On this episode, we discuss the FDA approval of Enspryng for TED, the new FDA Rare Disease Framework Initiative, Ultrangenyx’s second sold PRV this month, and using AI to reduce diagnostic delays.
FDA Approves Treatment for TED
The FDA has approved Enspryng, or satralizumab, for adults with thyroid eye disease. It’s the first subcutaneous treatment for the condition that patients can take at home.
Thyroid eye disease is a rare autoimmune disease that attacks the tissues around the eyes. It’s most often seen in people with Graves’ disease who have an overactive thyroid, hence the name of the disease.
The immune system targets cells in the eye socket. Because the eye socket is small, there’s nowhere for that tissue to go except forward. As a result, the person’s eyes bulge (proptosis). Patients can also have redness, pain, eyelid retraction, and double vision. In severe cases it can threaten eyesight.
Enspryng blocks the IL-6 receptor that is involved in inflammatory signalling. The drug was tested in two identical Phase 3 trials. In the first trial, the primary endpoint was not met, but in the second, it was: 53% of patients on Enspryng had at least a 2 mm reduction in eye bulging at 24 weeks, compared with 23% of patients taking placebo. Double vision improved in 61% of patients taking Enspryng versus 26% of patients taking placebo.
Enspryng was first approved in 2020 for neuromyelitis optica spectrum disorder, or NMOSD. Now it has a second indication in thyroid eye disease, with an FDA decision expected in January for MOGAD.
It’s a good example of how rare disease drug development can create opportunities beyond a drug’s initial approval.
FDA Rare Disease Framework Initiative
The FDA announced the Rare Disease Framework Initiative. The goal is to build a new, “fit-for-purpose” approach to reviewing rare disease therapies.
The traditional model of drug development was built around common diseases. You enroll hundreds or thousands of patients, randomize them to drug or placebo, and follow them to see if there is a difference between the two groups.
That works very well for common ailments like hypertension or diabetes. And for some of the more prevalent rare conditions, such as hemophilia or cystic fibrosis.
But when there are only a few hundred people in the whole country with a rare condition, developing a clinical trial can be challenging. People with rare diseases are highly heterogeneous, and it can be difficult to recruit a group of patients with similar symptom severity, age, comorbidities, etc., to create the bell-shaped population needed for statistical analysis.
That’s why things like natural history data, biomarkers, and newer statistical methods matter so much in rare diseases. And the FDA specifically lists in their initiatives support for advanced science, novel methods, and innovative statistical approaches as some of its goals.
The initial reaction has been positive. NORD called it an “important and welcome step forward,” and Leerink Partners noted that the FDA may be making the path to approval less of a guessing game.
But the real test is implementation. The rare disease community has heard a lot about policy changes, and it’s understandable to want to see those announcements translate into practical guidance for drug developers.
The FDA’s first public meeting on this initiative is set for October 28th.
Ultragenyx Gets $203M for PRV
Ultragenyx announced on October 14th that it has agreed to sell a rare pediatric disease priority review voucher for $203 million. That’s its second voucher sale in a week, following a $210 million sale announced on October 7th.
The first voucher came from a gene therapy to treat glycogen storage disease type Ia. This second, $203 million PRV was for a gene therapy to treat Sanfilippo type 3a.
In both diseases, the children are missing an essential enzyme and the gene therapy delivers a working copy of that enzyme. Also, both diseases affect very small populations of children. Without a PRV program in place, the odds of these treatments getting developed were likely very small.
This story also highlights an important part of the rare disease business model. Developing a therapy for a very small patient population of children can be expensive, and the financial return is often uncertain. A priority review voucher creates additional value beyond the drug itself.
These two sales bring in more than $400 million in about a week. That’s a meaningful source of capital, particularly for a company that recently reported a Phase 3 setback in Angelman syndrome, as we discussed in Episode 7.
The broader point is that rare disease innovation depends on more than scientific breakthroughs. Financial incentives can help companies keep investing in therapies for patients with few or no treatment options.
Using Artificial Intelligence To Reduce Diagnostic Delays
Researchers at Vanderbilt and the Undiagnosed Diseases Network have published a study in Nature Digital Medicine on an AI tool called RARE-PHENIX.
It focuses on diagnosing patients faster. Diagnosing a rare disease can take a long time. The authors of the paper note that the average delay is four to eight years, and those delays are filled with misdiagnoses, inappropriate treatments, and unnecessary tests.
RARE PHENIX is focused on organizing the symptoms that physicians write down when examining a patient. Those symptoms may include symptoms that may or may not be part of rare condition. Patients, even if they have a rare condition, can have other symptoms due to comorbidities, aging, etc. Looking for patterns in the plethora of symptoms can be difficult. RARE-PHENIX organizes that chaos. It first uses a large language model to convert those listed symptoms into standardized terms. And then it ranks them as to whether they are likely due to a rare condition or not.
Testing showed the model was twice as good as ChatGPT.
An interesting aspect of this model is that it mirrors how clinicians approach a diagnosis. That produced substantially better results than a single-prompt ChatGPT approach.
Just as importantly, the tool doesn’t make the diagnosis. It helps organize and prioritize the clinical clues for a physician to review. That’s a promising role for AI: helping clinicians see patterns they might otherwise miss while keeping clinical judgment a human process.
References
FDA approves Enspryng to treat thyroid eye disease
https://www.roche.com/media/releases/med-cor-2026-10-16
FDA Launches Rare Disease Framework Initiative to Develop Holistic, Fit-for-Purpose Framework for Rare Disease
Ultragenyx Enters into Agreement to Sell Rare Pediatric Disease Priority Review Voucher for $203 Million
Shyr C, et al. An artificial intelligence framework for end-to-end rare disease phenotyping from clinical notes using large language models. npj Digit Med. 2026
https://www.nature.com/articles/s41746-026-03275-0