Best indices of dual Scheimpflug/Placido tomographer for keratoconus detection.

ABSTRACT

PURPOSE: To evaluate the accuracy of different corneal parameters in detecting keratoconus using a dual Scheimpflug/Placido system (Sirius, CSO, Italy). METHODS: Sixty-eight eyes of 68 keratoconus patients and 77 eyes of 77 normal subjects were prospectively assessed in a diagnostic test study. The mean differences of the corneal parameters were compared using the independent t-test. The accuracy of curvature, elevation, pachymetry, and aberrometry indices, aqueous depth, and corneal volume in 10 mm was evaluated using the area under the curve (AUC), and the DeLong method was used for the comparison of AUCs. Multiple tests in a parallel manner and multiple logistic regression analysis were applied to determine the best predictor indices. RESULTS: All indices except aqueous depth and corneal volume were significantly different between the two groups (P < 0.001). Back keratoconus vertex (KVb) and front Baiocchi-Calossi-Versaci index (BCVf) had the highest sensitivity (for both 98.53%) followed by front symmetry index (SIf) (96.12%) and thinnest point of the cornea (88.24%) in elevation, aberrometry, curvature, and pachymetry parameters, respectively. The highest diagnostic ability was observed in KVb (AUC 0.993) and BCVf index (AUC 0.992) (DeLong > 0.05). Multiple test analysis showed a combination of indices with the highest accuracy that was similar to the performance of each one individually, and keratoconus was diagnosed correctly in 98.5% of the cases (R2 = 93%). CONCLUSION: The parameters extracted from Sirius can differentiate keratoconus from normal corneas with high accuracy without the need for complex computational algorithms. Elevation-based and combined aberrometry indices had the highest diagnostic power.