Abstract
Purpose: :
Refractive surgery aims to improve quality of life and requires high safety profile. Photorefractive Keratectomy (PRK) has already been shown as safe and efficient for these indications. The variations in interaction between the laser and tissue and the different healing process among patients makes it difficult to predict the exact outcomes. In order to reduce that predictive error the excimer laser manufacturers provide prediction systems based on data analysis of previous results (nomogram). The disadvantage of the standard, non-customized, prediction system is that parameters that are relatively fixed in a given system can not increase precision. Our goals were (A) to create a customized prediction system for PRK in a given system and (B) to check the theoretical differences in prediction accuracy between customized and the standard prediction systems.
Methods: :
The prospective cohort study included 78 patients (150 eyes) which underwent PRK and were followed up to 6 months postoperatively. After data collection, the patient files were randomly divided into 2 groups: 1) a group whose data we used for the creation of the prediction system; 2) a group for theoretical validation of the new system. The actual operations were done according to the surgeons' decision and not according to the new system.
Results: :
We found a significant theoretical improvement using a customized prediction system vs. using the standard system. The amount of the rediction errors in spherical equivalent effect of more then ±0.5D was reduced by 41.92% (from 38.30% using the manufacturer prediction system to 22.22% using the personalized predication system). For the sphere effect the quantity of errors larger then ±0.5D was reduced by 46.17% (from 27.66% to 14.89%). The cylindrical prediction showed better accuracy when measured both by magnitude and vector analysis.
Conclusions: :
The personalized prediction system is significantly more accurate than the standard prediction system. A prospective clinical study is required to further evaluate and validate this method. When completed, the personalized prediction system could be used to improvefuture surgical results. The system and the method can be applied to other refractive facilities, to improve their results.
Keywords: refractive surgery: PRK • refractive surgery: optical quality • computational modeling