I am trying to calculate a sample size for a study where we are comparing an xray to actual surgical placement of a device. In essence - we predict where something should end up from the xray and then after the surgery measure to see how far it is from what we predicted.

We want to get a sample size to show accuracy within 1 measurement (units not important)

past literature shows accuracy (to within 1 measurement) of 77%, 63% and 85% in three different studies.

What test do I use to calculate a sample size just with a percentage comparison?