Share This

In this multicentre, case-control study, the authors develop a new index, named Boosted Ectasia Susceptibility Tomography Index (BESTi), to detect early cases of keratoconus. This index was derived from multiple logistic regression analysis of 22 variables of tomography scans. The study population included 410 eyes with keratoconus (KC), 2296 healthy corneas (CG), and 187 eyes with very asymmetric ectasia and normal corneal topography and tomography corneas (VAE-NTT). Half of each of the three groups was randomly selected to form an external validation set (EV). BESTi was compared with the Belin-Ambrósio Deviation Index (BAD-D) in the same patients from CG and VAE-NTT group from the EV set, and the former exhibited a substantially greater area under the curve (0.912 versus 0.810; P < 0.0001). When the same comparison was applied to Pentacam Random Forest Index (PRFI), BESTi showed a significantly greater area under receiver operating characteristic (0.912 versus 0.870; P = 0.021). Therefore, BESTi had an 86% sensitivity and specificity of 83.97% for detecting early subclinical keratoconus, both greater than either the BAD-D or the PRFI. The study was limited by the lack of longer follow-up and the small samples size of the VAE-NTT group.

Novel artificial intelligence index based on Scheimpflug corneal tomography to distinguish subclinical keratoconus from healthy corneas.
Junior GCA, Guido RC, Silva HMB, et al.
2022 Epub ahead of print.
Share This
Mahmoud Ahmed

Royal Liverpool University Hospital, UK.

View Full Profile