January 2012
Feature Story

Anne Ford

If there’s one thing humans hate in high-stakes situations, it’s ambiguity. And the stakes don’t get much higher than they do in genetic testing.

“When a patient comes to a doctor for a genetic test, what they want out of that and expect is one of two things,” said Jeffrey A. Kant, MD, PhD, during a talk, “Testing, Reporting, and Assessment of Genetic Variants,” at last year’s AACC annual meeting in a session the CAP sponsored. “You have a positive result and information related to that for being at risk, or a negative result, in which case you’re not to worry about it.

“What they [patients] don’t want or expect are . . . results like, ‘You have an amino acid change, but we don’t know what it means,’ or ‘There’s some other molecular mechanism that hasn’t been ruled out,’” said Dr. Kant, professor of pathology and human genetics and director of the Division of Molecular Diagnostics at the University of Pittsburgh Medical Center, and vice chair of the CAP Council on Scientific Affairs.

So how can pathologists help reduce ambiguity in genetic testing? In part, by making it clear up front to clinicians (and occasionally patients) that the results of particular tests occasionally may not be clear-cut—and by working with clinicians to ensure the most appropriate test is done on the appropriate person in the first place. In illustrating the latter principle, Dr. Kant discussed model case studies of families affected by genetic diseases, among them cystic fibrosis and hereditary breast cancer.

Cystic fibrosis, Dr. Kant reminded his audience, is an autosomal recessive disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene; it is most common among Caucasians of Northern European ancestry and Ashkenazic Jews. More than 1,500 mutations in the CFTR gene have been identified. The American College of Obstetricians and Gynecologists and American College of Genetics recommend testing for the 23 most common mutations, including ΔF508, which accounts for about 70 percent of CF mutations in Caucasians, and W1282X, which is present in 46 percent of CF cases in Ashkenazic Jews.

“So let’s take a couple, pregnant for the first time, no family history of cystic fibrosis,” Dr. Kant said, “going through the standard pregnancy testing.” Imagine, he said, that the woman turns out to be a carrier of the ΔF508 mutation, while the man tests negative for all 23 of the mutations in the recommended screening panel, giving him a one in 240 chance of being a CF carrier, and their child in utero a one in 960 risk of being born with CF.

“Most couples with that type of relative risk will go ahead and have the child,” Dr. Kant continued. In this hypothetical scenario, the couple have a daughter who subsequently presents with bouts of pneumonia, chronic bronchitis, and a productive cough, and who is given a clinical diagnosis of CF based on the results of a sweat chloride test. Tested with the standard CF panel, she has the ΔF508 mutation. “Now, the question is: She must have another mutation from her father that’s not part of the panel. Is it important to identify this second mutation?” he asked.

Probably not in terms of managing this particular patient. “Because we know she has cystic fibrosis, she’ll be treated appropriately,” Dr. Kant pointed out. “It’s a clinical diagnosis. But the mutations (genotype) in the couple are important if the couple decides to have another child. Since it’s known that the father has a mutation of some sort, the likelihood of a second child having CF is one in four. Who would you recommend be tested”—the father, fetus, first child, or all three?

In this situation, the first child undergoes CFTR full-gene DNA sequence analysis and is found to have, in addition to the ΔF508 mutation, a “mutation which substitutes aspartic acid for histidine at position 1152,” Dr. Kant said. Subsequently, the father is tested and found to have the same mutation. (“A risk of testing is if Dad doesn’t have the mutation it’s likely you’ve disclosed non-paternity,” he added.) A second child who is clinically normal is tested postnatally and found to have the father’s, but not the mother’s, mutation.

“Now, Dad’s sister is pregnant, and she’d like to know her [child’s] risk,” Dr. Kant said. “What testing would you recommend, and would that recommendation be different if it were the mother’s sister? In this case here, because Dad has this rare mutation, any sort of testing that you would perform on the sister would need to include the mutation at position 1152, whereas testing of Mom’s relatives could probably be performed using the standard screening panel.”

Turning to hereditary breast cancer, Dr. Kant reviewed the facts: Five to 10 percent of breast cancer is associated with mutations of the BRCA1 or BRCA2 genes, and a woman with one of these mutations faces a 50 to 85 percent lifetime risk of breast cancer and a 20 to 40 percent lifetime risk of ovarian cancer. “The knowledge of the mutational status of the patient leads to management options” such as prophylactic surgery, chemoprevention, lifestyle changes, or a surveillance strategy, Dr. Kant said, “and of course, it offers the potential to test other family members.”

Whereas three so-called founder mutations identified in the Ashkenazic Jewish population allow for initial targeted testing of members of that group, mutations in other populations are “spread all over the gene,” Dr. Kant said, “and you’re pretty much stuck with doing full gene sequence analysis rather than a targeted test.”

Then, too, “two-thirds of families have mutations that are unique to that family, so you appreciate their significance by following the disease through the family over time. You don’t have other families to benefit you in your analysis. While most of the mutations can be picked up by sequencing, some of them are structural rearrangements, so there’s no single method that gets everything.”

He posited a situation in which the younger sister of a woman diagnosed with breast cancer at age 44 requests BRCA testing, saying that if she has a positive mutation, she will elect to have a prophylactic oophorectomy. Imagine that “I test her [with DNA sequencing], and the testing is negative, and the patient is told by her primary physician that her test is negative, so she doesn’t have to worry,” he said. “Four years later, she develops metastatic ovarian cancer. What happened?”

What happened is that her sister with breast cancer didn’t undergo genetic testing, which would have indicated that DNA sequence analysis in this family would be uninformative. It’s a valuable reminder, Dr. Kant said, that when DNA sequencing does not identify a BRCA1 or BRCA2 mutation, it may not be the case that there is not a familial mutation. It may be that there is a familial mutation that was not detectable by the methods used.

In other words, “if you get a normal result, it’s reassuring if testing is done in the proper context but doesn’t completely exclude a rare type of mutation,” he summed up. “Particularly deletion-duplication mutations where portions or all of a gene are lost or duplicated, which are typically not detectable by sequencing assays.” Also, rarely, DNA sequence variants within primer binding sites can lead to “dropout” of an allele with a mutation, and only the normal sequence information is obtained. “So you have to be aware.”