Clinical exome and whole-genome sequencing has provided firm diagnoses to many patients with hereditary disorders who were unable to get an answer before. But to ensure the success and cost-effectiveness of genome testing, a thorough review of patients' cases prior to testing appears to be equally important to the post-test analysis, according to findings from a whole-genome sequencing clinical research program for undiagnosed children undertaken by the Geisinger Health System.
In two cases researchers were considering to include in their study, they found that prior testing had already identified a causative mutation, though the patients' doctors had either not received that result or not recognized its significance for diagnosis. By not including those two patients and their families in the whole-genome sequencing study, the researchers avoided $30,000 in sequencing costs and many hours of personnel time for the interpretation and communication of results.
The researchers, led by Janet and Marc Williams at Geisinger's Genomic Medicine Institute in Danville, Pa., published an analysis of the time involved in pre-test case review and counseling for the study in the Journal of Genetic Counseling last month.
The ongoing study is funded by Geisinger, which serves about 2.6 million patients, mostly in Pennsylvania. It aims to use whole-genome sequencing in parent-child trios to discover the cause of disease in children with autism, developmental delay, intellectual disability, or congenital anomalies. The study began two years ago and has enrolled about 65 families with 75 children. So far, it has received sequencing results for 18 families, which are currently under review by an internal work group and a program oversight committee. Starting this spring, families will receive results related to their children's disorder, followed by the return of incidental findings later on. The goal is to complete the study by the end of this year.
Most children were referred to the study by their primary care provider at Geisinger. They had already undergone chromosomal microarray testing for copy number variants, and many had a variety of other diagnostic tests done, with no positive results, both at Geisinger and outside providers. For some patients, the list of potential candidate genes was too long to justify separate tests, so whole-genome sequencing seemed like a better alternative, Janet Williams, director of research genetic counselors at Geisinger's Genomic Medicine Institute, told Clinical Sequencing News.
Prior to enrolling patients, the Geisinger team reviewed their medical history in detail, including all prior diagnostic testing, both inside and outside of the Geisinger, a process that took seven hours per patient on average. During that review, they found two patients with test results that already established a definitive causal diagnosis, which neither the families nor their doctors knew about. As a result, those families were not enrolled in the study.
One patient had a chromosomal microarray test performed by an outside institution, which found a diagnostic copy number variant, but the result was never communicated to the patient's doctor or entered into the electronic health record. For the other patient, a Sanger sequencing-based single gene test had detected a mutation, but the ordering physician had not recognized that it was causative for the disease.
Missing diagnoses because test results are lost or not communicated "probably happens much more frequently than we would like to admit," Williams said. The second case also shows that physicians with no expertise in genetics sometimes miss the significance of a genetic test result, she added.
By not enrolling the two families that already had a diagnosis, the study saved $30,000 in sequencing costs – $5,000 per individual for the two trios – based on Geisinger's contracted rate for whole-genome sequencing with an outside laboratory. And while it took seven hours to review their medical history in detail, it saved "hours of personnel time" for pre-test counseling and consenting as well as interpretation and communication of the test results, the researchers wrote. In addition, they said, prior research has shown that precisely defining patient phenotypes before genomic testing "contributes to decreased time required for the interpretation of results."
Overall, it took between three months and nine months of pre-test work, from when a patient was referred to the study to when their sample was collected. While this amount of time is "beyond that of most clinical activities," the researchers wrote, the "investment is well-spent" in light of possible savings in cost and time later on.
While the number of diagnoses made from whole-genome sequencing for the first 18 families is not known yet, the results look "very promising" and the researchers expect the diagnostic yield to be "in the same ballpark" as that of other clinical exome or whole-genome sequencing studies, on the order of 25 to 30 percent, Williams said.
Families will also meet with the researchers on an annual basis to review their genome information. Over time, the number of definitive diagnoses will likely increase as more research is conducted.
Participants also consented to receiving incidental findings, in accordance with last year's recommendations by the American College of Medical Genetics and Genomics, but those are not the first priority. "We've decided that our primary focus is on returning results that pertain to the referral diagnosis and that we will have separate visits in the future with families regarding any incidental findings," Williams said.
All results reported back to families will go into their electronic health records.
The reason the study decided to use whole-genome sequencing, rather than exome sequencing, which many other programs for undiagnosed diseases have opted for, is the long-term value of having complete genome information. "Geisinger often sees families throughout their lifetime, often generations of a single family," Williams explained, and whole-genome sequencing data, though it might not all be interpretable today, will serve as a resource of information "that can be tapped over and over again." The genome data will be stored in a data warehouse, so it can be queried over time.
The study has been outsourcing sequencing to several service providers, including Knome, Personalis, and the HudsonAlpha Institute for Biotechnology. While Williams declined to comment on specific providers, she said they differed in their ability to meet promised turnaround times, their customer support, and in how they reported results.
Overall, what is important is "finding those laboratories that offer significant support, are available as you wish to talk to them, go over results with you in terms of web or phone, [and] encourage communication back and forth regarding results or reporting options," she said.
On the analysis side at Geisinger, ensuring "the availability of the informatics genetics personnel resources to appropriately manage the information coming back" has been important as well, and Geisinger is in the process of increasing its informatics staff, she said.
These and other lessons learned over the past two years will help inform the recently announced genomic sequencing study that Geisinger is conducting in partnership with Regeneron. The goal of that study is to sequence the genomes or exomes of approximately 100,000 Geisinger patients to find new associations between genes and disease.
The undiagnosed disease study serves as a "small demonstration project" that helps Geisinger to define protocols and processes, Williams said. However, the new study will be "on an entirely different scale" with "new challenges we will need to address in moving forward with this project."
The two studies have different goals – the Geisinger/Regeneron study will start with the genome and look for medically useful information, while the clinical study starts with a disease and uses the genome to look for its cause – but both projects will find it critical to integrate phenotypic information, Williams said. "It's one of the strengths of Geisinger that we have this very deep phenotypic record because of the use of electronic health records throughout the system."