Interesting. I'm glad to finally see some direct comparisons (I don't have a DVD burner or I would have done so myself).
It is surely too early to say with any certainty, but it looks as if the differences may not be systematic like they were with the C1/C2 scans. Every C1/C2 scan performed on a DVD-ROM drive that I saw dramatically under-reported the error counts, leading me to conclude that the DVD-ROM drives have some systematic problem that prevents them from properly reporting C1/C2 error levels.
However, in your two examples the "better" reader changes positions with the two different discs, and although the average levels are fairly different, the maximums aren't different by orders of magnitude. I would say it is possible that what you have shown may simply be attributable to the differing ability of each drive to read different discs. Thus, the DVD-ROM drives may be just as "accurate" as the burners for PI/PO testing: although they show different error levels, they may be showing the correct levels for each respective drive. In my mind this is different than the C1/C2 issue with respect to DVD-ROM drives where I feel that the reported numbers were lower than what the drive's internals actually encountered.