Many author guides I am aware of advise to only define something in the abstract, when it’s really needed. For example, Physical Review X advises (boldface mine):
In particular, the abstract and introduction of each article should be written so as to be understandable by a broad spectrum of readers. New terminology should be introduced only when clearly needed. It should be appropriate and, if possible, convey to the reader an accurate impression of its meaning.
Now, you have good arguments to define something in the abstract, but an overzealous editor or referee might still be against it due to interpreting guidelines as rules.
That being said, I have seen new terms introduced in abstracts many times and most of the times this was the better alternative to me – this also applies to interdisciplinary papers relating to medicine. I have rarely seen terms rigorously defined in abstracts though.
Some further remarks on this:
The main purpose of your paper is probably not to define these terms but to analyse the differences between them or similar. Most probably there also is a practical benefit of being able to make this distinction other than clarifying communication. This should be your motivation. Also, you should write your abstract (and paper) from this perspective and not from the perspective of introducing a new terminology – which is what your current abstract sounds like to me.
Do not start with the definition right away, but motivate it first. It’s not new language that is confusing (at least as long as we are talking about two terms only), it’s unmotivated new language. If you motivate the distinction first, the reader can associate something with the terms and thus better memorise them.
By all means, make clear that you are defining something when you do. Nothing is more confusing if you do not know whether something is new terminology or not. The context should make clear that the term is new and, if it’s not unseen in your community, italicise the newly introduced terms.
It’s not important to precisely define your concepts in the abstract – you can do that later. Rather give the main distincitive aspects, so that most readers get the central idea to your distinction. Of course, if the definitions are sufficiently concise, you can use them in the abstract.
To distill this into an example abstract, which does not exactly match your case, if I understood it correctly, but should illustrate the idea:
We show that phenomenon Z can occur in two distinct flavours, which exhibit relevant differences in behaviour in [application]: [definition X], which we denote as X, and [definition Y], which we denote as Y. […]