CART refers to a specific algorithm for training of a tree for classification or regression. Boosting typical refers to a technique which allows to train an ensemble of several trees. The added value of boosting is the higher performance of ensembles compared to individual trees.