NEW YORK (GenomeWeb News) – As translational medicine and genetic- and genomic-based diagnostic tests proliferate, some core laboratories that usually support basic research are looking to move into clinical testing. While such a move may make sense for some labs, there are a number of issues that should be considered and addressed, and understanding how such clinical labs are regulated is chief among them.
In the US, such clinical labs are regulated under the Clinical Laboratory Improvement Amendments of 1988, or CLIA. Certification and accreditation under CLIA aims to ensure that clinical labs meet certain quality standards, and many of the requirements are not necessarily something a research-focused lab would be doing as stringently.
"There are a lot academic core labs right now that are investigating what they need to do in order to become CLIA certified and essentially become centers for clinical diagnostic testing," said George Grills, the director of operations of core facilities at Cornell University's Life Sciences Core Laboratories Center. "This is actually something that is so widespread that there has been a lot of conversation about establishing best practice in terms of core labs that want to go into this arena and establishing a forum for communication for them as well."
Further, he noted that professional groups, like the Association of Biomolecular Research Facilities, where he serves as an executive board member, are interested in helping put together such guidelines.
Under CLIA regulations labs must compile and maintain a lot of documentation of the lab's processes, and diagnostic tests and the instruments they run on also have to meet certain specifications. Further, the results of such clinical tests also must be interpreted and reported to physicians.
In addition to the effort that such documentation entails, labs considering such a move to clinical testing should also be aware of what the potential demand is for the tests they plan to offer.
What is Required Under CLIA?
Under CLIA the US Centers for Medicare and Medicaid Services oversees laboratory testing of people, excluding research testing, to ensure that such testing is held to a certain standard. CMS estimates that CLIA regulations cover about 225,000 labs.
To open a clinical lab, the lab must first be certified by the state in which it is located. But then, it must also be accredited by certain approved organizations that have requirements stringent enough to meet CLIA requirements.
"Every state has some sort of process for labs to become certified to be able to open their doors as a clinical laboratory," Heidi Rehm, who is the director of the laboratory for molecular medicine at Partners Healthcare's Center for Personalized Genetic Medicine, told GenomeWeb Daily News. "My lab first got inspected by the Department of Public Health in Massachusetts, but then CLIA says you must be accredited on an ongoing basis by an approved organization."
Those accrediting organizations include the College of American Pathologists, the Joint Commission, and COLA, among others. In addition, two US states, New York and Washington, are accrediting bodies, as those states have higher levels of oversight.
The state inspection, Rehm said, included inspectors from the fire and public safety departments. Then, lab procedures and tests have to be validated and documented as part of the accreditation process.
Everything in a clinical lab must be documented. "It's really a huge process in document and protocol maintenance," Rehm said.
Mehdi Keddache, an assistant professor working in the Genetic Variation and Gene Discovery Core at Cincinnati Children's Hospital Medical Center, said that he conceptualizes the whole process as the master notebook of his lab. This overarching book lists where all the information about the lab can be found — what drawer it is in, what shelf it is on.
"When the inspector comes, the first thing that … she has done is looked through this book and then say, 'Alright, today I would like to see the record of temperatures.' So she opens the book, it's supposed to be in that drawer, so she pulls the drawer, and, yes, she finds the records," Keddache said. "It's a very procedural process."
Not only do refrigerator temperatures have to be logged, but so does nearly every detail of the lab, its staff, its instruments, and its procedures. "Everything that you do that touches clinical samples, there has to be a standard operating procedure for," Rehm said. There has to be a procedure for accepting samples, preventive maintenance of instruments must be recorded, training of staff has to be documented and kept up, and tests must follow validation protocols.
"A lot of things are very tightly regulated in a CLIA setting [that] you take for granted in a research setting," added Ulrich Broeckel, a professor at the Medical College of Wisconsin who oversees the Individualized Medicine Institute there.
Rehm noted that inspections often include a tracer analysis in which the inspector picks one case and then examines all the related documentation. "They follow everything that touches that case," she said. In such an instance, the inspector would want to see information including the training documents for the person who accepted the sample, the preventive maintenance for the machine that the sample was run on, and the validation for the test that was run.
"You are going through an extraordinary sort of process of quality assurance and quality control in order to maintain that as a clinical lab," she said.
Getting all that together can take time. Cornell's Grills recalled that when he was at Harvard the certification process for his lab took about a year; Broeckel said the process for his lab took about the same amount of time.
"The hardest part is the first time. The first time takes months because you are not used to performing clinical tests," Keddache told GWDN. After that, though, he said it gets a bit easier. "Once you have it, the year after year is like an incremental backup," he said. "The first time that you back up your entire hard drive, it takes forever, but then if you do it day by day, there is only so much work that you are doing in one day and there is only a little bit."
For labs just beginning to seek certification and accreditation, Keddache recommended starting off small and then growing to include more and more tests. Otherwise, he said, it's a lot to bite off at once in terms of all the necessary documentation and paperwork.
Testing, Testing
Tests offered by clinical labs are, naturally, subject to similar requirements of quality assurance and quality control. Clinical labs can offer commercial tests, but, as Rehm pointed out, many genetic and genomic labs offer laboratory-developed tests. For example, her lab offers a number of tests for inherited diseases and cancers.
For such LDTs, the development and validation of the tests are handled in house. "So I develop a method, and then I need a bunch of positive and negative controls that I run through it, and I can prove that it works and it gives accurate results, and I have to have a validation report that shows that," she said.
Developing such tests take time as they are put through stringent paces to ensure their robustness and reproducibility over time, on different machines, and with different technicians running the tests, among other variables. "This is one of the reasons that clinical tests are more expensive than research tests, because we end up having to do all of this," Rehm added.
Once the lab is set up and running, it's easier to develop and bring in new tests, especially from translational research labs. Broeckel's lab at Wisconsin offers mainly array-based tests for chromosomal abnormalities, though it is now moving toward more sequencing-based tests as well.
The research side of his lab is set up to follow the CLIA process, he added, to make the transition to clinical testing a bit smoother. "We basically already built the infrastructure for the research side to be able to quickly translate this into clinical applications," he said. "And so we are working on some targeted sequencing, maybe sequence panels, gene panels, using sequencing." Additionally, his lab is now familiar with the sort of documentation that is needed.
Similarly, Rehm noted that she keeps an eye on the technologies being used by the basic research core labs located on her floor, and sometimes she moves to bring them under the auspices of her lab by taking over the machine's preventive maintenance and maintaining the training records of the technician who runs the machine.
"We'll develop a test on that instrument and be able to offer it clinically and that means we didn't have to buy a new instrument, we already had trained staff. It made that transition much easier," she said, adding that that's how much of the technology in her lab — from Sanger sequencing to next-gen sequencing — has been added.
Being one of the first to bring in a new technology can introduce other wrinkles, as there may not be standardized guidelines on how to validate such instruments. In that case, there are different ways to interpret the regulations about standards, Keddache said.
One interpretation, he said, is quite stringent. For example, if a lab was one of the first to bring a next-gen sequencer in to the clinical space, it could always run the same set of PCR amplicons whose mutations were known from Sanger sequencing. "[It is] a foolproof validation because every time that you run the instrument, there's a sample that you run from the DNA to the mutation state, and it should give you the same result every time, so you know that you have your positive control," he said. "But, it is very costly."
The other approach would be to rely on other aspects of the testing process for validation. For example, the presence or absence of bands on the gel, or the presence or absence of PCR products during the library-prep step could be used as evidence toward validation. That, plus having evidence that the sequencer has always given high-quality base calls, could suffice as validation.
"It's really a matter of how you feel that your interpretation of the standard is respected and then the inspector will tell you, 'No, that is not good enough,'" he said. Then as technologies become more common and are brought into more labs, guidelines for using them will be established, as there are now guidelines for using next-gen sequencing in the clinical lab, he added.
Both New York State and the College of American Pathologists have developed checklists for using next-gen sequencing in the clinical lab and for clinical tests. The CAP requirements, for example, encompass good practices for all steps of the process from sample prep through to reporting results.
Delivering Results to Physicians
Once generated, results from these tests must be assembled into a report for physicians — and quickly. As Cornell's Grills noted, even though not every clinical test gives results that will immediately affect patients' mortality, patients will nonetheless be waiting anxiously for those results.
"You want reporting mechanisms in place that take the data, chew on it, and spit [it] out in a fashion that it gives the physician the data that they really need and no more than the data that they really need, unless they ask for more," Grills said.
There are, as Rehm pointed out, still a number of opinions and approaches to reporting results to physicians, and genetic and genomic test results are especially complex to report.
"We find lots of novel variants, and you have to have a whole workflow established for how you evaluate those variants and what you say about them, evidence-based classification rules, terminologies for how you classify variants, when do you call a report positive versus negative versus inconclusive, [and] how to structure those reports so the physicians easily follow," she said.
Indeed, Grills added that the recent Advances in Genome Biology and Technology meeting in Marco Island, Fla., touched not only on the best way to generate data, but also the best way to report that information to physicians. "That is not a small challenge," Grills said.
Rehm's lab has developed a software suite, called GeneInsight, to help streamline that interpretation and reporting process. It helps her lab keep up with all of the variants they have found, or may find in the future, and what the implications of those variants are and what the inheritance models of those diseases are, she said.
"You take all of these pieces of data and then use rules-based templating to structure a report based on the fields that are true for that case," she said. "We spent a lot of energy developing software to essentially make a highly manual process more efficient."
Partners HealthCare, where Rehm works, has signed an agreement with Illumina to incorporate this software suite into the MiSeq sequencing platform, she noted.
Further, the American College of Medical Genetics and Genomics' Laboratory Quality Assurance Committee, to which Rehm belongs, has recently put together guidelines regarding analytical standards and reporting guidelines for results from large targeted panels as well as from exome sequencing or genome sequencing of samples. Such guidances, she said, aim to help ACMG members "with processes that are, in fact, quite complex and difficult to navigate." The guidance is currently up for member comment, she added.
Business Issues to Consider
In addition to the technical requirements, determining whether to pursue CLIA certification or not is also a business decision for the lab.
Before putting in all the time and energy to document instrument maintenance and test validation steps, Broeckel said the lab should make sure that there is a market for the tests under offer. "I think it really needs some significant effort, and, when you do this, make sure that you actually develop this for a test where you do have customers," he said.
Keddache noted that his lab works with physicians at the Cincinnati Children's Hospital to bring in new tests. "I do receive requests from physicians to validate some tests that are outside of the research realm," he said.
He added that having the test performed in a CLIA lab is important for getting patients' insurance carriers to reimburse for the test.
Similarly, Broeckel suggested looking at the reimbursement rates for a particular test before offering it. "If you develop a diagnostic test, I think a careful point to look at is: What is the reimbursement landscape? And what's the demand for a test? And how do you get reimbursed for it?" he said. "Because you certainly can run very quickly into a deficit."
He noted that the tests that he currently offers — most of which are cytogenetic tests — are reimbursed, but that the situation for newer tests, such as a gene panel, is likely a bit more uncertain.
Becoming a clinical lab may work for some cores and not for others. Broeckel said he's spoken to groups that are considering more clinical work, but they are moving slowly. "A lot of labs realize that they want to make this decision very carefully," he added.
Keddache said the move made sense for his lab since it was located in a hospital. Similarly, Rehm's Center for Personalized Genetic Medicine is associated with Harvard Medical School and the Partners HealthCare System. Partners set aside about $80 million to support genetics and genomics work in both research and clinical arenas to support personalized medicine, she noted.
Cornell, Grills pointed out, has embarked on a similar project at its medical school and associated hospitals. It recently announced a "precision medicine" institute to bring together genetic and genomic tools to the clinic for personalized medicine. "Precision Medicine [at] Weill Cornell is essentially one of the initiatives that is going to move forward in terms of medical genomics using NGS," he said.
At the school's campus in Ithaca, though, the core labs there are not currently offering sequencing-based clinical tests, though Grills said that is something that two working groups at the school are examining. "And this had been an extremely important endeavor. I think everybody realizes that if they want to stay at the cutting edge, they really need these types of resources at their institution," he added.
However, Rehm said that labs whose main focus is research need to consider the extra pressure that comes with being a quality clinical lab. She referenced a research lab, which she declined to name, that launched a complex test as a clinical service for cheaper than she could offer it. But now, she said, physicians are returning to her lab to have that test performed. The interpretation and turnaround time was not quite what the physicians needed, she said.
"I actually think very highly of that lab. I think they are very good researchers, but the infrastructure you need to support a research lab process is actually very different than a clinical lab process," Rehm said.
"I think that's what they are sort of suffering from — their focus isn't clinical service, it's research, and so it is difficult to prioritize that service functionality compared to their research goals," she added. "People need to know what they are getting into when they decide to make this transition."