Just use set instead of a priority queue. It has the same functionality (keys are in sorted order), you can get first key by using *myset.begin(). And if you want to update some key called "key", then you just erase the key using, myset.erase(myset.find(key)), then add the updated version of the key — called "updated_key" — by using, myset.insert(updated_key). This is what I use when i implement dijkstra's algorithm.

you can also use __gnu_pbds::priority_queue: it has modify(iterator, value)