New Keratoconus Grading System Based on OCT: Threshold Adjustment for a Swept-Source OCT.

Abstract

Purpose: To assess the performance of the STEP keratoconus (KC) grading system using swept-source (SS)-OCT, and to compare the relative performance of SS-OCT and spectral domain (SD)-OCT for thickness measurements.

Setting: University hospitals in Germany and China.

Designs: Retrospective case-control study.

Methods: Rotating Scheimpflug Tomography (RSC, Pentacam) and SS-OCT (ANTERION) were used to examine 152 normal and 314 KC eyes from a German cohort; The STEP system (stroma thinnest thickness [ST] and epithelium standard deviation [EP]) was generated in SS-OCT and compared with RSC ABC system. Pentacam and SD-OCT (RTVue-XR) data from a Chinese cohort of 236 normal and 331 KC eyes were utilized to indirectly assess inter-device agreement between SS-OCT and SD-OCT using Bland-Altman plots and linear regression. Optimized thresholds for STEP 'ST' were determined for SS-OCT.

Results: In normal eyes, STEP 'ST' agreement with RSC 'C' was 91.45%, dropping to 28.98% in KC eyes. STEP 'EP' showed 58.55% and 60.53% agreement with RSC 'A' and 'B' in normal eyes, and 36.62% and 52.87% in KC eyes, respectively. SS-OCT's ST values were estimated to be higher than those of SD-OCT (estimated mean difference = 23.18 and 20.91 μm in the normal and KC eyes, respectively). Adjusted threshold of STEP 'ST' for SS-OCT was proposed.

Conclusions: SS-OCT's STEP 'EP' aligned well with RSC ABC system, while thresholds for SS-OCT's STEP 'ST' required adjustment for more accurate clinical application. These findings highlight important considerations for thickness measurement discrepancies between SS-OCT and SD-OCT, with broader implications for future clinical practice and research.