To optimize artificial intelligence (AI) algorithms to integrate Scheimpflug-based corneal tomography and biomechanics to enhance ectasia detection.
Multicenter cross-sectional case-control retrospective study.
3,886 unoperated eyes from 3,412 patients had Pentacam and Corvis ST (Oculus Optikgeräte GmbH; Wetzlar, Germany) examinations. The database included one eye randomly selected from 1,680 normal patients (N), and from 1,181 “bilateral” keratoconus (KC) patients, along with 551 normal topography eyes from very asymmetric ectasia patients (VAE-NT), and their 474 unoperated ectatic (VAE-E) eyes. The current TBIv1 (tomographic-biomechanical index) was tested, and an optimized AI algorithm was developed for augmenting accuracy.
The area under the receiver operating characteristic curve (AUC) of the TBIv1 for discriminating clinical ectasia (KC and VAE-E) was 0.999 (98.5% sensitivity; 98.6% specificity [cutoff 0.5]), and for VAE-NT, 0.899 (76% sensitivity; 89.1% specificity [cutoff 0.29]). A novel random forest algorithm (TBIv2), developed with 18 features in 156 trees using 10-fold cross-validation, had significantly higher AUC (0.945; DeLong, p<0.0001) for detecting VAE-NT (84.4% sensitivity and 90.1% specificity; cutoff 0.43; DeLong, p<0.0001), and similar AUC for clinical ectasia (0.999; DeLong, p=0.818; 98.7% sensitivity; 99.2% specificity [cutoff 0.8]). Considering all cases, the TBIv2 had higher AUC (0.985) than TBIv1 (0.974; DeLong, p<0.0001).
AI optimization to integrate Scheimpflug-based corneal tomography and biomechanical assessments augments accuracy for ectasia detection, characterizing ectasia susceptibility in the diverse VAE-NT group. Some VAE patients may be true unilateral ectasia. Machine learning considering additional data, including epithelial thickness or other parameters from multimodal refractive imaging, will continuously enhance accuracy.