Abstract
Purpose:
To evaluate the diagnostic performance of intereye asymmetry in corneal indices provided by the Pentacam tomograph for detecting keratoconus.
Methods:
Observational case series of 35 healthy subjects in group 1 and 40 keratoconic patients in group 2, evaluated with Scheimpflug tomography (Pentacam HR, Oculus Optikgerate GmbH, Germany). For data analysis, Student's t test was used to compare means and the area under the receiver operating characteristic curve (AROC) was used to compare the diagnostic performance of the intereye asymmetry in corneal descriptors for keratoconus detection. The statistical significance criteria used was p <0.05.
Results:
In group 1 vs group 2, mean age (33.5±11.4 vs 36.5±15.2 yrs) and gender distribution (60% vs 52.5% males) did not differ significantly. Intereye asymmetry for all variables in group 1 was not correlated to anisommetropia. Compared with group 1, intereye asymmetry was significantly greater in group 2 for all variables except for Dy, ARTMin and ARTMax. No patients in group 1 had an intereye difference in anterior keratometry > 0.5 D or in thinnest or apex pachymetry > 25 µm. Eight of the 19 variables had AROC > 0.93 for their asymmetry. The best 8 indices (AROC, cutoff, sensitivity and specificity) were D (0.99, >1.24, 95%, 97.1%), AvgPPI (0.97, >0.2, 92.5%, 97.1%), MaxPPI (0.97, >0.33, 92.5%, 94.3%), Dp (0.97, >1.37, 92.5%, 94.3%), anterior keratometry (0.97, >0.5 D, 90%, 100%), posterior keratometry (0.94, >0.1 D, 85%, 97.1%), MinPPI (0.94, >0.13, 87.5%, 91,4%) and Df (0.93, >1.17, 87.5%, 91,4%), respectively. Combined metrics of intereye asymmetry in 4 and 5 corneal descriptors had 100% sensitivity and 93% and 98% specificity, respectively.
Conclusions:
Intereye asymmetry is rare in normal corneas and very common in keratoconus patients. Almost half of the corneal descriptors analyzed had very high specificity and sensitivity when considered alone. A combined approach showed remarkable performance and might serve to discriminate true keratoconic patients from false positive cases that arise in clinical practice when examining single corneal indices.