Effect of Individualized and Adaptive Vision Training versus 6 hours of Patching in Children with Residual Amblyopia: A Randomized Clinical Trial

Abstract

Introduction

Despite refractive correction and patching, some patients have residual amblyopia. In the current study, we compared the effectiveness of daily 6-hours (6-h) intensive patching and a newly developed individualized and adaptive vision training (iAVT) protocol for residual amblyopia in children.

Methods

In a randomized clinical trial, 60 children aged 4 to under 11 years with residual amblyopia and visual acuity ranging from 0.2 to 0.8 logarithm of the minimum angle of resolution (logMAR) were assigned to receive either 50 sessions of iAVT training or 6 h of daily patching for 10 weeks.

Results

Visual acuity in the amblyopic eye improved more in the iAVT group than in the extended patching group (0.13 versus 0.07 logMAR; 95% confidence interval (CI) 0.010–0.100; p = 0.017) over 10 weeks. At 2 weeks, the iAVT group also showed a faster improvement compared with the patching group (0.09 versus 0.06 logMAR; 95% CI −0.004 to 0.067; p = 0.076). Meanwhile, no significant between-group difference was found in the area under the log contrast sensitivity function (AULCSF) change. However, the AULCSF of the iAVT group showed a marginally significant within-group improvement from base to week 10 (0.78 versus 0.89; 95% CI −0.225 to 0.001; p = 0.056), while no significant change was observed in the patching group. The iAVT group also reported higher quality-of-life scores after treatment, as monitored by the Pediatric Eye Questionnaire.

Conclusions

The iAVT provided a faster and more effective treatment for residual amblyopia compared with 6-h patching within the 10-week treatment period. These results suggest that iAVT may be an important new approach to treating residual amblyopia.

Trail Registration

Chinese Clinical Trial Registry, ChiCTR2300075594. Registered on 8 September 2023—retrospectively registered.

Key Summary Points

Why carry out this study?

After conventional optical correction, occlusion, and pharmacologic therapy, some children with amblyopia still fail to achieve normal visual acuity and develop residual amblyopia, for which evidence-based treatment options remain limited.

This study aimed to evaluate whether the iAVT protocol could further improve visual acuity in children with residual amblyopia compared with increasing patching therapy.

What was learned from the study?

Both iAVT and extended patching significantly improved amblyopic-eye visual acuity after 10 weeks, but iAVT achieved greater gains, higher compliance, and no adverse effects.

iAVT provided a safe, efficient, and engaging supplementary treatment for children with residual amblyopia.

Introduction

Amblyopia is a common cause of visual function impairment in both children and adults, resulting from abnormal early development of the central visual nervous system [1, 2]. It is characterized by deficits in best-corrected visual acuity (VA) [3], contrast sensitivity function (CSF) [4, 5], stereopsis [6–8], and eye–hand coordination [9, 10].

In clinical practice, amblyopia is typically treated with optical correction, occlusion, and medication [11]. While most children with amblyopia achieve normal visual acuity through these interventions, a significant subset of children fail to fully recover, resulting in residual amblyopia [12, 13]. For example, 6-month to 15-year follow-up results from the Pediatric Eye Disease Investigator Group (PEDIG) showed that 38.7–59% of amblyopic eyes in the patching group and 41.7–71% in the atropine group had visual acuity less than 0.2 logMAR after treatment [13–16]. Changing treatment modalities, increasing the intensity of the same treatment [12], or combining treatments [17–19] have become options for managing children with residual amblyopia. However, only increasing patching to 6 h significantly improved visual acuity (by 1.2 lines over 16 weeks) in children aged 4–7 years with residual amblyopia, compared with traditional 2-h patching (0.5 lines) [12]. At the same time, the effectiveness of intense patching in younger children is often limited by poor compliance rates due to the burden it can place on parents and children [20, 21].

Perceptual learning, long-lasting improvement in perceptual functions following repetitive training in perceptual tasks [22–24], has been shown to be effective to enhance visual functions in children and adults with amblyopia [25–28]. Polat et al. found that perceptual learning in children who did not comply with occlusion or failed to occlusion despite good compliance still improved their visual acuity in patients with amblyopia [28]. Liu et al. showed that binocular perceptual learning significantly improved visual acuity and stereopsis in amblyopia where masking was no longer responsive, although no improvement in contrast sensitivity function was observed [29]. There is still a lack of clinical evaluations of perceptual learning in residual amblyopia.

In this study, we conducted a randomized clinical trial involving children aged 4 to under 11 years with residual amblyopia, defined as a persistent reduction in visual acuity of the amblyopic eye below the normal level despite at least 12 weeks of full refractive correction and daily patching, with stable vision confirmed across two follow-up visits at least 6 weeks apart, to evaluate the efficacy and safety of an iAVT protocol in improving visual acuity for residual amblyopia, compared with increasing patching of 6 h. This study aims to provide new insights into optimizing treatment strategies for residual amblyopia, thereby improving long-term visual outcomes for affected children.

Methods

The study protocol and informed consent were approved by the Ethics Committee of the Affiliated Optometry Hospital of Wenzhou Medical University (2022-248-K-198). Written informed consent was obtained from the parents or guardians of each study participant. The retrospectively registered study (ChiCTR2300075594, registered on 8 September 2023) adhered to the principles of the Declaration of Helsinki and was monitored by an independent data and safety monitoring committee.

Inclusion Criteria

Patients with amblyopia were diagnosed by an ophthalmologist. Before randomization, patients underwent a minimum of 12 weeks of full refractive correction combined with daily 2-h patching. Patients were eligible for the randomized phase if their visual acuity remained stable (within 1 logMAR line of improvement) across two follow-up visits of at least 6 weeks apart. The main inclusion criteria were:

1. Age 4 to < 11 years

2. Clinically diagnosed amblyopia due to refractive error, strabismus, or both

3. At least 12 weeks of treatment with lenses and patching (at least 2 h per day)

4. No significant improvement in VA at two visits spaced at least 6 weeks apart (confirmed by retesting), with no change in spectacle correction during the intervening period

5. Best amblyopic eye VA of 0.3–0.8 logMAR, interocular difference (IOD) of ≥ 2 lines, or an amblyopic eye VA of 0.2 logMAR with 3 lines of IOD

6. Fellow-eye VA of 0.2 logMAR or better

7. Reported compliance of ≥ 10 h per week when prescribed 2 h of daily patching

Synopsis of Study Design

Participants were randomly assigned (1:1) to either the iAVT group or the 6-h patching group using a computer-generated randomization sequence prepared by an independent researcher not involved in recruitment. Group assignments were concealed in sequentially numbered, opaque, sealed envelopes, which were opened only after participants met the inclusion criteria and signed informed consent forms to ensure allocation concealment. Participants in the patching group were instructed to patch for at least 6 h per day or a cumulative total of 42 h per week for 10 weeks, and parents were asked to record daily patching duration, which was used to evaluate compliance. Participants in the iAVT group were asked to complete 5 training sessions per week, each lasting 20–30 min, for a total of 50 sessions over 10 weeks. To ensure treatment fidelity, participants did not receive any concurrent amblyopia therapies during the intervention period.

The primary efficacy measure was VA in the amblyopic eye, assessed without cycloplegia [12, 30] during the screening and follow-up periods using the Early Treatment of Diabetic Retinopathy Study (ETDRS) charts [31]. For children under 7 years, training sessions during the screening phase ensured familiarity with the test procedure. A trained examiner measured VA by identifying the smallest line of optotypes correctly recognized by the participant, with precision down to individual optotypes. Measurements were conducted at a 4-m distance, beginning with the right eye, followed by the left eye. Secondary measures included contrast sensitivity function, Titmus stereo acuity, compliance score, the Chinese version of the Pediatric Eye Questionnaire (PedEyeQ-CN) [32], which was used to assess quality of life, and other routine ophthalmic examinations. The CSF was measured by the Manifold Contrast Vision Meter (AST Inc., San Diego, CA). Considering the age applicability of the questionnaire, the PedEyeQ-CN was specifically tested in children aged 5–11 years at baseline and the final follow-up. All other measures were assessed at every follow-up visit, conducted at 2 weeks (± 3 days) and 10 weeks (± 3 days).

Two certified ophthalmologists, experienced in pediatric visual acuity testing, conducted all outcomes assessments under standardized conditions. Masking was not feasible because of the distinct nature of the interventions; nevertheless, all measurements followed identical procedures to ensure consistency and reliability.

The iAVT Protocol

The iAVT training was conducted in a dim-lit room at the patient’s home using a GAMMA-corrected Microsoft Surface Go 2 (10.5-inch, 1920 × 1280 pixels, refresh rate of 60 Hz) running the JAVA platform. The training distance varied with the participant’s visual acuity in the amblyopic eye, e.g., 1 m for VA of 0.7–1 logMAR or 2 m for VA ≤ 0.7 logMAR. At 2 m, each screen pixel subtended a visual angle of 0.003°.

Participants completed 50 sessions over 10 weeks (5 sessions per week). Sessions were organized into five consecutive 10-session blocks. At the very first session, participants underwent baseline threshold assessments for CSF, stereopsis, and motion. Within each 10-session block, session 1 and session 10 were reassessment (threshold) sessions in which no feedback was provided, while sessions 2–9 were training sessions with auditory feedback cues. Threshold estimates obtained in the reassessment sessions were used to update individualized stimulus parameters via a Bayesian adaptive algorithm (step size = 0.1).

CSF testing used sinusoidal gratings at six fixed spatial frequencies. At 2 m, the frequencies were 1.5, 3, 6, 12, 18, and 24 cycles per degree (cpd) (at 1 m: 0.75, 1.5, 3, 6, 12, and 18 cpd). Each CSF test consisted of 6 frequency blocks × 45 trials = 270 trials; gratings subtended 880 pixels in diameter and were presented either horizontally or vertically. Stereopsis tests comprised 45 trials per assessment with a starting disparity of 3600 arcsec; stimuli subtended 850 × 850 pixels and were randomly presented on the left or right side of the screen. Motion perception tests comprised 45 trials per assessment with a starting coherence of 0.99; coherence was adjusted adaptively according to responses.

During training, participants responded using a handheld controller. Stimulus durations were 0.167 s for CSF trials, continuous display until response for stereopsis trials, and 0.4 s for motion trials; the interstimulus interval was 0.8 s for all tasks. The training application automatically logged session start and end times, trial-level performance, and completion metrics for objective adherence monitoring.

Statistical Analysis

The primary outcome measure was the 10-week VA improvement in the amblyopic eye. The sample size for the randomized trial was calculated to include 50 participants (25 per group), which provides sufficient power (β = 0.8) with a type I error rate of 5% (α = 0.05) to detect a significant difference between the groups. Considering a potential dropout rate, the target enrollment was set at 60 participants. This calculation assumed a mean improvement of 1.2 ± 1.4 logMAR lines in the increasing patching group [12] and 2.3 logMAR lines in the iAVT training group [33].

The primary analysis used analysis of variance (ANOVA) to compare the VA improvement from baseline at 2 and 10 weeks between two different treatment methods. ANOVA was also used for the CSF results. The Rasch calibration of the PedEyeQ-CN questionnaire scores was calculated using a look-up table, and pre- and post-treatment domain scores within the same group were compared using paired t-tests. Stereo acuity measurement was successful in only 7 and 12 patients in the two groups, respectively, and thus the results is not analyzed in the following section.

Results

From May 2023 to November 2023, 60 patients were enrolled in the study and randomly assigned to the iAVT group (mean age = 7.6 ± 1.6 years, mean VA = 0.46 ± 0.15 logMAR) and the 6 h patching group (mean age = 7.1 ± 1.4 years, mean VA = 0.47 ± 0.17 logMAR). On the basis of a post-randomization review, two patients (one in the iAVT group and one in the patching group) were found to have systemic diseases and excluded from the data analysis. During the study, three patients (one in the iAVT group and two in the patching group) withdrew voluntarily, and two patients in the patching group were excluded owing to low compliance rates (less than 50%). Therefore, the primary and secondary outcome analyses were conducted on 53 participants who completed the assigned intervention and the 10-week assessment, constituting a per-protocol analysis set. Baseline details of all patients are presented in Table 1.

Table 1.

Baseline details of all randomized patients

	iAVT (n = 30)		Patching (n = 30)
	n	%	n	%
Male sex	19	63	14	47
Female sex	11	37	16	53
Age at randomization, years
4 to < 5	1	3	3	10
5 to < 6	3	10	3	10
6 to < 7	9	30	9	30
7 to < 8	6	20	9	30
8 to < 9	5	17	3	10
9 to < 10	3	10	3	10
10 to < 11	3	10	0	0
Mean (SD)	7.6 (1.6)		7.1 (1.4)
Amblyopia cause
Strabismus	8	27	6	20
Anisometropia	17	57	21	70
Mix (strabismus and anisometropia)	5	17	3	10
BCVA (logMAR) in the amblyopic eye at randomization
0.7 to ≤ 0.8a	4	13	4	13
0.6 to < 0.7	5	17	7	23
0.5 to < 0.6	4	13	5	17
0.4 to < 0.5	7	23	4	13
0.3 to < 0.4	8	27	7	23
0.2 to < 0.3	2	7	3	10
Mean (SD) logMAR	0.46 (0.15)		0.47 (0.17)
BCVA (logMAR) in the fellow eye at randomization
≥ 0.2a	4	13	5	17
0.1 to < 0.2	9	30	6	20
0 to < 0.1	11	37	14	47
< 0	6	20	5	17
Mean (SD) logMAR	0.04 (0.09)		0.04 (0.10)
Refractive error in amblyopic eye at enrollment, D
≥ +5.00	10	33	15	50
+4.00 to < +5.00	9	30	6	20
+3.00 to < +4.00	1	3	5	17
+2.00 to < +3.00	4	13	1	3
+1.00 to < +2.00	1	3	0	0
0 to +1.00	2	7	1	3
< 0	3	10	2	7
Mean (SD)	+ 3.59 (2.08)		+ 4.21 (2.36)
Refractive error in fellow eye at enrollment, D
≥ +5.00	2	7	2	7
+4.00 to < +5.00	4	13	1	3
+3.00 to < +4.00	0	0	8	27
+2.00 to < +3.00	6	20	5	17
+1.00 to < +2.00	5	17	5	17
0 to < +1.00	7	23	7	23
< 0	6	20	2	7
Mean (SD)	+ 1.65 (1.84)		+ 2.13 (1.71)

SD standard deviation, BCVA best correct visual acuity, logMAR logarithm of the minimum angle of resolution, D diopters

^aOne participant (patching group) had 0.3 logMAR in the fellow eye and a systemic disease at the time of randomization, completing the follow-ups but being excluded from the data analysis

In the iAVT group, 28 patients completed the 10-week primary outcome examination, with all participants (100%) finished the required 50 training sessions. In the patching group, 27 patients finished the final follow-up, and on the basis of total patching duration, compliance was judged as excellent (> 80%) in 22 patients, good (50–80%) in 3 patients, and fair (< 50%, dropped out) in another 2 patients (Fig. 1); ultimately, 25 patients were included in the analysis, with an average compliance rate of 94.2%.

Fig. 1.

Flowchart showing trial profile

Visual Acuity Improvement in the Amblyopic Eye

Both the iAVT training and 6-h patching groups exhibited visual acuity improvements from baseline to 10 weeks of treatment (Table 2 and Fig. 2). A two-way ANOVA analysis revealed that the iAVT significantly enhanced the VA of the amblyopic eye (F(1,51) = 6.328, p = 0.015) compared with 6 h of extended daily patching treatment, with a significant main effect of treatment time (F(1,51) = 5.899, p = 0.019). The interaction between group and treatment time was not significant (F(1,51) = 1.359, p = 0.249), suggesting that the magnitude of the difference in VA improvement between the iAVT and patching groups did not change significantly between 2 and 10 weeks (i.e., parallel effectiveness trends between the two groups).

Table 2.

VA in the amblyopic eye at 10 weeks (primary outcome)

	iAVT (n = 28)		Patching (n = 25)
	n	%	N	%
Change in VA from randomization
< 1 line worse	0	0	4	16
No change (0)	1	4	3	12
< 1 line improved	8	29	8	32
1 to < 2 lines improved	16	57	7	28
≥ 2 lines improved	3	11	3	12
Mean (SD) changes in line	1.3 (0.73)		0.7 (0.86)
Distribution of VA (logMAR)
0.7 to ≤ 0.8	0	0	1	4
0.6 to < 0.7	1	4	1	4
0.5 to < 0.6	6	21	6	24
0.4 to < 0.5	4	14	6	24
0.3 to < 0.4	6	21	5	20
0.2 to < 0.3	6	21	4	16
0.1 to < 0.2	4	14	2	8
0 to < 0.1	1	4	0	0
Mean (SD) logMAR VA	0.34 (0.16)		0.4 (0.16)

Fig. 2.

VA improvement of the amblyopic eye (A) and fellow eye (B) in the iAVT (blue) and 6 h patching (green) groups. Error bars represent the standard error. *p < 0.05

Furthermore, Pearson correlation analysis revealed that, in the 6-h patching group, the VA improvement in the amblyopic eye after 10 weeks exhibited a negative correlation with age (r = −0.45, p = 0.025) and a marginal correlation with baseline visual acuity in the amblyopic eye (r = 0.35, p = 0.086). Conversely, in the iAVT group, neither age (r = 0.011, p = 0.957) nor baseline visual acuity (r = 0.28, p = 0.153) showed significant correlations with visual acuity improvement, suggesting that iAVT offered consistent efficacy independent of these limiting factors in children with residual amblyopia (Fig. 3).

Fig. 3.

The correlation between VA improvement in the amblyopic eye and age (A and B) or baseline VA in the amblyopic eye (C and D) in the iAVT (blue) and 6-h patching (green) groups. Dashed lines represent linear regression. *p < 0.05

Specifically, at 2 weeks, the iAVT group (approximately 10 training sessions) exhibited an average VA improvement of 0.9 logMAR lines, whereas the 6-h patching group demonstrated an improvement of 0.6 logMAR lines. The difference between the groups was marginally significant (t(51) = 1.814, p = 0.076, Cohen’s d = 0.497, 95% CI −0.005 to 0.065). At 10 weeks, the iAVT group showed an average improvement of 1.3 logMAR lines, which was notably higher than the 0.7 logMAR lines improvement observed in the patching group (t(51) = 2.475, p = 0.017, Cohen’s d = 0.678, 95% CI 0.010–0.100).

Furthermore, a paired t-test revealed a significant difference in VA improvement at 2 and 10 weeks for the iAVT group (t(27) = −2.188, p = 0.038, Cohen’s d = 0.541, 95% CI −0.071 to 0.002), highlighting the sustained effectiveness of this training protocol. In contrast, the patching group reached a plateau after 2 weeks, with no significant further improvements observed (t(24) = − 1.21, p = 0.238, Fig. 2C).

We also evaluated the efficiency of both treatments by calculating how many hours needed to achieve 1 logMAR line VA improvement (e.g., total treatment time/VA improvement). The average training duration for the iAVT group was 23.3 h (50 sessions × 28 min per session), and the efficiency was 17.9 h per VA improvement (in the unit of logMAR line). The average patching duration for the extended patching group was 395.6 h, and the efficiency was 565.2 h per VA improvement. The efficiency of iAVT training is approximately 31.6 times greater than that of the daily extended patching.

Effect of Treatment on Visual Acuity in the Fellow Eye

Both the iAVT training and 6-h patching did not sacrifice VA in the fellow eye, with similar effects (group: F(1,51) = 0.229, p = 0.635, treatment time: F(1,51) = 2.053, p = 0.158, interaction: F(1,51) = 0.007, p = 0.936, Fig. 2B), demonstrating safety of these two treatments.

Contrast Sensitivity Function

We analyzed the contrast sensitivity change in terms of the AULCSF[18], a summary metric of the CSF function [34], across the full spatial frequency range (1–18 cpd) in the iAVT and 6-h patching groups. A two-way ANOVA revealed no significant difference in the change of AULCSF between the two groups (group: F(1,51) = 2.391, p = 0.128; treatment time: F(1,51) = 4.147, p = 0.021; interaction: F(1,51) = 1.486, p = 0.232). Subsequently, within-group analyses were performed to examine the time effect (Fig. 4). In the iAVT group, a one-way repeated-measures ANOVA showed a significant main effect of time (F(1,27) = 4.783, p = 0.015). Bonferroni-adjusted post-hoc comparisons indicated that visual acuity significantly improved from baseline to week 2 (95% CI 0.014–0.186; p = 0.018) and showed a marginal improvement from baseline to week 10 (95% CI −0.225 to 0.001; p = 0.056), whereas no significant difference was observed between week 2 and week 10 (p > 0.05). In contrast, the 6-h patching group showed no significant main effect of time (p > 0.05).

Fig. 4.

The AULCSF of the amblyopic eye at baseline, 2 weeks, and 10 weeks. Error bars represent the standard error. *p < 0.05

Pediatric Eye Questionnaire

A total of 27 and 23 patients were included in the iAVT and 6-h patching group, respectively. Statistical analyses revealed no significant differences between groups in the overall pre- and post-treatment scores across the three PedEyeQ-CN questionnaires. However, Wilcoxon tests showed that patients in the iAVT group demonstrated significant improvements in three dimensions of the child section of the PedEyeQ-CN: functional vision, bothered by eyes/vision, and frustration/worry. Additionally, parent and proxy sections improved significantly in social dimensions. In contrast, the patching group did not exhibit significant score improvements in any dimensions (Table 3).

Table 3.

Pediatric Eye Questionnaire average scores and p-values

	iAVT (n = 27)			Patching (n = 23)
	Pre	Post	p	Pre	Post	p
Child
Functional vision	68	77	0.020*	71	73	0.807
Bothered by eyes/vision	70	78	0.029*	75	78	0.807
Social	78	84	0.080	77	77	0.929
Frustration/worry	73	82	0.024*	74	71	0.807
Proxy
Functional vision	60	66	0.281	64	70	0.310
Bothered by eyes/vision	70	79	0.115	69	77	0.360
Social	68	76	0.010*	65	72	0.310
Frustration/worry	54	58	0.281	45	47	0.812
Eyecare	59	64	0.281	60	62	0.812
Parent
Impact on parent and family	58	65	0.117	72	73	0.857
Worry about child’s eye condition	35	46	0.072	40	48	0.216
Worry about child’s self-perception and interactions	60	70	0.123	58	68	0.216
Worry about child’s functional vision	46	59	0.072	53	57	0.458

One participant in the patching group did not complete the proxy session, resulting in a final sample size of 22 for this group. Multiple comparisons were controlled using the Benjamini–Hochberg false discovery rate (FDR) procedure at α = 0.05

*p < 0.05

Discussion

This study aimed to compare the effectiveness of iAVT and 6-h patching in treating residual amblyopia in children aged 4 to under 11 years. Our findings demonstrate that, although both treatments lead to significant improvements of VA in the amblyopic eye, the iAVT protocol is more efficient and pronounced in improving VA, as opposed to extended patching.

Wallace et al. [12] reported an improvement of approximately 1.2 lines in VA after 10 weeks of 6-h patching in patients with residual amblyopia, which is slightly greater than that of ours. We speculate that this discrepancy might be due to the younger average age (5.9 years versus > 7 years in ours) of patients in the PEDIG study, suggesting that younger patients might have greater visual plasticity [27, 35, 36].

Amblyopia treatment is known to be subject to compliance issues, especially for extended patching treatments [37]. In this study, 18.5% of patients in our 6-h patching group did not achieve 80% of the prescribed patching duration, which was slightly lower than that in previous study [12]. Similar to previous patching studies [12, 38, 39], 16% of patients in our study showed a decline in visual acuity after 10 weeks of patching, which is presumed to be related to patient compliance. This phenomenon is commonly observed in occlusion therapy studies and may reflect a subgroup of patients who remain unresponsive or less sensitive to prolonged patching. Future research is warranted to further explore the underlying causes of this treatment insensitivity. Digital therapies, on the other hand, tend to achieve better compliance than traditional patching, likely owing to shorter treatment durations (e.g., 1 h per day, 6 days per week for Luminopia [39]; 1.5 h per day, 5 days per week for CureSight [30]), or more engaging activities (e.g., gaming or watching video). Interestingly, 28 patients in our iAVT group adhered to the training (100% compliance), which may be related to its superior efficiency (20–30 min per day).

A decline in CSF, particularly at high spatial frequencies, is the hallmark of amblyopia [4, 5, 40]. The iAVT protocol was designed to provide intensive contrast detection training across a range of spatial frequencies, with an emphasis on higher frequencies where residual deficits are typically pronounced. Consistent with this rationale, the iAVT group exhibited a marginally significant overall improvement in CSF over time, suggesting that adaptive and high-frequency-oriented training can further enhance the contrast sensitivity in patients with residual amblyopia. The absence of a significant effect at week 10 may in part reflect the influence of an individual with markedly divergent follow-up performance. In contrast, no significant within-group change was observed in the 6-h patching group, which may reflect the limited impact of conventional occlusion therapy on contrast sensitivity once visual acuity has plateaued after standard treatment [41, 42]. This suggests that iAVT has the potential to substantially improve the CSF in patients with amblyopia.

The PedEyeQ-CN results suggest that iAVT may lead to partial functional and psychosocial benefits after training. Although no significant between-group differences were observed, children in the iAVT group showed within-group improvements in functional vision and vision-related emotional domains. However, these findings remain limited by the small sample size and the self-reported nature of the questionnaire data.

There were some limitations in this study. The amblyopia involved was mainly refractive amblyopia (60.4%). Since different types of amblyopia may have different pathogenetic mechanisms [43, 44], future studies shall investigate the effectiveness of iAVT by including more patients with different types of amblyopia. The analysis was performed per protocol rather than by intention-to-treat, owing to the small sample size and incomplete data from participants lost to follow-up. Although this may introduce potential bias, baseline characteristics remained balanced between groups, although future studies with larger samples are needed to confirm these findings. Moreover, the present trial evaluated outcomes after 10 weeks of treatment, which may not fully capture the long-term sustainability or retention of visual gains. Future studies shall also include longer-term follow-up evaluation of training retentions. Another limitation is that outcome assessors were not masked to group allocation because of the distinct nature of the interventions. However, all visual acuity measurements were performed by certified orthoptists following standardized ETDRS procedures, which likely minimized potential measurement bias. Although the trial was retrospectively registered, the study protocol was predefined and fully transparent, with no changes made after patient enrollment. Prospective registration in future trials is recommended to align with standard practice.

Conclusions

The iAVT provided safe, faster, and more effective treatment for residual amblyopia compared with 6-h patching in children aged 4 to under 11 years within the 10-week treatment period. It offers a potential and effective clinical practice option.

Author Contributions

Study concept and design: Junli Yuan, Chang-Bing Huang, Fang-Fang Yan, Xinping Yu, and Huanyun Yu. Data collection: Junli Yuan, Yuanyuan Chen, Lu Wang, Meiping Xu, Na Liao, Zhiyue Dai, Hui Chen, Binjun Zhang, Jinling Xu, Minghui Wan, Suzhong Xu, Yuwen Wang, Xiaolin Huang, Fuhao Zheng, Pingping Huang, and Jie Chen. Data analysis and interpretation: Junli Yuan, Yuanyuan Chen, Lu Wang, Chang-Bing Huang, and Fang-Fang Yan. Manuscript preparation: Junli Yuan, Yuanyuan Chen, Chang-Bing Huang, Fang-Fang Yan, Xinping Yu, and Huanyun Yu. All authors read and approved the final manuscript.

Funding

This research was supported by National Science and Technology Innovation 2030 Major Projects (2022ZD0204800), National Key Research and Development Program of China (2023YFC3604100), and National Natural Science Foundation of China (32071056 and 32100864). The journal’s Rapid Service Fee was funded by the authors.

Data Availability

Data are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Chang-Bing Huang have equity interest in Jiangsu Juehua Medical Technology Co., LTD. Junli Yuan, Yuanyuan Chen, Lu Wang, Meiping Xu, Na Liao, Zhiyue Dai, Hui Chen, Binjun Zhang, Jinling Xu, Minghui Wan, Suzhong Xu, Yuwen Wang, Xiaolin Huang, Fuhao Zheng, Pingping Huang, Jie Chen, Fang-Fang Yan, Xinping Yu, and Huanyun Yu have nothing to disclose.

Ethical Approval

The study protocol and informed consent were approved by the Ethics Committee of the Affiliated Optometry Hospital of Wenzhou Medical University (2022-248-K-198). Written informed consent was obtained from the parents or guardians of each study participant. The retrospectively registered study (ChiCTR2300075594, registered on 8 September 2023) adhered to the principles of the Declaration of Helsinki and was monitored by an independent data and safety monitoring committee.