Humer I think we are on the verge of a wave of innovation derived from the discoveries of the last 10 to 15 years in genomics and genetics [where the human genome, the underlying blueprint for human life, has been mapped by researchers]. This will allow us to integrate diagnostics with therapeutics so we can identify the right patients for the right medication at the right dose at the right time.

American Way How far in the future is that?

Humer Take Herceptin, a drug to treat breast cancer. When we started to explore that drug more than 10 years ago and did the first clinical trials, we had very low response rates. Some patients responded beautifully, others not at all. We were not quite sure where to go. Then we discovered that this is a drug that works in breast-cancer patients who overexpress [have too much of] a certain gene, and if we use that drug exclusively on that part of the population, which is somewhere between 15 and 20 percent of all breast-cancer sufferers, then we get response rates of 60 and 70 percent. So that is what is happening today.

American Way What's allowing you to make that shift? What tests are you doing?

Humer In the case of Herceptin, breast-cancer patients are tested on their expression of a gene called HER 2 [a so-called oncogene that regulates cell growth].

American Way How complicated is this test?

Humer To say it's a simple test you can carry out at home would be an exaggeration, but it is certainly a test that can be carried out in all major hospitals. Today in the United States, more than 90 percent of women with metastatic breast cancer get tested for that gene.

American Way What's your sense of the level of understanding of drug development? Do business leaders - oil company executives or bankers - understand the timelines and failure rates?