OBJECTIVES: Cost effectiveness analysis of MammaPrint in the diagnosis of early breast cancer from the Spanish NHS perspective.  METHODS: Markov model assuming a 60-year-old women cohort with negative-lymph node, positive estrogen receptors and negative HER2 breast cancer. Costs and effects of the treatment by identifying recurrence risk using A!O, MammaPrint or OncotypeDX were compared at 5 years, 10 years and lifetime horizons. Probability of low or high risk of recurrence by A!O was fixed 50/50 (ratios between 40/60 and 60/40 for sensitivity analysis). Risks predicted by OncotypeDX and  MammaPrint were assumed to be the same. Cost of chemotherapy (2.825€), recurrence (6,357€) and other direct health care costs were derived from local data. Three health states (free of recurrence, recurrence, death) were defined in a given period of time. RESULTS: 31% patients required adjuvant chemotherapy with MammaPrint classification (36.6% with OncotypeDX, 56.3% with A!O). MammaPrint showed a life expectancy of 23.55 years at lifetime Differences in chemotherapy costs, worse prognosis rates and a higher cost of the genomic profile (3,200€ vs. 2,675€) explained higher cumulative costs for OncotypeDX over both alternatives at any time. MammaPrint instead of OncotypeDX showed savings of at least 1,273€ after 5 years. MammaPrint resulted dominant (less costly and more effective) against OncotypeDx at any time horizon and would be cost-effective from the 6th year versus A!O (43,912€, 6,169€ and 287€ per QALY gained at 5, 10 years and lifetime; 30,000€/QALY gained threshold assumed). Sensitivity analysis confirmed base case results with MammaPrint remaining cost-effective until a willingness-to-pay threshold below 275€/QALY gained. Utility of recurrence, age at baseline and probability of A!O low risk were the key drivers at 10 years. CONCLUSIONS: MammaPrint in predicting risk of recurrence in these patients and avoiding chemotherapy overtreatment is a dominant strategy over OncotypeDX and it is highly cost-effective against A!O.