Greeks were cooking with gas (from the paper): “However, we note that the degree of manufacturing precision is remarkable, with standard errors in hole positions of only 0.028mm radially, and 0.129mm azimuthally. Budiselic et al. quote a standard deviation for their individual position measurements of 0.037 mm, so a good deal of the radial error may come from the measurements of the X-ray images themselves.”
As someone who spent a decade doing precision metrology with optical devices... there isn't a way to correctly measure that part with the precision they indicate in the measured values.
If I recall the precision is limited to by the wavelength of the light being used. If they are using X-Rays in the 0.01 nm to 10 nm range wouldn't that give them more precision then optical light in the 380 nm to 700 nm range?