There are protein markers that can be identified by PCR. It's hard to detect it in recently-infected trees since the concentration is not very high and maybe the branch or trunk you took the sample from hasn't been affected yet. So the earliest indicators tend to be running the test using insects from the yellow sticky traps. Having infected insects doesn't necessarily mean the tree is infected(if the tree's not been "bitten").
Once the symptoms are severe enough: fruits stay green and don't ripen, tastes like ____, it's a good idea to remove the tree since it's worthless. In CA, they still remove the trees to protect the others, since there are so few cases of HLB. Not sure if they do that in FL still, but it still seems like a good idea, unless the grower is abandoning the business and doesn't want to spend the cleanup money before they leave.
