I just did an experiment. I converted 118 Canon raw .CR2 files to DNG using Adobe DNG converter. These were images of flowers (lots of different, subtle colours), birds and kites (the things flown on the ends of pieces of string, not the birds). I imported all 236 raw and DNG files into Lightroom using default parameters and then exported them to JPEG 1300 pixels high. I then compared the pairs in Faststone Image Viewer, with the second image of each pair exactly overlaying the first one. This is the technique I use to identify differences between differently processed images, and the technique makes differences jump out at you, including quite minor differences.
I did not notice any differences in geometry, colour, tonal distribution or sharpness/detail in any of the pairs.
I then converted 73 Panasonic raw .RW2 files to DNG. These were images of flowers and invertebrates. I imported all 146 raw and DNG files into Lightroom. I wanted to see if they responded differently to processing and so processed them all as follows (these are not necessarily settings I would actually use btw, I was just using some non-trivial adjustments to see if I could induce a different response as between raw and DNG Versions).
I first used Autotone. This often increases the brightness more than I like, so I decreased the Exposure by 1 stop. Then Highlights -83, Clarity +21, Vibrance -12, Sharpening 100 Radius 1.0 Mask 60. They all looked rather dark at that stage, so I increased the Exposure back up by one stop. I then exported to JPEG, full size (compression 80) with Output sharpening Screen Standard. I then compared the pairs in Faststone at 100%, for each pair choosing a different position on which to centre the comparison, with the positions varying from around the centre to out towards the edges, and sometimes choosing subject areas and sometimes background or midground areas.
It was not a pretty sight (the images were from a noisy, small-sensor camera, without any luminance noise reduction and oversharpened for 100% viewing), but here too I did not notice any differences in geometry, colour, tonal distribution or sharpness/detail in any of the pairs.
It might not be the same for NEF images of course, but I wouldn't be surprised if it was.
If anyone would like to check whether I have missed differences I will upload all 384 JPEGs to Flickr.
btw, previous experiments have shown that if I use the same slider settings for a raw file in ACR in PSE14 as with Lightroom I get what appears to me to be exactly the same result.