Abstract
Purpose :
To develop an automated algorithm to predict the success probability of laser peripheral iridotomy (LPI) in eyes with primary angle closure suspect (PACS), using anterior chamber angle (ACA) characteristics of pre-treatment anterior segment optical coherence tomography (AS-OCT) scans.
Methods :
A total of 76 eyes with PACS underwent LPI and time-domain AS-OCT scans (temporal and nasal cuts) were performed before and 1 month after LPI. All the post-treatment scans were graded by a trained ophthalmologist to one of the following categories: (a) both angles open, (b) one of two angles open and (c) both angles closed. After LPI, success is defined as one or more angles changed from close to open on the AS-OCT images. Only AS-OCT scans with sufficient quality and if the scleral spur can be identified were selected for analysis. All the pre-treatment ASOCT scans were analyzed using the Anterior Segment Analysis Program to derive 76 ACA measurements which serve as features for a subsequent prediction algorithm.
Results :
In total, we included 76 eyes and 50 (65.8%) eyes fulfilled the criteria for success after LPI. Among the 76 ACA features, three features including iris thickness at the dilator muscle region [measured at half of the distance between the scleral spur and the pupillary margin], iris thickness measured at 2000 um from the scleral spur and trabecular iris space area (TISA 500) had the highest predictive score and they were selected using correlation-based subset selection method. These features were classified into two (‘successful’ and ‘unsuccessful’) categories using a Bayes classifier. The success of LPI in eyes with narrow angles can be predicted with 77.9% cross validation accuracy.
Conclusions :
Our study showed that using pre-treatment AS-OCT scans, an automated algorithm can show reasonable accuracy in predicting the success of LPI in improving ACA parameters in PACS eyes. Further analysis using newer AS-OCT parameters and algorithm could potentially guide ophthalmologists in deciding when to offer LPI as a prophylaxis for PACS
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.