Why is cross-validation not often performed with neural network models?

A. CV is not applicable with neural networks B. CV does not work well with neural networks C. there is no need to estimate test error with neural networks D. The models tend to be computationally expensive