Abstract
PURPOSE: Corneal nerve morphology and immune cells are critical biomarkers in the ocular surface. This study aimed to investigate the influence of immune cells on corneal nerve morphology and the clinical significance in diabetes-related dry eye disease (DED). METHODS: In the first part, 1075 in vivo confocal microscopy images containing dendritic cells or round cells were included as system validation. Key morphological parameters, including corneal nerve fiber density (CNFD), corneal nerve branch density (CNBD), tortuosity, and box-count fractal dimension (Boxdim), were measured before and after immune cells were excluded. In the second part, a pilot cross-sectional study was conducted involving control (26 eyes), DED without diabetes mellitus (DM) (34 eyes), and DED with DM (17 eyes) groups. The impact of immune cell exclusion on nerve metrics was assessed and correlated with clinical parameters, such as the fluorescein tear breakup time (TBUT), Ocular Surface Disease Index (OSDI), and corneal fluorescein staining (CFS) scores. RESULTS: Exclusion of immune cells resulted in significant reductions in CNFD, CNBD, Boxdim, and tortuosity. Compared to the control group, both DED without DM and DED with DM groups showed substantial reductions in CNFD, CNBD, and Boxdim, along with a significant increase in tortuosity. Moreover, the exclusion of immune cells enhanced the correlations between nerve metrics and fluorescein TBUT. CONCLUSIONS: Immune cells contribute to significant biases in the assessments of corneal nerve morphology, primarily false-negative results, in diabetes-related DED. Their exclusion improves the accuracy of nerve measurements, which may enhance the clinical evaluation of corneal nerve morphology. TRANSLATIONAL RELEVANCE: Advanced segmentation techniques addressing immune cell interference could improve diagnostic precision and inform treatment strategies for DED subtypes.