Preliminary pharmacogenetic evaluation of toxicity data in the prospective multicentre EORTC trial 40015 in metastatic colorectal cancer (CRC)
3072 Background: EORTC phase III study 40015 was initiated in 2003 to compare capecitabine (C) + irinotecan (CPT-11) vs 5FU/LV/irinotecan ± celecoxib in 1st line treatment of mCRC. The study was suspended after enrollment of 85 pts due to 8 fatal events not related to disease progression, 4 of which included thrombo-embolic events. Purpose: To test whether genetic polymorphisms involved in metabolism of the drugs are related to the increased toxicity observed in the study. Methods: 71 pts signed informed consent for genetic analyses and material was available for 58 pts. DNA was extracted from normal colonic mucosa or peripheral leukocytes. Polymorphisms were determined using PCR-based RFLP and direct sequencing. Genotypes known to be associated with increased toxicity (diarrhea, mucositis, leucopenia) were classified as unfavorable. Results: Unfavorable genotypes were distributed equally between C+CPT-11 and LV/5FU+CPT-11 arms. Baseline characteristics and treatment duration were similar in the pts with or without unfavorable genotypes. Unfavorable genotypes of thymidylate synthase (TS-5; TS-3) and UDP-glucuronosyltransferase 1A1 (UGT1A1) were associated with increased grade 3/4 toxicity. 18/35 (51%) pts with unfavorable UGT1A1 genotype experienced toxicity grade 3/4 compared to 3/23 (13%) pts with favorable genotype. 16/36 pts (44%) with unfavorable TS-5 genotype showed toxicity grade 3/4 compared to 5/22 (23%) pts with favorable genotype. Toxicity grade 3/4 was observed in 18/41 (44%) pts with unfavorable TS-3 genotype and 3/17 (18%) pts with favorable genotype. Increasing grade 3/4 toxicity rates were observed in the pts expressing 0/1 (2/16 [13%]), 2 (7/24 [29%]) or 3 (12/16 [75%]) unfavorable genotype(s) (p=0.0001). Among the 4 pts who died of a thrombo-embolic event, only one has been analysed at this stage and showed a Factor (V) Leiden mutation linked to 10-fold increased risk for thrombo-embolic events. Analyses are ongoing and complete data will be available for presentation. Conclusion: Our data suggest an association between polymorphisms of TS and UGT1A1 and toxicity. No differences of pharmacogenetic patterns were observed that could explain the increased rate of fatal events in the C/CPT-11 arm. [Table: see text]