Purpose
To evaluate the diagnostic performance of corneal indices provided by the Pentacam tomograph for detecting subclinical keratoconus.
Methods
Observational case series of 136 eyes from 136 healthy subjects in group 1 and 42 topographically-unremarkable eyes from 42 keratoconic patients in group 2, evaluated with corneal topography, aberrometry, and Scheimpflug tomography for Belin-Ambrosio D indices. For data analysis, Student's t test was used to compare means and receiver operating characteristic (ROC) curves were used to analyze the diagnostic performance of D indices for keratoconus detection. The statistical significance criteria used was p <0.05.
Results
Gender distribution (group 1 vs group 2, 42% vs 67% male, p <0.01), but not age (31.0 ± 8.4 vs 32.5 ± 11.4 years), was different between groups. Keratoconic corneae were centrally thinner (535 ± 30 vs 512 ± 77 µm, p <0.01), had similar average corneal power (44.26 ± 1.50 vs 44.40 ± 1.96 D) and more higher-order aberrations (0.24 ± 0.07 vs 0.66 ± 0.35 µm, p <0.0001) than control corneae. The area under the ROC curve, the sensitivity and specificity of Belin-Ambrosio D indices were: Df = 0.80, 57.1% and 89.0%, Db= 0.83, 52.4% and 99.3%, Dp= 0.92, 64.3% and 97.8%, Dt= 0.79, 42.9% and 94.9%, Dy= 0.80, 54.8% and 90.4%, D= 0.93, 73.8% and 98.5%. Comparing the global D index’s false negative (n = 11) and true positive cases (n = 31), there were no differences in age (28.0 ± 6.8 vs 34.0 ± 12.4 years), in corneal thickness (509 ± 40 vs 514 ± 88 µm) and corneal power (44.34 ± 1.44 vs 44.42 ± 2.13 D), but higher-order aberrations were increased in the latter (0.35 ± 0.19 vs 0.76 ± 0.33 µm, p <0.001).
Conclusions
Back corneal surface index (Db) was not clinically superior to the anterior surface index (Df) in sensitivity but provided greater specificity, unlike the pachymetric progression index (Dp) or the global D index that showed higher sensitivity and specificity. The false negative cases of the best index had lower corneal aberrations in the anterior surface than the true positive, suggesting that the first represent earlier stages of keratoconus progression. These data probably underestimate the diagnostic capability of the device due to the strict inclusion criteria used. Although the detection rate is good, there is significant margin for combining this technology with other methods and thus improve performance.
Keywords: 574 keratoconus •
479 cornea: clinical science •
552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)