There is no definitive test for EPM; even post-mortem it can be difficult to find. A large number of horses carry antibodies to EPM even if they are not clinically ill; so it can be difficult to assess the results of the tests.
The latest test that we are using at UF is an ELISA test developed at the University of Kentucky, that compares antibodies in the serum vs. the CSF fluid and comes up with a ratio. It is based on the response of the horse’s immune system to the immunogenic proteins on the surface of the protozoa, SAG 2,3,4. Apparently all of the ‘subtypes’ of protozoa have these proteins; while only 50% carry SAG 1 and a separate 50% carry SAG 5. Avoid tests that only look for SAG 1 or SAG 5 for this reason. By taking a ratio of the serum vs. the CSF, it can be determined if the antibody levels are high enough to indicate clinical disease.
Here’s the only link I can really find for it: http://www.equinews.com/article/epm-diagnostic-test-developed-for-horses
There is also the IFAT from UC Davis: http://www.vetmed.ucdavis.edu/ceh/docs/horsereport/pubs-HR27-2-bkm.sec.pdf
IIRC the issue with this one is that you can get immunofluorescence to subtypes that do not cause disease; therefore there is a high rate of false positives.
Marquis is still the gold standard in treatment; there are other options, such as diclazuril or sulfadiazine-pyrimethamine, but each treatment has its pros and cons.
Prognosis: it depends on the horse. The ones that do best are caught early and treated quickly. Those that have been lingering for months or years without treatment do not do as well. Either group can relapse with no warning.