@Mary Miller good questions and it's clear you're thinking about this the right way rather than just winging the layout.
On the matching question — the RF front-end on the MultiNav Pro+ is designed to be self-contained, meaning the LNA and PA are already handled internally with the matching done on-module. You're not in nRF52 bare-die territory where you need to sweat an external matching network. That said, you still need to treat your feedline seriously. A 50-ohm microstrip from the module's antenna port to your connector or antenna feed point is non-negotiable — any impedance discontinuity there will hurt you, and at 50km you're already working with a link budget that doesn't have much headroom to throw away. Get your trace width calculated properly for your stackup, keep it short, and avoid vias in that run if you can.
For external filtering at 2.4GHz, I wouldn't add a SAW or bandpass filter unless you're operating in a genuinely hostile RF environment (like next to a WiFi access point on a drone). The module's own front-end selectivity should be adequate for typical deployments, and adding a filter you don't need just adds insertion loss right where you least want it.
On thermal — at full TX power you will notice some warmth on the module, but in my experience it's not in the range where you need to treat it like a power amp with a heatsink. That said, if your enclosure corner is tight and you're stacking the sensors module close, I'd keep at least a few mm of airspace and make sure your ground pours have good copper fill around it rather than isolating it — better to conduct heat away than trap it. The sensors module doing temperature measurements right next to a warm RF PA is the more practical concern there honestly, so if you can orient them with a bit of separation or even a small cutout in the PCB that's worth considering.
For reference designs — best move is to post your carrier board outline in the Dev Hub and tag Moamen directly, there may be layout notes that haven't made it into the public docs yet.