I always forget the names of the tests, but basically the in-house one has a high percentage of false positives, the one done by the lab is the most accurate. The reason is that cats can come into contact with FeLV, but shed the virus, if they do that, they will be immune for life - it can take up to 3 months for them to shed the virus, although you don't know when they came into contact with it, so it could have been 2 months before testing. I'm fairly sure that if the lab one is negative, they are negative, it's a very low percentage of false negatives
I can't find the Uni website I used to use, but this might help
When a virus or any other foreign invader tries to enter the body, the immune system will produce several types of antibodies. These attack the invader and try to damage it so that it does not become overwhelming. For the leukaemia virus, this process (immunity) takes about four to 12 weeks. During this time the virus is only in the bloodstream. If immunity does not develop, it establishes in the bone marrow and the cat is infected for life.
The simple “in house” test is called the p27 test because it looks for a fragment of virus called the “p27 antigen”. This appears in the blood following infection. If a healthy cat tests positive, then repeating the test increases the accuracy. Sending a sample to a commercial laboratory for a different test makes it more accurate still. A cat that tests positive on two “in-house” tests is likely to be a true positive, but ideally this should still be confirmed by commercial testing.
However, testing while immunity is developing can show a cat in the process of overcoming infection as positive for FeLV, even if the cat is eventually going to develop immunity. Because of this, a single positive test result is not proof of infection in a healthy cat. If necessary, a special test can be done on healthy “positives” to look for antibodies of a type likely to be effective in providing immunity. These are called “virusâ€‘neutralising antibodies”.
False negatives can also occur occasionally, in which the test is clear when your cat is actually infected. To be truly confident that a healthy cat does not have the virus, a retest should be done after a 12-week interval. This detects those in the early stages of infection at the first test, which did not then develop immunity. A false negative can sometimes occur in a sick cat as well. If there is clinically a strong suspicion of FeLV infection in a “negative” sick cat, commercial testing may be worthwhile.
Occasionally, a cat is persistently positive on the p27 test but negative on the commercial tests. It is thought that some of these cats may have “latent” infection, with the virus hidden in their body (usually in the bone marrow). As long as the commercial tests are negative, they are not a risk for other cats, but the situation should be monitored.