First Author: S.Kocamış TURKEY
Co Author(s): H. Çakmak N. Çağıl
Purpose:
The aim of this study was to evaluate the sensitivity, specifity and the accuracy of Cone Location and Magnitude Index (CLMI) in the diagnosis of keratoconus by comparing this index with the other clinical parameters, used in the diagnosis of keratoconus.
Setting:
: Department of Ophthalmology, Medical School of Yıldırım Beyazıt University, Ankara, Turkey.
Methods:
301 keratoconus patient (158 male, 143 female) with a mean age of 26,19±7,9 and 395 refracitve surgery candidate (208 male,186 female) with a mean age of 30,88±7,57 as control group were enrolled into the study. Refractive errors, visual acuities, simulated keratometry readings (Sim K1, Sim K2), average and central keratometries ( Avr. K, Central K) thinnest corneal thickness (TCT), corneal wavefront aberrations (total high order aberrations (HOA), coma, vertical coma (3,-1), horizontal coma (3,1), spheric aberration (4,0) ) and CLMI derived from axial corneal topography maps of both groups were compared with the Student's t test. Parameters that had statistically significant differences within groups were analysed with ROC curves. Area under curves, best cut off points of the parameters, the sensitivity, specifity and the accuracy of the parameters at the best cut off points were investigated. Logistic regression analyze with Forward Wald method was employed in order to determine which clinical parameter was the best parameter for the diagnosis of keratoconus and also a logistic regression model was generated that enabled to calculate the odds ratio of keratoconus. Pearson's correlation analyze was performed in order to investigate the relationship between the CLMI and the other clinical parameters.
Results:
There was a statistically significant difference in spherical and cylindirical refractive errors, logMAR distance corrected visual acuities, Sim K1, SimK2, Avr. K, Central K, HOA, coma aberrations, (3,-1), (4,0), TCT, CLMI and there was no statistically significant difference in logMAR uncorrected viasual acuities and (3,1) between two groups. The area under curve of CLMI was 0,935. The parameter that has the greatest area under curve was CLMI with a best cut off point of 1,82. At this cut off point, CLMI had the best sensitivty (89%), specifity (94%) and accuracy (92%) for the diagnosis of keratoconus. The CLMI was the strongest parameter that could differentiate keratoconus from normals in the logistic regression model. Only employement of CLMI could diagnose keratoconus with the accuracy of 92,8% in the logistic regression model. With the participation of Avr. K and (4,0) in the model, the accuracy increased to 94,2%. CLMI had a strong correlation all with central K (r=0,669), Avr. K (r=0,690), TCT (r=-0,589), HOA (r=0,858), coma aberrations (r=0,881), (3,-1) (r=-0,814) and (4,0) (r=-0,526). CLMI had the strongest correlation with coma aberrations, HOA and (3,-1).
Conclusions:
The CMLI was the most efficient single variable in the diagnosis of keratoconus, having a rebust and high sensitivity, specifity and accuracy levels which were well beyond the other variables, accepted as strong pillars of conventional diagnosis decision.
Financial Disclosure:
No
