NEW YORK (GenomeWeb) – Improving diagnosis for children with rare diseases is one of the major goals of the recently-founded Children's National Rare Disease Institute (CNRDI) in Washington, DC, and both the structured use of clinical information and targeted genomic testing will likely help with that.
The institute, which was announced last month, is part of Children's National Health System and evolved from its existing Division of Genetics and Metabolism. It plans to focus exclusively on advancing the care and treatment of children and adults with rare genetic diseases.
"We want to try to pull together the best diagnostic practices and the best clinical practices, but continually improve those," said Marshall Summar, chief of genetics and metabolism at Children's National and the leader of CNRDI, in an interview.
CNRDI currently has 13 clinical geneticists, 12 genetic counselors, and two biochemical nutritionists, and plans to add more staff, in particular in process management and IT. The institute expects to see more than 8,500 patients per year, based on the number of clinical visits to the Genetics and Metabolism clinic.
Having worked in clinical genetics for more than 25 years, Summar said, he noticed that the field of rare or orphan diseases has begun to develop into its own field of medicine. In the US, rare diseases are defined as disorders that affect fewer than 200,000 individuals. Almost 7,000 rare diseases — about 80 percent of them genetic in origin — affect more than 25 million Americans.
"They are a group of patients with complex conditions but very small numbers of patients [for each disease] because they tend to have extreme genetic changes that are not that common in the general population," Summar explained. "The Rare Disease Institute evolved out of a need for a home for these types of patients, both from a diagnostic standpoint and also from a long-term care model [point of view] after they are diagnosed."
The first step is to come up with an accurate and, whenever possible, molecular diagnosis for patients, he said. "The molecular testing is the beginning of the process. Then you have to figure out: What are you going to do?"
Clinical information to guide diagnostics
Overall, the diagnosis of rare diseases tends to take several years. In the past, the average time to diagnosis was about seven years, Summar said, but it has recently improved to about four to five years across the field. While metabolic diseases are often faster to diagnose because they have biochemical markers, other disorders still send patients on a "diagnostic odyssey" that the institute hopes to shorten.
One approach CNRDI plans to take is to structure a patient's clinical information at the front end and use it in a systematic fashion to help decide what diagnostic testing is going to be most successful.
"Right now, we spend a lot of time bringing patients in, getting a rudimentary clinical history, then we will send out a wide swath of genetic testing, and then we try to retrofit that to their condition," Summar said. "And as you know, there is a lot of interpretation in that, [and] there is a high risk of false negatives as well as false positives."
This could potentially be avoided by using clinical information to guide the testing upfront. For example, a patient may present with a delay in development and seizures, and information about additional symptoms could help narrow the testing to a specific panel of genes.
"We've been working with this concept already for developing algorithms and flows for doing diagnosis," Summar said, adding that work on this is in the design phase and focuses on clusters of symptoms.
"Right now, we're picking what our largest group of presentations are and developing approaches around those, but basing them around the symptom complex rather than the final diagnosis," he explained. Until now, he said, "a lot of times, we base the [diagnostic] workup around what the answer would be if we knew the answer. But since we don't know the answer, we're taking a step back, and instead are trying to group these [patients] around presentation complexes."
On a research basis, the institute has also started to work with facial recognition technology to pre-screen patients with suspected chromosomal disorders, like Down syndrome or DiGeorge syndrome. That technology was developed by researchers at the Sheikh Zayed Institute for Pediatric Surgical Innovation, which is part of Children's National, and its use in diverse populations was recently described in the American Journal of Human Genetics. Summar said the technology is currently undergoing validation for FDA approval before it will be used clinically. Screening patients with facial recognition technology would only cost a few dollars and could predict whether patients might benefit from FISH or microarray-based testing, he said, which could cost several thousand dollars.
Overall, rare diseases fall into two broad categories, Summar said — those caused by inborn errors of metabolism, which have historically been diagnosed by biochemical testing, and those with other causes. Almost all patients in the latter category receive molecular testing these days, he said, including chromosomal microarrays, single-gene tests, gene panels, and exomes, and testing is often tiered. But even for metabolic diseases, molecular genetic testing is often performed in addition, after a biochemical diagnosis has been made, because it may allow doctors to predict the clinical phenotype and to test other family members for the mutation.
Digital panels
Children's National does not favor diagnostic whole-exome or genome sequencing, Summar said, because the chance of false positives is high. "As a friend of mine once put it, if you do a whole exome, the answer is always 'yes', you just don't know if it's the right 'yes' or not, because you always find changes and variations whenever you do these things," he said. In addition, much of the cost of genetic testing is related to the interpretation, and "the more genes you're looking t, the more time you have to spend on the interpretation, backchecking, things like that."
For that reason, Children's National is using so-called digital panels. The test currently involves sequencing all Mendelian disease genes – Summar said this will soon be replaced with the entire exome – but analyzes only pre-defined panels, depending on the patient's condition, for example, genes involved in seizures or non-dysmorphic developmental delay.
Genes can easily be added to or dropped from the panels, and if no answer is found, the search can be widened to other genes without the need to draw a new blood sample, extract new DNA, or run another test.
"One of the things we want to do at the Rare Disease Institute is make that even more robust," Summar said. "At the end of the day, the idea is that you would like to have a molecular diagnosis. For most things now, you can get pretty close to that. There is still a group that you can't. What we want to do is make sure that we are using and targeting that testing as efficiently as we can."
That also means keeping the cost of testing down. "One of the biggest barriers right now is getting [insurance] authorization for testing," Summar said. "We spend a tremendous amount of time as clinicians getting authorization." The institute will try to change that by providing targeted testing in order to convince payers that "we're trying to be good stewards with the money, we are not just throwing a wide net but a more focused net."
To help patients who remain undiagnosed, Children's National has teamed up with Regeneron Pharmaceuticals. Last fall, the partners announced a three-year exome sequencing study of patients with undiagnosed rare diseases that aims to enroll 3,000 pediatric patients and their families in the first year. Regeneron also plans to use the results in its drug development efforts.
Improving treatment and care
Improving the diagnostic process, though, is only part of CNRDI's mission. "Making the diagnosis is great, but then you've got to do something about it," Summar said.
At the moment, care for patients with rare diseases is often fragmented across different medical specialties. To change that, the institute has partnered with the National Organization for Rare Disorders (NORD), a charity that represents more than 250 patient advocacy groups. The idea is to compile what are considered the best standards of care for patients with a certain disease, publish open-source treatment protocols, track patients' outcomes in longitudinal studies, and improve their care over time, said Summar, who chairs NORD's board of directors.
The model for this approach is how the treatment for cystic fibrosis has been improved in the past . "Cystic fibrosis [researchers and clinicians] pulled together what was thought to be the best treatments at the time but then continually evaluated that and evolved that model such that now, they almost tripled the life expectancy of a patient with cystic fibrosis," Summar said.
NORD, in partnership with and funded by a grant from the FDA, has already developed a natural history studies program for a variety rare diseases, building registries that are based on a common platform where patient data can be deposited. So far, it has set up 30 registries and the program is growing rapidly, he said.
The registries will be of use to both patients and researchers because they may uncover previously unknown information and connections. "As you know, typically, for most rare diseases, there is an initial publication around a small group of patients, and then there is not often very much after that," Summar explained. "What we need to do now is add depth to that, so we know what happens to patients over time, what the co-morbidities are that we didn't expect."
"From the patient's perspective, it provides sort of a safe place to go," he said. "A place where people know what their condition is, have an idea of what the best standards are currently for caring for that, and can work with their other healthcare providers to organize that and get that care set up."
Molecular diagnostic results are collected in the registries when available and might be helpful in several ways. On the one hand, they can help uncover disease mechanisms, based on genes that are affected, and suggest targets for new therapies, or existing therapies that might be effective. On the other hand, they can predict what other symptoms a patient may develop, for example, if they have a mutation in a gene that may affect several organ systems. This can help develop a monitoring and care plan for the patient, Summar said.
CNRDI already has more than 25 ongoing investigator-sponsored clinical trials for therapeutic development, as well as industry-sponsored ones. It is also conducting research into new treatments for patients with inborn errors of metabolism, based on their genetic variants.
"But we've got so much work to do," Summar said. "We have about 600 treatments for 7,000 diseases. So, the more people we can get involved in this, the better."