Purpose
To evaluate the concordance between Scheimpflug tomography and Ocular Response Analyser (ORA) testing in topographically unremarkable corneas for keratoconus detection
Methods
Observational case series of 206 eyes of 206 refractive surgery candidates evaluated with corneal topography for inclusion (only eyes with average corneal power power [ACP]≤47.75 D and corneal higher-order aberrations [HOA]≤0.650 µm) and then analyzed with Scheimpflug tomography (OCULUS Pentacam HR, Optikgeräte GmbH, Germany) and biomechanical testing (Ocular Response Analyzer, Reichert Ophathalmic Instruments, Depew, NY). Main outcome measurements were the Pentacam’s ectasia detection indices (Df, Db, Dp and D, cutoff ≥1.6) and previously validated combinations of ORA measurements (DifCRF and LF2, J Refract Surg. 2013 Sep;29(9):596-602). Statistical significance was set at p<0.05.
Results
Mean age (±SD) was 32.5±8.0 years and 43% of the subjects were male. Mean average corneal power was 44.5±1.5 D, HOA 0.263±0.093 µm, Df 0.53±1.11, Db -0.15±0.92, Dp 0.51±0.94, D 0.98±0.89, DifCRF 0.08±1.58 mm Hg and LF2 0.27±0.26. Actual positive rate (%) was Df 16.5, Db 5.3, Dp 10.2, Dt 11.2, Dy 10.2, D 21.8, DifCRF 30.6 and LF2 20.4. HOA were positively correlated with ACP (r=0.157, p=0.02), Dy (r=0.191, p=0.01), D (0.153, p=0.03) and LF2 (0.146, p=0.04). Concordance (both tests either positive or negative) was 62.1% for D and DifCRF, 68.4% for D and LF2, 66% for Dp and DifCRF and 76.2% for Dp and LF2. Only D-LF2 and Dp-LF2 double-positive cases had significantly (p<0.01) higher HOA (0.367±0.154 and 0.412±0.172 µm, respectively) than their double-negative counterparts (0.246±0.067 and 0.249±0.070 µm, respectively). Eyes with LF2>0.5 did not significantly differ in mean D and Dp from their counterparts, but eyes with LF2>0.7 did (Dp 0.47±0.91 vs 0.99±1.10, p=0.03; D 0.94±0.85 vs 1.42±1.14, p=0.03). LF2 reached highest concordance with Dp (86.9%) and D (76.2%) with a >0.7 cutoff, and conversely, the latter showed highest concordance with LF2 (89.3% and 86.4%, respectively) with a ≥2.0 cutoff.
Conclusions
Striving for high sensitivity for ectasia detection by a single method comes at the cost of reduced specificity. Concordance between ORA testing and Pentacam Scheimpflug tomography is acceptable, but there is the prospect of lowering the false positive rate by combining both technologies with modified diagnostic criteria.