Background: Companion biomarkers for targeted therapies have increased the expectation that biomarkers can improve health outcomes or potentially save health resources without compromising patient outcomes. However, few countries provide health economic assessment methods guidance (e.g. health technology assessment guide) specifically for co-dependent technologies such as companion diagnostics. Aim: This thesis aims to explore good practices for evaluating companion biomarker tests as part of health economic assessments of their co-dependent targeted therapies in cancer. Scope of the study: Cancer biomarkers for targeted therapies investigated in this thesis are restricted to companion biomarkers, classifying patients into responders and non-responders for a specific targeted therapeutic agent. Methods: Four research activities were designed: two systematic literature reviews (SLR) and two health economic models. The first SLR (Chapter 2) was conducted to demonstrate the impact of companion biomarker tests on the cost-effectiveness of targeted therapies, focusing on metastatic colorectal cancer (mCRC). The second SLR (Chapter 3) considered all cancer areas. It investigated current and best practice for modelling and incorporating companion biomarker tests when assessing the cost-effectiveness of targeted cancer therapies. The findings from these two SLRs were then applied to the cost-effectiveness modelling of a novel candidate companion biomarker test, Heat Shock Protein 27 (HSP27) expression (Chapter 4). The final work (Chapter 5) developed a practical guide to modelling companion biomarker tests as part of economic evaluations of corresponding targeted therapies; a global model was constructed and provided as a worked example coupled with step-by-step guide for readers to follow. Results: The first SLR study showed that the use of companion biomarker tests saved some costs however, the saving was not high enough to change materially the cost-effectiveness of co-dependent therapeutic agents. The second SLR found that there was inconsistency in the methods for evaluating companion biomarker tests in the appraisal of co-dependent agents. The cost-effectiveness analysis of HSP27 expression showed conflicting results depending on the structure of the comparative analysis. Finally, the modelling guide coupled with a worked example of a global model demonstrated how to model characteristics of companion biomarker tests in economic evaluations of test-guided therapies. Conclusion: This thesis highlights the need to reach a consensus on the methods of evaluating companion testing technologies as part of economic evaluations of their corresponding test-guided therapies. Built upon the consensus, a methods guide for co-dependent technologies needs to be developed and introduced, providing a coherent and unified guidance on good practices, reference case, evidentiary standards and data requirements for economic evaluations.