No.  You should study patients with abdominal pain and suspected appendicitis, since this is who you will be ordering the test on.  Using only patients with definite disease and those with completely normal abdomens is too easy, and makes the test look unrealistically good!