Provided are a method for predicting sensitivity to a molecularly targeted drug, a method for selecting patients for which administration of the drug is determined to have a high probability of success, and a reagent used in said methods. Specifically provided are: a method for using a biological sample derived from a cancer patient in order to predict responsiveness to cancer treatment by a compound that inhibits the mitogen-activated protein kinase (MAPK) signal transduction pathway, said method being characterized by measuring whether [beta]-catenin in the biological sample has at least one mutation selected from the group consisting of (i) an activating mutation and (ii) a mutation in which an aspartic acid residue is substituted by a serine residue at the 287 position a method for selecting a subject for cancer treatment using said compound a cancer treatment method that includes administering the compound to a patient selected by the aforementioned method and a reagent used in said methods.
