This content is only available for ASCRS Members
This content from the 2020 ASCRS Virtual Annual Meeting is only available to ASCRS members. To log in, click the teal "Login" button in the upper right-hand corner of this page.
Purpose
To compare the ability of tomographical parameters, biomechanical parameters and the integration of both approaches for discriminating ectatic disease.
Methods
A retrospective study involving a total of 1295 eyes divided into 736 normal eyes (group 1), 321 KC eyes (group 2), 113 unoperated ectatic eyes from patients with very asymmetric ectasia (group 3), who presented fellow eyes (125 eyes) with normal topographic maps (group 4). For groups 1 and 2, only one eye per patient was selected randomly for the inclusion in the study, in order to avoid the bias of the relation between eyes. Tomographic parameters included ISV, BADD, PRFI and TKC. Biomechanical parameters included SPA1, Inverse radius, DA Ratio, and CBI. The TBI was also evaluated.
Results
With a cut-off value of 0.79, TBI had 100% sensitivity and specificity to detect frank ectasia cases (AUC = 1.0 in groups 2 and 3); however, for the correct characterization of eyes with standard topography having no definitive signs of ectasia from patients with clinical ectatic disease in the fellow eye, an optimization of cut-off value was necessary, and a value of 0.29 provided 90.4% sensitivity with 4% false-positive results (96% specificity; AUC = 0.985). Comparing all groups, the AUC of the TBI to detect ectasia (groups 2, 3 and 4) was 0.992 and had a statistically higher AUC (DeLong, p<0.001) than all other parameters tested for every analysis performed.
Conclusion
The integration of biomechanical data and corneal tomography with artificial intelligence data augments the sensitivity and specificity for screening and enhancing early diagnosis of patients with corneal ectasia. The AUC of the TBI was statistically higher than all other analyzed parameters, including the CBI.
To compare the ability of tomographical parameters, biomechanical parameters and the integration of both approaches for discriminating ectatic disease.
Methods
A retrospective study involving a total of 1295 eyes divided into 736 normal eyes (group 1), 321 KC eyes (group 2), 113 unoperated ectatic eyes from patients with very asymmetric ectasia (group 3), who presented fellow eyes (125 eyes) with normal topographic maps (group 4). For groups 1 and 2, only one eye per patient was selected randomly for the inclusion in the study, in order to avoid the bias of the relation between eyes. Tomographic parameters included ISV, BADD, PRFI and TKC. Biomechanical parameters included SPA1, Inverse radius, DA Ratio, and CBI. The TBI was also evaluated.
Results
With a cut-off value of 0.79, TBI had 100% sensitivity and specificity to detect frank ectasia cases (AUC = 1.0 in groups 2 and 3); however, for the correct characterization of eyes with standard topography having no definitive signs of ectasia from patients with clinical ectatic disease in the fellow eye, an optimization of cut-off value was necessary, and a value of 0.29 provided 90.4% sensitivity with 4% false-positive results (96% specificity; AUC = 0.985). Comparing all groups, the AUC of the TBI to detect ectasia (groups 2, 3 and 4) was 0.992 and had a statistically higher AUC (DeLong, p<0.001) than all other parameters tested for every analysis performed.
Conclusion
The integration of biomechanical data and corneal tomography with artificial intelligence data augments the sensitivity and specificity for screening and enhancing early diagnosis of patients with corneal ectasia. The AUC of the TBI was statistically higher than all other analyzed parameters, including the CBI.
View More Presentations from this Session
This presentation is from the session "SPS-106 Keratoconus: Measurements, Treatments, New Technology" from the 2020 ASCRS Virtual Annual Meeting held on May 16-17, 2020.