A phase II study of temozolomide in patients with advanced aerodigestive tract and colorectal cancers and methylation of the O6-methylguanine-DNA methyltransferase promoter.

Responses of patients with gliomas to temozolomide are determined by O(6)-methylguanine-DNA methyltransferase (MGMT) and mismatch repair (MMR) pathways. This phase II study (NCT00423150) investigated whether MGMT promoter methylation predicts response in patients with advanced aerodigestive tract and colorectal cancers (CRC). Tumor and serum samples were screened for MGMT promoter methylation. In methylation-positive patients, 150 mg/m(2) temozolomide was administered daily on a seven-day-on, seven-day-off schedule for each 28-day cycle. The primary efficacy endpoint was response rate (RR). MMR status was determined by a microsatellite instability assay. Among 740 patients screened, 86 were positive for MGMT promoter methylation and enrolled. Nineteen percent of the screened population (137/740) had confirmed tissue and/or serum MGMT promoter methylation, including 25% (57 of 229) for CRC, 36% (55 of 154) for esophageal cancer, 11% (12 of 113) for head and neck cancer, and 5% (13 of 242) for non-small cell lung carcinoma. Among patients with valid methylation results in both tissue and serum samples, concordance was 81% (339 of 419). The majority of enrolled patients (69 of 86; 80%) had microsatellite stable cancer. Overall RR was 6% (5 of 86 partial responses); all responders had microsatellite stable cancer. Temozolomide resulted in low RRs in patients enriched for MGMT methylation. MGMT methylation status varied considerably in the patient population. Although serum methylation assay is an option for promoter methylation detection, tissue assay remains the standard for methylation detection. The low RR of this cohort of patients indicates that MGMT methylation as a biomarker is not applicable to heterogeneous tumor types, and tumor-specific factors may override validated biomarkers.