How AI is Transforming Treatment for Rare Diseases?

I recently came across a powerful article in The New York Times about the miraculous recovery of a young man named Joseph. This article offers a glimpse into the future of rare disease treatment - an area long overlooked by pharmaceutical companies - now being transformed by artificial intelligence. By uncovering the hidden potential of existing drugs at remarkable speed, AI is raising new hopes for life-saving breakthroughs.

Here’s the story.

Just over a year ago, Joseph was preparing to die. His doctors had run out of options – until AI found one more. It gave him an unexpected second chance.

At the age of 37, Joseph was battling POEMS syndrome, a rare and potentially fatal malignant blood disorder that had ravaged his body. He was too sick to receive a stem cell transplant – one of the only treatments that could have put him in remission.

In desperation, his wife reached out to a researcher across the country who offered not a new experimental drug, but an unconventional combination of existing drugs recommended by artificial intelligence.

That drug regimen - chemotherapy combined with immunotherapy and steroids - saved Joseph's life.

This regimen, not yet tested for POEMS, was not devised by the doctor – but spit out by an artificial intelligence model.

What is a rare disease?

The National Institutes of Health defines rare diseases as those affecting fewer than 200,000 people in the United States. But there are more than 7,000 known rare diseases, affecting an estimated total of 25 to 30 million Americans and 27 to 36 million Europeans. Yet, more than 90% of rare diseases have no approved treatment.

Why do rare diseases need AI’s help most?

Pharmaceutical companies rarely invest in rare diseases because developing a new drug is expensive - often taking over a decade and billions of dollars. The potential return on investment just doesn’t justify the cost for such small patient populations.

But drug repurposing flips the equation.

Old drugs have well-known, well-described safety and efficacy profiles and established mechanisms of action. These drugs are already FDA-approved and many of them are inexpensive generics. The repurposing of old drugs allows researchers to build on existing data, which expedites the development time and increases the chances of success. Physicians and researchers can move faster- and AI helps them move even faster still.

The hidden power of old medecines

Generally speaking, when a drug is developed, it is intended for a single, specific use. In reality, many drugs can have much broader effects than those initially sought. Sometimes those “side effects” turn out to be their most powerful characteristics.

Take minoxidil, developed to treat hypertension, which was repurposed and is now commonly used as a treatment for hair regrowth; some patients treated for hypertension complained of excessive hair growth. Sildefenil (Viagra) was also originally developed to treat hypertension; its users began to complain of unexpected erections, so the product is now prescribed to treat erectile dysfunction. These success stories, and many others, are the result of observation and chance.

But what if we didn't have to rely on luck?

This is where machine learning comes in. The first step is to identify one or more genetic abnormalities that are causing the dysfunction in a particular patient. The AI model can then be fed information and comb through tens of thousands of existing drugs, medical studies and patient records, looking for patterns and potential drug-disease matches. Once a potential new use is identified, the drug is tested in preclinical models and then in clinical trials to assess its efficacy and safety in the targeted rare disease. This is called ‘drug repurposing’ and AI plays an important role in it.

AI: a medical detective

Instead of taking years to explore whether one drug might treat one disease, AI can run millions of comparisons in days. For example, Every Cure, a nonprofit founded by Dr. David Fajgenbaum, has created a platform that matches about 4,000 known drugs against over 18,000 diseases, assigning them a potential effectiveness score.

Today, this approach is so promising that many companies around the world are now using AI for drug repurposing. They include small companies such as BenevolentAI, Atomwise Inc , Insilico or Cloud Pharmaceuticals but also large pharmaceutical companies such as Astra Zeneca, Pfizer, Novartis and Janssen, which are engaged in this type of research.

The results are extraordinary. Chemotherapy combined with immunotherapy and steroids suggested by AI brought Joseph back from the brink. A drug for Parkinson’s disease has helped children with a rare movement disorder walk and talk.

Thanks to the versions of the technology developed by Dr. Fajgenbaum’s team and others, drugs are quickly being repurposed for conditions including rare and aggressive cancers, fatal inflammatory disorders and complex neurological conditions. And often, they’re working.

These discoveries are no longer accidents. They are the result of AI exploring neglected corners of medical research, clinical data and historical case studies. As Dr. Fajgenbaum puts it, there’s a “treasure trove of medicine” hiding in plain sight.

Challenges and the road ahead

Of course, AI isn’t perfect. Not all predictions are breakthroughs. Some drugs fail. Human judgment, ethical oversight, and rigorous clinical trials still play an essential role. But the potential upside is undeniable.

And while drug repurposing won't make big money for pharmaceutical companies (many drugs are off-patent and cheap), it offers something even more valuable: cures for people who didn't have any.

Thanks to recent funding and growing public interest, projects such as Every Cure are scaling up. They’re bringing together researchers, clinicians, and data scientists to turn the world’s libraries of old drugs into engines of new possibilities.

Final Thoughts

Joseph's story is more than a medical miracle. It is a window into a future where technology will help us care in a faster, smarter, more humane way through drug repurposing.

AI didn't invent a new drug, it allowed doctors to see the ones they already have in a new light! AI won't replace doctors, but it can provide them with better tools, sharper insights and new lifelines to offer their patients. For the millions of people facing rare diseases, often in silence and isolation, this is about much more than innovation.

This is about hope, rediscovered through the detours of old research, which is just waiting to be revealed in a new light.