When Rare Becomes
Recognizable

“Thirty years,” says Parth Shah, who is also assistant professor of pathology and laboratory medicine at the Geisel School of Medicine at Dartmouth. “How much are we missing?”

Parth Shah has since identified and treated more patients with PROS, as well as with other rare diseases, through a comprehensive genomic sequencing platform. Today, DHMC is among the first U.S. hospitals to offer comprehensive genomic sequencing to patients, making it a leader in how rare disease diagnoses are done.

A Need for Speed

PROS is one of more than 10,000 known rare diseases, which together affect roughly 30 million Americans, according to the National Organization for Rare Disorders. Half are children, three in 10 of whom won’t see their sixth birthday.

Despite remarkable advances in treating rare diseases over the past several decades, diagnosis remains the field’s greatest bottleneck. These patients still wait an average of five years for an accurate diagnosis, and 60% are misdiagnosed at least once before learning what they have, according to a 2024 analysis published in the European Journal of Human Genetics.

“Rare disease is rare because we don’t test for it,” Parth Shah says.

By definition, a rare disease affects fewer than 200,000 people at any given time. Individually, these conditions are uncommon. Collectively, they represent one of medicine’s greatest diagnostic challenges. Patients often spend years cycling through specialists and inconclusive tests while conditions worsen, organs sustain irreversible damage, and disabilities become permanent.

In parallel, without enough diagnosed patients, clinical trials struggle to enroll enough participants, slowing the development of new therapies. As a result, clinical trials for rare diseases run nearly twice as long as those for common diseases, according to a 2026 white paper by Arya Consulting Partners.

Traditional genetic tests scan for a narrow list of known mutations. If a patient’s mutation is excluded, the test will come back negative, and their diagnostic odyssey continues. Parth Shah’s comprehensive genomic sequencing casts a much wider net.

DHMC runs whole-exome sequencing, which reads all 18,000-plus protein-coding genes, alongside whole-transcriptome sequencing, which captures what those genes are actively doing at any given moment. Think of DNA as a face—the stable, permanent record. The transcriptome is the expression, constantly shifting, reflecting what’s happening in the body right now.

“You can look at a transcriptome and say this cell is normal or abnormal,” Parth Shah explains.

This approach can also help hone treatment approaches. When Parth Shah’s team analyzed a cohort of small cell lung cancer patients treated with tarlatamab, a newly approved immunotherapy, they found that while only 40% of patients responded in clinical trials, their research suggested transcriptome sequencing could predict response with up to 90% accuracy—more than doubling precision in patient selection.

DHMC is among the few academic medical centers in the country running both exome and whole-transcriptome sequencing routinely on patients, versus just in research settings. Parth Shah helped develop the custom transcriptome assay that many large academic medical centers across the country have since adopted.

“What is new at DHMC is that those genes are no longer being shipped to labs across the country to be studied,” Parth Shah says. “Now it’s all in-house. We are cataloging them for every patient who walks in.”

The speed of that in-house process can transform care for many diseases, including for acute myeloid leukemia (AML). Patients with AML once waited weeks in the hospital to learn their prognosis. Now, Parth Shah’s team has identified gene fusions with 100% sensitivity, he says, while uncovering mutations previous testing missed—delivering precise information about a patient’s diagnosis and treatment options in 24 to 36 hours. So far, six AML patients of his have already gotten treatments they would never have received as a result of conventional testing alone.

The sequencing platform also creates institutional memory. As Parth Shah sequences each patient, their genetic profile is cataloged in a HIPAA-compliant database. When physicians at DHMC were looking for patients with a specific rare lung cancer mutation that would allow them to be offered treatment through a clinical trial, Parth Shah queried the database. Five minutes later: “Yep, I have three.”

Newborn Sequencing for Life

But what if Parth Shah’s PROS patient had been diagnosed at birth, long before two strokes and 30 years of suffering?

That possibility drives Nidhi Shah. In addition to her role as a clinical geneticist at DHMC, she is the principal investigator for Dartmouth Health Children’s on the nationwide BabySeq project, which sequences newborns’ entire genomes in the first weeks of life using a separate whole-genome platform designed to detect the thousands of genetic disorders that can affect children from birth onward.

“We sequence right after birth, then return information as it becomes relevant throughout [their] life,” explains Nidhi Shah, who is also an assistant professor of pediatrics at Geisel. Because a genome doesn’t change, one test at birth becomes a lifelong reference: how a toddler will metabolize medications, what cancer risks might emerge later, whether a young adult carries genes they might pass to their future children.

One newborn tested positive for biotin deficiency, a condition that can cause seizures and developmental delays. The preventive treatment is biotin, a daily vitamin. “If you just take the biotin, you don’t ever have seizures,” Nidhi Shah says. The child is now healthy.

ZebraMD is an artificial intelligence (AI) platform that scans electronic health records (EHRs)—previous visits, lab results, family history—alongside wearable data, home monitoring devices, and environmental and geospatial information to identify symptom patterns that match known rare diseases and prompt physicians to consider diagnoses they might otherwise miss. When a match surfaces, AI analyzes the full chart against published clinical guidelines and alerts the treating physician, either in their inbox or at the point of care. The platform also connects patients to relevant clinical trials, support organizations, and financial assistance programs.

“Even highly trained specialists can’t memorize 10,000 different diseases,” Schmolly says.

The University of California, San Francisco (UCSF)— which has extensively reviewed zebraMD’s platform for security and patient data compliance and is where Schmolly went to medical school and first began to develop zebraMD—and NewYork-Presbyterian Weill Cornell Medical Center are piloting the system. In a study published in January 2025 in the Journal of the American Medical Informatics Association, zebraMD’s original algorithm, co-created with collaborators at UCSF, identified porphyria patients with 89–93% accuracy and caught 71% of cases 1.2 years earlier than doctors did. The platform now has prediction algorithms for more than 350 rare diseases.

“If we can diagnose even one person who otherwise wouldn’t have been diagnosed at all, that’s success,” Schmolly says.

As pilots at UCSF and Weill Cornell wrap up, Schmolly launched a standalone browser-based version, usable by both physicians and patients without EHR integration, for a public beta this September aimed specifically at medically underserved communities in northern New England. The standalone tool can connect to patient portals and accept medical records in any format. Later this year, it will connect to national health exchanges, giving it reach across all EHR systems used in rural New Hampshire and Vermont.

The Cost of Diagnosis

Even with better tools, a diagnosis doesn’t always open a door to treatment. Nearly 95% of rare diseases have no FDA-approved therapy, according to the National Center for Advancing Translational Sciences, in large part because it is nearly impossible to enroll enough patients in clinical trials to validate new drugs. That’s the deeper promise these diagnostic tools carry. More diagnoses mean more patients eligible for trials. More trials mean more therapies.

Bereft of treatment options, rare diseases are inherently more costly to manage, up to 10 times more expensive than common conditions, The American Journal of Managed Care reports. And despite impacting a fraction of the population, rare diseases are responsible for nearly half of all direct medical costs, according to a study published in the Orphanet Journal of Rare Diseases.

“Cures only get funded at scale when the economics work,” says Emma Dean, PhD, MS, a health economist and assistant professor at The Dartmouth Institute for Health Policy & Clinical Practice. “And the hidden kingmaker is not the drug company. It’s the payer.”

Insurance companies, Medicare, and Medicaid decide who and what gets covered. Even when treatments exist, they’re not guaranteed to be covered by medical plans. Off-label use, like repurposing a breast cancer drug for PROS, can foil reimbursement. “Payers often require prior authorization, and patients may face higher out-of-pocket costs even when the drug works,” Dean explains.

One treatment for a condition recently added to New Hampshire’s newborn screening panel costs $800,000 per year, says Nidhi Shah. But “if you start treating them before things go wrong,” she adds, “you don’t need to treat them for as long.”

Seeing the Stripes

Parth Shah’s PROS patient has recalibrated his perspective. “It changed my whole point on how I wanted to deliver care,” he says. “I wouldn’t be surprised if in the next four to five years almost everybody’s being sequenced. We’re doing it now, bench to bedside, in real time.”

Already, DHMC tests every leukemia patient in its care, and now completes rapid genomic testing for critically ill babies in neonatal intensive care. The hospital also conducts molecular autopsies for New Hampshire: sequencing children who die with no known cause, providing both closure for families and valuable genomic data.

For patients born tomorrow with biotin deficiency, with PROS, or with any condition in that fortunate 5% with available treatments, they may get something previous generations never had: answers before it’s too late.

The diseases won’t disappear. But the diagnostic odyssey, the years of inconclusive tests and delays that rob people of money, comfort, and often their lives—that is already changing for the better.