advertisement
Abstract #106903 Published in IGR 23-3
Quantification of Hypotony Maculopathy Using Spectral-domain Optical Coherence Tomography
Dumas R; Lacourse M; Kassem R; Lesk MR; Costantino SJournal of Glaucoma 2023; 32: 287-292
PRCIS: We provide a free-to-use, open-source algorithm to quantify macular hypotony based on OCT images. This numerical approach calculates a metric that measures the deviations of the Bruch's membrane from a smooth ideal retinal layer. PURPOSE: Hypotony maculopathy is a recurrent complication of glaucoma surgical interventions, in which extremely low intraocular pressure triggers changes in the shape of retinal layers. Abnormal folds can often be observed in the retina using standard fundoscopy, but OCT is particularly important to appreciate the severity of symptoms at different depths. Despite the need for metrics that could be used for informed clinical decision to evaluate progression and resolution of macular hypotony, algorithms that quantify the retinal folds are not available in the literature or included in clinical imaging equipment. The purpose of this work is to introduce a simple algorithm that can be used to assess hypotony maculopathy from OCT B-Scans and volumes, as well as a free, open-source implementation. METHODS: The pipeline we present is based on a straightforward segmentation of the Bruch's membrane complex. The principal idea of the quantification is to compute a smoothed version of this complex and analyze the deviations from an ideal interface. Such deviations are then measured and added to create a metric that characterizes each OCT B-Scan. A full OCT volume reconstruction is thus characterized by the average metric obtained from all planes. RESULTS: We tested the metric we propose against the assessment of three experts and obtained a very good correspondence, with Pearson correlation coefficients higher than 0.8. Furthermore, agreement with automatic analysis seemed better than between experts. We describe the pipeline in detail and illustrate the results with a group of patients, comparing baseline images, severe hypotony maculopathy, and a variety of outcomes. CONCLUSION: The tool we introduce, and openly provide, fills a clinical gap to quantitatively grade hypotony maculopathy. It offers a metric of relatively simple interpretation that can be used to help clinicians in cases where regression of symptoms is not obvious to the naked eye. Our pilot study demonstrates reliable results, and an open-source implementation facilitates easy improvements to our algorithm.
Full article