By Allison Proffitt
June 2, 2026 | BOSTON—Rare disease is, by its very nature, a puzzle. It’s not the satisfying kind that you can solve in one rainy afternoon, but the kind where the pieces arrive years apart, from different laboratories, different clinics, different patient experiences, and different legislative incentives.
Editor's Note: This article was originally published on Bio-IT World News. To read the original article, visit here.
It’s a puzzle that needs community.
In the opening plenary session at the 25th annual Bio-IT World Conference & Expo on Tuesday, we hosted an intimate series of paired conversations, each designed to place a different set of puzzle pieces on the table, examining that puzzle from four distinct vantage points — patient experience, AI-assisted diagnostics, venture investment, and policy reform — with each conversation delivering a bit more clarity to the emerging picture.
The Patient Perspective: Living with MG
The session opened with a conversation between Susan Ward, founder and executive director of clinical Trial Advocacy Partners (cTAP), and Tom Bartlett, an ambassador for MG Uniter and a patient living with myasthenia gravis (MG).
Bartlett, a 15-year Bio-IT World veteran who spent nearly two decades at Apple and later led global sales for a life sciences company, described the abrupt onset of his disease in 2019 during a tennis match. “Things weren't working,” he said. Unlike most rare disease patients — who wait an average of seven years for a correct diagnosis — Bartlett was identified relatively quickly, thanks to a primary care physician who thought to test for MG.
MG is an autoimmune disorder in which the body attacks the receptors that allow the brain to communicate with muscles. Bartlett described its effects with candor: a voice that degrades over hours of speaking, legs that must be rationed against a strict energy budget, and a four-hour recovery period required after four hours of activity. “If I try to work full time, that would mean I would sleep 24 hours a day. The math doesn’t work.”
He also spoke to the state of clinical data collection in rare disease. Twice-yearly visits to his neurologist — which he described as a stroke of luck, given that his specialist is among the best in the world — produce an assessment via pen-and-paper questionnaire that dictates the next steps in his care. When his Apple Watch recently flagged a high probability of falling, his neurologist's response was blunt: there was nothing she could do with data that wasn’t clinical.
“So it’s like, okay, how do we change that?” Bartlett said. The exchange captured a tension the session would return to repeatedly: the gap between the data patients generate and the data clinicians can use.
Ward noted that this gap is structural, not incidental. Rare diseases lack the billing infrastructure that drives the development of structured data modules in electronic health records. Unstructured notes, inconsistent standards across geographies, and the fundamental heterogeneity of rare diseases — in which two patients with the same diagnosis may present very differently — compound the challenge at every step.
“You need a really rich and deep ontology, a really rich phenotype,” Ward said. “And you need to be able to also have the context for what those data mean in the context in which they’re derived.”
Still, she closed with a note of conditional optimism. “The technology is there. What needs to happen is to think about it orthogonally and put the pieces together in a different way.”
The Diagnostic Engine: From Seven Years to Fourteen Hours
The session’s second conversation paired William Van Etten, founder of StarfleetBio, with Sebastien Lefebvre, head of technology, data, and AI at Aurelis Insights — a researcher with more than a decade of experience building AI-assisted diagnostic platforms for rare disease.
Lefebvre opened with data. Drawing on a 2022 report conducted with RARE-X using Orphanet and the U.S.-based OMIM database, he described a landscape of approximately 12,000 distinct rare diseases when counted at the level of individual gene-condition combinations. A majority have a genetic underpinning; a similar proportion have at least three documented phenotypes. "Theoretically speaking, 80% of these 12,000, if they're genetically underpinned, we can start helping out from a digital standpoint," he said.
The platform his team built, called Rare Answers, combines genomic sequencing with phenotypic extraction from medical records — mapped against the Human Phenotype Ontology — to generate a patient signature that can be matched against the known disease landscape. Working with Rady Children's Hospital in San Diego beginning in 2017, the team brought whole genome sequencing and interpretation time down to 19 hours from blood sample to diagnosis. Since then, further developments have brought the time down to just a handful of hours.
The practical stakes were illustrated by a single case: a newborn girl seizing from her first day of life and placed on standard anticonvulsant medication. Whole genome sequencing identified a sodium channel variant. The treatment was switched accordingly. “Condition stabilized,” Lefebvre said. “Kid is normal now.”
He framed the broader diagnostic delay as something between a tragedy and a solvable problem. "There's no reason for it," he said. "A kid in an ICU can be diagnosed and treated under 14 hours. It doesn't need to take seven years."
Van Etten then turned the conversation to the data problem that precedes diagnosis: the privacy risks inherent in commercial genomic sequencing. His new company, StarfleetBio has developed an iPhone application that stores and processes an individual's genome locally, encrypted with a key held only by the user's device. The genome never leaves the phone in a readable form.
“Your personal genome is out there,” Van Etten said of the current commercial model. “It’s not like you can change a password and get a new credit card.”
The app, developed in collaboration with BioTeam, which Van Etten cofounded, and the University of New Hampshire's Hubbard Center for Genomic Studies, also includes a feature explicitly designed for rare disease research: users can toggle a switch to indicate interest in participating in a clinical trial or cohort study. Researchers can then send queries to devices — receiving aggregate responses about cohort eligibility — without ever accessing individual genomic data. “What you're sending is a question,” Van Etten said. “Your phone answers the question and sends it back.”
The Investment Lens: Capital, Computation, and the Bedside
The evening’s third conversation drew the most expansive arc, ranging from the history of biomedical modalities to the future of medical genetics. Catherine Brownstein, scientific director of the Manton Center for Orphan Disease Research at Boston Children's Hospital and Harvard Medical School, interviewed Morgan Cheatham, a partner and head of healthcare and life sciences at Breyer Capital who is also completing his training as a pediatric medical genetics physician.
The Manton Center, Brownstein explained, is a philanthropically funded virtual center to which patients can be referred from any continent, any U.S. state, and virtually any department within Boston Children's — specifically for cases where no diagnosis has been reached. Its research program is unusual in allowing patients to remain identifiable, enabling researchers to return to families with new findings as science advances.
Working with OpenAI, the center deployed agentic AI models in a zero-shot configuration — no specialized fine-tuning — to conduct retrospective analyses of undiagnosed patients. “We were able to diagnose patients that had gone years, decades, their entire lives in some cases without being diagnosed,” Brownstein said.
Cheatham extended the finding into a broader argument about the role of frontier technology in medicine. The most important modalities of recent decades — antisense oligonucleotides, RNA medicines, in vivo CAR-T — were not validated in large population trials. They were validated in rare disease cohorts. “It was thanks to the help of rare [disease] patients who were willing to participate in trials that allowed us to show the efficacy, the safety, and the durability of these modalities,” he said, pushing back on what he characterized as a post-GLP-1 industry tendency to fixate on large-indication opportunities.
Brownstein offered her own version of the same point, drawn from years spent studying hypophosphatemic rickets — a rare skeletal disorder. “People would say, why are you studying hypophosphatemic rickets?” she recalled. “And it's because these extreme cases, these rare presentations of disorders where you don't know the underlying etiology, they inform the common diseases,”
Cheatham’s investment thesis at Breyer Capital tracks this logic. The fund is structured as a single-LP evergreen vehicle — a deliberate departure from the standard venture fund cycle, which he argued is poorly suited to the timelines of genuine biomedical breakthroughs. Current areas of focus include AI applied to diagnostic imaging and endoscopy, in vivo bioengineering as a more cost-scalable alternative to ex vivo cell therapies, and tools for standardizing clinical documentation.
On the question of AI in therapeutics, he offered a gentle corrective to industry hype. “A patient can’t take an AI model. A patient needs a drug to see effect, to heal.” He described walking the hematology wards at Boston Children's on a day of poor air quality, watching admissions stack up with sickle cell patients in vaso-occlusive crisis — patients for whom approved gene therapies technically exist but whose treatment journey involves months of hospitalization, myeloablative conditioning, and conversations about fertility preservation at age 12. “We need to ground ourselves in what those lived realities are.”
Asked where AI would take the field of medical genetics, Cheatham predicted that genetics — long a consult service, called in late and rarely the custodian of the intervention — would be progressively absorbed into other specialties. Neurologists and cardiologists, he argued, would increasingly become dual-board certified in genetics. “We’re the nerds in the back of the hospital in the dusty room that hasn't been renovated in 20 years,” he said. “Over the next five or the next ten, we may emerge in a much more interesting position.”
The Policy Front: Rewriting the Right to Try
The final conversation returned to Van Etten, now in the interviewer’s role, speaking with Dylan Livingston, founder and president of the Alliance for Longevity Initiatives — a nonprofit advocacy organization focused on legislation to extend healthy human lifespan.
Livingston described an unlikely entry into the longevity field: a Harry Potter-reading twelve-year-old whose father brought home a description of Aubrey de Grey from a regenerative medicine conference. He came back to the field during COVID, in his early 20s spending his “safer-at-home” time with his 92-year-old grandfather and struck by the stark age gradient of risk and disease severity. “It all kind of comes back to aging.” he said.
His legislative focus is the Right to Try framework — a federal law enacted in 2018 that permits patients with terminal illnesses to access unapproved therapies that have passed Phase 1 clinical trials. In practice, Livingston argued, the law has been largely unused: the definition of terminal illness is narrow, and a lack of liability protections for physicians and manufacturers created a "nightmare" that discouraged participation.
In 2023, the Alliance succeeded in passing legislation in Montana that expanded Right to Try eligibility to all patients, regardless of health status. The rationale was dual: to protect patients who were already obtaining experimental therapies at overseas clinics from doing so without adequate safeguards, and to capture the data those treatments generate for the benefit of the broader research ecosystem. “By going to these clinics abroad, we lose the potential data collection that would be very beneficial to clinical trials,” Livingston said.
The Montana model requires prescriptions from two physicians and review by an IRB. A 2% annual tax on treatment centers creates a fund for patients who lack the financial means to access therapies. A parallel bill — HB 1734 — is currently advancing in the New Hampshire legislature, where it has passed the House and is under consideration in the Senate.
Livingston illustrated the stakes with a case he is currently working on: a father whose son died from a rare mitochondrial disease, who then sequenced his two surviving children and found the same mutation. “He said, ‘To hell with the clinical trial data. I want my kid to have the drug,’” Livingston said. Montana’s framework offers a potential path.
He acknowledged the counterarguments — concerns about access being limited to wealthy patients, and fears of bad actors pushing unapproved therapies — but addressed both. Democratization of cost follows the curve of every technology, he argued, while the requirement that treatments have passed Phase 1 “shakes out the snake oil.” More broadly, he cast the state-level expansion of Right to Try as a form of federalist competition that could eventually inform federal policy. “The innovation is going to come from the states competing against each other to create the best model for approving and giving access to these therapeutics.” And while Livingston acknowledges that FDA has given no indication that it would consider Right to Try use alongside traditional clinical trial data, that doesn’t mean the investors won’t be watching. “The way I see this is [Right to Try use] will make investors a lot more informed in choosing which companies to back and which ones to not back,” he said.
A Puzzle in Progress
The session closed with the metaphor it opened with — a puzzle whose pieces are still arriving. What the evening made clear is that the pieces are now arriving faster: from pipelines that can diagnose in hours what clinicians missed for decades, from genomic privacy architectures that could enable privacy-preserving research at scale, from capital structures designed to outlast a fund cycle, and from legislative experiments in states that are choosing to move before the federal government does.
But we don’t have the picture on the box, and it will be up to the community to see how all the pieces connect.