The purpose of this study was to define the maximum tolerated dose of erlotinib and characterize its pharmaco-kinetics and safety profile, alone and with temozolomide, with and without enzyme-inducing antiepileptic drugs (EIAEDs), in patients with malignant gliomas. Patients with stable or progressive malignant primary glioma received erlotinib alone or combined with temozolomide in this dose-escalation study. In each treatment group, patients were stratified by coadministration of EIAEDs. Erlotinib was started at 100 mg orally once daily as a 28-day treatment cycle, with dose escalation by 50 mg/day up to 500 mg/day. Temozolomide was administered at 150 mg/m2 for five consecutive days every 28 days, with dose escalation up to 200 mg/m2 at the second cycle. Eightythree patients were evaluated. Rash, fatigue, and diarrhea were the most common adverse events and were generally mild to moderate. The recommended phase 2 dose of erlotinib is 200 mg/day for patients with glioblastoma multiforme who are not receiving an EIAED, 450 mg/day for those receiving temozolomide plus erlotinib with an EIAED, and at least 500 mg/day for those receiving erlotinib alone with an EIAED. Of the 57 patients evaluable for response, eight had a partial response (PR). Six of the 57 patients had a progression-free survival of longer than six months, including four patients with a PR. Coadministration of EIAEDs reduced exposure to erlotinib as compared with administration of erlotinib alone (33%-71% reduction). There was a modest pharmacokinetic interaction between erlotinib and temozolomide. The favorable tolerability profile and evidence of antitumor activity indicate that further investigation of erlotinib is warranted.

BACKGROUND The epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor erlotinib (also known as Tarceva or OSI-774) has shown promising response rates in malignant gliomas. We investigated the association between expression of EGFR and downstream signaling components and the response of malignant gliomas to erlotinib in a phase I trial of erlotinib administered either alone or with the alkylating agent temozolomide. METHODS Expression of EGFR and ligand-independent EGFRvIII mutant proteins and of phosphorylated protein kinase B (PKB)/Akt in specimens from glioma patients were assessed by immunohistochemistry. EGFR gene amplification was evaluated by fluorescence in situ hybridization. Mutations in PTEN and EGFR were assessed by polymerase chain reaction amplification and sequencing. Response was evaluated by sequential magnetic resonance imaging every 2 months. The Cochran-Mantel-Haenzel test was used to assess associations between biomarker status and response. All statistical tests were two-sided. RESULTS Of 41 glioma patients, eight responded to treatment. Response to erlotinib was associated with EGFR expression (P = .07) and EGFR amplification (P = .08). These associations were stronger and statistically significant among the 29 patients initially diagnosed with glioblastoma multiforme (P = .03 and P = .02, respectively). Among six responders with sufficient tumor tissue, none had EGFRvIII mutations. None of the 22 tumors with high levels of phosphorylated PKB/Akt responded to erlotinib treatment, whereas eight of the 18 tumors with low levels of phosphorylated PKB/Akt responded to erlotinib treatment (P < .001). The level of phosphorylated PKB/Akt was also associated with time to progression (P < .001). CONCLUSIONS Among glioma patients, those with glioblastoma multiforme tumors who have high levels of EGFR expression and low levels of phosphorylated PKB/Akt had better response to erlotinib treatment than those with low levels of EGFR expression and high levels of phosphorylated PKB/Akt.

1This clinical research has been funded by NIH P01-NS42927; NIH P50-CA097257; General Clinical Research Center at University of California, San Francisco, M01-RR00079, and Genentech, Inc. Some data from this study have been presented at the 12th European Cancer Conference (Prados et al. 2003a) and at the American Society for Clinical Oncology 2003 Annual Meeting (Prados et al. 2003b). Dong Xie holds Genentech stock. Sean K. Kelley holds Genentech common stock and nonqualified stock options.

Inhibitors of the epidermal growth factor receptor (EGFR), both small molecule pharmacologic agents and monoclonal antibodies, have entered the treatment of a myriad of human malignancies. Nevertheless, no study to-date has established a correlation between response to anti-EGFR treatment and expression of EGFR. We sought to assess whether expression of EGFR predicts tumor response to erlotinib (OSI-774), an EGFR inhibitor that has shown promise in the treatment of gliomas. Patients with stable or progressive malignant gliomas were treated in a dose-escalating phase I trial of erlotinib administered either alone or with temozolomide. Although the primary endpoint was toxicity, patients treated at progression were evaluated for response with MRIs every two months. There were 8 responders, 6 treated with erlotinib alone, and 2 with erlotinib plus temozolomide. We acquired tissue from 39 of the 52 patients enrolled in this trial who were evaluable for response. EGFR expression level was assessed by immunohistochemistry (IHC) and EGFR gene amplification was evaluated by fluorescence in situ hybridization (FISH). A strong association was noted between diffusely positive protein expression as measured by IHC and EGFR gene amplification. Table 1 delineates the relationship between radiographic tumor responses to treatment and EGFR protein expression as assesses by IHC. Despite the fact that patients were treated at various dose levels, preliminary analyses, shown in Table 1, reveal a significant association between expression levels of EGFR as measured by IHC and response to treatment (p = 0.04, Wilcoxon rank sum test). FISH analyses, although completed on only 18 patients to-date, suggest a trend towards an association between EGFR gene amplification and response to treatment (p = 0.18, Fisher exact one-sided test). In a Phase I trial of the EGFR inhibitor erlotinib (OSI-774) alone or with temozolomide for patients with recurrent malignant gliomas, preliminary evidence suggests that the expression levels of EGFR as assessed by IHC correlates with response to treatment. Evaluations of EGFR gene amplification by fluorescence in situ hybridizations (FISH) are ongoing, as are analyses of additional patients enrolled in this trial in the hope that characterizing the signaling profile of each individual tumor may predict its response to erlotinib.