Corneal Biomechanical and Tomographic Signatures for Keratoconus

Objectives: This study aimed to identify tomographic and biomechanical biomarkers that differentiate mild/moderate from advanced/severe keratoconus (KCN) based on the ABCD classification. The secondary objective was to assess the diagnostic performance of previously reported biomarkers for distinguishing KCN from KCN-free controls and to compare classifications derived from cohort-specific and previously established cutoff values, evaluating their consistency and potential generalizability in the Romanian KCN population. Methods: Adult patients who presented to the Oculens Clinic between 1 January 2024 and 1 September 2025 for clinical care were eligible for enrolment in this observational study. Diagnosis of KCN was made following current practice, and all participants underwent ophthalmologic examination followed by corneal tomography and biomechanical measurements. Results: Forty-four subjects with KCN and 38 controls, with a median age of 26.5 years. Patients with mild/moderate KCN (23 patients) were similar in terms of age and sex to those with advanced/severe KCN (21 patients). The values of Kmax showed the best performance in terms of AUC (0.908, 95%CI = [0.824 to 0.992], Se = 76.19 [52.83 to 91.78], Sp = 86.96 [66.41 to 97.22], cutoff = 54.15), with an excellent discriminatory ability in the identification of patients with mild/moderate KCN compared to advanced/severe KCN. Low TL values indicate advanced/severe disease, with the highest values of true positives (18), true negatives (22), and the smallest value of false negatives (3) but with a moderate AUC (0.880 [0.755 to 1.0]). The performances of previously reported cutoff values on the Romanian cohort were tested in 44 patients with KCN and 38 KCN-free individuals. The discriminatory abilities of the previously reported indices, except for KI and BAD-D which were overfitted, were retained in the current cohort, and all exhibited excellent (AUC lower bounds higher than 0.9) or good (AUC lower bounds higher than 0.8) discriminatory performances. The agreement between the current cutoff values and the classification generated by the previously reported cutoff values is sustained by almost perfect agreement in the classification as KCN or KCN-free (Cohen’s kappa values ≥ 0.81, and raw agreement ≥ 90%). Conclusions: Our findings support further evaluation of Kmax and corneal thinnest location as markers of keratoconus severity and as potential tools for disease detection. The high concordance observed between classification based on cohort-specific and previously established thresholds suggests consistent performance of these biomarkers within the study population. These parameters may represent promising tools for the management of patients with keratoconus, pending further validation.