Partial Dependence Plot: Incorporating two predictors simultaneously

[dmresearch15]

Is it possible to elucidate the predicted response by incorporating two predictors simultaneously, as opposed to just one? I am keen on generating a 3-d/contour map if the values {predictor1, poredictor2, predicted response using train/test data} are at our disposal.