Here's my comparison. Photo #1 was DSLR scanned, Canon T2i with Magic Lantern and an EF 100 2.8 macro. Tethered to LR the exported stitched, and inverted in PS. I used a lightly modified version of Iamthejeffs technique. It was then saved back into LR as a tiff and balanced with color correction and light levels.
LR5, PSCS6. Bronica S2, Portra 160NC cut down from 70mm rolled into a 220 roll. Expired in the early 00s. C-41 home dev in a wrung out kit somewhere around roll 30, kit recommends 8.
Below is the same photo using NLP and the current LR. Everything was done in LR.
This final one is the NLP exported as a TFF so I can tease just a little bit more out of it.
Of the three I think #2 is the nicest looking one while #3 is probably closest to real life. #1 looks sharper but on zooming in it is quite a bit softer.
Pros of using NLP is everything is done in one program, right out of the inversion the photo is 90% there and you get a sharper image. Cons are cost, you'll need a current computer, and you lose access to LRs native controls unless you export as a TIFF.
Other method I use makes massive files through two programs switching back and forth and skin tones are awful to work with. However I do have more control to a point. And it was free. It also helped me learn the ins and out's so I can take full advantage of NLP. I'd suggest learning alternate methods first and then moving on to NLP when you've pushed the method you've learned as far as it can go.
P.s. Sorry about the massive photos.