Hastings Center Report

The phrase "personalized medicine" has a built-in positive spin. Simple genetic tests can sometimes predict whether a particular individual will have a positive response to a particular drug or, alternatively, suffer costly and debilitating side effects. But little attention has been given to some challenging issues of justice raised by personalized [End Page 16] medicine. How should we determine who would have a just claim to access particular treatments, especially very expensive ones? How effective do those treatments need to be?

If there were a thick, bright line separating minimal responders from maximal responders, then we could allocate these treatments in a fair and cost-effective way. But there is no bright line. The reality is more like a ragged edge—some people will clearly benefit a lot, some people will clearly not benefit at all, and many people will benefit somewhat. Daniel Callahan, cofounder of The Hastings Center, identified the ragged edge as a fundamental problem of medical advancement twenty years ago, but it is particularly relevant to the decisions that are beginning to be made—and will increasingly need to be made—as personalized medicine becomes established. How can we decide fairly who gets which expensive therapies when patients just barely on the "wrong" side of the line can offer some good reasons to think they have just as much right to those resources as those slightly on the "right" side?

In July, a panel of the Food and Drug Administration recommended revoking approval of bevacizumab (Avastin), a targeted biologic drug for breast cancer that costs one hundred thousand dollars a year. The panel determined that the drug did not benefit patients because it extended life by only a few months. But in one recent trial, there were marked differences in survival among patients with specific genotypes.¹

The study compared treatment with pacilitaxel (Taxol), a chemotherapy agent, to Taxol and Avastin. Overall, median survival in the two groups was virtually indistinguishable: 25.2 months versus 26.7 months. Avastin increased median survival by six weeks. However, median survival with Avastin was 49.7 months for the subgroup with the VEGF AA/AA genotype, 46.5 months for individuals with the VEGF-1154AA genotype, 37 months for those with the VEGF-2578AA genotype, 30.2 months for patients with the VEGF AA/GA genotype, 27.1 months for those with the VEGF CA/GA genotype, and only 21.7 months (less than the survival on Taxol alone) for individuals with the VEGF CC/GG genotype. In this sample, the population most likely to benefit from Avastin was the one with the fewest patients: AA/AA represented only 7.6 percent of the cohort.

These findings have not been validated by other studies and therefore should not be considered in clinical practice, said Bryan P. Schneider, the lead author of the original study, in an e-mail message to me after the FDA panel's decision on Avastin. But if they do get independent validation, that would raise the ragged edge question: Should the drug retain approval only for the small percentage of patients who are likely to benefit the most? Ragged edges mean rough justice.

The European Medical Agency recently approved panitimumab (Vectibix) and cetuximab (Erbitux), two expensive biologics, as first-line therapies with chemotherapy for patients with metastatic colorectal cancer who have no mutations in the codon 12 and 13 of the KRAS gene. Research shows that the drugs can prolong life by two years in some patients without these mutations, compared with only a few weeks for all patients with metastatic colon cancer. About 35 to 40 percent of patients with metastatic colon cancer have the genetic characteristics that predict response to the drugs, whereas 40 percent have a KRAS mutation that predicts no response.

The United States could save several hundred million dollars a year if it, too, restricted access to these drugs. Such a decision would seem to be morally reasonable. However, future research could complicate this decision. Individuals without the KRAS mutations do not all show the same response. Only some achieve maximal gains in life expectancy, whereas others gain a few extra...