Reliability of Macular Ganglion Cell-Inner Plexiform Layer Thickness Measurements Across Scan Protocols in Spectral-Domain Optical Coherence Tomography

Objective: We aimed to assess whether modified OCT acquisition parameters improve GCIPL measurement reliability. Methods: Participants with multiple sclerosis (PwMS) and age- and sex-matched healthy controls (HC) underwent OCT (Spectralis OCT, Heidelberg Engineering, Heidelberg, Germany) at baseline and after two and four weeks. At each visit, five macular scan protocols were acquired: standard (49 lines, high-speed [HS], automated real-time tracking [ART]: 9), high-ART (49 lines, HS, ART: 50), high-lines (97 lines, HS, ART: 9), high-resolution (49 lines, high-resolution [HR], ART: 9), and maximum (97 lines, HR, ART: 50). Reliability was assessed using intraclass correlation coefficients (ICC). Absolute test–retest reproducibility was quantified using the mean absolute difference (MAD). Results: Thirty-eight eyes from nine PwMS (mean age 34.1 ± 8.0 years, 44.4% female) and ten HC (31.7 ± 11.1 years, 50.0% female) were included. At baseline, mean GCIPL thickness ranged from 70.8 µm to 71.5 µm across protocols, demonstrating excellent inter-protocol agreement (ICC 0.99; 95% confidence interval [CI]: 0.98–0.99; p < 0.001) and only marginally higher values with increased ART. Test–retest reliability was excellent for all protocols, demonstrating marginally lower absolute measurement variability of high-ART and high-lines protocols (MAD 0.26–0.27; 95% CI: 0.21–0.32), while temporal agreement remained excellent and comparable across acquisition settings. Mean acquisition time ranged from 10.6 ± 1.6 s for the standard protocol to 231.9 ± 36.4 s for the maximum protocol. Conclusions: All OCT acquisition protocols demonstrated excellent inter-protocol and test–retest reliability for GCIPL measurements. The high-lines protocol provides the most favourable balance between measurement reliability and acquisition time, supporting its potential utility for longitudinal GCIPL monitoring.