PURPOSE: To assess the effect of signal strength (SS) on concordance rate or agreement on retinal nerve fiber layer (RNFL) thickness measurement (measurement agreement) and its color code classification (classification agreement) between Stratus optical coherence tomography (OCT) and Cirrus OCT. PATIENTS AND METHODS: From 300 participants who underwent both Stratus and Cirrus OCT scans at the same visit (SS ≥ 6), intraclass correlation and weighted κ coefficients were calculated as indicators for RNFL measurement agreement and classification agreement between 2 OCTs. RESULTS: In inferior and superior quadrants, agreements were excellent (intraclass correlation) or good (κ) and neither the measurement nor the classification agreement was statistically different between participants grouped by SS difference between Stratus and Cirrus OCTs (inter-OCT SS difference). In nasal and temporal quadrants, agreements were low and there were groups with inter-OCT SS difference whose classification agreement between 2 OCTs was not better than the chance (κ=0). CONCLUSIONS: When comparing RNFL thickness estimated by Cirrus OCT with that measured by Stratus OCT in temporal and nasal quadrants, special attention should be given, as the agreements or concordance rate was low, suggesting that neither the measured value nor its probability algorithm classification would be interchangeable. In inferior and superior quadrants, agreements were good or excellent and not affected by SS difference as long as SS ≥ 6.
PURPOSE: To assess the effect of signal strength (SS) on concordance rate or agreement on retinal nerve fiber layer (RNFL) thickness measurement (measurement agreement) and its color code classification (classification agreement) between Stratus optical coherence tomography (OCT) and Cirrus OCT. PATIENTS AND METHODS: From 300 participants who underwent both Stratus and Cirrus OCT scans at the same visit (SS ≥ 6), intraclass correlation and weighted κ coefficients were calculated as indicators for RNFL measurement agreement and classification agreement between 2 OCTs. RESULTS: In inferior and superior quadrants, agreements were excellent (intraclass correlation) or good (κ) and neither the measurement nor the classification agreement was statistically different between participants grouped by SS difference between Stratus and Cirrus OCTs (inter-OCT SS difference). In nasal and temporal quadrants, agreements were low and there were groups with inter-OCT SS difference whose classification agreement between 2 OCTs was not better than the chance (κ=0). CONCLUSIONS: When comparing RNFL thickness estimated by Cirrus OCT with that measured by Stratus OCT in temporal and nasal quadrants, special attention should be given, as the agreements or concordance rate was low, suggesting that neither the measured value nor its probability algorithm classification would be interchangeable. In inferior and superior quadrants, agreements were good or excellent and not affected by SS difference as long as SS ≥ 6.
Authors: Margaret R Strampe; Luai Eldweik; Benjamin C Chaon; Sarah Maki; Tyler Wieland; Celine Satija; Collin McClelland; Michael S Lee Journal: Neuroophthalmology Date: 2019-10-29