Personalized treatment of primary angle closure

Abstract

PURPOSE. To develop a personalized approach to selecting a treatment method for primary angle closure (PAC) based on a comparison of the predicted hypotensive effect of laser peripheral iridotomy (LPI) and lens extraction (LE).METHODS. This prospective study included 60 patients (60 eyes) with PAC aged 41 to 80 years. LPI was performed on 30 eyes and LE on 30 eyes. All patients underwent swept-source optical coherence tomography (SS-OCT) of the anterior and posterior segments of the eye before and 1 month after treatment. A set of 37 clinical and anatomical parameters of each patient was evaluated. The hypotensive effect of the surgery was considered to be the magnitude of the decrease in intraocular pressure (IOP) after the intervention relative to the baseline. The principal component regression method was used to develop regression models predicting the hypotensive effect of treatment.RESULTS. An innovative step-by-step algorithm for choosing a treatment method for PAC (LE or LPI) has been developed. The first step is to measure and take into account 4 clinical and anatomical parameters of the patient: gender, IOP, axial length of the eye (AL), and anterior chamber depth (ACD). The second step is to calculate the predicted difference in hypotensive effect (Ind_Short) using the formula Ind_Short=B0+B1·Gender+B2·IOP+B3·AL+B4·ACD, where B0=16.8; B1=-0.28; B2=0.24; B3=-0.65; B4=-2.36; male gender is 0 (zero) and female gender is 1. The third step is to compare the obtained result with zero: LPI is the preferred method if Ind_Short is less than 0, otherwise LE.CONCLUSION. The developed step-by-step algorithm involving the use of machine learning methods allows a personalized approach to the treatment of primary angle closure.