If the appropriate timely surgery is not performed on someone who has

Such a great question--it requires care with the exact terms of the conclusion, as a lot of the hardest CR questions do, so start there (OR ELSE): the conclusion promises that doctors can largely (but not necessarily completely) avoid unnecessary appendicitis surgeries without performing any fewer necessary surgeries. The issue with the premise supporting the conclusion is that it says that the new test only correctly diagnoses appendicitis in 98% of the cases. This raises a question: if the test isn't perfectly accurate, should you trust its results at all? How can you guarantee that all necessary surgeries still get performed?

When it comes to accuracy in testing, it's important to know whether the test, when wrong, will yield a false positive result (that is, the person is not sick but the test says he or she is), a false negative result (the person is sick but the test says he or she isn't), or both at various times. If you need this test to ensure what the conclusion promises--that even though the test is only 98% accurate, doctors will still not perform any fewer necessary surgeries--then the test needs to diagnose 100% of the people for whom the surgery is necessary. That means that the test needs to show a positive result every time a person has appendicitis. Of course, you know that the test is only 98% accurate, so the only way to know for sure that nobody who needs the surgery will fall through the cracks is if the 2% of inaccurate tests are all false positives; that is, the test correctly catches all cases that are actually positive, but in about 2% of cases, it also comes out positive when the person in fact does not have appendicitis. That will lead to some unnecessary surgeries, to be sure, but go back to the conclusion: it only promises that unnecessary surgeries will be largely reduced. That will still be achieved in this case, as this test will reduce the current 20% rate of unnecessary surgeries to presumably something that's 2% or less.

A. the misdiagnoses produced by this scan are always instances of attributing appendicitis to someone who has not had it
This answer reflects what was said above: when the test is wrong, it's a false positive every time. This is the only way to ensure that nobody who needs the surgery will have been misdiagnosed as not needing the surgery. Winning time.

B. the misdiagnoses produced by this scan are never instances of attributing appendicitis to someone who has not had it
This is the opposite of (A): if the misdiagnoses are never false positives, then they must always be false negatives. These false negatives, if taken as true negatives, will lead to necessary surgeries not being performed, which contradicts the conclusion. No thanks, (B). Get lost.

C. all of the patients who are diagnosed with this scan as having had appendicitis have actually had appendicitis
Like (B), this answer requires a little bit of inferring/deduction to understand. If all of the patients whom the scan diagnoses as positive are in fact positive, the question becomes who's caught up in the 2% of tests that are inaccurate. All positive tests are correct, so inaccuracy must come in some of the negative tests. And for a negative test to be inaccurate implies that the person being tested should have gotten a positive result rather than a negative one. These people will sadly slip through the cracks in this case, and their appendices might kill them (thanks for nothing, horror test). More importantly, this means that doctors will be performing fewer necessary surgeries than they should, so this hurts the conclusion.

D. every patient who is diagnosed with this scan as having had appendicitis always has more than one of the symptoms associated with appendicitis
Cool, (D). Cool. This is one of those answers that drops buzzwords from the argument (diagnosed, symptoms, appendicits) but doesn't say anything important with them. Whether the people have symptoms at all (leaving aside the more than one part for the moment) is irrelevant to this conclusion--you just need the test to be as accurate as possible, even in the absence of symptoms. And there's certainly no reason to believe that people who show up positive on the test should always need to have more than one of the symptoms anyway.

E. the only patients who are misdiagnosed using this scan are patients who lack one or more of the symptoms that are generally associated with appendicitis
You do want to know which patients are diagnosed by the scan; however, you really want to know whether the people who are misdiagnosed actually have appendicitis or not. Knowing whether they have symptoms or not isn't all that useful--the whole point of the scan is for it to be a more accurate marker of appendicitis than the symptoms were anyway, and part of the implication of this whole setup is that we can't really trust the symptoms to begin with. So explaining the accuracy of the scan in terms of how it related to the symptoms displayed doesn't really move the story forward at all.