Login to Your Account

Genomes in the Clinic: Have A Nickel? Need an Elephant?


By Anette Breindl
Science Editor

BOSTON – The parable of the blind men and the elephant is sometimes applied to whole genome sequencing. But at a lively session on "The Science of Uncertainty in Genomic Medicine" James Evans of the University of North Carolina told his audience that there is also a personal finance rule about elephants that is apt when considering the clinical use of genomic information.

That rule is that 'A nickel for an elephant is only a good bargain if you have a nickel and you need an elephant.'

At the American Association for the Advancement of Science's annual meeting in Boston Friday, Evans and his co-panelists Robert Green and Giovanni Parmigiani had a wide-ranging discussion about whether the medical community, collectively, needs the whole-genome elephant – or has a nickel for it.

Evans, overall, is a skeptic. "Most of the time, we don't need that elephant," he told the audience.

Evans stressed that the cost-benefit ratio is very different in basic research, in which uncertainty is what drives the important questions. But as far as clinical uses are concerned, the genomics elephant is being looked at by, well, the blind men.

But "in clinical medicine, stakes are high and we hurt people when we act in the face of uncertainty that we cannot clear up. And "in genomics today, there's uncertainty at every turn." Most genomic data "is not information in any useful sense so far."

But if it isn't useful, that doesn't mean it's necessarily harmless. Evans pointed to PSA testing and hormone replacement therapy as examples where the medical community "did far more harm than good" by jumping the gun on data of dubious information value.

In his talk, he distinguished between using sequencing sick individuals versus basically healthy ones. For sick individuals, especially for those with "enigmatic conditions" and clues suggesting a primarily genetic cause, sequencing can be very useful. And Evans thinks there will be "unique and powerful opportunities" to apply whole genome sequencing in cancer diagnostics.

But "I don't really see its utility in being applied massively in an indiscriminate way."

Sequencing healthy people and giving them information usually does more harm than good, he argued, because "as humans we are evolutionarily predisposed to overreact to false positives." It is better to treat every rustle in the bushes as tiger than to ignore the rustle that really is a tiger.

That predisposition, combined with a high false positive rate in genomic analysis, means that "serious decisions are made on the basis of alleged mutations" – ranging from years of screening, to risk-reducing surgery, to abortion.

Evans and his colleagues have suggested that genes be classified into three categories or bins, with only mutations in bin one – those that are both significant and actionable – being obligatory to report by clinicians to patients. (See BioWorld Insight, Oct. 1, 2012.)

Evans said clinicians should essentially ignore bin three, which is where most human genes fit – genes where there is no known direct association between a gene and disease. "I don't think it behooves [patients] or the medical establishment," he said, "to spend a lot of time worrying" about mutations in the latter category.

To Robert Green, though, Evans' opinions smack just a little of paternalism.

Those arguing that there is no value to people's knowing their own genomes, Green said, are saying "the same kinds of things we said when we said a few years ago that 'oh, we can't tell people that they have cancer.'"

"The culture of genetics," Green added, "is to protect people from information." But Green disagrees with that approach.

He is the director of Genomes2People, which calls itself a research program "focused on the judicious integration of genomic research into personalized medicine and clinical practice."

Green described studies this group has done about the effects of revealing negative medical information that people cannot do anything about – such as their being in possession of the high-risk variant of the ApoE4 allele, which predisposes to Alzheimer's disease.

Green and his team both looked at whether people would want to receive such information, and whether they would be able to handle bad news about their genetic status.

"At the bottom of this question is, would people go and jump off the roof? And that was a very scary question to ask," he told the audience.

His team found that "people handled the information very well" and that many did try to change their behaviors in ways that could mitigate their risk – although along with changes in diet, exercise, and legitimate medical care, there was also "a dark side to silver lining": many patients with high risk "went online and ordered unregulated supplements."

Such treatments get to the heart of the second part of what makes an elephant a good deal: it has to be affordable. And Evans argued strongly that in most cases, the treatments based on genomics are not worth their money – not just the Internet snake oil, but the legitimate treatment that the doctor ordered, too.

There is indeed a libertarian argument to be made that people can take as many poorly predictive tests and use as many treatments of dubious value based on those tests as they want.

But Evans pointed out that Internet purchases aside, many poor tests – think PSA again – and the treatments based on them are paid for by everybody, with money that in his opinion could be better spent elsewhere than on genomics.

"Whether it's Medicare or Blue Cross Blue Shield," he said, "we really are all in this together."