Segmentation Errors of Retinal Nerve Fibre Layer and Implications in Glaucoma
Dr. Chennamsetty Alekhya, Dr. Vinay Nangia, Dr. Sarang Lambat
Suraj Eye Institute, Nagpur, India
Glaucoma is a progressive optic neuropathy, characterized by the loss of retinal ganglion cells (RGCs) and their axons, resulting in thinning of the neuro retinal rim of the optic nerve head (ONH) and retinal nerve fiber layer (RNFL) as well as the development of visual field (VF) defects.1 The RNFL thickness is an important parameter along with the optic disc to be considered while assessing, diagnosing, managing and following up a case of glaucoma. Spectral Domain Optical Coherence Tomography (SD OCT) imaging is used for quantitative analysis of RNFL thickness. Inbuilt software algorithm in SD OCT detects anterior and posterior boundaries of RNFL based on the change in reflectivity at each of these interfaces. After segmentation, two- dimensional thickness maps are generated. If the software fails to delineate the retinal boundaries correctly (i.e. segmentation break- down), it will result in thickness measurement error which may lead to incorrect diagnosis of glaucoma.
A male, 62 years of age, reported to us with the chief complaints of diminution of vision in both eyes since 5 years. He had a Best Corrected Visual Acuity (BCVA) of CF 3m, N36 in both eyes. Slit lamp biomicroscopy showed early cataractous changes in right eye and immature senile cataract in Left eye. Intraocular pressure was 16mmHg in both eyes. Fundus examination showed 0.75 vertical cup disc ratio (VCDR) and inferior rim thinning in both eyes (Fig.1). SDOCT showed severe RNFL thinning in inferotemporal quadrant, moderate RNFL thinning superotemporally and globally in right eye (Fig.2). Left eye SDOCT showed severe RNFL thinning in inferotemporal quadrant, moderate RNFL thinning temporally and globally (Fig.3). Automated segmentation errors were noted in left eye (Fig. 3). After manual correction of segmentation error, it showed only inferotemporal RNFL thinning in left eye (Fig.4). The patient was advised Brimonidine eye drop 0.2% twice daily based on baseline intra ocular pressure and the extent of RNFL thinning in both eyes.
OCT is currently the most precise and reliable imaging technique for RNFL thickness measurements.2 Because of it’s ability to quantify changes in RNFL thickness, and good reproducibility, OCT helps us to monitor progression of RNFL thinning. However, automated image segmentation algorithms can fail to accurately delineate the layers of the retina, and 19.9–46.3% of scans contain at least 1 segmentation artifact.4 These artifacts have been associated with several findings, such as image decentration, epi-retinal membranes, long axial length, poor visual acuity, cataract and advanced glaucoma. So, there is need for inspection and manual refinement of automated segmentation errors.
Correct assessment of the RNFL in glaucoma patients may improve our ability to monitor for progressive RNFL thinning. The awareness about possible segmentation errors is important. Recognition of measurement errors and their correction will improve the accuracy of OCT interpretation of RNFL thickness and should be an integral part of OCT scan analysis.3
- Artes PH, Chauhan BC et al. Longitudinal changes in the visual field and optic disc in glaucoma. 2005; 24(3):333–354. DOI:10.1016/j.preteyeres.2004.10.002
- Georgieva et al. Optical Coherence Tomography – Segmentation Performance and Retinal Thickness Measurement Errors. European Ophthalmic Review, 2012; 6(2):78 82 DOI:10.17925/EOR.2012.06.02.78
- Aojula et al. Segmentation error in spectral domain optical coherence tomography measures of the RNFL thickness in idiopathic intracranial hypertension. BMC Ophthalmology (2017) 17:257 DOI 10.1186/s12886-017-0652-7
- Mansberger et al. Automated Segmentation Errors When Using OCT to Measure RNFL Thickness in Glaucoma Am J Ophthalmol 2017; 174:1-8. DOI: 10.1016/j.ajo.2016.10.020