Several STM32 MCUs have an extra "SPDIFRX" peripheral, most of their SAIs (serial audio interface) have an SPDIF TX option (I use that a lot for audio debugging, quite easy to use).
The main problem might be to retrieve and use the SPDIF RX source clock, I have never tried that, but from my experience that might cause the biggest trouble. (I like to use a sample rate converter.)
Which MCU you will finally use might also depend on how heavy the audio processing will be and the sampling rate. That's why I would start with an F7 or H7 Nucleo board.
Connecting optical IO to a Nucleo is not that hard, as it's usually only 3 pins (signal, VCC, GND).
