Most of the focus on predicting response to EGFR inhibitors has been on identifing molecular markers that are associated with major response to this kind of treatment. But we know that there is a group of patients who get no benefit from these expensive drugs, and in these patients, EGFR inhibitors would just lead to side effects and keep them from a potentially more effective therapy for them. As you can see, the overall survival and progression-free survival curves from the BR.21 trial (abstract here) show that there are a group of patients clustered on the right side of the curve who do no better than the placebo group. In addition to detecting which patients are going to be the greatest beneficiaries, we need to identify which patients will get no benefit at all from EGFR tyrosine inhibitors like Tarceva.

Ras (rhymes with mass) is a gene and protein product that are involved in molecular signalling and very important in many cancers. There are several related genes in the family, and most of the focus is on k-ras. Mutations are seen in a significant minority of lung cancers and are far more likely to be seen in smokers than in non-smokers (abstract here). As you can see from the tables below, ras mutations are consistently more common in smokers than in non-smokers in multiple case series, and they are essentially mutually exclusive with EGFR mutations that have been associated with a high response rate to EGFR tyrosine kinase inhibitors like Iressa and Tarceva.

(click to enlarge)

The available evidence also suggests that patients with ras mutations don't respond to EGFR inhibitors. Looking at the BR.21 trial of tarceva vs. placebo, the patients with an identified k-ras mutation had a worsening of survival if they received tarceva (abstract here), compared to recipients of placebo (remember, with a hazard ratio, numbers lower than 1 mean a benefit, and numbers above 1 translate to a harmful effect of the treatment, in this case a 67% worse survival):

Similarly, Dr. Miller looked at EGFR and k-ras mutations from tumors in a trial of Tarceva in bronchioloalveolar carcinoma (BAC)(abstract here) and found that there were no responders among the patients with k-ras mutations, vs. a 30% response rate for patients who don't have this mutation:

Some other analyses that have looked at ras mutations and EGFR inhibitors have also supported this finding.

Testing for ras mutations isn't commonly done yet, but it is commercially available in some labs. Perhaps it's because we're reluctant to find a reason not to treat a patient with a certain therapy when the options are limited, or maybe we really should be doing this test routinely in order to help us clarify the patients in whom this approach is just a waste of time and money. I'd be interested to know whether patients would want to know before trying tarceva whether they shouldn't bother with it, or whether the hope of a new therapy would lead people to prefer a "don't ask, don't tell" policy, even if it means eliminating a treatment from a short list of agents that has a demonstrated survival benefit for patients with advanced NSCLC who have progressed through at least one line of chemo.