PURPOSE: To compare optic disc parameters when using manual disc detection versus automated software on RTVue-100 Fourier-domain optical coherence tomography. METHODS: Nonrandomized comparative study using images obtained from a tertiary-care Glaucoma Clinic from August, 2008 to May, 2009. Four hundred thirty-four patients with mean age of 67.1 years (±13.4) underwent OCT imaging with RTVue-100. Optic disc margins were manually defined by a single, trained operator and redefined with automated software that estimates the position of the retinal pigment epithelium edges. Bland-Altman plots and intraclass correlations measured agreement between the 2 detection methods for various optic disc parameters. RESULTS: Bland-Altman plots/intraclass correlations showed good agreement in the following optic disc parameters: optic disc area, cup area, rim area, cup-to-disc area ratio, cup-to-disc ratio, and retinal nerve fiber layer thickness. There was no significant difference in measurements when comparing hand versus automated disc contour determination (P>0.05). CONCLUSIONS: Automated determination of the disc margin was in high agreement with subjective, manual assessment and should be used in clinical practice as it can make optical coherence tomography scanning more efficient and standardized, as disc detection would only be done by the software and not biased by multiple operators.
PURPOSE: To compare optic disc parameters when using manual disc detection versus automated software on RTVue-100 Fourier-domain optical coherence tomography. METHODS: Nonrandomized comparative study using images obtained from a tertiary-care Glaucoma Clinic from August, 2008 to May, 2009. Four hundred thirty-four patients with mean age of 67.1 years (±13.4) underwent OCT imaging with RTVue-100. Optic disc margins were manually defined by a single, trained operator and redefined with automated software that estimates the position of the retinal pigment epithelium edges. Bland-Altman plots and intraclass correlations measured agreement between the 2 detection methods for various optic disc parameters. RESULTS: Bland-Altman plots/intraclass correlations showed good agreement in the following optic disc parameters: optic disc area, cup area, rim area, cup-to-disc area ratio, cup-to-disc ratio, and retinal nerve fiber layer thickness. There was no significant difference in measurements when comparing hand versus automated disc contour determination (P>0.05). CONCLUSIONS: Automated determination of the disc margin was in high agreement with subjective, manual assessment and should be used in clinical practice as it can make optical coherence tomography scanning more efficient and standardized, as disc detection would only be done by the software and not biased by multiple operators.
Authors: Alexander Pinhas; Rachel Linderman; Shelley Mo; Brian D Krawitz; Lawrence S Geyman; Joseph Carroll; Richard B Rosen; Toco Y Chui Journal: PLoS One Date: 2018-05-24 Impact factor: 3.240