The question is: is the signal-to-noise ratio good enough? If so a cheap cable that passes the data is every bit as good as an expensive one, so long as the packets arrive intact at the other end.
Ethernet already does a lot to counter noise. The signals are differential pairs (so instead of having ground and signal, you have signal+ and signal-). The wire pairs are twisted, which keeps them in close proximity. Interference tends to be common mode noise (so for two wires close together it will affect the signal in each wire almost the same), and differential amplifiers are designed to only amplify the difference between the two wires and will therefore reject common mode noise. Each end also has an isolating transformer, and each end has proper termination (to avoid things like reflections which can bugger up signal integrity). It takes a significantly terrible out-of-spec twisted pair cable to make ethernet stop working.
Incidentally, the signalling for 100baseTX ethernet only has a fundamental frequency of 31.25MHz (naively people would expect 1MHz per 1Mbps but this is not so). 100baseTX uses a 3 level (in other words +1, 0, -1) non return to zero signalling (in other words, a 1 will cause the signal to change level and a 0 will cause the signal to remain at the current level - or it might be the other way around, it's a long time since I did this stuff). Each 4 bits is encoded into a 5 bit symbol designed to prevent long runs of 0s (which would cause the signal level to remain constant for too long). Lots of people call an ethernet connection a "broadband" connection, but it's not, it's baseband (hence the "base" in 100baseTX).