Year: 2026
Type: article
Abstract: PURPOSE: Accurate application of electroretinography (ERG) remains challenging due to the heterogeneity of retinal disorders. Despite its potential as a diagnostic tool, the clinical use of ERG is limited by the expertise required for interpretation and the error-prone manual extraction of waveform data. At scale, waveform analysis in academic research and clinical trials is made difficult by these challenges. This study performed external validation and quality assessment of automated compared to manual extraction and collection of raw ERG waveforms. METHODS: Raw waveform data was collected using International Society for Clinical Electrophysiology of Vision (ISCEV) standard ERG studies from the ophthalmology practice at Columbia University. A recently developed automated batch exporting tool (Espion, Diagnosys LLC, Lowell, MA) was used to retrieve and compile data and compare with manually collected data for 357 patient sessions. RESULTS: The automated and manually exported data had 99.75% ± 0.17% similarity using a Jaccard similarity test, suggesting the automated tool is both rapid and highly accurate. We identified several inconsistencies leading to errors in the automated batch export. The most common errors in compiled batch exported data included: improperly merging two ERG files taken on the same day and lack of accounting for missing data. CONCLUSIONS: Use of batch export with manual oversight for sources of potential error can allow for faster, more effective data gathering and analysis of large-scale ERG waveform datasets in academic and clinical trial settings.
Source: Documenta Ophthalmologica
Institution: Columbia University Irving Medical Center
Language: English
Subfield: Ophthalmology
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
Estimated APC paid: $4,090
OpenAlex ID: W7236292407
