There just aren't many G9/10 cases, maybe 2% or so of all PCa. You can look at many studies, and you rarely find one with a significant result identified for those cases. There just aren't enough to show up in the math.
I've looked at many studies over the years since my case has dominant type 5, though I'm not finding handy links like Tall Allen provides (he probably has a good one on this topic too). But I have on a couple of occasions seen studies identifying primary type 5 cells as a particular risk factor. Looking at this seems a bit like our fascination with horror movies, or amusement parks. We seem to like to scare ourselves in a managed way. We know the movie is just a movie. We know the amusement rides are safe (well, mostly). We know our PCa treatment is all we can do, and looking for better/worse outcomes based on available metrics lets us scare ourselves a bit anyway.
I don't know how much it really means. It's probably a distinction without a difference. There are undoubtedly other genetic characteristics that are more important for the risk of recurrence. The Gleason grade is useful, but I wouldn't lose sleep over the distinctions between any of the grades with type 5 cells. They all merit the "kitchen sink" treatment!
For what it's worth, this is a study where they analyzed the SEER database. It's retrospective, so subject to selection biases and so on. It does analyze a large number of cases in the 4+5, 5+4, 5+5 groups. The primary 5 showed higher hazard ratios than the secondary 5. It's not a great study, and I don't think any treatment decisions would be changed by it.
Risk of Death from Prostate Cancer with and without Definitive Local Therapy when Gleason Pattern 5 is Present: A Surveillance, Epidemiology, and End Results Analysis