Binary trees in js or content query

I’m creating a list of synonyms to apply to a search.
Lets say there will be about 200.
Several of them may apply at the same time.

This might not be so easy to predict, but would it be faster to implement a “alphabetical” binary tree in js, or do a content query for matching synonyms?

I guess serverside caching would be a factor too.

One could cache the binary tree. Searches into the binary tree, or content searches.
Depending upon how many entries there are in the cache, it would at some point be faster not to cache.

200 entries might not sound as much, but lets say 3 words are entered into the search, that would mean 3*200=600 possible matches, and so on.

Some algorithm suggestions: