SAN FRANCISCO — Among the many challenges of implementing a clinical sequencing pipeline for diagnostics or to guide treatment, the actual sequencing step is quickly becoming a minor consideration, according to a number of speakers at Cambridge Healthtech Institute's Clinical Genome Conference here this week.
Several researchers and clinicians who have either already implemented or are looking to implement clinical sequencing pipelines said that the major challenges are the pre- and post- sequencing aspects such as patient counseling, analysis and interpretation, and reimbursement.
"The sequencing is almost the most trivial part," Mark Boguski, associate director of pathology at the center for biomedical informatics at Harvard Medical School told Clinical Sequencing News here this week. Boguski is part of a team that is working to implement a whole-genome sequencing pipeline within Beth Israel Deaconess Medical Center.
As a pilot, the team plans to sequence 50 patients each from six different cancer etiologies, beginning with breast cancer. The other cancers are still being decided upon but will likely include melanoma, lung, and prostate cancer.
Sequencing will be outsourced to an as-yet-unidentified vendor that can do the sequencing in a CLIA-validated setting.
The team, led by Peter Tonellato, who founded the Laboratory for Personalized Medicine at Harvard, is currently working on building relationships with surgeons. As part of this effort, Tonellato has been sitting in on Beth Israel's tumor board meetings to figure out how best to deliver genomic sequence data.
The multidisciplinary tumor board at Beth Israel reviews four to six new cases per week, Tonellato explained in a presentation at the conference. The board reviews a host of data for each patient, including family history, imaging data, and information from the patient's primary care physician. They "have to digest that, incorporate it into their thinking, and figure out how to approach the patient about the best treatment option," Tonellato said.
Figuring out how to add whole-genome information to that already complicated process is tricky. Tonellato said that the end product will be a report that could be presented to the patient's physician on an iPad or other tablet device.
He acknowledged that the introduction of the whole-genome sequencing effort is further complicated by the availability of a number of single-gene and targeted sequencing-based panels for breast cancer — including Myriad's BRACAnalysis, Washington University's targeted gene panel, and Asuragen's AsuraSeq panel — as well as a host of non-sequencing-based tests that screen for breast cancer-related variants. In total, he said, there are some 37 multianalyte tests that test for variants associated with breast cancer that are either predictive of risk, subtype, or drug response.
For whole-genome sequencing to be cost-effective it can't be "layered on all these other tests," Tonellato said. To garner both clinician and insurance company acceptance, it will have to be done as a first-line test, meaning it will have to be accurate, quick, and as simple as possible for clinicians to interpret.
It will also have to be cost-effective. Tonellato noted that if a physician conducts around 10 different multianalyte tests, costs will easily run upwards of $19,000. Sequencing a tumor/normal pair, meantime, runs around $25,000 and "that gives you everything," Tonellato said. "This demonstrates that we're in the ballpark."
He said the team plans to initiate sequencing for the pilot project this year. Boguski added that eventually the program could be turned into a clinical service.
While the Beth Israel team is jumping right into whole-genome sequencing of cancer patients, the Fox Chase Cancer Center, which originally planned to launch a whole-exome and transcriptome sequencing service for cancer patients, said last year that it planned to scale back to do targeted sequencing on Life Technologies' Ion Torrent PGM (CSN 12/21/2011).
At this week's conference, executive director Jeff Boyd gave an update on the center's activities.
He explained that bioinformatics challenges, turnaround time, reimbursement challenges, and cost of whole-genome sequencing — coupled with the fact that the majority of the information was not immediately clinically useful — led the center to switch to a targeted approach.
Using the Ion AmpliSeq Cancer panel, costs have now been reduced to $3,500 per patient, as opposed to the $15,000 it would have cost to do exome and transcriptome sequencing on the SOLiD machine as originally intended.
Boyd said that the center plans to begin the sequencing program this year.
From Pilot to Clinical Service
The Medical College of Wisconsin, which has been a leader in using whole-genome sequencing for the clinical diagnosis of rare diseases, has now expanded its initial pilot program to sequence the genomes of children with rare, undiagnosed diseases into a full clinical service, said Howard Jacob, director of the Human and Molecular Genetics Center at the Medical College of Wisconsin.
During the pilot phase, 60 cases were submitted for approval and 18 were accepted. The 18 were chosen based on their likelihood of being a Mendelian disease, whether all other diagnostic options had been exhausted, and whether a diagnosis was likely to improve the patient's outcome.
Now that the pilot phase has been completed, Jacob said that the vast majority of the submitted cases that had not previously been approved have now been approved. Additionally, one physician, after not immediately receiving approval from the MCW program, had the patient's exome sequenced at Baylor College of Medicine, said Jacob, demonstrating that this is a service that physicians want. Physicians are saying, "get out of my way," Jacob said.
The program has also made significant headway on the reimbursement front. In February 2011, the team received a letter from an unnamed insurance company stating it would reimburse for whole-genome sequencing as a first-line test in cases where it would be more cost-effective than conducting a series of single-gene tests. Since then, Jacob said four out of 10 cases that have been submitted for reimbursement have received it.
Jacob said that within the pilot program, one of the hardest lessons has been deciding when whole-genome sequencing has failed to find an answer. He described one case in which a patient presented with fevers. Initial analysis of whole-genome sequencing data failed to identify a cause. The team has since been following up on the case, evaluating candidate variants and genes, so far to no avail. "We need criteria to determine when ... we fail," he said.
Another challenge has been providing counseling related to testing results. Currently, families receive between six and 10 hours of counseling spread out over three different sessions with about 18 percent of the time devoted to discussing the types of results the family wants returned. Going forward, this will not be sustainable, Jacob said.
Moving forward, he said the center's goals are to figure out how to link the genomic data to patients' electronic medical records as well as to biomedical literature; to figure out the best way to deliver a genomic report to physicians in a meaningful way; and to do whole-genome sequencing, de novo assembly, and analysis and clinical annotation all in one day.