When you take an LSAT containing even a few questions that you've already seen before, your resulting score will be inflated. Your score will be an overestimation of your true performance.
There are at least two factors that play a role in this. First, that you've seen a question previously increases your chances of getting that question right while spending less time on it. Second, as a flow-over of the saved time effect, you now find yourself with more time to spend on the other questions, thereby increasing your chances of getting those right too.
Even if you've seen just one question from the whole test, that will still have an inflating effect on your score. You have to properly adjust your score down to reflect your true performance. Of course, there is no way to accurately do that. You just have to say to yourself "well, crap - I had already seen 3 LR questions from this test and I got a 164, so my real score is lower than a 164." How much lower? No one knows. You would have to make too many uncertain assumptions to answer that question. Just be okay with the information that your real score would have been lower. Having said that, the diagnostic is not at all useless. At least it sets an upper bound on what your score would have been had you not seen any of the questions before. There is, of course, also the Blind Review that you can still do for that test.
We are paranoid about ruining recent LSAT PrepTests for students because we know how important taking clean, timed, practice LSATs are for students to improve their score. This is why we save the recent LSAT PrepTests and provide them to you to use as diagnostics, along our video explanations for each of the questions on the test. This is why we confine our curriculum (where we cover LSAT theory and explain how to approach various question types) to only LSATs from PrepTest 35 (October 2001) and lower. Everything higher than PrepTest 35 is fair game for timed, practice LSATs.