OBJECTIVES: The identification of potential biomarkers in oncology have led to the development of companion diagnostics to better select cancer treatment and consequently, to increase success rates and potentially reduce costs. The objective of this study was to explore the existing evidences regarding the cost-effectiveness of companion diagnostics in cancer therapy. METHODS: An exhaustive literature search previously performed using 4 electronic databases from January 1967 until December 2012 allowed retrieving 15,374 studies. Among them, 19 studies assessed the economic impact of a companion diagnostic to guide therapy or to predict treatment-related toxicity compared with the “no testing” strategy. Cost-effectiveness ratios and study parameters were extracted from these studies and costs were reported in US$2012. RESULTS: Breast cancer was the main indication of the companion diagnostics evaluated. Cost-utility analyses (cost per QALY) were performed in 79% of studies. Other studies reported results as a cost per case or a cost per life-year gained. Cost of companion diagnostic was comprised between $US57 and $US4,469. Cost-effectiveness ratios of companion diagnostics varied from dominant to approximately $US162,000 per QALY. Specifically, using a $50,000 per QALY threshold, 9 studies suggested that companion diagnostics are cost-effective or could be cost-effective in specific conditions. Furthermore, few companion diagnostics were dominated by the “no testing” strategy, depending on the method of testing used. In addition, two studies have established a maximum cost for the companion diagnostic to be cost-effective. CONCLUSIONS: Most of analyzed studies suggest that companion diagnostics are or could be cost-effective in specific conditions, despite a wide cost-effectiveness ratios range. Ratios obtained mainly depend on cancer type, method of testing and cost of companion diagnostic.