Early detection of brain metastases increases survival in patients with cancer, since image-guided radiosurgery is the most widely used treatment. As support for the qualitative diagnosis made by radiologists, Computer Assisted Diagnosis provides a quantitative and reproducible analysis. This article reviews the methods for an automatic detection of brain metastases in contrast-enhanced T1-weighted magnetic resonance imaging. Model-based methods detect metastases due to their high degree of similarity with models representing their morphology, mainly templates. On the other hand, methods based on brain symmetry and intensity search intensity differences between both brain hemispheres with respect to the symmetry axis. Model-based methods are more commonly used because they allow the detection of metastases of a wider range of measures.