Lessons learned: quality analysis of optical coherence tomography in neuromyelitis optica

INTRODUCTION: Optical coherence tomography (OCT)-derived retina measurements are markers for neuroaxonal visual pathway status. High-quality OCT scans are essential for reliable measurements, but their acquisition is particularly challenging in eyes with severe visual impairment, as often observed in neuromyelitis optica spectrum disorders (NMOSD). OBJECTIVE: To investigate OCT quality issues in real-world data from the international Collaborative Retrospective Study on Retinal OCT in Neuromyelitis Optica (CROCTINO). METHODS: We evaluated the quality of peripapillary and macular OCT scans, using Heidelberg Spectralis SD-OCT, Carl Zeiss Cirrus HD-OCT, or Topcon SD-OCT across 22 centers. Experienced graders applied OSCAR-IB criteria for OCT quality. Eyes were classified as severely visually impaired or not based on a 1.0 logMAR cut-off. Quality outcomes were compared using the Chi-square test. RESULTS: A total of 3075 OCT scans (1630 peripapillary, 1445 macular) from 539 people with NMOSD and related conditions were evaluated. Macular scans were rejected more often than peripapillary scans due to quality issues (20.1% vs. 14.5%, p < 0.001). Rejection rates were higher in eyes with severe visual impairment (peripapillary: 28.9%, macular: 41.6%) compared to eyes without severe visual impairment (peripapillary: 10.7%, p < 0.001; macular: 14.6%, p < 0.001). CONCLUSION: Our study revealed that approximately one in six scans was rejected due to low quality, with higher rejection rates in eyes with severe visual impairment. As scan quality can bias quantitative outcomes and artificial intelligence applications, these findings emphasize the unmet need for standardized OCT practices tailored to NMOSD and other conditions involving severe visual impairment.