These results are actually not very different. Just that in your second scan, the second layer is slightly more difficult to read, higher PIE levels, but still nothing to be concerned with.
Don't be fooled by small differences in quality scores. In @1ECC scans (like it's the case here), a single PIF spike of 4, like in your second scan @3.6GB, makes the QS drop to 90. Without it, it would have been 93. Single spikes are not important when scanning, it's larger areas that are relevant.
The "score" reported by CDSpeed is just an indication. Sometimes it can get in the lower 80's with perfectly good discs, simply because of one or two single spikes.
Your scans are fine, nothing to worry about.