There was a paper in which researchers performed viable bacterial count for the microorganism with varying concentrations of the antibiotic to obtain the IC50, they also pointed that IC50 calculated from this method is more precise than MIC.
I don’t know the exact protocol but here is the link for the abstract