Doctors Told Joseph He Was Going to Die. Then AI Saved His Life

A little over a year ago, Joseph Coates, a 37-year-old man from Renton, Washington, faced a grim reality. He had been battling POEMS syndrome, a rare blood disorder that left him with debilitating symptoms, including numb hands and feet, an enlarged heart, and failing kidneys. Doctors informed him that he had reached a critical point where the only remaining choice was whether he wanted to spend his last days at home or in the hospital.

During this time, Joseph’s condition deteriorated, leading to frequent procedures to drain excess fluid from his abdomen. His health was too poor to qualify for a stem cell transplant, which was one of the few remaining treatment options that could have offered him a chance at remission. Joseph felt completely hopeless, stating, “I gave up. I just thought the end was inevitable.”

However, his girlfriend, Tara Theobald, refused to accept this fate for Joseph. She sent an email to Dr. David Fajgenbaum, a doctor on the other side of the country whom they had met at a rare disease summit a year prior, asking for help.

The next morning, Fajgenbaum responded with a suggestion: an unusual combination of chemotherapy, immunotherapy, and steroids, which had not been tested before for Coates’ condition. Within a week of starting this new treatment, Joseph began to respond positively. In just four months, he was healthy enough to undergo a stem cell transplant, and today, he is in remission.

The lifesaving treatment protocol did not originate from Fajgenbaum or any other doctor—it was generated by an artificial intelligence model designed to identify potential therapies from existing medications.

Using AI to Find New Treatments

In laboratories worldwide, researchers are now harnessing the power of AI and machine learning to repurpose established medications for the treatment of rare diseases. This approach, known as drug repurposing, is not new; however, the integration of AI is significantly speeding up the process and broadening the possibilities for individuals suffering from rare conditions.

Fajgenbaum's team at the University of Pennsylvania, along with multiple global research initiatives, is utilizing machine learning to quickly repurpose drugs for various conditions, including aggressive cancers, rare inflammatory disorders, and complex neurological issues. The preliminary successes have raised an important question: How many more potential cures might be overlooked in the vast archives of existing medications?

Donald C. Lo, a former head at the National Center for Advancing Translational Sciences, highlighted the abundant opportunities present in the pharmaceutical world, stating, “There’s a treasure trove of medicine that could benefit many other diseases. We just didn’t have a systematic way of exploring it.”

In the U.S., a disease is classified as rare if it affects fewer than 200,000 people. Despite the existence of thousands of rare diseases affecting millions globally, over 90% currently lack approved treatments, primarily because pharmaceutical companies are reluctant to invest in developing new drugs for small patient populations.

This is precisely why drug repurposing has become an attractive option for discovering treatments for rare diseases. Dr. Marinka Zitnik, a Harvard researcher studying the application of computer science in medicine, remarked, “AI essentially puts rocket boosters on that.”

Finding New Uses for Existing Medications

Drug repurposing is somewhat commonplace in the medical world. For instance, minoxidil, initially developed to treat high blood pressure, is now known for treating hair loss. Similarly, Viagra, originally intended for a cardiac ailment, gained fame as a treatment for erectile dysfunction. These examples demonstrate that many existing drugs have additional uses that may not have been previously identified.

Dr. Fajgenbaum's journey into drug repurposing began with his own health struggles. After being diagnosed with a rare subtype of Castleman disease at age 25, he faced significant challenges in finding effective treatments. In desperation, he scoured medical literature and began investigating unconventional therapies, ultimately discovering that a generic medication called sirolimus could successfully manage his condition, keeping it in remission for over a decade.

Recognizing the potential of existing research, Fajgenbaum founded a nonprofit organization called Every Cure in 2022. The mission of this organization is to leverage machine learning to analyze vast quantities of pharmaceutical data for more rapid drug repurposing.

Other research teams in the U.S. and around the world are undertaking similar initiatives. For example, in Alabama, researchers utilized an AI model to propose an unconventional treatment for a 19-year-old experiencing chronic vomiting, which turned out to be remarkably effective. Additionally, several cases have been documented where known medications were repurposed to treat rare genetic and neurological disorders.

Challenges and Future Prospects

As promising as these developments are, drug repurposing presents its own set of challenges. Identifying the efficacy of treatments can be difficult, and careful physician oversight is crucial in determining when to attempt novel therapies suggested by AI. Both Dr. Zitnik and Christine Colvis emphasize that while AI can significantly aid in discovering new treatment options, the quality of evidence and physician judgment remain paramount.

In Joseph Coates’ case, Dr. Wayne Gao initially hesitated before implementing the AI-recommended treatment due to its unconventional nature, fearing it might worsen Coates' condition. Yet, with no other options available and considering Joseph’s young age, Dr. Gao felt obligated to proceed with the treatment. The AI-driven model had suggested what seemed like a risky approach, but it ultimately saved Coates’ life.

Almost a year after this pivotal moment, Joseph and his girlfriend returned to Philadelphia to express their gratitude to Dr. Fajgenbaum. With newfound health and vitality, Joseph shared his experience, stating he felt “just fine” after undergoing such a challenging journey.

This story is just one of many emerging examples highlighting how AI can positively impact medical treatment. The advancements made possible by such technology show promise in reshaping healthcare for those facing rare and difficult-to-treat diseases.

health, AI, medicine