Early diagnosis of subclinical keratoconus by wavefront parameters using Scheimpflug, Placido and Hartmann-Shack based devices.

2020 
PURPOSE: To compare the accuracy of three devices (Pentacam, Sirius and OPD-Scan III) to differentiate subclinical keratoconus from normal corneas by wavefront parameters. METHODS: Two hundred and seventeen patients were enrolled in three groups [68 normal, 79 subclinical keratoconus (SKCN) and 70 KCN eyes] in this prospective diagnostic test study. Wavefront indices were evaluated between the groups using Pentacam, Sirius and OPD-Scan III. The accuracy of the parameters was determined by measuring the area under the receiver operating characteristic curve (AUC) for each group. RESULTS: Front Baiocchi-Calossi-Versaci (BCV) index with Sirius (sensitivity = 87.7%, specificity = 83%, AUC = 0.887), front Vertical Coma (Z3(-1)) with Pentacam (sensitivity = 75%, specificity = 100%, AUC = 0.857) and Corneal Z3(-1) with OPD-Scan III (sensitivity = 100%, specificity = 78.6%, AUC = 0.857) had the highest AUC values for the diagnosis of subclinical KCN. In the KCN group, the highest AUC values were obtained for front higher-order aberration (HOA), front/back Z3(-1) and front Secondary Vertical Coma (Z5(-1)) with Pentacam (sensitivity = 100%, specificity = 100%, AUC = 1.00 for all three), front root mean square values per unit area (RMS/A), HOA, Residual HOA, BCV, RMS Trefoil and RMS Coma with Sirius (sensitivity = 100%, specificity = 100%, AUC = 1.00 for all) and Corneal HOA, RMS total Coma and Z3(-1) with OPD-Scan III (sensitivity = 100%, specificity = 93%, AUC = 0.96 for all three). CONCLUSION: Corneal wavefront indices generated from different devices have acceptable validity for differentiating normal cornea from the early form of KCN, and this can be very useful for preoperative screening before refractive surgery. The front BCV with Sirius was the most accurate parameter for diagnosis of SKCN followed by Z3(-1) with Pentacam and OPD-Scan III.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    34
    References
    7
    Citations
    NaN
    KQI
    []