If you have a nominal (i.e., 2 level) response and a model that has been fit to explain the response then you can get a ROC curve but it is a two step process (to the best of my knowledge). In such situations I save my fitted model to my data table and then using fit y by x with the model score as the x and the response as the y. Then you will have a logistic fit and you will find the ROC curve as an option.