PURPOSE: Monoclonal antibodies against the epidermal growth factor receptor (EGFR), such as cetuximab, have led to significant clinical benefits for metastatic colorectal cancer (mCRC) patients but have also increased treatment costs considerably. Recent evidence associates KRAS and BRAF mutations with resistance to EGFR antibodies. We assessed the cost-effectiveness of predictive testing for KRAS and BRAF mutations, prior to cetuximab treatment of chemorefractory mCRC patients. EXPERIMENTAL DESIGN: A life-long Markov simulation model was used to estimate direct medical costs (€) and clinical effectiveness (quality adjusted life years, QALYs) of the following strategies: KRAS testing, KRAS testing with subsequent BRAF testing of KRAS-wildtypes (KRAS/BRAF), cetuximab treatment without testing. Comparison was against no cetuximab treatment (reference strategy). In the testing strategies, cetuximab treatment was initiated if no mutations were detected. Best supportive care was given to all patients. Survival times/utilities were derived from published randomised clinical trials. Costs were assessed from the perspective of the Swiss health system.RESULTS: Average remaining life-time costs ranged from €3'983 (no cetuximab) to €38'662 (no testing). Cetuximab treatment guided by KRAS/BRAF achieved gains of 0.491 QALYs compared to the reference strategy. The KRAS testing strategy achieved an additional gain of 0.002 QALYs compared to KRAS/BRAF. KRAS/BRAF testing was the most cost-effective approach when compared to the reference strategy (incremental cost-effectiveness ratio: €62'653/QALY). CONCLUSION: New predictive tests for KRAS and BRAF-status are currently being introduced in pathology. Despite substantial costs of predictive testing, it is economically favourable to identify patients with KRAS and BRAF wildtype status.