If latency is not an issue, get any inexpensive USB SPDIF device and either create an Aggregate Device (which you have to do anyhow) or switch devices when SPDIF is needed. Not sure what exactly "monitoring" refers to here.

Otherwise, the smallest RME solution would be a Digiface USB, but that's only got optical and you may need a converter to coax (no big deal).

Anything else may be overkill (external interface with SPDIF optical or format converters from MADI to AES). What about using the MADI cards' phones output as line outputs?

Regarding "overkill" .. Or - in nicer words - everything can also be a new chance for consolidation and new features.

Replace cards against two (newer) HDSPe MADI FX which (will or do) support Room EQ.

If three MADI channels would be enough, then you could even think about replacing the two HDSPe MADI against one HDSPe MADI FX. This gives you additionally:
- 2x AES I/O
- FX chip
- Room EQ
- the special driver for the card optimizes CPU utilization. Driver/CPU resources will be allocated in chunks of 8 I/O ports and only if at least one port of eight is in use.
- as an option the 3rd port can be changed against an optical module (separately available for purchase).

If you want you could also think of connecting e.g. an ADI-2 Pro FS R BE (or ADI-2/4 Pro SE) for monitoring via AES port.
They also offer coax SPDIF in case it would still be needed for whatever purpose. The monitoring I would perform through the ADI-2 ...
USB connectivity would only be needed in case you want to use the new AD-2 Remote software to control the device or for flashing a new firmware. Otherwise pure stand-alone mode would be sufficient.

HDSPe MADI card could be sold to cross-finance. You should still get around 70-75% of street price for them.