MR Imaging of the Anterior Visual Pathway in Primary Open-Angle Glaucoma: Correlation with Octopus 101 Perimetry and Spectralis Optical Coherence Tomography Findings

Abstract

Purpose: To investigate structural changes in the visual pathway measured by magnetic resonance imaging (MRI) and its relationship with the clinical severity of glaucoma in primary open-angle glaucoma (POAG) patients.

Materials and methods: The study included 28 patients with POAG and 26 age- and sex-matched healthy volunteers. All the subjects underwent spectral domain optical coherence tomography (OCT) of the peripapillary retina nerve fiber layer (RNFL). The optic nerve diameter (OND), chiasma height (Ch), and lateral geniculate nucleus height (LGNh) were measured bilaterally using a 1.5-Tesla MRI system.

Results: The mean values of the OND and LGNh were significantly lower in the POAG group (OND: right p = 0.043 and left p = 0.048; LGNh: right p = 0.008 and left p = 0.025). The OND was not correlated with the clinical stage of glaucoma, but it was correlated with the ipsilateral RNFL thickness. The Ch was correlated with the ipsilateral clinical stage (right r = -0.536, p = 0.004; left r = -0.537, p = 0.004) and average RNFL thickness (RNFLav) (right r = 0.655, p < 0.001; left r = 0.626, p < 0.001). The sum of bilateral clinical stages and left clinical stages showed significant correlations with the right and left LGNh and the sum of both the right and left LGNh. The left RNFLav and the sum of the right-left RNFLav were significantly correlated with all LGNh measurements.

Conclusions: 1.5-Tesla MRI can detect structural changes in the visual pathway early in the course of glaucoma. Thin optic nerve can be a risk factor for glaucoma. The Ch and LGNh seem to be correlated with the clinical stage of glaucoma and RNFL thickness. In particular, LGN can be a target of glaucomatous damage.

Keywords: Glaucomatous optic nerve damage; lateral geniculate nucleus; magnetic resonance imaging; octopus perimetry; optic chiasm; retinal nerve fiber layer thickness.