There's a few things going on here. First, ST would have a few negative things to say about using a Nucleo board in a product. Then they'd have many more things to say about using a Nucleo board in a medical device. But the gist of the conversation would be "NO, ABSOLUTELY NOT, STOP NOW".
When developing a medical device, your design file would have the prototype device information and the analysis that you did to determine that the choices you made were appropriate for use. Your FMEA and requirements would indicate how you managed to choose a '439 as opposed to other processors.
No amount of finger pointing at ST would keep the FDA from saying that you have to justify your use of common off-the-shelf hardware. You didn't design it, why did you decide to not design it, where is your design justification that decided that ST's design is appropriate for use, and why would you decide to go against their terms of use and use their board in a product?
Now, if you were to choose a properly SIL rated device that is appropriate for your purpose, an operating system that has been certified for use in a medical device, and justified it all properly ("We chose an ARM Cortex-R device because the life of my grandmother is more important than simplicity and expediency")
Next, get ahold of your testing house and have a technical information session with them to determine what testing is appropriate for your device, but the less knowledge you have about your requirements, the more it's going to cost.
All of this must be documented or it didn't happen, and will be tied into your 510k or PMA. When you go for your electronics regulatory inspection, have all of your boxes complete, checked and rechecked or you will get torn apart when you go back to be reinspected.
(A medical device takes about $10M and 6 years to get to market. In that time your investors cannot get a profit. Medical VCs are very special people. These are just the rules. Medical devices are hard and must be taken seriously.)
