Limitations of pupil tracking in refractive surgery: systematic error in determination of corneal locations

The goal of this investigation was to show the theoretical limitations of pupil tracking in refractive surgery. The parallax error associated with localizing corneal positions by tracking the subjacent entrance pupil center was quantified.

Optical ray-tracing in a schematic model eye was performed to determine the geometric parallax error. The calculations required several assumptions regarding ocular geometry, eye movements, and eye tracker position. Various parameter combinations were evaluated to assess the potential range of error to be expected in clinical practice.

Tracking error can amount to 30% (or more for eye trackers mounted closer than 500 mm to the eye) of the detected lateral shift. Thus, if the eye tracker registers a lateral shift of the entrance pupil of 0.2 mm away from the tracking reference axis, the point of interest located on the cornea would essentially be 0.26 mm away from this reference axis. A laser pulse fired at that moment would be systematically displaced by 60 microm. Our results depended on geometric parameters of the eye and the tracking device. Based on conservative assumptions regarding these geometric parameters, partial compensation could be realized by adding a certain percentage to the modulus of each eye tracker reading.

The fact that corneal displacement was generally underestimated by up to 30% of the measured entrance pupil shift demonstrates the severity of the parallax effect.