Mori Diego Martinez, Kuchhangi Ashika, Tame Jessica, Cooper Karen, Hajkazemshirazi Leila, Indaram Maanasa, Keenan Jeremy D, Oatts Julius T
University of California (D.M.M., A.K., J.T., K.C., L.H., M.I., J.K., J.T.O), San Francisco, California, USA.
University of California (D.M.M., A.K., J.T., K.C., L.H., M.I., J.K., J.T.O), San Francisco, California, USA; Francis I. Proctor Foundation (J.K.), San Francisco, California, USA.
Am J Ophthalmol. 2025 Jan;269:266-272. doi: 10.1016/j.ajo.2024.08.042. Epub 2024 Sep 6.
A sensorimotor examination is the gold standard for strabismus diagnosis and quantification but requires a highly skilled examiner and may be limited by a child's cooperation. Virtual reality (VR) employs eye-tracking technology to monitor eye position and may be able to measure strabismus. The aim of this study was to assess a prototype VR-simulated alternate cover test to detect and measure strabismus.
Prospective, masked diagnostic test study.
Patients aged 5-18 years with visual acuity of 20/80 or better were prospectively enrolled to undergo strabismus measurements using a VR simulated alternate cover test (Olleyes, Inc., Summit, NJ) followed by an alternate cover test performed by a masked pediatric ophthalmologist or orthoptist. The main outcome measure was correlation between gold standard and VR-obtained strabismus measurements (in prism diopters [PD]) in primary gaze at near using Pearson correlation coefficients and Bland-Altman analysis with limits of agreement (LOA). A secondary measure was the diagnostic accuracy for the VR headset to detect strabismus.
A total of 85 participants were enrolled, mean ± standard deviation age was 10.8 ± 3.8 years, 45.9% (39/85) male. 40.0% (34/85) had strabismus: 17.7% (15/85) esotropia, 22.4% (19/85) exotropia, and 5.9% (5/85) vertical strabismus. 52.9% (18/34) of strabismus was intermittent. The overall correlation between VR and gold standard strabismus measurements was moderate but significant (r = 0.42, 95% CI 0.22, 0.58, P < .001), and correlation was strong for esotropia and constant deviations (r = 0.74, 95% CI 0.38, 0.91, P = .001 and r = 0.74, 95% CI 0.39, 0.91, P < .001, respectively). In participants with horizontal strabismus, Bland-Altman analysis showed a mean difference between standard and VR measurements of 3.55 ± 8.33 PD for esotropia (upper and lower LOA 19.89, -12.78 PD) and 17.15 ± 11.20 PD for exotropia (LOA 39.09 and -4.79 PD). Sensitivity for detecting strabismus was low: 27.6% (95% CI 12.7, 47.2), but specificity was high: 87.5% (95% CI 75.9, 94.8).
A prototype VR simulated alternate cover test showed a moderate but significant correlation with the gold standard sensorimotor examination and correlation was strong in those with esotropia and constant deviations. While the level of agreement demonstrated by this novel VR technology is promising, further improvements are needed before clinical deployment. However, this study demonstrates that VR has the potential to expand our ability to detect, measure, and monitor strabismus.
感觉运动检查是斜视诊断和定量的金标准，但需要技术高超的检查者，且可能受儿童配合度的限制。虚拟现实（VR）利用眼动追踪技术监测眼睛位置，或许能够测量斜视。本研究的目的是评估一种用于检测和测量斜视的VR模拟交替遮盖试验原型。
前瞻性、盲法诊断试验研究。
前瞻性纳入5至18岁、视力为20/80或更好的患者，使用VR模拟交替遮盖试验（Olleyes公司，新泽西州萨米特）进行斜视测量，随后由一位盲法儿科眼科医生或视光师进行交替遮盖试验。主要结局指标是在近距离原发性注视时，金标准测量与VR获得的斜视测量结果（以棱镜度[PD]表示）之间的相关性，采用Pearson相关系数和具有一致性界限（LOA）的Bland-Altman分析。次要指标是VR头戴设备检测斜视的诊断准确性。
共纳入85名参与者，平均年龄±标准差为10.8±3.8岁，45.9%（39/85）为男性。40.0%（34/85）患有斜视：17.7%（15/85）为内斜视，22.4%（19/85）为外斜视，5.9%（5/85）为垂直斜视。52.9%（18/34）的斜视是间歇性的。VR与金标准斜视测量结果之间的总体相关性中等但具有显著性（r = 0.42，95%可信区间0.22，0.58，P <.001），内斜视和恒定性斜视的相关性较强（分别为r = 0.74，95%可信区间0.38，0.91，P =.001和r = 0.74，95%可信区间0.39，0.91，P <.001）。在水平斜视参与者中，Bland-Altman分析显示，内斜视的标准测量与VR测量之间的平均差异为3.55±8.33 PD（LOA上限和下限分别为19.89，-12.78 PD），外斜视的平均差异为17.15±11.20 PD（LOA为39.09和-4.79 PD）。检测斜视的敏感性较低：27.6%（95%可信区间12.7，47.2），但特异性较高：87.5%（95%可信区间75.9，94.8）。
一种VR模拟交替遮盖试验原型与金标准感觉运动检查显示出中等但显著的相关性，在内斜视和恒定性斜视患者中相关性较强。虽然这种新型VR技术显示出的一致性水平很有前景，但在临床应用前还需要进一步改进。然而，本研究表明VR有潜力扩展我们检测、测量和监测斜视的能力。
