1. Resolution. They measure the number of "line widths per picture height." This tells us how much detail the lens resolved from the sensor's perspective. Hypothetically, a 6MP sensor would only resolve half the detail of a 12MP sensor of the same size with the same lens, so if you test with a 12MP camera, the number of LW/PH will be higher than on a 6MP sensor. For reasons that I don't understand well enough (yet) to explain it to someone else, this doesn't scale linearly, so you can't take the test results from a 6MP sensor and simply double it to get comparative results for a 12MP. For one thing, the lens might not be good enough, but there are other factors.
That's why I said you can look at the trend, i.e. look at he bar graphs and check where the resolution falls in the bad/average/good/excellent bands
2. De-centering roughly means that the image doesn't fall entirely level on the sensor (I may have this slightly wrong, but that's the effect). At smaller apertures, where you have more DOF, this may not matter too much, but at f/2.8 on a tele lens, this can be a significant issue. That's why I mentioned that, in that review, they're getting a resolution drop at f/2.8, but not really at smaller apertures, and they specifically mention the de-centering, so to my mind, that may be the cause (given that the lens returns excellent resolution otherwise). I'm speculating, of course, but I've seen shots from that lens and I've seen other reviews, and it's very good.
De-centering can happen to any lens by any manufacturer. My Canon 28-200mm has it (which is why I don't use it). My Sigma 10-20 got de-centered after about six months, most likely due to my (mis)handling of it. Sigma replaced it under warranty, no questions asked.