The Effect of Ametropia on Glaucomatous Visual Field Loss

Deep Learning Evaluation of Glaucoma Detection Using Fundus Photographs in Highly Myopic Populations

OBJECTIVES

This study aimed to use deep learning to identify glaucoma and normal eyes in groups with high myopia using fundus photographs.

METHODS

Patients who visited Tri-Services General Hospital from 1 November 2018 to 31 October 2022 were retrospectively reviewed. Patients with high myopia (spherical equivalent refraction of ≤-6.0 D) were included in the current analysis. Meanwhile, patients with pathological myopia were excluded. The participants were then divided into the high myopia group and high myopia glaucoma group. We used two classification models with the convolutional block attention module (CBAM), an attention mechanism module that enhances the performance of convolutional neural networks (CNNs), to investigate glaucoma cases. The learning data of this experiment were evaluated through fivefold cross-validation. The images were categorized into training, validation, and test sets in a ratio of 6:2:2. Grad-CAM visual visualization improved the interpretability of the CNN results. The performance indicators for evaluating the model include the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity.

RESULTS

A total of 3088 fundus photographs were used for the deep-learning model, including 1540 and 1548 fundus photographs for the high myopia glaucoma and high myopia groups, respectively. The average refractive power of the high myopia glaucoma group and the high myopia group were -8.83 ± 2.9 D and -8.73 ± 2.6 D, respectively (p = 0.30). Based on a fivefold cross-validation assessment, the ConvNeXt_Base+CBAM architecture had the best performance, with an AUC of 0.894, accuracy of 82.16%, sensitivity of 81.04%, specificity of 83.27%, and F1 score of 81.92%.

CONCLUSIONS

Glaucoma in individuals with high myopia was identified from their fundus photographs.

Association between contrast sensitivity function and structural damage in primary open-angle glaucoma

AIMS

To evaluate the association between contrast sensitivity function (CSF) and glaucomatous structural damage in primary open-angle glaucoma (POAG).

METHODS

A cross-sectional study was performed with 103 patients (103 eyes) aged 25-50 years who had POAG without any other ocular disease. CSF measurements were obtained by the quick CSF method, a novel active learning algorithm that covers 19 spatial frequencies and 128 contrast levels. The peripapillary retinal nerve fibre layer (pRNFL), macular ganglion cell complex (mGCC), radial peripapillary capillary (RPC) and macular vasculature were measured by optical coherence tomography and angiography. Correlation and regression analyses were used to assess the association of area under log CSF (AULCSF), CSF acuity and contrast sensitivities at multiple spatial frequencies with structural parameters.

RESULTS

AULCSF and CSF acuity were positively associated with pRNFL thickness, RPC density, mGCC thickness and superficial macular vessel density (p<0.05). Those parameters were also significantly associated with contrast sensitivity at 1, 1.5, 3, 6, 12, 18 cycles per degree spatial frequencies (p<0.05) and, the lower the spatial frequency, the higher the correlation coefficient. RPC density (p=0.035, p=0.023) and mGCC thickness (p=0.002, p=0.011) had significant predictive value for contrast sensitivity at 1 and 1.5 cycles per degree, with adjusted R 2 of 0.346 and 0.343, respectively.

CONCLUSIONS

Full spatial frequency contrast sensitivity impairment, most notably at low spatial frequencies, is a characteristic change in POAG. Contrast sensitivity is a potential functional endpoint for the measurement of glaucoma severity.

Structural Abnormalities in the Papillary and Peripapillary Areas and Corresponding Visual Field Defects in Eyes With Pathologic Myopia

Purpose

To identify structural abnormalities in the papillary and peripapillary area in eyes with pathologic myopia (PM) and normal IOP and to determine their relationship to visual field (VF) defects.

Methods

One hundred eight eyes of 70 patients with PM were retrospectively studied. The disc-centered swept source optical coherence tomographic images and the Goldmann VF recorded within 1 year of the optical coherence tomographic examination were analyzed. Four structural abnormalities were identified: lamina cribrosa (LC) defects, ridge protrusions, intrachoroidal cavitations (ICC), and prelaminar schisis. The correspondence of the VF defects with the structural abnormalities was assessed.

Results

The mean age, axial length, and optic disc area of the 108 eyes were 58.7 ± 10.0 years, 31.1 ± 2.4 mm, and 4.7 ± 2.2 mm2, respectively. Eighty-five of the 108 eyes (78.7%) had at least one abnormality and 49.4% (42/85) had two or more abnormalities. LC defects, ridge protrusions, ICC, and prelaminar schisis were detected in 47.2%, 33.3%, 21.3%, and 30.6% of the eyes, respectively. VF defects at the corresponding areas of these structural abnormalities were seen in 63% of the eyes with LC defects, 39% of the eyes with ridge protrusions, and 21% of the eyes with ICC.

Conclusions

Four kinds of structural abnormalities with corresponding VF defects are commonly observed in the papillary and peripapillary region of eyes with PM. The presence of these abnormalities suggests a possibility of functional damage.