College of American Pathologists
Printable Version

  Popping the question:
  Is it a genetic test?


cap today

October 2004
Cover Story

Karen Titus

Theres an easy answer to a hard question labs are struggling with: Is (fill in the blank) a genetic test?

This question, no stranger in clinical labs, is starting to take up a new haunt—anatomic pathology. Take testing for hereditary nonpolyposis colorectal cancer (Lynch syndrome), which was the topic of a lively discussion on the Association for Molecular Pathology’s listserv this spring. The disease is relatively uncommon, and linking an initial colon cancer tumor to HNPCC is not critical, in and of itself. But monitoring these patients is crucial, as they are at risk for developing other types of tumors as well as a second colon cancer, and their families may need to be checked for the disease. "So this is a good example of a test that could have profound implications. But it’s still not clear how best to test these patients," says Jennifer Hunt, MD, director of molecular anatomic pathology and assistant professor, Department of Pathology, University of Pittsburgh Medical Center.

Tumor microsatellite instability testing, one option, looks for the results of errors in DNA mismatch repair. Microsatellite instability is present in almost all patients with HNPCC, but it’s not specific to the disease. It’s also possible to look for the germline mutation in three key mismatch repair genes using DNA sequencing, but the high specificity of this method is offset by its lower sensitivity and higher complexity. Finally, labs can turn to immunohistochemistry to look for underexpression of the encoded repair enzyme in the tumor cells, which implies a gene mutation. This too is a specific, but less sensitive, method.

"In reality, we end up doing a combination of these three tests," says Dr. Hunt. "And the idea of, What is a genetic test? It’s incredibly important, because we should be thinking about it before we request these assays. In AP especially, when we order an immunostain, we never think about informed consent, and we’re not thinking about the genetic ramifications."

Gene sequencing, clearly, is a genetic test. MSI reveals a potential for having the genetic disease, but not whether it’s actually present. And "the immunostain is somewhere in the middle," says Dr. Hunt. The question, it turns out, has as many shades as a Rembrandt masterpiece. So many subtleties, worthy of endless rumination and discussion. No wonder so many labs are stumped by the question.

And yet, as is often the case, the best and not completely glib answer may be, "Does it really matter?"

Not as much as you may think.

"Why are you even asking if it’s a genetic test?" queries Jan Nowak, MD, PhD, a pathologist at Advocate Lutheran General Hospital, Park Ridge, Ill., in response to the aforementioned MSI/HNPCC quodlibet. "Why would anyone ask if it’s a genetic test?"

Dr. Nowak is not ignorant of the emerging debate. He’s a pensive though hesitant speaker on the subject, who describes his views as "not fully cooked" before recounting how he and his lab colleagues launched MSI testing for colon tumors at their institution.

At Lutheran General, the lab offers the standard panel recommended by the revised Bethesda Guidelines for HNPCC and MSI (J Natl Cancer Inst. 2004;96:261-268). After putting together the assay, the lab embarked on educating the clinicians. Much of the teaching took place at the regular GI tumor board, says Dr. Nowak, which went a long way toward raising, if not resolving, the issues.

"What I’ve come to realize is that because this test is done on DNA, and this testing does have a true genetic aspect—meaning some of these patients may have HNPCC, and that’s part of the motivation of doing it—some clinicians and surgeons see this as a genetic test," he says. "And judging from the comments I hear, I think they present it to their patients that way—which is, I think, inappropriate."

In conversation, Dr. Nowak emphasizes repeatedly that his thoughts about genetic testing are not final. "I keep feeling compelled to express my hesitancy," he says. But he’s scarcely the only one whose views could be considered al dente.

No one quite knows what to think in part because the question obscures other pivotal lines of inquiry. Asking the question, Is ____ a genetic test? is like dropping a cluster bomb into the lab, with the real impact occurring away from the original target, in the areas of reimbursement, informed consent, and regulatory oversight.

And yet the question isn’t going to go away, at least not anytime soon.

On a scientific level, deciding whether a test is genetic is a complicated though not impossible task. Dr. Nowak, who’s chair of the AMP clinical practice committee, reports there’s been extensive discussion of the MSI issue in that group, and that a subcommittee may draft guidelines for laboratories. One reason is to help those who run molecular labs understand the ins and outs of MSI testing, information they can then pass onto their clinical colleagues. Secondly, Dr. Nowak says, there’s a big need to educate surgical pathologists.

Others second that notion. Even many mainstream pathologists may be unaware of the MSI discussions, since they’re mostly taking place among molecular specialists at this point. But not for long. "We are going to be detecting germline mutations in patients from their AP material," says Dr. Hunt. "Whether we should or not is another question. And how we want to do that is a huge question."

As MSI/HNPCC demonstrates, the discussions twist like a pretzel.

"MSI is not really a pure genetic test," says Wayne Grody, MD, PhD, professor, UCLA School of Medicine, Departments of Pathology and Laboratory Medicine, Pediatrics, and Human Genetics, and director of the molecular pathology laboratory. "You’re not looking at the primary genetic mutation of the disease, but rather an effect of the mutation." Of course, it happens to be an effect at the molecular level, thus giving the test the appearance of being genetic because it’s performed on DNA. But in essence, pathologists are using MSI to look at a molecular phenotype.

Microsatellites are one of the "hot spots" for error formation. "All you’re doing is detecting the outcome of having the error, the actual mutation in the enzyme," says Dr. Grody. The difference between a somatic mutation and the germline mutation may be a subtle one, but it’s important. The alleles uncovered by MSI testing serve as surrogate signs of a possible familial colon cancer gene mutation. "So is it kind of a secondary genetic test?" Dr. Grody asks.

Further complicating matters, some sporadic colon cancers have the same defect. "So even if you’ve done this so-called genetic test, the result of it doesn’t necessarily indicate a genetic disease," says Dr. Grody, who’s also advisor to the CAP Molecular Pathology Committee.

Because MSI is more reasonably considered a downstream phenotype, which may or may not be linked to a genetic form of colon cancer, some pathologists don’t consider MSI to be a genetic test. On the other hand, the immunohistochemistry might well be classified as such, even though it’s not a DNA test. "But you’re finding bad proteins, which implies a bad gene," says Dr. Grody. "And the next step would be sequencing to actually find the mutation."

HNPCC, say those embroiled in these discussions, is probably the most complex test example to crop up in the debate. A simpler one is factor V Leiden, plainly a genetic test. There’s no doubt about heritability, there are implications for family members, and it could be done as a predictive test in an asymptomatic person. Yet the mutation is common; the test is ordered by hematologists and internists; and the risks of having the defect are "soft," as some put it. So even with red flags waving, it’s routinely treated as a regular hematology test.

As well it should be, pathologists are quick to add. But there are good reasons to toss out old assumptions about genetic tests, especially as the search for germline mutations broadens.

Dr. Grody points to the p53 gene, which is mutated in many tumors somatically but which can also be inherited in a familial cancer syndrome. If p53 testing is part of the pathology workup of a tumor, either by itself or as a part of a large array of tests, and a mutation is found, the question arises: Was a genetic test performed without patient consent? Or, because the mutation is likely restricted to the tumor, is it not a genetic test?

What about an assay that’s used for cancer testing in one instance and to determine infection in another, such as the EBV viral load assay? Are either or both tests genetic?

How will new information change matters? In fragile X testing, for example, only the so-called full mutation was thought to be important. But with the appearance of new clinical manifestations, it’s becoming evident that the pre-mutation—formerly thought to be relatively trivial—is a predisposer to disease. "So now we have different indications for testing, and different outcomes, and it changes what’s required for counseling," says Margaret Gulley, MD, associate professor and director of molecular pathology, University of North Carolina, Chapel Hill, and chair of the CAP Molecular Pathology Committee.

It will also be harder to dole out simple answers as more becomes known about the human genome and cancer. It’s not far-fetched to think that pathologists will encounter many gene changes that can be either inherited or acquired. Or they may come across a genetic component linked to 10 different downstream somatic changes but which wouldn’t have occurred without the inherited genetic mutation. And as more tumor tests emerge, using higher density microarrays involving hundreds, perhaps even thousands, of genes, labs will find many, many mutations—how will it be possible to tell which are inherited and which are specific to the tumor?

Think too long or too hard about these questions, and it’s easy to feel like you’re at a Zen retreat, puzzling out the sound of a tree hitting the floor of a tenant-free forest.

But there’s nothing philosophical about matters of patient consent, and possible insurance or employment discrimination, which are among the reasons the is-it-or-isn’t-it question gets raised in the first place. And labs are firmly grounded in real-world concerns as they chew on some possible answers.

"It’s really important to get the definitions right for genetic testing," says Dr. Gulley. Notice her use of the plural—definitions. Like everyone else, she acknowledges the futility of trying to arrive at one, simple answer.

Her starting point consists of inheritable factors. That would mean most of the testing done in molecular pathology labs is not genetic, since the focus is looking at acquired somatic mutations in cancer or at the genome of infectious agents.

Dr. Gulley doesn’t limit the definitions to DNA testing. "It could be RNA or protein," she says, "although the further away you get from DNA testing, the more likely it is that other factors besides inheritance play a role."

But every case is unique, she acknowledges. "Even serum cholesterol could be considered a genetic test," she says.

Does she consider it to be so? She pauses long and hard before replying with a "no." Then she laughs, explaining: "I can’t really answer that, because it depends on the context. Sometimes cholesterol would be included, and sometimes it wouldn’t be." And that’s why her institution—and doubtless most others—lacks a set policy for the question, What is a genetic test?

Even when tests do fit a definition—whatever it may be—Dr. Gulley advocates further refinement. "Divide them by their impact," she suggests. That would place factor V Leiden in one category of genetic test, and Huntington disease in another. "We need to use a practical and rational approach, instead of lumping genetic testing into one bin, and saying we need consent for all genetic tests. It’s not realistic, it’s not practical, there aren’t enough counselors, and it’s just not needed."

Factor V Leiden and MSI have both spent time in the hot seat at Evanston (Ill.) Northwestern Healthcare, as laboratorians tried to decide whether either (or both) required patient consent, says Karen Kaul, MD, PhD, director of the molecular diagnostics laboratory.

"Broadly, everything we do in the lab is a genetic test if you define it as something that uses DNA," says Dr. Kaul, who’s also professor, pathology and urology, Northwestern University Feinberg School of Medicine, Chicago. "However, I would say we make a distinction between germline and somatic tests. Much of what we test for is not transmittable. Much of tumor testing, to look for mutations or gene rearrangement or microsatellite instability, is a function of the tumor." None of which should require informed patient consent, she says.

Mutations for Huntington disease, inherited predisposition to breast cancer, and other serious, heritable illnesses, on the other hand, would. And factor V Leiden? "I think most people would agree that it’s not in the same category," says Dr. Kaul. "In fact, there are biochemical or functional assays that will give you much the same information, and they certainly don’t require informed consent. So factor V appears to be heading in the same direction—not needing consent—even though it is, technically, a heritable mutation."

Requiring consent for factor V Leiden testing "would be ridiculous," in Dr. Kaul’s view. "It would add so much headache, paperwork, expense, you name it." Ditto for other tests that skirt the boundaries of genetic testing. "You can take it to an extreme," says Dr. Kaul, "and argue that testing for germline mutations that cause hypercholesterolemia requires consent. But if you just do a cholesterol test, are you going to require consent for that too? Where do you want to draw the line?"

For all the soul-searching, labs will probably be fairly united on where that line should be drawn. Their bigger concern is where federal interlopers will start their scribbling. "In the past we’ve been made nervous by regulatory agencies that lump somatic and germline mutations together," says Dr. Kaul. "We’re talking about extremely high levels of oversight and consent for both types of tests, and I think that is inappropriate."

Like Dr. Kaul, Dr. Gulley argues for a more restrictive use of the term "genetic testing," limited to heritable factors, especially when it comes to writing legislation or similar regulations.

Are pathologists’ concerns being heard by outside parties? "Up to a point," says Dr. Grody. "It’s always been sort of a sore point with us. At least we have a seat at the table now, thanks to Dr. Leonard."

That would be Debra Leonard, MD, PhD, vice chair of laboratory medicine at Weill Medical College of Cornell University, New York, and a member of the Secretary’s Advisory Committee on Genetics, Health, and Society. SACGHS is essentially a more recent incarnation of the Secretary’s Advisory Committee on Genetic Testing, which got its start when advisors to the Human Genome Project began worrying about the project’s ethical, legal, and social fallout.

The first committee’s definition of genetic tests was as broad as the Bay of Bengal. "It included basically anything that had to do with DNA, including somatic mutations," says Dr. Leonard, who was not part of that committee. Labs were understandably concerned about such an all-encompassing definition; the current committee has not said it will use that definition, nor has it said it won’t.

In fact, says Dr. Leonard, SACGHS has spent little time hammering out definitions. Instead, members have delved into several other topics they’ve deemed more crucial.

First up: genetic exceptionalism—is genetic information different from other types of medical information? Not surprisingly, this question also lacked a simple answer. "It’s something you couldn’t really decide in and of itself," says Dr. Leonard. "It has to be done in the context of how you discuss the information." The committee, echoing many pathologists, essentially said in some cases such information is different, and in some cases it isn’t.

SACGHS has since turned to reimbursement. "Unless these tests are appropriately paid for, no one is going to have access," says Dr. Leonard. Which effectively sends a torpedo right into the side of the "Is it a genetic test?" question.

In Dr. Leonard’s view (she makes it clear she’s expressing her opinions, not those of the committee), there’s another issue that further undermines the is-it-or-isn’t-it question: protection against discrimination based on genetic information. "If we have that protection, then I don’t think there will be as much concern," she says. This has become one of SACGHS’ highest priorities. Equally important, she says, is deciding when informed consent and the services of a genetic counselor are required—which, again, may not hang strictly on the definition of a genetic test.

By now, the original question is losing some of its urgency.

Dr. Grody, for one, would like to sidestep the question because it implies a separate, important category of medical information that, for all practical purposes, shouldn’t exist. "I’ve tried to argue against isolating genetic testing and genetic disease as somehow being more important than all the other areas of medicine," he says.

Putting genetic tests in the penthouse has been driven somewhat by ethicists, he suggests. Concerned by past examples of genetics-based abuses that have led to ethnic discrimination, they’ve turned their attention to eliminating possible future abuses. A worthy goal, but perhaps a bit misguided, some say. "The big risk, which everyone talks about the most, is insurance discrimination. And we have very little evidence that there’s a major problem with that in this country," Dr. Grody says. Lack of insurance, on the other hand, "is a far bigger problem."

Moreover, the is-it-or-isn’t-it question, with its negative spin, doesn’t acknowledge the profoundly positive benefits of genetic testing. "It’s easy for people to get wrapped up in the emotion, and go overboard to try to prevent misuse of genetic testing, instead of focusing on all the wonderful outcomes that can occur," says Dr. Gulley. In the end there may be very few circumstances in which pathologists or others need to worry about the implications of genetic testing. Huntington disease is the obvious example, cited by everyone, but it tops a nearly empty (so far) list.

But even if the question has no final, useful answer, pathologists still need to ask it. "In setting up our MSI test here, our own pathologists never thought of it as a genetic test until I raised the issue," says Dr. Grody. He’s not saying it is such a test. "But it ought to occur to us that it’s somewhat genetic. It’s certainly more genetic than just looking at the colon cancer under the microscope." Pathologists need to recognize the gradations on the scale between somatic and inherited, and ask at which points it might be necessary to require consent and involve a genetic counselor.

In other words, a little Socratic self-examination in the lab is in order. Pathologists who don’t ponder the potential ethical questions genetic tests raise and how to address them in the lab, Dr. Leonard says, "are sticking their heads in the sand."

Karen Titus is CAP TODAY contributing editor and co-managing editor.