An Artificial Intelligence-Based Prognostic Model for Prediction of Functional Glaucoma Progression From Clinical and Structural Data.

Macular Thickness

To analyse retinal nerve fibre layer (RNFL) thickness in eyes with compression of the optic chiasm by a pituitary adenoma. RNFL thickness was analysed with optical coherence tomography (OCT) and compared to visual field measurements using high-pass resolution perimetry (HRP). Sixteen eyes from eight patients with pituitary adenoma were studied. All had bitemporal visual field depression caused by compression of the optic chiasm. Patients were submitted to an ophthalmic examination more than 14 months after surgery (seven patients had undergone trans-sphenoidal and one trans-cranial adenomectomy). The examination included HRP, fundus photography and measurement of the peripapillar RNFL thickness using OCT. In spite of temporal visual field depression, not all eyes showed reduced RNFL thickness by OCT. This was also true for some eyes in which RNFL was judged to be reduced on fundus photographs. Contrary to our expectations, RNFL thickness in the nasal quadrant was normal in nine of the 16 eyes. Corresponding figures for the superior, inferior and temporal quadrants were eight, six and five, respectively. The overall RNFL thickness, as measured by OCT, did not correlate well with neural capacity, which is an index of remaining retino-cortical neural channels in HRP. RNFL thickness as measured with OCT was reduced in most, but not all, eyes with temporal field depression caused by chiasmal compression. The pattern of RNFL loss did not correlate well with the visual field defect. Sensitivity of RNFL thickness measurement in OCT was low. The method has limited value in the diagnosis of pituitary tumour compression.

This study aimed to investigate normal values and interocular differences in retinal nerve fibre layer (RNFL) thickness, using optical coherence tomography (OCT) and Heidelberg retinal tomography (HRT), in 5-16-year-old children born at full-term with normal birthweights. Fifty-six children with normal visual acuity and refraction were examined with Stratus OCT and HRT. Three examinations were performed in each eye. One eye in each child was randomized for analyses of normal values. Findings in 54 eyes were evaluated. Mean values of RNFL thickness were calculated. Coefficients of variance and intraclass correlations were calculated. The correlation between right and left eyes and the limits of difference were determined for both methods. Mean RNFL thickness was 98.4 μm (standard deviation [SD] 7.88 μm) assessed with OCT and 213.0 μm (SD 54.0 μm) assessed with HRT. No correlations between age or gender and RNFL thickness were found. The coefficients of variance were 2.9% and 5.6% for OCT and HRT, respectively, and intraclass correlations were 0.85 and 0.88, respectively. The limits of difference between the two eyes ranged from -9 μm to 9 μm with OCT and from -109 μm to 87 μm with HRT. Both OCT and HRT can be used in children aged 5-16 years, but OCT provides less variability in determinations of RNFL thickness, both in repeated examinations of the same eye and in comparisons between the two eyes. The present study provides values for normal RNFL thickness in healthy children which can be used to make comparisons with values in children with optic nerve diseases.

To compare retinal nerve fiber layer (RNFL) and macular thickness measurements obtained with the Stratus optical coherence tomography (OCT) and OPKO/OTI OCT devices. Included in the study were 59 eyes of 30 participants. All measurements for each eye were done on the same day with both devices. Student's paired t-tests were used to compare the central macular thickness and RNFL measurements of the Stratus OCT and OPKO/OTI OCT. Pearson correlation was used to assess the relationship between the devices. Coefficient of variation (COV) was calculated to assess intersession repeatability. Using both the Stratus OCT and OPKO/OTI OCT, respectively, the measured mean average RNFL thicknesses were 98.9±11.1 µm and 115.1±9.6 µm (P=0.001), and the measured mean central retinal thicknesses (CRT) were 196.2±18.8 µm and 204.5±21.1 µm (P<0.001). Measured by the two devices, the RNFL thickness values were correlated in all quadrants, as were the retinal thickness values except the inferior outer sector. COV for average RNFL and CRT thickness were 2.9% and 4.6% for Stratus OCT, and 2.1% and 4.2% for OPKO/OTI OCT, respectively. We found good reproducibility of RNFL and retina thickness measurements for both Stratus OCT and OPKO/OTI OCT devices. However, even though the two OCT systems provided statistically correlated results, the values for both RNFL and macular thickness were statistically different. RNFL and macular thickness measurements with the OPKO/OTI OCT were higher than that of the Stratus OCT; therefore, the two OCT systems cannot be used interchangeably for the measurements of RNFL and macular thickness.

To compare the optical coherence tomography retinal nerve fiber layer and macular thickness measurements for detection of progressive axonal loss following acute traumatic optic neuropathy in a longitudinal study. Three patients with unilateral traumatic optic neuropathy were evaluated sequentially after trauma. Macular and retinal nerve fiber layer thickness measurements were obtained using optical coherence tomography weekly for five weeks and around the twelfth week after trauma. All patients showed progressive macular and retinal nerve fiber layer thickness reduction. The mean retinal nerve fiber layer thickness on the first week was 114 microm and reduced sequentially over the first five weeks and was 46 microm on the twelfth week. For macular parameters, the mean average thickness on the first week was 248 microm and also reduced over the first five weeks and was 218 microm on the twelfth week. When compared to the initial measurement, macular thickness average reduction rate at the 12th week was 14% while peripapillary retinal nerve fiber layer thickness average reduction rate was 59%. Although both measurements reduce significantly after trauma, retinal nerve fiber layer thickness measurements show greater and faster retinal neural reduction if compared to macular thickness measurements in traumatic optic neuropathy.

Prediction of Visual Field Progression with Baseline and Longitudinal Structural Measurements Using Deep Learning

Glaucoma patients had a reduction in the inner annulus peripapillary choroidal microvascular density (PCMD) that became worse as the glaucoma severity progressed, which might provide new evidence supporting the vascular theory. To compare PCMD among normal tension glaucoma (NTG), primary open angle glaucoma (POAG), and healthy controls using optical coherence tomography (OCT) angiography (OCTA). The study included 40 POAG, 25 NTG, and 33 healthy controls. All subjects underwent OCT and OCTA testing. The inner annulus and outer annulus PCMD, as well as peripapillary vessel density (VD), were calculated. One-way analysis of variance was used to compare the vascular parameters of the 3 groups. Pearson correlation analysis or Spearman correlation test was used to evaluate the correlation between PCMD and glaucomatous severity factors. The spatial positional relationship between PCMD and corresponding peripapillary retinal nerve fiber layer (RNFL) thickness and visual field (VF) mean deviation (MD) was also assessed. The average and 4 quadrants of the inner annulus PCMD and peripapillary VD in the 2 glaucomatous groups were significantly lower than in normal eyes ( P <0.05). Strong correlations were found between inner annulus PCMD and VF MD, peripapillary VD, and RNFL in POAG patients. Similarly, the inner annulus PCMD in NTG patients was strongly correlated with peripapillary VD and RNFL (all r >0.5). Strong positional correlations were found between inner superior quadrantal PCMD and RNFL thickness in both POAG and NTG patients ( r =0.566 and 0.731, respectively). Likewise, inner inferior quadrantal PCMD exhibited a strong correlation with RNFL thickness in POAG patients ( r =0.608). Strong positional correlations were also found between inner superior PCMD and VF MD in both POAG and NTG patients ( r =0.589 and 0.622, respectively). Inner inferior PCMD exhibited a moderate correlation with VF MD in both POAG and NTG patients ( r =0.487 and 0.440, respectively). The study found that the inner annulus PCMD decreased to varying degrees in NTG and POAG patients. The inner annulus PCMD was closely related to the structural and visual function parameters of glaucoma in both NTG and POAG. Furthermore, inner PCMD demonstrated a spatial correlation with corresponding RNFL thickness and VF MD.

To evaluate the structural, microvascular, and functional progression of normal tension glaucoma (NTG) with or without high myopia by examining longitudinal changes in optical coherence tomography angiography (OCTA) and visual field (VF) parameters. We evaluated 61 NTG eyes and classified 25 of the eyes with axial lengths (ALs) of ≥26 mm as highly myopic. We assessed the rate of change in OCTA parameters, namely radial peripapillary capillary (RPC) vessel density (VD), parafovea VD, deep parafovea VD, retinal nerve fibre layer (RNFL) thickness, and ganglion cell complex thickness. We evaluated the correlation of the rate of change in OCTA parameters with VF loss and AL. Among the 61 NTG eyes, rates of loss of RPC VD, parafovea VD, deep parafovea VD, and RNFL thickness were significantly different from zero despite the nonsignificant rate of change in VF mean deviation (MD). Changes in these OCTA parameters did not differ significantly in highly myopic NTG eyes. The rate of change in VF MD was significantly correlated with the rate of change in parafovea VD in highly myopic and non-highly myopic NTG eyes. In highly myopic NTG eyes, AL was negatively correlated with the rates of loss of RNFL thickness, VF MD, and VF PSD. NTG eyes with a relatively stable VF exhibited loss of VD and RNFL thickness. VF progression in NTG was correlated with decreasing parafovea VD, indicating a structure-function correlation. Greater AL may indicate faster VF loss and RNFL thinning in highly myopic NTG eyes.

The aim of this study was to test the correlation between Fourier-domain (FD) optical coherence tomography (OCT) macular and retinal nerve fibre layer (RNFL) thickness and visual field (VF) loss on standard automated perimetry (SAP) in chiasmal compression. A total of 35 eyes with permanent temporal VF defects and 35 controls underwent SAP and FD-OCT (3D OCT-1000; Topcon Corp.) examinations. Macular thickness measurements were averaged for the central area and for each quadrant and half of that area, whereas RNFL thickness was determined for six sectors around the optic disc. VF loss was estimated in six sectors of the VF and in the central 16 test points in the VF. The correlation between VF loss and OCT measurements was tested with Spearman's correlation coefficients and with linear regression analysis. Macular and RNFL thickness parameters correlated strongly with SAP VF loss. Correlations were generally stronger between VF loss and quadrantic or hemianopic macular thickness than with sectoral RNFL thickness. For the macular parameters, we observed the strongest correlation between macular thickness in the inferonasal quadrant and VF loss in the superior temporal central quadrant (rho=0.78; P<0.001) whereas for the RNFL parameters the strongest correlation was observed between the superonasal optic disc sector and the central temporal VF defect (rho=0.60; P<0.001). Although FD-OCT RNFL and macular thickness measurements were both correlated with VF loss, the correlation was stronger with quadrantic macular than with RNFL thickness measurements in patients with temporal hemianopia. Such measurements could potentially be used to quantify neuronal loss in patients with chiasmal compression.

Purpose: To assess the reproducibility of retinal nerve fiber layer (RNFL) and macular thickness by spectral domain optical coherence tomography (SD-OCT) when the same investigator does scan thrice in a span of one hour without reference to the previous scan, is able to get similar results or not, without using the repeat function.Methods. In this prospective observational study 200 subjects who fulfilled the inclusion &amp; exclusion criteria were scanned 3 times as per pre-defined guidelines at 0 minutes, 30 minutes; 60 minutes on the same day, by the same investigator using SD-OCT for measurements of RNFL and macular thickness &amp; observations were statistically analyzed &amp; correlated.Results. In RNFL thickness, temporal sector shows the worst reproducibility as compared to other sectors. The RNFL thickness was greatest in superior quadrant and thinnest in temporal quadrant. Female values were significantly higher than males in RNFL superior and RNFL symmetry. For macular thickness, temporal sector (mid-zone) showed the worst reproducibility and in outer-zone, Inferior sector showed the worst reproducibility. It also shows that macular thickness was thinnest at the central zone (innermost 1 mm ring), thickest within the inner 3 mm ring and diminished peripherally.Conclusion. RNFL and macular thickness measurements by SD-OCT by the same observer at 0 minutes, 30 minutes and 60 minutes were very reproducible except in the sectors specifically mentioned. The greater the thickness of RNFL in any sector, the better will be the reproducibility in that sector. For macular thickness, temporal sector (mid-zone) showed the worst reproducibility &amp; with an increase in age the macular thickness measurements decreases.

To evaluate the effects of ocular rotation on parapapillary retinal nerve fiber layer (RNFL) thickness measured by spectral-domain optical coherence tomography (SD-OCT). Eighty-eight normal and 205 glaucomatous eyes were studied. RNFL thickness was measured by 3D OCT. Ocular rotation angle was measured from a fundus image obtained by a non-mydriatic fundus camera equipped with 3D OCT. The average, hemi-superior, and hemi-inferior RNFL thicknesses as well as those in the 4 quadrants and the 16-segmented superotemporal (ST-1) and inferotemporal (IT-4) sectors were compared both before and after correcting for ocular rotation. Receiver operating characteristic curves and the areas under the curve (AUC) for the RNFL thicknesses were calculated on the basis of the data from glaucomatous and control eyes. The relationships between RNFL thickness and retinal sensitivity in the corresponding visual field were analyzed using a Humphrey field analyzer. Correction for ocular rotation did not affect the AUCs of the hemi-superior and hemi-inferior RNFL thicknesses. RNFL thicknesses in all of the quadrants and in ST-1 and IT-4 were significantly changed by correcting for ocular rotation. The correlations between all RNFL sectors and retinal sensitivity were not changed by correcting for the ocular rotation angle. Ocular rotation compensation affected RNFL thickness measurement with 3D OCT. However, the effect was clinically negligible in the diagnosis of glaucoma.

Retinal Nerve Fiber Layer Imaging with Spectral-domain Optical Coherence Tomography: Patterns of Retinal Nerve Fiber Layer Progression

Background: Phacoemulsification is a common cataract operation nowadays. During phacoemulsification, variation in intraocular pressure (IOP) may occur, which might change the retinal nerve fiber layer (RNFL) thickness. This study was aimed to evaluate the change in peripapillary RNFL thickness and mean deviation (MD) of visual field after phacoemulsification in chronic primary glaucoma and non-glaucoma patients. Methods: Cohort prospective study was done on 26 patients (13 chronic glaucoma eyes and 13 non-glaucoma eyes) who underwent phacoemulsification. The changes in peripapillary RNFL thickness and MD of visual field were measured as the primary outcome. Comparison between pre- and post-surgery was analyzed with paired t-test, while unpaired t-test was used for comparison between groups.Results: There were no significant changes in RNFL thickness on both groups. Average RNFL thickness in glaucoma group before and after phacoemulsification were 94.9±20.0 μm and 99.1±21.3 μm, respectively (p&gt;0.05). Average RNFL thickness in non-glaucoma group were 100.2±11.1 μm and 101.7±6.8 μm, respectively (p&gt;0.05). Glaucoma patients yielded decreasing mean deviation (MD) of visual field, but it was not statistically significant (p=0.071). In contrast, the MD of visual field after surgery was significantly increased in non-glaucoma group (p=0.005).Conclusion: Phacoemulsification tended to increase peripapillary RNFL thickness in glaucoma or non-glaucoma patients. The visual field tended to decrease in glaucoma patients, but was significantly increased in non-glaucoma patients.

目的 应用光学相干断层成像术(optical coherencetomography,OCT)测量青光眼患者的视网膜神经纤维层(reti-nal nerve fiber layer,RNFL)厚度和黄斑厚度,比较两者在青光眼早期诊断中的作用.方法 应用Stratus OCT测量62例(101眼)正常人和41例(64眼)青光眼患者的RNFL厚度和黄斑厚度,比较正常人和青光眼患者之问,正常人和早期青光眼患者之间上、下、鼻、颞四个象限与平均RNFL厚度、黄斑区内外环各象限厚度和总黄斑体积等参数有无统计学差异;用受试者工作特征曲线下面积(area under the receiveoperating characteristic curve,AROC)评价各个参数在青光眼早期诊断中的作用.结果 青先眼患者各象限、平均RNFL厚度和各分区黄斑厚度、总黄斑体积均比正常人减少,差异有显著性(P＝0.014～0.000),视乳头旁平均RNFL的AROC最大(0.961),其次为下方RNFL(0.928)和上方RNFL(0.924).黄斑下方外环区AROC最大(0.876).早期青光眼患者各象限、平均RNFL厚度和黄斑外环各分区、总黄斑体积较正常人减少,差异有显著性(P＝0.021～0.000),而黄斑内环各区与正常人差异无显著性.视乳头旁平均RNFL的AROC最大(0.877),其次为上方RNFL(0.783)和下方RNFL(0.767).黄斑下方外环区AROC最大(0.728).结论 测量RNFL厚度较测量黄斑厚度具有更强的青光眼早期诊断价值.黄斑厚度测量为青光眼的早期诊断提供了一种新的手段。

To compare retinal nerve fiber layer (RNFL) and macular thickness measurements using time-domain optical coherence tomography (TD OCT) and 2 Fourier-domain OCTs (FD OCT). Sixty eyes of 60 normal participants underwent eye examination followed by OCT evaluation with the time-domain Stratus OCT (Carl Zeiss Meditec, Inc. Dublin, CA) and 2 FD OCTs-the Cirrus HD-OCT (Carl Zeiss Meditec, Inc. Dublin, CA) and the 3D OCT-1000 (Topcon Corporation, Tokyo, Japan). Each patient was scanned 3 times on each machine by the same physician examiner. RNFL and macular thickness measurements were compared across the 3 devices using repeated measures analysis of variance (ANOVA) and multiple comparison t tests with Bonferroni adjustments. The coefficients of variation (CV) and the intraclass correlation coefficients were calculated to assess reproducibility. Bland-Altman plots were constructed to assess the level of agreement between every 2 OCTs. All 3 machines revealed similar patterns of regional differences in RNFL and macular thicknesses. Macular thickness was consistently greatest on the Cirrus and lowest on the Stratus device. However, RNFL measurements were less consistent. The overall average RNFL thickness was 4 μm greater on the Stratus than on the 3D OCT-1000, which was in turn 5 μm greater than on the Cirrus device. Bland-Altman plots showed poor-to-moderate agreement in terms of macular and RNFL thickness measurements between each 2 of the OCT devices. Both FD OCTs provided more reproducible readings on macular measurements than the Stratus, as evidenced by the lower CV and higher intraclass correlation coefficients. However, RNFL measurements on the 2 FD OCTs were not necessarily more reproducible than those from the Stratus OCT. Thickness measurements are not interchangeable among different OCT devices because of the poor-to-moderate interdevice measurement agreement. FD instruments yield more reproducible macular but not RNFL thickness measurements.