OBJECTIVES Biomarkers associated with treatment efficacy and safety can be used to develop pharmacogenomic tests. However, translation of the evidence into clinical practice is difficult due to barriers to development and the difficulty to demonstrate clinical and economic utility. Health technology assessment (HTA) methodology may be used to inform decisions at many points during the product lifecycle. We assessed the role of HTA in the development phase of pharmacogenomic tests in multiple myeloma. METHODS Early-stage HTA was conducted separately for two clinical applications of biomarkers: 1) a safety-based companion diagnostic, and 2) a prognostic test. We reviewed the methods that were useful in answering the questions about the health benefits and costs of the intervention and comparators strategies in each scenario were reviewed. RESULTS An evidence-based approach was applied for both scenarios. Using literature reviews and interviews with relevant stakeholders, questions about the target patient population, the intervention and standard care could be answered. Some issues arose such as variation in standard care and the definition of risk categories. Evidence on the prognostic value of some biomarkers was also found to be inconsistent. Decision-analytic modelling could be applied to synthesize all available evidence and estimate the health benefits and costs of the intervention and standard care options. The influence of the diagnostic accuracy of the test and assumptions about the patient population and treatment outcomes could be evaluated. Not all questions regarding the generalizability of the results to other settings could be answered by these techniques. CONCLUSIONS The role of early HTA for companion diagnostics is to identify the gaps in evidence on the clinical and economic value of biomarkers, variation in standard care, and key stakeholders and their preferences. This information is necessary for forecasting the issues that will be encountered when positioning a pharmacogenomic test in the market.