Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitors Plus Chemotherapy: Case Closed or Is the Jury Still Out?

The cases for and against giving combinations of epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) such as gefitinib and erlotinib concurrently with chemotherapy for first-line treatment of advancedstage non–small-cell lung cancer (NSCLC) have just become even more complex. In this issue of the Journal of Clinical Oncology are two additional pieces of evidence— one favoring the prosecution’s case (EGFR TKIs plus concurrent chemotherapy is a bad idea: “guilty as charged”) and one for the defense (the evidence against the combination is all “circumstantial”: this is a case of the defendant, erlotinib, being “in the wrong place at the wrong time”). Let’s examine the evidence from these two studies. Herbst et al present data from the TRIBUTE trial of paclitaxel/carboplatin plus concurrent erlotinib or placebo, showing that erlotinib added nothing to chemotherapy for any outcome measure in the overall study population. These data are strikingly reminiscent of those published previously in the gefitinib-based INTACT (Iressa NSCLC Trial Assessing Combination Treatment) trials and the similarly designed TALENT trial of gemcitabine/cisplatin plus or minus erlotinib. Altogether, these four trials testing the concept of chemotherapy plus concurrent EGFR TKI account for 4,421 patients, a substantial investment of patient resources. While erlotinib added nothing to efficacy, it did contribute to toxicity. The authors report 48 deaths attributable to an adverse event: 32 (10.2%) of 322 deaths in the erlotinib arm versus 15 (4.4%) of 340 in the placebo arm. So the question at hand is why were TRIBUTE and TALENT negative studies? Similarly, why did gefitinib not add benefit to chemotherapy in INTACT 1 and INTACT 2? After all, we know that chemotherapy by itself produces objective responses and prolongs survival in NSCLC when compared with supportive care alone. In a recently reported trial in patients with NSCLC failing chemotherapy (BR21) erlotinib monotherapy resulted in objective responses and prolonged survival when compared with placebo. Thus, it is logical to conclude that the antitumor activity of these agents should be at least additive (ie, 1 1 should equal 2). (Personal communication, Alisha and Zachary Stickney, April 18, 2005.) However, in the four large clinical trials cited above, 1 1 appears to equal only 1. How could this be? Although there are a number of possible explanations, two hypotheses that seem most likely have been proposed: (1) Lack of patient selection for the target (EGFR) explains the results; or (2) Chemotherapy and EGFR TKIs, given concurrently, result in a negative interaction (antagonism?) in at least a subset of patients treated. The first hypothesis, that negative results are explained by failure to “enrich” the study population(s) by a biomarker predictive of response, clearly has merit. One need only look back to the clinical trials of trastuzumab plus chemotherapy in breast cancer to see the rationale for predetermining a sensitive population most likely to benefit from such a combination. Mathematical probability suggests that if trastuzumab plus chemotherapy had been tested in a nonselected population, no benefit of trastuzumab would have been discernable. In favor of this explanation, the authors of TRIBUTE report a survival benefit for the combination in patients categorized as “never-smokers,” a group previously reported to have a higher response rate to erlotinib or gefitinib as single agents, and as noted below, a group with a higher likelihood of harboring an EGFR-activating mutation. It is here that the study by Eberhard et al describing EGFR and KRAS mutations in a subset of 274 patient tumors from TRIBUTE becomes relevant. A large number of potential biomarkers for sensitivity or resistance to EGFR TKIs have been proposed, including protein expression of JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 23 NUMBER 25 SEPTEMBER 1 2005