Reproducibility of videokeratographic digital subtraction maps after excimer laser photorefractive keratectomy

Background: Digital subtraction photokeratography can best identify topographic changes after excimer laser photorefractive keratectomy (PRK). To evaluate the reproducibility of these topographic maps, the authors used a topographic modeling system to generate multiple subtraction maps from different combinations of technically acceptable preoperative and postoperative maps for eyes that underwent PRK. The assigned patterns for each patient then were evaluated for consistency. Methods: Seven hundred twenty-two individual subtraction maps were generated for 64 eyes that underwent PRK. A mean of 11.3 maps were generated for each eye. The topography of each map was individually classified as normal, central island, peninsula, or asymmetric. All maps within a set (consisting of examinations for I patient at a single postoperative interval) then were examined as a unit to determine the overall topographic classification for that set of maps. Each set in which each constituent map had the same topographic assignment as the set was considered 'nonvariant,' whereas those sets in which one or more individual subtraction maps had different topographic assignments were considered 'variant.' Results: Of the 64 sets, 33 (52%) were variant and 31 (48%) were nonvariant. Conclusions: Any one subtraction map produced by the topographic modeling system may not be a reliable indicator of the excimer effect.