[Q] How to interpret variable impact in a regression model with transformed variables?

Hey everyone, I have a Cox survival regression model and due to non linear martingale residuals I had to add extra variables that are quadratic and cubic transformations of a variable already used in the model. How do I interpret such variable? As an example I have:

>exp exp lower .95 upper .95
>axil_nodes 1.332 0.7506 1.1861 1.4965
>I 0.986 1.0142 0.9782 0.9938
>I 1.000 0.9998 1.0001 1.0003

I get that normally, increasing ‘axil_nodes’ by one, would increase the chance of event by 1.332, but how do I interpret it, if I have transformations of this variable as other variables in the same model?