Diagnostic accuracy of indigenously developed computer-based binocular vision assessment.

CONTEXT: The increased prevalence of nonstrabismic binocular vision anomalies (NSBVA) has given rise to the need for cost-effective screening and diagnostic tools. AIMS: The aim of the study is to assess the efficacy of an indigenously developed computer-based binocular vision assessment software (Train Your Eyes®) in screening NSBVA. SUBJECTS AND METHODS: Subjects who visited the binocular vision clinic of a tertiary eye care center with asthenopic symptoms between January 2019 and January 2020 were included in the study. Patients with other ocular comorbidities and stereopsis poorer than 500 arc seconds were excluded. All subjects underwent a comprehensive eye examination followed by binocular vision assessment using both the manual and computer-based methods. STATISTICAL ANALYSIS USED: Receiver operating characteristic (ROC) curves were utilized to choose the cut-off points that maximize the sensitivity and specificity.
RESULTS: The mean (standard deviation) age of 88 subjects was 22 (4.5) years with 34 males. Based on the conventional manual assessment, 71 (81%) were diagnosed to have NSBVA and 17 (19%) had normal binocular vision. Based on the ROC analysis, the following cut-off points are proposed: 14 prism diopter (PD) for near positive fusional vergence amplitudes, 4.5 PD for near negative fusional vergence amplitudes, 4.5 cycles per minute (cpm) for binocular accommodative facility, and 3.5 cpm for monocular accommodative facility. All the binocular vision parameters demonstrated statistical significance in the ROC analysis (P < 0.05).
CONCLUSIONS: The software-based screening tool was found to be highly sensitive in identifying NSBVA and thus could be used as a potential screening tool in the clinic and community. Copyright:

Introduction

Binocular vision and accommodative anomalies are termed as “Nonstrabismic Binocular Vision anomalies” (NSBVA) represent the most common visual disorder compared to refractive errors among clinical pediatric populations.[12] The prevalence of NSBVA among schoolchildren in Southern India was reported to be 30.8%.[3] Based on BAND study, the minimum battery of tests required to screen NSBVA in the community set-up included phoria for distance and near, near point of convergence with red filter, and monocular accommodative facility through manual assessment.[4] The expected findings or normative data for binocular vision and accommodative testing used in the diagnosis and classification of NSBVA also vary by ethnicity.[567]

Computer-based orthoptics or vision therapy enables the standardization of the testing and treatment protocols.[6] In the recent past, the use of computers has seen a steady upsurge, increasing the accessibility to provide computer orthoptics for diagnosing and giving therapy programs. The convergence insufficiency treatment trial utilized computer-based vision therapy (home therapy solutions – HTS Inc.,) programs to provide in-office and home-based vision therapy. Accommodative and vergence parameters have shown clinically significant improvement with home vision therapy (HTS) program that utilizes randot stereogram utilizing an operant conditioning paradigm. Eight various other studies have also reported a reduction of asthenopic symptoms following computer-based vision therapy.[8910]

There is still paucity in the literature about the diagnostic accuracy of computer-based software compared to conventional manual binocular vision assessment. Considering the increased prevalence and need for diagnosing NSBVA, this study aimed to assess the diagnostic accuracy and efficacy of an indigenously developed computer-based orthoptic program (Train your eyes®).

Subjects and Methods

Subjects

This prospective study was carried out between January 2019 and January 2020. Subjects with asthenopic symptoms and age group between 8 and 35 years who visited the binocular vision clinic of a tertiary eye care center in Southern India were included in the study. Patients with Best Corrected Visual Acuity (BCVA) greater than 6/9, N6 were included in the study. With stereopsis poorer than 500 arc seconds and other ocular comorbidities were excluded. The study was approved by the Institutional review board of Vision Research Foundation and adhered to the Tenets of Declaration of Helsinki.

A sample size of 61 subjects was estimated based on the primary outcome measure of sensitivity considering 80% sensitivity and 10% precision.[11]

Manual conventional assessment

All subjects underwent a comprehensive eye examination followed by a detailed binocular vision assessment using the standard manual method. Binocular vision parameters included sensory evaluation, distance and near phoria measurement, near point of convergence, near point of accommodation, dynamic retinoscopy (monocular estimation method), negative and positive relative accommodation, vergence amplitudes for distance and near, vergence and accommodative facility.[12]

Computer-based binocular vision assessment

The indigenously developed computer-based Train Your Eyes (TrYE) software was intended to aid as a quick screening tool in the diagnosis of NSBVA. The software has been designed to execute a binocular vision assessment, in-office vision therapy, and home-based vision therapy. It uses Random Dot Stereogram (RDS) as a stimulus for testing vergence parameters combined with red-blue anaglyph glasses. Standard ± 2.00 DS accommodative flippers are utilized with Landolt C targets as stimulus to test for accommodative facility. The oculomotor functions of Saccades and Pursuits are assessed using a Tumbling E as a target. The software is calibrated for testing at a distance of 40 cm.

Before performing the computer-based assessment, the subjects were informed as to the different targets that would be used during the assessment and the responses that are expected out of them. The manual and computerized assessments were carried out by masked experienced examiners.

For the vergence testing, the RDS stimulus is initially presented at a disparity of zero, and as the patient gives a positive feedback, the disparity is increased by one prism diopter (PD) steps for convergence and by 0.5 PD steps for divergence. Moreover, for a negative feedback, the disparity gets reduced by one step and the disparity further reduces until a positive response is keyed in. The vertical fusional ranges were also assessed using the same principle.

For accommodative facility, the patient would have to use the accommodative flippers and clear the targets (Landolt C) displayed on screen. The patient would be presented with four such targets which have to be cleared out by the patient while viewing through the lenses of the accommodative flippers and a response is to be keyed into the keyboard. Upon completion, the patient is required to flip to the other side of the flipper and repeat the same procedure for 1 min.

The designed software was validated by comparing the near fusional amplitude ranges (positive and negative) and accommodative facility (monocular and binocular) with the conventional manual assessment.

Data analysis

The results of TrYE and the standard examination were recorded in a Microsoft Excel Spreadsheet and analyzed using Statistical Package for the Social Sciences (SPSS) Version 20.0 for sensitivity and specificity. Receiver operating characteristic (ROC) curves were plotted for the outcome measures of fusional vergence parameters and monocular and binocular accommodative facility.

Results

A total of 88 consecutive subjects were recruited among which 34 were males. The mean (standard deviation) age of the subjects in our study was 22 ± 4.5 years. The range of phoria for distance was orthophoria to 20 PD exophoria (median with interquartile range [IQR]: 0[0] and near phoria ranges from 6 PD esophoria to 30 PD exophoria (median with IQR: 0 [−4–0]), respectively. The median and IQR for near positive and negative fusional vergence amplitudes and monocular and binocular accommodative facility using manual assessment and TrYE software are shown in Table 1. Based on the diagnostic criteria by Scheiman and Wick,[12] 71 subjects (81%) were diagnosed to have NSBVA of which 38 subjects (43%) had accommodative infacility and 21 subjects (24%) had convergence insufficiency with accommodative dysfunction [Figure 1].

Table 1

Median and interquartile range for binocular vision parameters

Test parameters (near)	Median (IQR)
	Manual assessments	TRYe software assessments
Negative fusional vergence (PD)	14 (10.0-18.0)	4 (2.5-9.6)
Positive fusional vergence (PD)	30 (19.5-40)	25 (10.5-26)
Monocular accommodative facility right eye (cpm)	5.5 (2.4-12.0)	2 (1.3-4.0)
Monocular accommodative facility left eye (cpm)	6 (2.0-12.0)	2 (1.5-3.6)
Binocular accommodative facility (cpm)	6 (2.5-9.6)	4 (2.7-5.5)

IQR: Interquartile range, PD: Prism Diopter, TRYe: Train your eyes

Figure 1

Percentage of nonstrabismic binocular vision anomalies. CI: Convergence insufficiency, AIF: Accommodative facility, CE: Convergence excess, FVD: Fusional vergence dysfunction, IDS: Intermittent divergent squint

ROC curve analysis showed statistically significance for cut-off points of 4.5 PD and 14 PD for near negative (P = 0.007) and positive fusional vergence (P = 0.034) amplitudes, respectively. The sensitivity, specificity, and area under curve (AUC) for negative fusional vergence as 70.3%, 60.8%, and 0.67 and for positive fusional vergence as 63.6%, 74.2% and AUC as 0.651, respectively [Figure 2a and b].

Figure 2

Receiver operating characteristic curve for (a) negative fusional vergence amplitude and (b) positive fusional vergence amplitude

Similarly, ROC analysis for accommodative facility showed statistically significance for cut-off points of 3.5 cycles per minute (cpm) (P = 0.001) for right eye and left eye (P < 0.001) and 4.5 cpm for both the eye accommodative facility (P = 0.006) [Figure 3a-c].

Figure 3

Receiver operating characteristic curve for (a) right eye accommodative facility; (b) left eye accommodative facility and (c) both eye accommodative facility

The sensitivity, specificity, and AUC for right eye accommodative facility was 79.7%, 58.56%, and 0.723, left eye accommodative facility was 84.4%, 62.5%, 0.794 and for both eye accommodative facility was 69.4%, 67.7% and 0.721, respectively.

Monocular and binocular accommodative facility and negative fusional vergence amplitude had higher specificity and sensitivity among the five diagnostic parameters. The results of the AUC along with the 95% confidence interval ranges, sensitivity, specificity, positive and negative likelihood ratios for each parameter, and the respective cut-off points obtained with ROC curves are shown in Table 2.

Table 2

Receiver operating characteristic curve analysis for binocular vision parameters

Test parameters (near)	Cut off	Sensitivity (%)	Specificity (%)	AUC	P (AUC)	LR+	LR−
Negative fusional vergence	4.5 (PD)	70.30	60.80	0.67	0.007	1.79	0.48
Positive fusional vergence	14 PD)	63.60	74.20	0.651	0.034	2.47	0.49
Monocular accommodative facility right eye	3.5 (cpm)	79.70	58.56	0.723	0.001	1.92	0.49
Monocular accommodative facility left eye	3.5 (cpm)	84.40	62.50	0.794	0.000	2.26	0.25
Binocular accommodative facility	4.5 (cpm)	69.40	68.70	0.721	0.006	2.22	0.45

AUC: Area under the curve, LR+: Positive likelihood ratio, LR−: Negative likelihood ratio, PD: Prism Diopter, cpm: Cycles per min

Discussion

The conventional method of measuring fusional vergence amplitudes and accommodative facility requires trained practitioners and also involves considerable time commitment. Computer orthoptics simplifies the process, brings in objectivity to the assessment, reduces the time considerably, and also could serve as a screening tool due to the increased burden of NSBVA. This is even more relevant in the current COVID-19 scenario where increased concerns about time spent in the clinics prevail. The automatized measurements using a computerized program increase the measurement's repeatability by reducing the measurement errors in assessing binocular vision functions and thus improve the accuracy of diagnosing binocular vision anomalies.[13]

Computer-assisted orthoptics started in the year 1987 for both diagnosis and treatment.[14] It has been well established that RDS targets to measure fusional vergence amplitudes are devoid of monocular cues and also has shown clinically significant improvements in vergence parameters.[1516] A computer-based accommodative facility tested using flippers, and anaglyph with squares as targets was compared with the standard clinical method and found to have little or no effect of accommodative stimulus in the testing software.[17]

The ROC analysis in this study showed good sensitivity and specificity for the near-fusional ranges and accommodative facility, and this can be tried in the place of the manual minimum test battery screening tool as proposed by the BAND study.[4] TrYE software-based binocular vision assessment has an added advantage over the conventional method. It can complete the testing for near-fusional vergence parameters and accommodative facility in 8–10 min and does not require a trained person to run the test. Moreover, the results will be displayed immediately after completing each test along with a pass/fail based on the established cut-off points.

Computer-based screening program is convenient for both examiner and for the patients, and much needed considering the current COVID-19 pandemic situation, computerized tools offer the option to provide teleconsultation and vision therapy. A recent study also supports that the robust model of telehealth vision therapy as a valuable tool for the clinician to provide vision therapy.[18]

Conclusion

The indigenously developed computer software-based screening tool TrYE demonstrated good sensitivity and specificity to screen for NSBVA. This software has its potential application in the clinics and community as a feasible screening and vision therapy tool for NSBVA.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.