The New York State Department of Health this summer released guidelines laying out the requirements for the development and validation of next-generation sequencing-based cancer diagnostic tests in New York laboratories, its first set of recommendations pertaining specifically to NGS.
The guidelines, which cover somatic variant detection for molecular oncology testing, were issued in August by the department's Clinical Laboratory Evaluation Program, or CLEP, which monitors the quality of testing conducted by clinical laboratories and blood banks in the state as well as out-of-state labs that accept clinical samples from New York.
The document complements existing department guidelines for molecular oncology tests, following "the same basic principles for validating most other complex molecular diagnostics procedures," and are expected to evolve "as the field matures and gains experience," according to the document.
Erasmus Schneider, director of CLEP's oncology section at the department's Wadsworth Center in Albany, told Clinical Sequencing News that the need for the guidelines arose because next-gen sequencing is making its way into clinical diagnostics and the department had received requests from laboratories wanting to know how to validate their tests in order to gain New York State approval.
Input for the guidelines came from meetings Schneider and his colleagues attended where the clinical use of NGS was discussed and from feedback from interested parties on a draft document.
NGS-based molecular oncology testing differs from most other tests in this area because it interrogates panels of tens to hundreds of mutations in parallel rather than individually. Schneider said the only other high-throughput genotyping assay currently used for oncology mutation detection is the Sequenom MassArray assay, but that this "has not been widely adopted" by clinical laboratories in the state, in part because of the big investment required. "I think most people are much more familiar with sequencing-based assays," he said.
So far, the majority of clinical laboratories offering somatic mutation detection for cancer in New York have expressed an interest in NGS-based assays, Schneider said. Overall, 104 labs in the state hold a permit for molecular oncology testing, but not all of them offer somatic mutation detection, he added.
No NGS-based oncology tests have been approved yet, but several applications are pending.
Cyrus Hedvat, director of the diagnostic molecular pathology lab at Memorial Sloan-Kettering Cancer Center, said that the guidelines provide "a solid starting point for labs in the process of developing clinical next-generation sequencing-based tests."
His lab is currently developing an assay for solid tumors that runs on Illumina's MiSeq and uses the firm's TruSeq Amplicon Cancer Panel, which includes 212 amplicons from 48 cancer genes (CSN 10/17/2012).
Among other requirements, New York State's guidelines call for a laboratory to define the minimum coverage required for calling a variant with high confidence and for calling an amplicon normal. According to Hedvat, this is similar to guidelines published by the College of American Pathologists this summer (CSN 8/1/2012), but New York offers "no specific guidance on the coverage depth required or the minimum allelic frequency to report a positive result."
Whether or not a positive result needs to be confirmed by an independent reference method depends on how thoroughly it has been previously validated, according to the guidelines. For example, if at least 10 positive samples for a specific amplicon have been verified independently, such confirmation is no longer required going forward.
"This will be a significant advantage for common mutations in genes with hotspots such as EGFR, KRAS, and BRAF, but more problematic for a tumor suppressor such as P53," Hedvat said.
He also said that the guidelines are unclear about how many positive controls need to be included in each run, and whether they have to cover mutations from all regions tested. The guidelines do state that a positive control has to be included during validation "and periodically thereafter," and suggest using a control containing "multiple known somatic alterations of each kind to be detected."
With regard to reports, the guidelines state that they should include all detected somatic variants, even those of unknown significance. "This could result in a long list of variants that could be difficult for a clinician to interpret," Hedvat said, but on the other hand, if later publications make such variants clinically interpretable, the lab would not need to issue an updated report.
The guidelines do not list specific requirement for data analysis, he pointed out, other than saying that the initial validation should be done with a single version of all analysis software, and that software updates affecting key parameters must be revalidated.
In the meantime, in response to feedback from stakeholders, CLEP is revising the guidelines in order to "clarify a few issues and make it more explicit what we're looking for," Schneider said, adding that the update is scheduled to be completed by the end of November.
He also said that another section of CLEP is currently working on guidelines for NGS-based germline mutation testing but did not provide a timeline.
While New York State's guidelines may be "a little more specific of what labs need to do" than existing guidelines, Schneider said, he does not believe the differences are all that great. " Validation of a test has certain principles that must be followed, and it doesn't really matter what the test is," he said. "Some details are obviously technology-specific, but I don't think we are that much different from what others have put out."