Genomics Can Help Diagnose Rare Pediatric Diseases

Genomic analysis can lead to a diagnosis in a substantial proportion of children with rare, undiagnosed diseases, according to results from the Deciphering Developmental Disorders study.

Of 13,449 kids included in the multicenter study conducted in the U.K. and Ireland, 41% received a diagnosis using exome sequencing and microarray analyses, reported Caroline Wright, PhD, of the University of Exeter in England, and colleagues in the New England Journal of Medicine.

While the paper does not discuss whether those children went on to treatment, the authors stated that the study shows "how the fusion of clinical expertise, genomic science, and bioinformatics can drive diagnosis and discovery in families in which standard, phenotypically driven diagnostic approaches have failed."

In an accompanying editorial, Jennifer Posey, MD, PhD, and James Lupski, MD, PhD, of Texas Children's Hospital in Houston, agreed that the findings "show how a genome-driven approach in combination with detailed clinical phenotyping improves diagnostic yield over that with the previous standard of care."

"Although much work remains to be done in improving genomic diagnostics and in understanding disease biology, it is clear that clinical genomics in medical practice is making a difference for patients with rare genetic disorders and their families and physicians," Posey and Lupski wrote.

Stephen Kingsmore, MD, DSc, the president and CEO of Rady Children's Institute for Genomic Medicine in San Diego, who wasn't involved in the study, agreed that the paper "cements the idea that this approach, decoding the human genome, is very effective for diagnosing children with diseases that might be genetic ... and ought to be something that happens frequently in such children early in their disease course."

He noted that the methods used in the study, exome sequencing and microarray analyses, were novel when the study started recruiting a decade ago but are somewhat antiquated by now. Today, whole-genome sequencing that can be done quickly and inexpensively would be used instead.

"For those of us in the field, it doesn't really say anything that new, and uses an old technology," he said. "I think we already knew all of this, but tragically, we haven't communicated it as effectively to our colleagues."

That may have something to do with the notion that medical genetics has traditionally been a "naming" profession that has unfortunately earned the heuristic, "diagnose and adios."

"Medical practitioners have never gotten past that idea," Kingsmore said. "But now, most of the time when you make a diagnosis, it changes the management of the child."

For their study, Wright and colleagues included 13,449 children with severe, difficult-to-diagnose developmental disorders -- along with their families -- who were recruited from April 2011 through April 2015 at 24 regional genetics services in the U.K. and Ireland. The median age was 7 at the time of recruitment, with 42% being female and 16% being of non-European ancestry. More than 200 clinicians were involved in the project.

Children had neurodevelopmental disorders, congenital anomalies, abnormal growth measurements, dysmorphic features, unusual behavioral phenotypes, and genetic disorders with an unknown molecular basis.

Thus far, Wright and colleagues have identified a total of 19,285 potentially pathogenic sequence and structural variants among participants.

Among the children who did receive a diagnosis, 76% had a pathogenic de novo variant, and another 22% had variants of uncertain significance in genes that were strongly linked to monogenic developmental disorders.

In assessing factors contributing to the likelihood of getting a diagnosis, they found the strongest to be being part of a parent-offspring trio (OR 4.7, 95% CI 4.16-5.31). Other factors tied to a greater chance of diagnosis included the presence of severe intellectual disability or developmental delay (OR 2.41, 95% CI 2.10-2.76) and being the only affected family member (OR 1.74, 95% CI 1.57-1.92), among other factors.

Patients were less likely to get a diagnosis if they were born prematurely (OR 0.39, 95% CI 0.22-0.68), had in-utero exposure to antiepileptic drugs (OR 0.44, 95% CI 0.29-0.67), had mothers with diabetes (OR 0.52, 95% CI 0.41-0.67), or were of African ancestry (OR 0.51, 95% CI 0.31-0.78).

The researchers and editorialists noted that the latter finding points to the need to increase the participation of underrepresented groups in research.

The study was limited by the genotyping approach used, as it could not assay most non-coding variants and could not detect all complex structural variants or tissue-specific mosaicism. Also, their analytic approach was insensitive to incomplete penetrance, they noted.

In their editorial, Posey and Lupski said that when it comes to genome sequencing in diagnosing disease, "'best practices' have often been more readily defined according to the requirements of healthcare systems and payers rather than the findings from any true, broad assessment of diagnostic utility."

Kingsmore added that there's "geographic inequity" in terms of access to genome sequencing for diagnosing rare disease. Not every family lives near a large children's hospital that's equipped to do genomic analysis.

"There's no reason why we can't change this," Kingsmore said, noting that with the COVID-related increase in telehealth, it's possible for families to send in blood or spit samples for sequencing. "We can now decode a genome in a day and we can do it for $100. It's not an economic problem anymore. It's a matter of willpower."

Comment