Perspective: AI’s Role in Combating Castleman’s Disease
Rare diseases impact thousands of people worldwide, often leaving patients with limited treatment options and little hope for recovery. Many of these conditions are so poorly understood that drug development stalls before it begins. However, a new study from the University of Pennsylvania offers a glimpse into a future where life-saving treatments are discovered not through years of trial and error, but by revisiting medicines we already have—with the help of cutting-edge scientific tools.
A Deadly Diagnosis and a Last Chance
The patient, facing idiopathic multicentric Castleman’s disease (iMCD), had run out of options. iMCD is a brutal, rare immune disorder where the body attacks itself in what’s called a cytokine storm. Organs swell, inflammation spreads, and the body’s own defense system becomes its worst enemy. Fewer than 5,000 Americans are diagnosed each year—and fewer still survive the worst flares.
This patient had reached hospice care. But instead of planning goodbyes, his doctors made one more try—this time, with help from a machine learning platform designed to comb through 4,000 existing drugs in search of a life-saving match. What they found was remarkable: adalimumab, a drug already used for conditions like rheumatoid arthritis and Crohn’s disease.
The Power of Repurposing
Adalimumab targets a protein called tumor necrosis factor (TNF), which, the team discovered, played a significant role in this patient’s disease. Elevated TNF levels had been driving his flare, and by blocking it, there was a chance—just a chance—that they could quiet the storm.
Against the odds, the gamble worked. Nearly two years later, that patient is in remission.
What struck me most wasn’t just the outcome but the principle behind it. This wasn’t about inventing a new drug. It was about repurposing what we already have. About realizing that diseases might look different but sometimes share the same biological switches.
As someone who’s watched rare disease steal years from a loved one, I couldn’t help but imagine what it would have meant to have this kind of approach back then. To know that answers don’t always require new frontiers—they might be sitting quietly on a pharmacy shelf.
When the Scientist is Also the Patient
There’s a deeply human layer to this study. Dr. David Fajgenbaum, the lead researcher, knows iMCD intimately—not just as a scientist but as a patient. Diagnosed in medical school, he nearly died five times before discovering his own life-saving repurposed therapy over a decade ago.
I’ve always believed that the best science comes from those who know what it means to wait at a bedside, to feel the crushing weight of limited options. Fajgenbaum isn’t just studying hope but he’s lived it.
His experience inspired the creation of Every Cure, a nonprofit that seeks to use data, machine learning, and sheer persistence to find untapped uses for existing medications. It’s a mission born from survival; and a future where rare disease doesn’t automatically mean helplessness.
Rethinking the Way We Search for Cures
This research reminds me that medicine doesn’t always have to mean a race for new drugs. Sometimes, it’s a matter of asking new questions about the old ones.
We live in a world where there are over 4,000 FDA-approved drugs—each with known safety profiles and decades of patient data. Yet most of these drugs remain locked in their original uses. Imagine what could happen if we systematically asked: Could this treat something else?
For rare diseases especially, this is a lifeline. Big pharmaceutical companies rarely invest in developing drugs for conditions that affect only a few thousand people. It’s too risky, too costly. But repurposing changes that equation. It offers a way to bridge the gap between what exists and what’s needed—without waiting decades for new drug approvals.
The Human Cost of Rare Disease and the Hope Ahead
For every person like the patient in this study, there are thousands of others whose lives hang in the balance. I’ve sat through enough waiting rooms to know that desperation doesn’t always make the headlines. But it’s there—in the quiet prayers, the long stares at lab results, the empty chairs once filled by someone you loved.
This study isn’t just a victory for one man or one disease. It’s a beacon for the entire rare disease community, which so often feels invisible. It proves that when we invest in creative thinking—pairing AI tools with biological insight and human empathy—we can change the narrative.
Looking Forward
Penn’s team is already pushing forward. Their next target is testing a JAK1/2 inhibitor, another repurposed drug, in iMCD patients. But their bigger mission is clear: make drug repurposing routine, not rare.
There’s a powerful lesson here, one I wish my grandmother’s doctors had been able to act on years ago: that sometimes, the answers are already here. What’s missing is the right lens, the right people, and the courage to try.
My Final Thoughts
I keep thinking about that patient—how he was packing for hospice, planning for the end. And then, thanks to a team of doctors who refused to give up, he got the chance to start over. That’s the kind of story that stays with you.
We talk so often about the future of medicine; gene therapy, precision drugs, personalized care. But sometimes, the most revolutionary thing is simply remembering that every pill on the shelf once started as someone’s long shot. And maybe those pills still have stories left to write.
Citation: Fajgenbaum, D. et al. (2025). AI-Guided Drug Repurposing for Idiopathic Multicentric Castleman’s Disease. New England Journal of Medicine. https://www.pennmedicine.org/news/news-releases/2025/february/ai-tool-helps-find-life-saving-medicine-for-rare-disease