OBJECTIVES: Health economic (HE) modeling of companion diagnostics (CDx) shares certain principles with pharmacoeconomic modeling, however it does have underlying challenges. The aim of this study is to review methodologies used in the HE modeling of CDx for targeted cancer therapy. METHODS: A literature search was conducted in EMBASE^® to identify HE studies on CDx for targeted cancer therapy. Full journal articles were selected following the PRISMA statement and the modeling methodologies were reviewed. RESULTS:  A total of 367 abstracts were identified and 40 studies on CDx HE modeling were selected for review. The number of HE modeling on CDx increased from 1 in 2004 to 8 in 2016, including lung cancer (22), breast cancer (7), colorectal cancer (7), melanoma (1), ovarian cancer (1), gastric cancer (1) and chronic myeloid leukemia (1). 37 studies were cost-effectiveness analysis and 3 other studies reported only economic outcomes. All 37 studies compared targeted therapy (with vs. without CDx) with standard of care (SoC) and 16 also compared different CDx strategies. Cost-effectiveness analysis was conducted with a time horizon ranging from 1 year to lifetime, and markov model or decision-tree was commonly used with 5 using both models. All studies included CDx and treatment costs from a payer or health system’s perspective and 5 were conducted from a societal perspective. Limited studies specified the selection of CDx diagnostic performance or considered impacts of false positive/negative results. The majority of studies concluded that compared with SoC, using CDx to guide targeted therapy could improve health outcomes and increase treatment costs. It was stated that cost-effectiveness of a CDx is sensitivity to treatment costs of a targeted therapy. CONCLUSIONS: With the increased adoption of targeted cancer therapy with CDx, best practice in HE modeling of CDx will be critical to assist the economic evaluation of both CDx and targeted therapies.