Medical Device Daily Contributing Writer

As the cost of whole genome sequencing drops, its routine use in the clinic is moving from someday to someday soon.

But as such routine use comes nearer, many questions remain about how much of the data in a genome sequence is useful information – and how much of that information patients want to know, and doctors need to deliver.

One school of thought holds that patients want to know all they can. But "I think that is somewhat of a trope. People want control [over their medical information]. But I don't think that everyone wants to know everything. In fact, it's been well documented that not everyone wants to know everything," Jonathan Berg told Medical Device Daily.

At the same time, Don't-Ask-Don't-Tell is not any better as a medical strategy than as a military one. Part of a physician's duty to his or her patients is to warn them of impending harm, if there are ways to prevent or address that harm. And so, "if we, as clinicians, have management strategies, then we feel we have an obligation to make that information known," Berg said.

In some ways, the problem is an old one. It has similarities to decisions physicians make about when to address a patient's weight. Someone who is 100 pounds overweight should clearly be doing something about it. But what about 10 pounds?

But in other ways, genome data is different.

Risk variants, unlike weight, cannot be changed. Plus, there are lots of them. Lots and lots.

And given the rarity of Mendelian disorders, "except for one or two of them, they are completely incidental," Berg said. Overwhelming patients and clinicians with every variant that could be a risk factor only makes it more likely that the needle in the haystack will be lost.

Berg and his colleagues at the University of North Carolina at Chapel Hill have developed a new approach to identifying which gene variants need to be discussed with patients. They published their approach in the Sept. 20, 2012, issue of Genetics in Medicine.

In their paper, lead author Berg and his team sorted roughly 2,000 gene variants that are implicated in Mendelian disorders. Those variants, Berg said, are the sort of information that would be expected to show up in whole-genome sequencing for a disorder, but that are not directly related to that disorder – the so-called incidentalome.

One important question is whether such variants are actionable, that is, whether there is anything to do about them from a clinical point of view. That bin, Berg said, ends up being "pretty small – we all know that there are not that many Mendelian disorders we can do anything about."

For genes that do not come with a definite plan of action, the question of whether to report them hinges partly on how likely they are to become medically actionable, as well as how much distress is likely to result from knowing one's carrier status. Certain genetic variants – Huntington's disease is a famous one – make it all but inevitable that the individual will develop a serious disease. "To some people," Berg said, "that knowledge is devastating" – essentially robbing them of what good time they have left.

Berg stressed that the approach his team has outlined is not a definitive method. There is "a grey area" – for example, variants that are actionable, but sometimes very frightening to their carriers, such as BRCA mutations. Other gene variants, such as variants that cause hemochromatosis, have a very low penetrance, meaning that most people with the risk variants will never develop iron overload. Good arguments can be made for telling patients about such risk variants – or for not telling them.

But, he added, in some sense, that is the point. "As a field, we are on the cusp of thinking about how to do this . . . this [paper] is our contribution to the discussion."

No Comments