feature request: Chi Square as Decision Tree Splitting Criteria

Hi,

Is it possible to have a tree learner/predictor with a Chi square criterion to select the variable at each step ?

Best regards

Fabien