I've seen an extraordinarily good, very large (30"x40"?) optical enlargement from a disc negative - it was prepared for the retirement party of a Canadian Kodak employee who retired just a few months short of his 50th anniversary with the company. Disc film and cameras were current at the time.
It required just about every trick (masking and ??) that Eastman Kodak had available in their Rochester labs, but it wasn't fake or the result of any interpolation.
I expect it involved some of the approach that was used for the Colorama photos from 35mm that were displayed in New York's Grand Central station.
The problem with any of there "how many megapixel" questions is that they require you to take the output from one process and procedure, convert it to the output of another process and procedure, and then evaluate the resulting conversion in relation to the normal, native output from the latter process and procedure. The conversion steps are critical, and inevitably distort the comparison.
If you are going to be printing digitally, then it is relatively tough (but still possible) to achieve high quality from small film. If you are going to be printing optically, it is relatively easy to achieve high quality.
It is, however, also relatively easy to end up with lousy results - film or digital it doesn't matter - if your technique and equipment are lousy.
I don't often print both optically from a negative and print digitally from the same negative. When I have done that though, the results tend to be more different in qualities than different in relative quality.
One thing I am sure of though - anyone who can only extract the same quality from a 35mm negative as they can from a 4 megapixel sensor either is lousy with film, or lousy with digitizing film, or has a really lousy testing methodology.