I agree with Fred, the logic says a Gleason score could never go down. To have a G7 biopsy there must have been at least one tiny bit of Gleason pattern 4 in the samples. In the path lab they can now examine all of the prostate, they 'must' see that area of pattern 4 again so it cannot go down, and if they happened to see some pattern 5 it would go up. You might say well if they spotted a lot of pattern 1 or 2 that would bring it down? No, Gleason score is about the maximum and secondary maximum in two samples it is not about averages.
So now if we follow the mathematical logic, we have to ask. Was the pattern 4 in one place and tiny, and completely removed by the biopsy needle? Extremely unlikely, about 0% chance that that is the reason. Maybe the cancer just got better? Again extremely unlikely, yes the immune system attacks cancer, but only in the early stages once it has got hold it won't die on its own, about 1% chance this is the explanation.
So our maths and logic has not given us the answer.
Now if you go on the internet and look up pictures of biopsy samples you will see how ordered Gleason patter 1 is and how disordered pattern 5 is. If you then look at more samples you can start saying what you think the pattern is, and then check with what the expert says. 80% of the time you will agree with the expert on what a pattern 2,3 or 4 looks like, but some will be marginal, maybe pattern 3 maybe pattern 4.
So I would guess that all the samples from biopsy and pathology looked about the same and it was a bit more disorders than patten 3 but not quite as bad as a typical pattern 4. One pathologist decided to call a 4 the other called a 3. Hence Gleason score 8 and 6, because of the difference in the person looking at the sample not the samples themselves.