Though low in number, massive stars play an essential role in the early and present Universe, due to their enormous energy output and production of metals. Moreover, the end-products of massive star evolution are among the most spectacular events/objects in our Cosmos: Supernovae, (long-duration) Gamma-Ray Bursters and Black Holes. Unfortunately, the evolution of massive stars is not as well understood as required for reliably predicting the different evolutionary channels, and problems already exist during the Main Sequence, related to mass loss and rotation. To constrain present evolutionary predictions, the observational analysis of CNO surface abundances is a prime tool, giving insight into rotationally-induced mixing processes and angular momentum transport. Such analyses, however, are hampered by many difficulties, increasing the measurement uncertainties significantly, compared to the low-mass case. In this talk, I will outline these difficulties, also with focus on problems arising from physically complex line transitions, and discuss the results and implications from more recent abundance studies of hot massive stars.