A child came to Vanderbilt University Medical Center in Nashville, Tennessee earlier this year with an illness that resembled progeria, a rare genetic disorder in which children manifest symptoms of old age. Her symptoms were more severe, though, and the test for a mutation in the LMNA gene, which would confirm a progeria diagnosis, came back negative. In an effort to identify this perplexing condition, she was accepted to Vanderbilt as a patient in the Undiagnosed Diseases Network (UDN), a collaboration of scientists and clinicians founded by the National Institutes of Health (NIH) to try to solve the most mysterious medical cases. DNA sequencing of this child revealed a mutation in ZMPSTE24, a gene that modifies the LMNA protein. “That just immediately looked like the diagnosis,” said John Phillips III, MD, David T. Karzon professor of pediatrics and co-principal investigator at the Vanderbilt Center for Undiagnosed Diseases. 

Through the UDN, the clinicians learned there was a mouse model of the ZMPSTE24 mutation and that the mouse had responded to therapy with two medications that happen to be Food and Drug Administration-approved. The patient has just started treatment with both drugs. “That’s a long way from telling someone, ‘Your child has a very serious disease and we didn’t know what it was,’” Phillips said. 

While success stories like this abound, critics argue that the diagnosis rate in the UDN is low—around 25%—and that diagnosis doesn’t help when there is no known treatment. “The goal seemed to us to be diagnosis for the sake of diagnosis itself,” said Jonathan Brower, MD, a plastic surgery resident at Brown University’s Alpert Medical School in Providence, Rhode Island who co-authored a letter critiquing the program (JAMA 2016;315:1903–4). “Diagnosis of these patients is valuable only if it results in a treatment, which is not offered by the UDN. That was where we saw the disconnect.” 

Clinicians involved in the UDN see it differently. Patients who aren’t diagnosed sometimes learn about the mechanism of their disorder, which can help their doctors treat symptoms, said Anastasia L. Wise, PhD, program director in the division of genomic medicine at the National Human Genome Research Institute in Bethesda, Maryland, and UDN co-coordinator. “Even if there’s not necessarily a treatment for many patients, just ending the diagnostic odyssey and knowing what they’re dealing with can be helpful,” she said. 

Growth and Progress of the UDN

The precursor to the UDN, NIH’s Undiagnosed Diseases Program (UDP), started in 2008. Patients with mysterious ailments could apply to spend a week at the NIH Clinical Center in Bethesda, Maryland, and have a team of specialists use the latest research techniques—often whole genome or whole exome sequencing—to try to find a diagnosis. 

As of May 2014 the UDP had received more than 3,000 applications and accepted 750 patients. Between 25% to 50% of cases had resulted in a diagnosis, and some of those diagnoses were previously undescribed. That     year NIH announced it would expand the program, dedicating $117 million over 5 years to fund the UDP plus six other clinical sites around the United States to be known as the UDN. In addition to the network’s seven clinical sites, it now has two DNA sequencing core facilities, a metabolomics core facility, a model organisms screening center, and a central biorepository. UDN began accepting patients in September 2015. 

UDN’s diagnosis rate is not available yet because data from the first year is under review, Wise said, but she expects it to be similar to the UDP. She went on to explain that a relatively low rate is to be expected in these programs, because the UDN patients have already had exhaustive clinical evaluation with multiple specialists before enrolling. Also, some diseases are so rare that the suspected diagnosis cannot be confirmed until a second case is reported or an animal model proves the mechanism of action. 

More than 900 patients applied to the UDN in its first year and about 100 have had clinical evaluations. Of about 33 patients who have completed the UDN program at Vanderbilt, seven have received full diagnoses so far, said John Newman, MD, Elsa S. Hanigan professor of pulmonary medicine and co-principal investigator, adding that these are “completely new diagnoses that have baffled physicians for any number of years.” At least five of those diagnosed cases are treatable, he said, “and we’ve helped them get on new treatment, so there really are stunning successes.” 

The Role of Genomic Sequencing and Data Sharing

Whole genome and whole exome sequencing have been key to many of the diagnoses made through the UDP and UDN. An analysis of the first 2 years of the UDP found that 12 diagnoses were made solely based on clinical findings, while 19 were based on molecular findings (Genet Med 2012;14:51–9). 

Back in 2008, patients typically had to enroll in a program like the UDP to have access to whole genome or whole exome sequencing. Now, as these techniques are increasingly available clinically, patients sometimes have whole exome sequencing before they apply to the UDN. Some medical centers are even creating their own undiagnosed diseases programs outside of the NIH network. 

Wendy K. Chung, MD, PhD, directs the DISCOVER Program at Columbia University in New York City, which uses whole exome and whole genome sequencing for making difficult diagnoses independent of a Duke Medicine and Columbia University Medical Center UDN site. “This used to take an army of researchers,” said Chung, who is the Herbert Irving associate professor of pediatrics. “Because the science has become easier, it has become integrated into clinical care.” 

She predicts that, much like chromosomal microarrays, whole genome and whole exome sequencing will eventually become a commodity, available at any hospital. “Instead of this being at the end of a very long diagnostic journey, it is evolving into a frontline test,” Chung said. “I see yields as high as 50 percent with frontline sequencing, which is great. You can just do one test and save significant cost, time, and frustration.” 

Sequencing is not the only factor behind improved diagnosis of mysterious diseases—data sharing has also made its mark. UDN sites follow common protocols and database formats so data can be analyzed across sites and with partners internationally. Also, patients may opt to create a website through the UDN to connect with those searching for a related condition. 

Patients who leave the DISCOVER program without a diagnosis sometimes find an answer later as more data becomes available, Chung said. As an example, she cited one recent undiagnosed patient who became pregnant, prompting her mother to call the DISCOVER program to find out if there was any new information on the case. 

“We sort of panicked and looked at it again and, sure enough, there had been a publication about the disorder that wasn’t there when we first did the analysis six months ago,” Chung said. “It was literally only six months ago. And thank goodness, we actually now had an answer for her mom. We performed prenatal testing and it came full circle.” 

Chung believes that everyone in medicine, especially laboratorians, can help with undiagnosed diseases by sharing information. “I think the more that we share data and get collectively smarter, the better we can do for patients around the world,” she said. 

Wise also urged laboratorians to stay alert for cases that seem unusual or difficult to diagnose as potential candidates for the UDN program. Labs may call the helpline at 1-844-746-4836 or email UDN@hms.harvard.edu to discuss a case, refer a patient, or discuss research collaboration. “We’re a resource that is available,” Wise emphasized. 

Julie Kirkwood is a freelance writer who lives in Rochester, New York. +Email: julkirkwood@gmail.com