Detection of Longitudinal Visual Field Progression in Glaucoma Using Machine Learning

To evaluate the use of global indices summarizing the current status of a patient's visual field as predictors of their future rate of change. Ninety-five subjects with early or suspected glaucoma were studied, of whom 50 exhibited glaucomatous optic neuropathy (GON) at baseline. Subjects underwent seven annual visual field tests. Results from the first test in the sequence were used to predict their subsequent rate of change. Two global indices were considered: mean deviation (MD) and pattern standard deviation (PSD). Using multiple linear regression, baseline MD predicted subsequent slope of MD significantly better than baseline PSD predicted subsequent slope of PSD (p = 0.017). Using multiple logistic regression, a worse initial MD was predictive of being in the worst tertile for subsequent change in MD (pseudo-R2 = 0.33 for subjects with GON at baseline and 0.31 for those without). Worse initial PSD was not predictive of being in the worst tertile for subsequent change in PSD (pseudo-R2 = 0.09 with GON at baseline, 0.10 without). Among patients with otherwise similar clinical profiles, a worse visual field at baseline, as measured by the global index MD, is predictive of a more rapid future rate of change. This should be taken into account when clinical decisions are made concerning management of patients who already have some visual field damage at presentation.

PurposeTo quantitatively evaluate the association of intraocular pressure (IOP) reduction with visual field defect (VFD) progression in normal tension glaucoma (NTG) under medical therapy.Patients and methodsClinical data for 622 eyes of 311 primary open-angle glaucoma and NTG patients were collected from April 2006 to March 2016. Of these patients, those with normal IOP, glaucomatous VFD by Anderson’s criteria, corrected visual acuity ≥0.7, >5 years of medical therapy, ≥5 visual field tests at least five times at 12-month intervals, visual field testing reliability coefficients of ≤33%, and mean deviation (MD) exceeding below −20 dB in the initial visual field test were included in this retrospective data analysis. MD and IOP data were collected at baseline and after 5 years. Following MD categorization into stages I to IV, stage transition matrices were generated using a Markov model to evaluate VFD changes. Eyes were divided based on IOP reduction (0%, 10%, 15%, 20%, 25%, 30%) from baseline. VFD aggravations were compared using the Markov model and MD slopes with IOP reduction rates as cutoff values.ResultsOverall, 132 eyes of 132 NTG patients fulfilled the eligibility criteria. MD decreased significantly (P<0.0001) at 5 years vs baseline. During follow-up, visual field stage using the Markov model was constant in ~60%, with transitions in ~40%. IOP decreased significantly (P>0.001) at 5 years vs baseline. Though MD slopes did not differ significantly between each of the groups that achieved the various IOP reduction cut-off values and the corresponding group that did not achieve the cut-off values, a significant difference (P=0.0432) in VFD was found between the group that achieved the 25% cut-off value and the group that did not when evaluated using the Markov model.ConclusionIn NTG patients, VFD aggravation was significantly suppressed in groups with IOP reductions of ≥25% from baseline.

Glaucoma and Associated Visual Acuity and Field Loss Significantly Affect Glaucoma-Specific Psychosocial Functioning

Background/aimsTo identify clinical characteristics and factors associated with microcystic macular edema (MME) in patients with primary open-angle glaucoma (POAG).MethodsWe included 315 POAG eyes between 2010 and 2019 with good-quality macular...

A New Index to Monitor Central Visual Field Progression in Glaucoma

To validate a method for visual field (VF) progression analysis, called ANSWERS (Analysis with Non-Stationary Weibull Error Regression and Spatial Enhancement), which takes into account increasing measurement variability as glaucoma progresses and spatial correlation among test locations. ANSWERS outputs both a global index of progression and a pointwise estimate of rate of change at each VF location. ANSWERS was compared with linear regression of mean deviation (MD) and permutation of pointwise linear regression (PoPLR). Visual field series of up to 2 years from the United Kingdom Glaucoma Treatment Study were used. This consists of 9104 Swedish Interactive Thresholding Algorithm Standard 24-2 VFs. ANSWERS and PoPLR rate of change were used to predict the VF at the next visit using subseries that were within 7, 13, 18, or 22 months from the baseline. The comparison was carried out on the statistical sensitivity, specificity, and accuracy of predicting future VF. Across all subseries, statistical sensitivity of ANSWERS in detecting VF deterioration was significantly better than the linear regression of MD and PoPLR, especially in short time series. Prediction accuracy of ANSWERS was better than PoPLR at all series lengths, and the improvement was particularly marked in shorter series. Seventy-five percent of VF series were better predicted by ANSWERS compared with PoPLR. The average prediction error of ANSWERS was 15% lower than that of PoPLR. ANSWERS is more sensitive to detect VF progression and predicts future VF loss better than linear regression of MD and PoPLR, especially over short observation periods. (http://www.isrctn.com number, ISRCTN96423140.).

Sector-specific Association of Intraocular Pressure Dynamics in Dark-room Prone Testing and Visual Field Defect Progression in Glaucoma

PurposeThis concise review aims to explore the potential for the clinical implementation of artificial intelligence (AI) strategies for detecting glaucoma and monitoring glaucoma progression.MethodsNonsystematic literature review using the search combinations “Artificial Intelligence,” “Deep Learning,” “Machine Learning,” “Neural Networks,” “Bayesian Networks,” “Glaucoma Diagnosis,” and “Glaucoma Progression.” Information on sensitivity and specificity regarding glaucoma diagnosis and progression analysis as well as methodological details were extracted.ResultsNumerous AI strategies provide promising levels of specificity and sensitivity for structural (e.g. optical coherence tomography [OCT] imaging, fundus photography) and functional (visual field [VF] testing) test modalities used for the detection of glaucoma. Area under receiver operating curve (AROC) values of > 0.90 were achieved with every modality. Combining structural and functional inputs has been shown to even more improve the diagnostic ability. Regarding glaucoma progression, AI strategies can detect progression earlier than conventional methods or potentially from one single VF test.ConclusionsAI algorithms applied to fundus photographs for screening purposes may provide good results using a simple and widely accessible test. However, for patients who are likely to have glaucoma more sophisticated methods should be used including data from OCT and perimetry. Outputs may serve as an adjunct to assist clinical decision making, whereas also enhancing the efficiency, productivity, and quality of the delivery of glaucoma care. Patients with diagnosed glaucoma may benefit from future algorithms to evaluate their risk of progression. Challenges are yet to be overcome, including the external validity of AI strategies, a move from a “black box” toward “explainable AI,” and likely regulatory hurdles. However, it is clear that AI can enhance the role of specialist clinicians and will inevitably shape the future of the delivery of glaucoma care to the next generation.Translational RelevanceThe promising levels of diagnostic accuracy reported by AI strategies across the modalities used in clinical practice for glaucoma detection can pave the way for the development of reliable models appropriate for their translation into clinical practice. Future incorporation of AI into healthcare models may help address the current limitations of access and timely management of patients with glaucoma across the world.

Artificial Intelligence Classification of Central Visual Field Patterns in Glaucoma

PurposeTo investigate parapapillary choroidal microvasculature dropout (MvD) in branch retinal vein occlusion (BRVO) patients and compare them with open-angle glaucoma (OAG) patients using optical coherence tomography angiography (OCT-A).MethodsIn total, 85 eyes of BRVO patients and 85 eyes of OAG patients, matched by age, spherical equivalent, and baseline mean deviation (MD) of the visual field (VF), were assessed. MvD was defined as complete loss of microvasculature within the choroidal layer on OCT-A. Linear regression analysis was used to obtain the slope of the MD change of the VF.ResultsThe presence of MvD on OCT-A was significantly more frequent in OAG eyes (63.1%) compared to BRVO eyes (31.8%). BRVO eyes with MvD showed worse baseline MD of the VF than BRVO eyes without MvD (−10.19 ± 8.50 and −7.77 ± 6.46 dB, respectively; P = 0.045). The presence of MvD was the only factor significantly associated with MD change of the VF in OAG eyes. Lower baseline average RNFL thickness, greater MvD angle, and lower macular superficial vessel density were significantly associated with MD change of the VF in BRVO eyes.ConclusionsOCT-A of the parapapillary area showed choroidal microvasculature impairment in both BRVO and OAG patients. However, the frequency was higher in glaucoma patients with similar degrees of VF damage, which suggests that the glaucomatous process contributes to MvD development. The effect of MvD on VF change was different between BRVO and OAG, suggesting that the underlying pathogenesis may also be different.

Previous work has investigated whether a significant regression slope in the first 2 years for the summary index Mean Deviation (MD) is predictive of rapid (≤-2 dB year-1 ) glaucomatous visual field progression. This work assumed six visual fields were obtained as per management guidelines, but in clinical practice commonly only two or three fields are measured. We used simulation methods to investigate how reducing test frequency influences the prediction of rapid visual field progression, along with the influence of including criteria based on regression slope. We simulated visual field series (N = 100 000) spaced annually in the first 2 years and then biennially. We calculated positive and negative predictive values (PPV & NPV) for detecting rapid progression, based on a criterion of a significant negative regression slope of any magnitude, or of a magnitude less than a particular limit. We performed a second simulation using test frequency and disease prevalence parameters from a dataset of 255 glaucoma patients from The University of Tokyo Hospital, to check the validity of our method. Prediction values at 2 years were slightly less than those obtained using six visual fields. An addition of an appropriate slope based criterion materially improved PPV, with little detrimental effect on NPV. Simulated prediction values for the Tokyo dataset were similar to those determined empirically. Infrequent visual field testing does not dramatically alter predictive values at 2 years, but does substantially delay when significant progression may first be detected.

Estimating the Reliability of Glaucomatous Visual Field for the Accurate Assessment of Progression Using the Gaze-Tracking and Reliability Indices

Artificial Intelligence-Guided Endpoint Selection for Neuroprotection Trials in Glaucoma.

To evaluate the effect of trabeculectomy (Trab MMC) on visual field (VF) progression in eyes with glaucoma and high myopia. Patients diagnosed with primary open-angle glaucoma or exfoliation glaucoma who underwent Trab MMC as the first glaucoma surgery along with ≥3 VF tests preoperatively and postoperatively were enrolled. High myopia was defined as an axial length ≥26.5 mm. Postoperative reductions in intraocular pressure (IOP) were assessed by survival analysis using IOP measurements obtained preoperatively. The longitudinal trends of the outcome measures were evaluated using linear mixed models. Thirty-five eyes of 32 patients were included in this study, including 22 eyes of 20 patients in non-highly myopic group and 13 eyes of 12 patients in highly myopic group. IOP decreased after Trab MMC, and the survival rate did not differ significantly in relation to axial length. Linear mixed-model analyses suggested that the inhibitory effects of Trab MMC on the rate of mean deviation (MD) changes were significant in the non-highly myopic group (-0.53 ± 0.15 dB/year preoperatively to -0.16 ± 0.13 dB/year postoperatively; P = 0.004), but not in the highly myopic group (-0.66 ± 0.19 dB/year preoperatively to -0.48 ± 0.18 dB/year postoperatively; P = 0.32). Trab MMC reduced IOP in both highly myopic and non-highly myopic eyes, and IOP reduction was very similar in both groups. The VF deterioration rate decreased in both groups, but the change was weaker and nonsignificant in the highly myopic group.

Purpose: To determine the effect of cataract extraction on visual field damage in glaucoma patients with cataracts. Methods: Sixty-two eyes of fifty glaucoma patients (primary open-angle glaucoma, angle-closure glaucoma, and normal-tension glaucoma) were divided into three groups according to the MD (mean deviation) value from the Humphrey automatic visual field test. The PSD (pattern standard deviation) values were analyzed before and after the cataract surgery. Additionally, the visual acuity, intraocular pressure and the number of antiglaucomatous agents were compared. Results: All of the three groups, showed significant improvement in visual acuity. However, there was only a significant decrease in intraocular pressure at three months postoperatively. When comparing the mean MD values of each group, there was a statistically significant difference between the mild defect group and the severe defect group, but no statistically significant difference in the moderate defect group. There was also no statistically significant difference in PSD values in the any of the groups before and after cataract surgeries. Conclusions: In glaucoma patients with cataracts, there were no significant differences in PSD values before and after cataract surgery. Additionally, to perform an adequate cataract operation in advanced glaucoma patients without the progression of visual field damage, lowering intraocular pressure and improving visual acuity is helpful.