Healthcare AI Guy Weekly Newsletter | 12/5

Our Picks ✨

Highlights if you’ve only got 2 minutes…

1/

ChatGPT’s latest model can diagnose complex healthcare cases

ChatGPT celebrated its one-year anniversary last week, quickly becoming the fastest-growing consumer application in history post-launch… The latest model version, GPT-4, has improved problem-solving abilities and has shown promise in diagnosing complex healthcare cases in recent studies. Two of them are highlighted below:

• GPT-4 to diagnose complex clinical cases: A study from the New England Journal of Medicine found that GPT-4 correctly diagnosed over half of complex clinical cases. Each case included a medical history along with six multiple-choice options. The most common diagnoses included infectious disease, endocrinology, and rheumatology.

• GPT-4 displays radiology ability: A new study from Microsoft found that GPT-4 demonstrates strong performance on radiology tasks like summarization and disease classification, achieving top results.

It’s great to see the measured progress of AI in healthcare and the good news is these models will also only keep getting better as multimodal data gets incorporated.

2/ 

Startup Jona using AI for microbiome medicine research

Healthtech vet Leo Grady believes the microbiome is going to revolutionize healthcare. The microbiome is related to various diseases ranging from gut health to chronic conditions, such as autoimmune diseases, obesity, metabolic health, and food allergies.

That’s why Grady, launched Jona, a startup that provides an at-home microbiome profiling kit and then uses AI to analyze an individual's gut microbiome to deliver actionable, personalized insights.

The startup developed a proprietary large language model (LLM) to interpret an individual's gut microbiome and summarize the scientific literature linking their microbiome to different diseases, conditions, symptoms and allergies. Then Jona generates a personalized interactive report with recommendations on food, diet and lifestyle modifications. The kit costs $385 and is now available to both consumers and providers.

This is a tech-enabled approach to lifestyle medicine in a way and seems like a logical use case for AI in healthcare. Might have to try one of these myself and report back... (link)

3/

AMA unveils seven principles for use of AI in healthcare

The American Medical Association (AMA) unveiled seven new principles for the development, deployment, and utilization of AI in healthcare. The goal here is to maximize the benefits and minimize the potential harms of AI tech in healthcare, specifically for patients and providers. Key concepts include:

1. Oversight: AMA supports a whole-of-government approach to governance.

2. Transparency: AMA wants mandates detailing key characteristics and information in the design, development and deployment of AI.

3. Disclosure and documentation: the principles call for adequate disclosure and documentation when AI directly affects patient care, access, medical decision-making, communications or medical records.

4. Generative AI: healthcare orgs should adopt policies addressing potential negative impacts before adoption/use.

5. Privacy and security: implement safeguards to ensure responsible handling of personal information.

6. Bias mitigation: promote a fair, inclusive and discrimination-free healthcare system

7. Liability: AMA said it will continue to advocate for limitations on physician liability related to the use of AI-enabled technologies, consistent with existing legal approaches to medical liability.

Hopefully, this helps the industry inch closer to establishing a unified governance framework in order to progress AI in healthcare. Still a lot of work to be done so we hope lawmakers also don’t overburden or convict AI of a pre-crime and slow much-needed innovation. (link)

Tool Box 🧰

Latest on business, consumer, and clinical healthcare AI tools and partnerships…

🔧 Nvidia’s medical imaging: Nvidia's new cloud-based APIs, built on its open-source Monai framework, expedite the development of specialized AI models for medical imaging, potentially revolutionizing healthcare AI. It will be easier and faster for developers to create and deploy AI models. (link)

🔧 Oura ring AI upgrades: Oura recently upgraded its smart ring to combat stress — using health metrics and an AI-powered journal to detect spikes, identify triggers, and suggest interventions, including mindfulness prompts from Headspace. (link)

🔧 AI revolutionizing microscopic views: A new paper details how AI-powered image analysis tools are changing the game for microscopy data, with new models reaching unprecedented accuracy heights. The algorithms can automatically segment and trace cells in all types of data, aided by growing troves of annotated images. (link)

🤝 NextGen EHR + Nabla: Nabla Copilot was selected to power NextGen’s new Ambient Assist feature within its EHR, which automatically generates subjective, objective, assessment, and plan (SOAP) notes from patient-provider conversations. (link)

🤝 Mass General Brigham + Nuance: Mass General Brigham is one of the first health systems to use a radiology ‘copilot’ from Microsoft subsidiary Nuance. Its ‘PowerScribe Smart Impression’ technology drafts impressions for radiologists, saving them up to a minute per read. (link)

Deal Desk 💸 

Spotlight on latest capital raises, M&A and investments…

📈 PhysicsX: a developer of AI technology designed to automate engineering processes and accelerate accurate physics simulations around human health raised $32M in Series A funding. General Catalyst led the round and was joined by others. (link)

📈 Cradle: a startup that uses generative AI to design proteins raised $24M in Series A funding. Index Ventures led, and was joined by Kindred Capital. (link) 

📈 Mytos: a biotech platform based in the UK that automates cell manufacturing raised $19M in Series A funding. Buckley Ventures led the round and was joined by others. (link)

📈 Biolexis Therapeutics: a company based in UT using AI for metabolic drug discovery raised $10M in Series A funding from Clarke Capital. (link)

📈 Pepper Bio: an AI-powered platform designed to improve drug discovery and development raised $6.5M in seed funding. NFX led the round and was joined by Silverton Partners, Merck Digital Sciences Studio, Mana Ventures, and others. (link)

Other Relevant News 🔍

News, podcasts, blogs, tweets, resources, etc…

• Forbes 30 under 30 in healthcare (link)

• House lawmakers agree AI data privacy protections needed (link)

• Study shows AI priority #4 for group practices (link)

• Experts unclear on patient compensation when AI is wrong (link)

• Regulators in blue states increasing scrutiny of insurers’ AI use (link)

• Clinical innovation mindset (link)

• The future of AI in biology with Dr. George Church (spotify)

• Top 10 Healthcare Industry Predictions For 2024 by Sachin H. Jain (link)

• Mayo Clinic investing $5B on high-tech medical campus redesign (link)

Visuals of the Week 📸

Funny memes, cool pics, and interesting data from around the web…