Skip to main content
Radiolab

The Medical Matchmaking Machine

61 min episode · 2 min read
·
David Fagenbaum

Episode

61 min

Read time

2 min

Topics

Relationships, Investing, Fundraising & VC

AI-Generated Summary

Key Takeaways

  • Drug Repurposing Economics: Twenty to thirty percent of all prescriptions written in the US are off-label because FDA approval locks drugs to specific diseases at fixed prices. Companies maximize profit on first approval, creating no financial incentive to study additional uses for generic medications.
  • AI-Driven Medical Discovery: Machine learning algorithms trained on thousands of known drug-disease relationships generate 75 million probability scores (0.01 to 0.99) ranking every FDA-approved drug against every disease. This identifies repurposing candidates humans cannot systematically evaluate, like sirolimus for Castleman disease or pembrolizumab for angiosarcoma.
  • Biomedical Knowledge Graphs: Mapping every gene, protein, pathway, and disease as connected nodes reveals treatment opportunities through shared biological mechanisms. Fagenbaum identified mTOR activation in his blood samples, traced connections to sirolimus (rapamycin), and achieved eleven-year remission after five near-fatal relapses.
  • COVID Drug Success Model: Dexamethasone and tocilizumab, both repurposed drugs, saved millions of lives during the pandemic. Dexamethasone was initially recommended against for COVID but reduced mortality by thirty-five percent when tested, demonstrating how systematic repurposing evaluation uncovers life-saving treatments faster than new drug development.
  • Public Algorithm Release Strategy: Every Cure will release their Matrix algorithm publicly within nine months, allowing researchers and physicians to access the same drug-disease probability scores the medical team uses. The tool generates research hypotheses requiring laboratory validation and clinical trials, not direct treatment recommendations for patients.

What It Covers

Doctor David Fagenbaum survives five near-death relapses from rare Castleman disease by discovering rapamycin through self-experimentation, then builds AI system scoring 75 million drug-disease combinations to identify repurposing opportunities for all 18,000 known diseases.

Key Questions Answered

  • Drug Repurposing Economics: Twenty to thirty percent of all prescriptions written in the US are off-label because FDA approval locks drugs to specific diseases at fixed prices. Companies maximize profit on first approval, creating no financial incentive to study additional uses for generic medications.
  • AI-Driven Medical Discovery: Machine learning algorithms trained on thousands of known drug-disease relationships generate 75 million probability scores (0.01 to 0.99) ranking every FDA-approved drug against every disease. This identifies repurposing candidates humans cannot systematically evaluate, like sirolimus for Castleman disease or pembrolizumab for angiosarcoma.
  • Biomedical Knowledge Graphs: Mapping every gene, protein, pathway, and disease as connected nodes reveals treatment opportunities through shared biological mechanisms. Fagenbaum identified mTOR activation in his blood samples, traced connections to sirolimus (rapamycin), and achieved eleven-year remission after five near-fatal relapses.
  • COVID Drug Success Model: Dexamethasone and tocilizumab, both repurposed drugs, saved millions of lives during the pandemic. Dexamethasone was initially recommended against for COVID but reduced mortality by thirty-five percent when tested, demonstrating how systematic repurposing evaluation uncovers life-saving treatments faster than new drug development.
  • Public Algorithm Release Strategy: Every Cure will release their Matrix algorithm publicly within nine months, allowing researchers and physicians to access the same drug-disease probability scores the medical team uses. The tool generates research hypotheses requiring laboratory validation and clinical trials, not direct treatment recommendations for patients.

Notable Moment

Fagenbaum's uncle received a terminal angiosarcoma diagnosis with two months to live. Despite physician resistance, testing revealed ninety-nine percent of tumor cells expressed PD-L1. Pembrolizumab treatment achieved nine-year remission, and the approach became standard care without formal clinical trials.

Know someone who'd find this useful?

Episode Transcript

This holiday season, find something extraordinary. The Genesis Signature event is your moment to experience design that draws you in, power you can feel, and space that fits up to seven when you need it. And every Genesis comes with SiriusXM, three months of ad free music, plus sports, news, and more. Find the Genesis that's right for you at the Genesis Signature event, happening now. And make Genesis a memorable start to your season. Building a portfolio with Fidelity Basket Portfolios is kinda like making a sandwich. It's as simple as picking your stocks and ETFs, sorta like your meats and other topics, and managing it as one big juicy investment. Now that's pretty good. Learn more at fidelity.com/baskets. Investing involves risk, including risk of loss. Fidelity Brokers Services LLC, member NYSCSIPC. Getting more as a my Lowe's Pro rewards member is easier than ever with the Lowe's app. Download it today and earn 500 points the first time you log in. Plus, your digital wallet helps you scan, save, earn, and access what you need to manage it all in one place. Download the Lowe's app and take advantage of your pro benefits today. Lowe's. We help. You save. Valid 12125 to 1726. Offer valid for first login per organization only. Loyalty program subject to terms and conditions. Additional restrictions apply. Visit lowes.com/terms for more details subject to change. Wait. You're listen k. Alright. K. Alright. You're listening listening to RadioLab. From W n y w n y six. See? Yep. Hey. I'm Latif Nasr. This is Radio Lab. And today, I wanna share a conversation that I had with a guy named David Faganbaum. He's a doctor and a professor at the University of Pennsylvania, and there's a combination of reasons why I think his personal story is so extraordinary and why I wanted to share with you. Part of it is this staggering series of crises that he faced in his personal life starting when he was in university. Part of it is kinda his personality, like, how he what there was in him that made him stand up to these crises in a really particular way. And then part of it is the way that he he took his response to those crises, and now he's scaling it up using one of the most controversial technologies around AI. The result of all of this is that he is right now in the middle of doing something wildly ambitious, something I find kinda miraculous, also maybe troubling. Either way, it it is definitely going to change the medical system down to the level of the pills that that you put in your mouth. That said, I just found my conversation with David so fascinating, and and his personal backstory in general, I found it so dramatic that I wanted to let it unfold at its own pace without jumping too quickly to the end. So here we go. Yeah. And then I'll hit join studio. Latif? …

Get the full transcript (13,502 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all Radiolab transcripts →

You just read a 3-minute summary of a 58-minute episode.

Get Radiolab summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links.

Tools

  • by Every Cure

    Every Cure will release their Matrix algorithm publicly within nine months, allowing researchers and physicians to access the same drug-disease probability scores the medical team uses.

More from Radiolab

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best Science Podcasts (2026) — ranked and reviewed with AI summaries.

Read this week's Investing & Markets Podcast Insights — cross-podcast analysis updated weekly.

You're clearly into Radiolab.

Every Monday, we deliver AI summaries of the latest episodes from Radiolab and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime