PURPOSE: To discriminate between subclinical fellow eyes of patients with keratoconus (scKCE) and normal eyes (NE) using swept-source Fourier-domain anterior segment optical coherence tomography (SS-OCT). SETTING: University Medical Center Hamburg-Eppendorf, Germany. DESIGN: Retrospective cross-sectional study. METHODS: NE (both eyes with a KISA% Index <60; N = 213) and scKCE (KISA% Index <60, but an index of >100 in the fellow eye; N = 61) eyes were grouped according to the KISA% Index. Topography and tomography data were obtained by SS-OCT (Casia SS1000) and reviewed retrospectively. Receiver operating characteristics were performed for Casia SS1000 integrated automated parameters and newly computed indices. Additionally, discriminant functions were calculated for the automated single parameters, the computed indices, and the combination of both. RESULTS: An accuracy ≥0.7 was reached for 28 (8 automated and 20 newly computed parameters) of 785 parameters. Fourier analysis posterior higher-order irregular astigmatism revealed the highest accuracy with 0.81. By combining automated parameters with the newly computed indices in 1 discriminant function, an area under the curve of 0.92 with a corresponding accuracy of 0.85 was achieved. CONCLUSIONS: SS-OCT is able to discriminate between NE and scKCE with high accuracy. Automated single parameters and newly computed indices were combined to further improve this accuracy.
PURPOSE: To discriminate between subclinical fellow eyes of patients with keratoconus (scKCE) and normal eyes (NE) using swept-source Fourier-domain anterior segment optical coherence tomography (SS-OCT). SETTING: University Medical Center Hamburg-Eppendorf, Germany. DESIGN: Retrospective cross-sectional study. METHODS: NE (both eyes with a KISA% Index <60; N = 213) and scKCE (KISA% Index <60, but an index of >100 in the fellow eye; N = 61) eyes were grouped according to the KISA% Index. Topography and tomography data were obtained by SS-OCT (Casia SS1000) and reviewed retrospectively. Receiver operating characteristics were performed for Casia SS1000 integrated automated parameters and newly computed indices. Additionally, discriminant functions were calculated for the automated single parameters, the computed indices, and the combination of both. RESULTS: An accuracy ≥0.7 was reached for 28 (8 automated and 20 newly computed parameters) of 785 parameters. Fourier analysis posterior higher-order irregular astigmatism revealed the highest accuracy with 0.81. By combining automated parameters with the newly computed indices in 1 discriminant function, an area under the curve of 0.92 with a corresponding accuracy of 0.85 was achieved. CONCLUSIONS: SS-OCT is able to discriminate between NE and scKCE with high accuracy. Automated single parameters and newly computed indices were combined to further improve this accuracy.
Authors: Marcella Q Salomão; Ana Luisa Hofling-Lima; Louise Pellegrino Gomes Esporcatte; Bernardo Lopes; Riccardo Vinciguerra; Paolo Vinciguerra; Jens Bühren; Nelson Sena; Guilherme Simões Luz Hilgert; Renato Ambrósio Journal: Int J Environ Res Public Health Date: 2020-03-23 Impact factor: 3.390