INCIDENCE, PROGRESSION, AND RISK FACTORS OF EPIRETINAL MEMBRANES IN THE ELDERLY

The Classification of Stages of Epiretinal Membrane using Convolutional Neural Network on Optical Coherence Tomography Image

BACKGROUND AND OBJECTIVE

The gold standard for diagnosing epiretinal membranes is to observe the surface of the internal limiting membrane on optical coherence tomography images. The stages of the epiretinal membrane are used to decide the condition of the health of the membrane. The stages are not detected because some of them are similar. To accurately classify the stages, a deep-learning technology can be used to improve the classification accuracy.

METHODS

A combinatorial fusion with multiple convolutional neural networks (CNN) algorithms are proposed to enhance the accuracy of a single image classification model. The proposed method was trained using a dataset of 1947 optical coherence tomography images diagnosed with the epiretinal membrane at the Taichung Veterans General Hospital in Taiwan. The images consisted of 4 stages; stages 1, 2, 3, and 4.

RESULTS

The overall accuracy of the classification was 84%. The combination of five and six CNN models achieves the highest testing accuracy (85%) among other combinations, respectively. Any combination with a different number of CNN models outperforms any single CNN algorithm working alone. Meanwhile, the accuracy of the proposed method is better than ophthalmologists with years of clinical experience.

CONCLUSIONS

We have developed an efficient epiretinal membrane classification method by using combinatorial fusion with CNN models on optical coherence tomography images. The proposed method can be used for screening purposes to facilitate ophthalmologists making the correct diagnoses in general medical practice.

Sphingosine-1-phosphate expression in human epiretinal membranes

The abnormal posterior vitreous detachment (PVD) is speculated as an important mechanism of the development of the epiretinal membrane (ERM). However, there is only limited information about the molecular mechanism. Sphingosine-1-phosphate (S1P) is a mediator of the mechanosensitive response in several cell types that may have a role in the pathogenesis of ERM during abnormal PVD. Therefore, we evaluated the expression of S1P in the human ERM and the role of S1P in cultured human Muller glial cells. Among 24 ERM specimens, seven specimens (29.2%) exhibited S1P expression. Patients with secondary ERM or ellipsoid zone defects, which suggest abnormal PVD presented a significantly higher S1P+ cell density (secondary ERM: 128.20 ± 135.61 and 9.68 ± 36.01 cells, p = 0.002; EZ defects: 87.56 ± 117.79 vs 2.80 ± 8.85, p = 0.036). The addition of S1P increased the migrative ability and expression of N-cadherin and α-SMA in human Muller glial cells, suggesting S1P is a potential causative molecule for the development of ERM during abnormal PVD.

Epiretinal Membrane Surgery in Eyes with Glaucoma: Visual Outcomes and Clinical Significance of Inner Microcystoid Changes

OBJECTIVE

To investigate visual acuity (VA) outcomes and OCT-based biomarkers of vision outcomes in eyes with glaucoma undergoing pars plana vitrectomy (PPV) for idiopathic epiretinal membrane (ERM).

DESIGN

Retrospective, consecutive case-control series. A previously described ERM grading scale was utilized for OCT analysis.

SUBJECTS

Eyes with glaucoma undergoing PPV for idiopathic ERM.

INTERVENTION

PPV with membrane peel (MP) surgery.

MAIN OUTCOME MEASURES

The primary outcome was VA at postoperative month 6. Outcomes were compared to a contemporary, matched control group of eyes without concurrent glaucoma undergoing PPV for idiopathic ERM.

RESULTS

A total of 103 eyes from 103 patients with ERM and glaucoma were followed for a mean (± standard deviation) of 656 (± 421) days after PPV with MP surgery. Glaucoma was classified as open angle in 98 (95.1%) eyes and closed angle in 5 (4.9%) eyes. Visual acuity improved from 0.72 ± 0.48 (20/105) to 0.55 ± 0.51 (20/71) at 6 months and to 0.50 ± 0.56 (20/63) at final follow-up (P < 0.001 for both the time points). Eyes with preoperative inner microcystoid changes (n = 59; 57.3%) had significantly worse preoperative VA, postoperative VA at month 6, and final VA compared to eyes without inner microcystoid changes (P = 0.028, 0.004, and 0.007, respectively). Eyes were then compared to a matched control group of 139 eyes without glaucoma. Eyes with ERM and glaucoma had a higher rate of microcystic changes both before surgery (P < 0.001) and at postoperative month 6 (P < 0.001), and had a worse VA at 6 months (P = 0.03) and final follow-up (P = 0.04) compared to control eyes without glaucoma. Advanced disc cupping was the only factor independently correlated with worse 6-month (P = 0.01) and final (P = 0.007) VA in multivariate analysis.

CONCLUSIONS

Preoperative inner microcystoid changes on OCT were present in over half of eyes with ERM and concurrent glaucoma, and may be a poor prognostic OCT biomarker. Eyes with ERM and concurrent glaucoma experienced worse vision outcomes compared to eyes with ERM alone, particularly those with advanced disc cupping.

Prevalence of epiretinal membrane in the phakic eyes based on spectral-domain optical coherence tomography

The prevalence of epiretinal membrane (ERM) and associated factors in the phakic eyes have not been fully elucidated yet. This cross-sectional study included 2,354 phakic eyes without retinal diseases or surgical history. Ocular parameters, such as uncorrected distance visual acuity (UDVA), corrected distance visual acuity (CDVA), spherical equivalent (SE), intraocular pressure (IOP), white-to-white corneal diameter (WTW), mean keratometric value (Km) of total corneal refractive power at 4-mm diameter (TCRP4), astigmatism of TCRP4, total corneal irregular astigmatism (TCIA), pupil diameter, axial length (AXL), anterior chamber depth (ACD), lens thickness (LT), and posterior vitreous detachment (PVD) were compared between ERM group and control group. Additionally, an age-matched control group was selected by individual matching and compared with the ERM group to eliminate the confounders. Multiple logistic regression analysis was performed to evaluate the factors associated with the presence of ERM. Among 2,354 eyes, 429 eyes (18.2%) had ERM based on spectral-domain optical coherence tomography. The ERM group showed higher prevalence of PVD, worse CDVA, higher astigmatism of TCRP4, higher TCIA, smaller pupil size, longer AXL, and thicker LT than control group (P < 0.001, P < 0.001, P = 0.011, P < 0.001, P = 0.023, P < 0.001, and P < 0.001, respectively). Only PVD, CDVA, SE, astigmatism of TCRP4, TCIA, and AXL maintained the significance when compared with the age-matched control group (P < 0.001, P = 0.026, P < 0.001, P = 0.001, P = 0.003, and P < 0.001, respectively). Multivariate logistic regression analysis showed that age, PVD, CDVA, and TCIA were independently associated with the presence of ERM (P < 0.001, P < 0.001, P = 0.011, and P = 0.002). The prevalence of ERM detected using SD-OCT was 18.2% in the middle aged phakic population. Eyes with TCIA, in addition to older age and PVD, were more likely to have ERM.