Morphological And Functional Alternations Of The Central Visual Pathway In Primary Open-angle Glaucoma

Background and purposeGlaucoma is the leading cause of irreversible blindness worldwide, characterized byprogressive loss of retinal ganglion cells (RGCs) and degeneration of their axons in opticnerve (ON).Primary open-angle glaucoma (POAG) is the most common type of glaucoma.Although RGCs death is the major pathological hallmark of POAG, findings from animalexperiments, human autopsy and glaucoma patient studies indicated that the whole visualpathway may be involved in glaucomatous damage, including the RGCs, the lateralgeniculate nucleus (LGN) and the primary visual cortex (V1). POAG is currentlyconsidered the complicated optic neuropathy caused by environmental and genetic factors,but the pathogenesis of POAG remains elusive.Recently, studies on the disruption of the visual pathway in POAG have been widelyconcerned by using magnetic resonance imaging (MRI) method. Based on the neuroimagingtechnology, the alternations of visual pathway in POAG can be revealed in both thestructure (i.e. grey matter and white matter) and function. However, the previous MRIstudies are mainly focused on the advanced or severe stages of POAG and the alterations ofcentral visual pathway in the early stage of disease are still unknown. In addition,glaucomatous damage of the visual cortex has seldom been studied by usingmulti-parameter shape analysis and resting-state functional magnetic resonance imaging(R-fMRI) method. Thus, the objective of this study was to investigate possiblymorphological and functional changes of central visual pathway in POAG by usingmulti-modal MRI.Material and methodThe present study was divided into four parts:1. Thirty-six patients with bilateral POAG patients and forty matched health controlswere enrolled in this study.3D high-resolution structural T1images were obtained by using a3-Tesla MR scanner with an8-channel phased-array head coil. Cortical thickness wasprocessed using the automated CIVET pipeline (version1.1.9; Montreal NeurologicalInstitute at McGill University, Montreal, Quebec, Canada) to assess the possible changesbetween the patients and controls, and the correlation between the retinal nerve fiber layer(RNFL) thickness and cortical thickness was also investigated. Finally, the patients weresplit into mild and severe sub-groups for investigation of the relationship between POAGstage and cortical changes.2. Twenty normal controls (NC),20mild (MP) and17severe (SP) POAG patientswere recruited and scanned using the3-Tesla MR scanner. Cortical thickness analyses withregions of interest (ROI) was performed using FreeSurfer (version5.3.0,http://surfer.nmr.mgh.harvard.edu) to assess the cortical changes (V1, V2and V5/MT+)among the three groups. Finally, the associations of cortical thickness with RNFL thicknessand mean deviation (MD) of visual field were analyzed.3. Twenty normal controls (NC),19mild (MP) and17severe (SP) POAG patientswere recruited and scanned using magnetic resonance imaging. Multi-parametricmorphologic analyses (cortical thickness, volume, surface area and mean curve) with ROIwere performed by using FreeSurfer pipeline to assess the cortical changes (V5/MT+,anterior and posterior subregions of V1and V2) among the three groups. The correlationsbetween cortical thickness and clinical measurements (RNFL thickness and MD of visualfield) were also analyzed.4. Twenty-one patients with bilateral early stage of POAG patients and20matchedhealth controls were enrolled in this study. R-fMRI data and3D high-resolution structuralT1images were obtained using the3-Tesla MR scanner. Fractional amplitude of lowfrequency fluctuation (fALFF) method were applied to investigate the functional changesbetween patients and health controls. Functional connectivity was calculated by using theseed voxel correlation approach (precuneus, V1, V2and V5/MT+areas).Results:1. Compared with the normal controls, POAG patients showed significantlydecreased bilateral cortical thickness in the calcarine sulci (right BA17, left BA17and BA18) and in some smaller other regions including the left middle temporal gyrus (BA37) andthe fusiform gyrus (BA19)(P <0.005, uncorrected). The left and right cortical thickness of the calcarine sulcus correlated positively with the RNFL thickness (left, r=0.44, P=0.01;right, r=0.38, P=0.03).The significant difference was also found in cortical thickness ofthe calcarine sulcus and RNFL thickness between the mild and severe sub-groups. RNFLthickness (mild versus severe group), left:77.3±13.0vs.59.2±18.1μm (P=0.001); right:73.3±19.0vs.59.2±14.5μm (P=0.020). Cortical thickness (mild vs. severe), left:2.59±0.10vs.2.46±0.18mm (P=0.014); right:2.70±0.14vs.2.56±0.16mm (P=0.009).2. Compared with the NC group, the SP group had significantly lower thickness in theleft V1area and bilateral V2and V5/MT+visual cortices, whereas the MP group onlyexhibited a significant reduction of cortical thickness in the V5/MT+area. Although thethickness of these visual cortices was reduced in the SP group relative to the MP group, onlythe differences in the bilateral V2regions were statistically significant. The mean RNFLthickness in the POAG patients was positively correlated with cortical thickness in the V2(r=0.38, P=0.02) and V5/MT+(r=0.44, P=0.006) regions. Additionally, the mean MDwas also related to cortical thickness in the V1(r=0.34, P=0.04), V2(r=0.42, P=0.009)and V5/MT+(r=0.37, P=0.03) areas.3. Compared with the NC group, the SP group had significantly lower thickness in theanterior subregions of V1, V2and the V5/MT+visual cortices. Unexpectedly, corticalthinning in the posterior subregion of V2was statistically significant in the SP groupcompared with NC and MP group. In addition, the gray matter volume in posteriorsubregion of V2and was significantly decreased in the SP group compared with the NC (P=0.001) and MP (P=0.034) groups. Mean curvature in V5/MT+areas was significantlyincreased in the SP group compared with the NC group (P=0.031). There were nodifferences in surface area in all of the investigated visual areas among the three groups.Cortical thickness of V5/MT+was positively correlated with the mean RNFL thickness (r=0.42, P=0.01). In SP group, cortical thickness of the anterior subregion of V1alsoexhibited a correlation trend with the RNFL thickness (r=0.46, P=0.08).4. Significant between-group difference in fALFF was observed in left precuneus(BA7) and left angular gyrus (BA39/40)(P <0.05, AlphaSim corrected). In contrast, noregions in healthy controls showed higher fALFF than in POAG patients (P>0.05).Compared to normal controls, the early stage of POAG patients showed significantlydecreased functional connectivity with ROI in multiple brain regions including frontal gyrus (BA9, BA10and BA11), frontal eye field (BA8), left anterior cingulate gyrus (BA24), rightmiddle temporal gyrus (BA39) and left cerebellum posterior lobe (P <0.05, AlphaSimcorrected).Conclusion1. Cortical thickness analysis revealed cortical thinning in the visual-related cortex inPOAG patients, and the cortical thickness was correlated positively with the RNFLthickness. The results indicate that cortical thickness may be a sensitive measure fordetection of cortical alterations of the visual system in POAG. However, the whole-braincortical thickness based on vertex-based analysis could not reveal the changes of anatomicalor functional regions.2. This study provides direct in vivo evidence of the progressive thinning in the visualcortex in patients with POAG. The most significant novel finding is that cortical thicknessof the bilateral V5/MT+areas decreases in the early stage of POAG. These cortical changesindicated progressive neurodegeneration and early disruption of the visual cortex in POAG.Finally, we showed that the thickness of visual cortex correlated well with RNFL thicknessand visual field defects in POAG patients, which suggested that the thickness of visualcortex is helpful in detecting POAG pathology and could serve as a novel indicator ofdisease severity.3. This is the first study that provides direct in vivo evidence of the morphologicalterations in the visual cortex in POAG. The results indicated early disruption of the visualcortex in POAG and the glaucomatous process may be complex and heterogeneous. Ourfindings suggested that the thickness of visual cortex is more sensitive than the othermorphologic parameters.4. The dorsal attentional network may be affected inpatients with early POAG. Theprecuneus and V5/MT+may be the key regions and the dorsal visual pathway plays acritical role in the pathogenesis of early POAG.