So i think I came up with a decent solution that works within the constraints of my project environment...
Every node in the graph has a unique uuid value. During the query to delete a particular node we can pass in an empty option {} to the Elastic Search index using the apoc.es.put since we can pass in that particular node's uuid to update the ES index. After the query is run every node that gets deleted has its respective ES index set as an empty object, {}. Then, using curl and the Delete by query API we match every object from that index where the designated required property, uuid, does not exist and delete the those index objects.
I have done some initial testing on the concept and it seems to work pretty darn well. So I figured I would share in case others encounter an issue like this.