't wait to see the diverse applications it will bring. #AI #Healthcare #CTScans #Radiology #MedicalImaging #MachineLearning #ArtificialIntelligence #ComputedTomography

[Louis Blankemeier @loublanks]

🧙 Excited to introduce Merlin, a vision language foundation model for 3D computed tomography 🐈‍⬛🩻

Trained to understand 3D abdominal CT scans using supervision from:
💾 Structured electronic health records (1.8+ million codes)
🗒️ Natural language radiology reports (6+ million tokens)

Paper: https://t.co/ljneXOpKFC

🧵 1/10

[Louis Blankemeier @loublanks]

🧙 Excited to introduce Merlin, a vision language foundation model for 3D computed tomography 🐈‍⬛🩻

Trained to understand 3D abdominal CT scans using supervision from:
💾 Structured electronic health records (1.8+ million codes)
🗒️ Natural language radiology reports (6+ million tokens)

Paper: https://t.co/ljneXOpKFC

🧵 1/10

[Zhongnan FANG @zhongnanf]

Frequency-selective control of cortical and subcortical networks by central thalamus https://t.co/02OY1Y6NhD