Suprachoroidal Injection of Wharton’s Jelly-Derived Mesenchymal StemCells in Inherited Retinal Diseases: A Prospective Exploratory Single-Arm Study

1. Introduction

Inherited retinal diseases (IRDs) comprise a clinically and genetically heterogeneous group of disorders characterized by progressive dysfunction and degeneration of photoreceptors, retinal pigment epithelium (RPE), and other retinal neuronal populations. Their clinical manifestations include night blindness, peripheral visual-field loss, central visual impairment, and, in advanced cases, profound irreversible blindness. Disease progression varies according to the affected gene, inheritance pattern, residual retinal structure, and age at onset [1,2].

Retinitis pigmentosa (RP), Stargardt disease, and PPH2 dystrophy represent distinct phenotypic entities within the spectrum of hereditary retinal degeneration. RP is generally characterized by rod-predominant degeneration followed by cone dysfunction and progressive peripheral field constriction. Stargardt disease primarily affects the macular region and may result in progressive central visual loss. These conditions share the lack of broadly effective treatments for many affected individuals, particularly when patients are not candidates for mutation-specific gene therapy or when photoreceptor degeneration is advanced [1,3].

Cell-based therapies have emerged as potentially mutation-independent strategies for retinal neuroprotection, tissue repair, or cellular replacement. Candidate products include retinal progenitor cells, bone-marrow-derived mononuclear cells, CD34-positive cells, mesenchymal stem/stromal cells (MSCs), human embryonic stem-cell-derived RPE cells, and induced pluripotent stem-cell-derived retinal cells. Proposed mechanisms include replacement of damaged cells, secretion of neurotrophic and anti-apoptotic factors, immunomodulation, attenuation of inflammatory signaling, and transfer of biologically active molecules through extracellular vesicles [2,4,5,6,7,8,9,10,11,12].

Wharton’s jelly-derived mesenchymal stem cells (WJ-MSCs) are of particular interest because they can be obtained from umbilical cord tissue and exhibit immunomodulatory and trophic properties. Preclinical studies suggest that WJ-MSCs may preserve retinal structure, support photoreceptor survival, and delay outer nuclear layer degeneration, although electrophysiological improvement has not been consistently demonstrated. Early clinical studies involving WJ-MSCs in retinal degeneration have reported variable functional outcomes and generally favorable short-term safety profiles, but the evidence remains limited by small samples, heterogeneous routes of administration, and lack of control groups [13,14,15,16,17].

The route of cell delivery is an important determinant of therapeutic targeting and safety. Intravitreal injection is technically accessible but may limit exposure of the outer retina because of the internal limiting membrane and vitreous environment. Subretinal injection offers direct access to the RPE–photoreceptor interface but requires vitreoretinal surgery and carries risks related to retinal detachment, hemorrhage, inflammation, and graft-associated complications. Suprachoroidal injection may provide targeted exposure of the choroid, RPE, and outer retina while avoiding direct penetration of the retinal tissue [2,4,6,7,18].

A systematic review and meta-analysis of 21 prospective studies involving 382 patients and 496 eyes reported early visual-acuity improvement after cell-based therapy in RP and Stargardt disease. In RP, approximately 49% of eyes showed visual-acuity improvement at 6 months, decreasing to approximately 30% at 12 months. The pooled visual-acuity improvement at 6 months was approximately −0.12 logMAR, whereas the difference was not statistically significant at 12 months. Suprachoroidal administration was identified as a potentially promising route for RP, although the overall evidence quality was limited and randomized trials were lacking [3].

In addition to visual acuity, visual-field testing, OCT, and electroretinography provide complementary information about treatment response. Visual-field mean deviation (MD) estimates global sensitivity loss, while pattern standard deviation (PSD) reflects localized field irregularity. OCT can identify structural changes such as retinal thickening, outer-retinal preservation, fluid accumulation, epiretinal proliferation, or segmentation artifacts. ERG provides an objective measure of retinal electrophysiological function through amplitude and latency parameters under scotopic, photopic, and flicker conditions [1,2,3].

However, interpretation of ERG changes requires attention to test–retest variability. Electrode placement, retinal adaptation, patient cooperation, stimulus conditions, and technical factors can produce substantial variation between examinations. In the present exploratory analysis, a 30% threshold was applied to classify changes as beyond presumed test–retest variability. This threshold was assumed for exploratory purposes and was not derived from a dedicated repeatability study using the specific device and protocol.

The present study evaluated preliminary 45-day structural, functional, electrophysiological, and safety outcomes after suprachoroidal injection of WJ-MSCs in patients with genetically confirmed IRDs. We hypothesized that the intervention could produce measurable changes in visual-field sensitivity, retinal structure, and ERG parameters while maintaining short-term ocular safety [3,14,16,19].

2. Materials and Methods

2.1. Study Design and Setting

This was a prospective, open-label, single-arm clinical trial with preliminary 45-day outcomes. The study was conducted at Universidade Santo Amaro (UNISA), São Paulo, Brazil.

Twenty eyes of 10 patients were included in the analysis. Patients had genetically and multimodally confirmed hereditary retinal dystrophies, including RP, PPH2 dystrophy, and Stargardt dystrophy.

The study cohort was recruited in June 2026 and evaluated at baseline and 45 days after suprachoroidal WJ-MSC injection.

Baseline examinations were performed before treatment, and the post-treatment evaluation was conducted 45 days after the intervention.

The study included both eyes of each participant. Because measurements from the two eyes of the same patient are correlated, the statistical analysis was performed using paired observations and the study design considered within-patient correlation.

2.2. Participants

Patients were eligible if they had an established diagnosis of hereditary retinal dystrophy confirmed through multimodal ophthalmic imaging and genetic testing.

Inclusion criteria

• Established hereditary retinal dystrophy confirmed by multimodal imaging;
• Confirmatory genetic testing;
• No previous ocular treatment for the retinal disease;
• BCVA ≤20/30;
• Availability of baseline and 45-day follow-up examinations;
• Ability to provide written informed consent.

Exclusion criteria

• Previous ocular treatment;
• Unclear or unconfirmed ocular diagnosis;
• BCVA better than 20/30;
• Active ocular infection or inflammation;
• Coexisting ocular disease likely to affect visual function;
• Incomplete baseline or post-treatment records;
• Inability to undergo the required examinations.

The study population included patients with RP, PPH2 dystrophy, and Stargardt dystrophy.

2.3. Ethical Approval and Informed Consent

The study was conducted in accordance with the Declaration of Helsinki. The protocol was approved by the Research Ethics Committee of Universidade Santo Amaro under CAAE 97710926200000081.

Written informed consent was obtained from all participants using the standardized informed consent form approved by the UNISA Research Ethics Committee.

2.4. Cell Product and Intervention

The intervention consisted of a single suprachoroidal administration of human Wharton’s jelly-derived mesenchymal stromal cells (WJ-MSCs). The cells were obtained from human umbilical cord Wharton’s jelly and were supplied as a commercially manufactured cryopreserved cell product (stemCORD; StemCorp Serviços Biomédicos Ltda., São Paulo, Brazil; batch no. 18340000260226). According to the manufacturer’s documentation, the cells were isolated from umbilical cord tissue obtained from healthy donors and underwent in vitro expansion during the manufacturing process using the CellBind culture system (Corning, Corning, NY, USA). Therefore, the term “passage 0” in the present study refers specifically to the absence of additional study-specific in vitro expansion after receipt and preparation of the cryopreserved product, rather than to the absence of prior manufacturing expansion.

The cryopreserved product was supplied in 2-mL vials and stored in liquid nitrogen at approximately −196 °C, with a declared content of 8 × 10⁶ cells per vial. The cryopreservation formulation contained StemPro™ MSC SFM XenoFree medium and clinical-grade dimethyl sulfoxide (DMSO). According to the manufacturer’s certificate of analysis, product characterization included donor serological screening, G-banding karyotype analysis, flow-cytometric immunophenotyping, endotoxin testing, mycoplasma testing, and sterility testing. The manufacturer also documented the use of fetal bovine serum, DMSO, antibiotics, and antifungal agents during the manufacturing process, with the possibility of residual amounts being present in the final product.

Immediately before administration, the cryopreserved cell preparation was thawed and processed according to the institutional protocol. The final cell suspension was adjusted to a dose of 1 × 10⁶ cells in 0.1 mL for each treated eye. No additional in vitro cell expansion or culture was performed as part of the study-specific preparation. The same standardized surgical protocol was used for all participants by the same qualified ophthalmic surgeon.

For suprachoroidal administration, a 30-gauge needle was incorporated into a modified peripheral intravenous catheter segment that functioned as a depth-limiting guide. Approximately 1 mm of the sharp distal portion of the needle was exposed beyond the catheter tip to facilitate controlled scleral penetration while limiting excessive advancement toward the vitreous cavity. The cell suspension was loaded into a 1-mL syringe connected to the protected 30-gauge needle assembly (Figure 1).

After topical ocular anesthesia and 5% povidone-iodine antisepsis, the device was positioned approximately 4 mm posterior to the limbus. The needle-guide assembly was gently advanced through the sclera to access the suprachoroidal space (Figure 2). A total volume of 0.1 mL containing 1 × 10⁶ WJ-MSCs was then administered slowly over approximately 10 s. Following injection, the needle was withdrawn carefully, and the ocular surface and injection site were examined for immediate complications. Post-procedure topical antibiotic drops were prescribed for 5 days.

The suprachoroidal route was selected to facilitate delivery of the cellular product toward the choroid–retinal pigment epithelium–outer retinal interface while avoiding direct subretinal or intravitreal injection. All patients underwent standardized postoperative clinical monitoring and were evaluated for ocular inflammation, infection, intraocular pressure changes, hemorrhage, retinal or choroidal complications, and other treatment-related adverse events during the 45-day follow-up period.

2.5. Ophthalmic Assessments

The following examinations were performed at baseline and 45 days after treatment:

• BCVA;
• IOP;
• OCT;
• Automated visual-field testing;
• Electroretinography;
• Clinical examination for ocular inflammatory and infectious signs.

2.6. Visual Acuity

BCVA was recorded using Snellen notation and converted to logMAR values.

Lower logMAR values indicated better visual acuity. A negative post-treatment change indicated improvement.

Counting-fingers visual acuity was assigned a reference value of 2.0 logMAR solely for quantitative statistical analysis. This approximation does not constitute a standardized continuous visual-acuity measurement and was interpreted cautiously.

2.7. Intraocular Pressure

IOP was measured in mmHg using standardized tonometry. When bilateral IOP values were reported in the clinical records, the recorded value was applied to both eyes.

The analysis evaluated both group-level change and individual values that might indicate clinically relevant pressure elevation.

2.8. Optical Coherence Tomography

OCT was performed using the RS-330 Retina Scan Duo system (Nidek, Japan).

Central foveal thickness (CFT) was recorded in micrometers. OCT images were assessed for central retinal thickness and should also be reviewed qualitatively for intraretinal fluid, subretinal fluid, cystoid spaces, epiretinal membrane, vitreomacular traction, hyperreflective foci, segmentation artifacts, and changes in outer-retinal architecture.

An increase in CFT was not interpreted automatically as regeneration because retinal thickening may also reflect edema, inflammation, traction, or measurement artifact.

2.9. Automated Visual-Field Testing

Automated visual-field examinations were performed using the Humphrey Field Analyzer II, series 700 (Carl Zeiss, Germany), with 24-2 and/or 10-2 protocols.

The following indices were analyzed:

• Mean deviation (MD), in dB;
• Pattern standard deviation (PSD), in dB.

Reliability criteria included fixation losses, false-positive responses, and false-negative responses. Tests not meeting the institutional reliability criteria were excluded from the corresponding paired analysis.

A positive change in MD was considered favorable because it represented a shift toward less negative values. PSD was analyzed as a structural-functional descriptor without prespecifying that a small numerical change represented clinical improvement.

2.10. Electroretinography

ERG was performed using the RETIport system (Roland Consult, Germany). The analyzed parameters were:

• Scotopic 0.01 b-wave latency;
• Scotopic 0.01 b-wave amplitude;
• Photopic 3.0 a-wave latency;
• Photopic 3.0 b-wave latency;
• Photopic 3.0 a-wave amplitude;
• Photopic 3.0 b-wave amplitude;
• 30-Hz flicker P1 latency;
• 30-Hz flicker N1–P1 amplitude.

For amplitude parameters, an increase was interpreted as directionally favorable. For latency parameters, a decrease was interpreted as directionally favorable.

2.11. ERG Test–Retest Variability Criterion

Because a dedicated repeatability study for the RETIport device and the present protocol was not available, an exploratory 30% threshold was assumed.

A change greater than 30% relative to the baseline value was classified as beyond the assumed test–retest variability. Changes within ±30% were considered stable because they could plausibly reflect measurement noise associated with electrode positioning, adaptation, patient cooperation, and technical factors.

For each ERG parameter:

%Δ = [(post-treatment value − baseline value) / baseline value] × 100

For amplitude parameters:

• >30% increase: trend toward improvement;
• >30% decrease: trend toward decline;
• Change within ±30%: stable.

For latency parameters:

• >30% decrease: trend toward improvement;
• >30% increase: trend toward decline;
• Change within ±30%: stable.

This threshold was exploratory and should not be interpreted as a validated minimally clinically important difference.

2.12. Outcome Measures

Primary outcome

The primary exploratory outcome was the change in visualfield MD from baseline to day 45.

Secondary outcomes

Secondary outcomes included:

• Change in BCVA in logMAR;
• Change in IOP in mmHg;
• Change in CFT in micrometers;
• Change in visual-field PSD;
• Change in ERG amplitudes and latencies;
• Ocular inflammatory signs;
• Ocular infectious signs;
• Clinically relevant IOP variation.

2.13. Statistical Analysis

Statistical analyses were performed using Python 3.12.3 and SciPy 1.17.1.

Continuous variables were summarized using means and standard deviations. Pre- and post-treatment values were compared using paired Student’s t-tests and Wilcoxon signed-rank tests.

A two-sided p value <0.05 was considered statistically significant. Because the study evaluated several structural, functional, and electrophysiological endpoints in a small cohort, the results were considered exploratory and no causal efficacy conclusion was drawn.

Both eyes were included and correlated within the patient-level study structure. For confirmatory analyses, a mixed-effects model with a patient-level random effect would be preferable.

Patient 9 was retained in the primary analysis because exclusion of an extreme observation after reviewing the results could introduce post hoc selection bias. Its influence was acknowledged and interpreted through individuallevel analysis.

2.13.1. Statistical Analysis Report

Parameters analyzed: BCVA (Best-Corrected Visual Acuity)· IOP (Intra-ocular Pressure) Visual Field (VF: MD and PSD) · OCT (Central Foveal Thiness)

This report presents the comparative statistical analysis between assessments performed before and after the application, organized by patient and by examination instrument. For each data set, detailed per-patient tables are presented, followed by the group's statistical summary (paired t-test and Wilcoxon signed-rank test for paired samples, appropriate for small samples and potentially non-normal distributions). Patient names have been replaced with sequential numeric identifiers (Patient 1-10, in the same order as the original document) to protect patient identity.

Methodological note: visual acuity values recorded as 'counting fingers' (CF) were converted to a reference logMAR equivalent (2.0) solely to enable quantitative statistical treatment; this is a widely used approximation in the ophthalmology literature, not a direct measurement. Cells with no record in the original document (e.g., 'unavailable' or missing data) were excluded from the corresponding paired calculation.

2.14. Safety Analysis

Safety outcomes included:

• Ocular inflammatory signs;
• Ocular infectious signs;
• IOP variation;
• Other ocular or systemic adverse events potentially related to the intervention.

Patients were clinically monitored throughout the 45-day follow-up.

3. Results

3.1. Best-Corrected Visual Acuity

Mean BCVA improved numerically from 0.849 ± 0.750 logMAR at baseline to 0.740 ± 0.718 logMAR at day 45, representing a mean change of −0.109 logMAR.

The difference was not statistically significant by the paired Student’s t-test (t = 1.7275, p = 0.1003) or the Wilcoxon signed-rank test (W = 4, p = 0.0904) (Table 1).

The largest individual improvements were observed in Patient 3’s right eye, from 20/200 to 20/30, and Patient 5’s left eye, from counting fingers at 2 m to 20/200.

Visual acuity was converted from Snellen notation to the logMAR scale (log of the Minimum Angle of Resolution), where lower values indicate better visual acuity. Thus, a negative delta (post minus pre) represents improvement (Table 2 and Table 3).

Significance tests (paired):

• Paired t-test: t = 1.7275, p = 0.1003 → difference not statistically significant (p ≥ 0.05).
• Wilcoxon signed-rank test: W = 4, p = 0.0904 → difference not statistically significant (p ≥ 0.05).

Interpretation: there was a mean reduction in logMAR (mean improvement in visual acuity) after injection, but the difference did not reach statistical significance in the paired tests, which may reflect either a true absence of a group-level effect or the small sample size (10 patients / 20 eyes) combined with high individual variability — some patients showed marked improvement (e.g., Patient 3, RE: 20/200 → 20/30) while others remained stable.

3.2. IOP — Intraocular Pressure (mmHg)

IOP values reported as both eyes were considered equal for OD and OS in the corresponding pair (Table 4 and Table 5).

Significance tests (paired):

• Paired t-test: t = -1.045, p = 0.3092 → difference not statistically significant (p ≥ 0.05).
• Wilcoxon signed-rank test: W = 27.5, p = 0.3343 → difference not statistically significant (p ≥ 0.05).

Interpretation: mean IOP showed a slight increase (approx. +0.5 mmHg) after application, without statistical significance. Clinically, values remained within the normal range (10–21 mmHg) for most patients, with the exception of isolated cases in Patient 1 (22 mmHg post) and Patient 4 (20/22 mmHg post), which warrant follow-up.

3.3. OCT — Central Foveal Thickness (µm).

Table 6. OCT – central foveal thickness (µm) mean Δ values pre and post-injection.by patient.

Patient	Eye	Pre (µm)	Post (µm)	Δ (µm)	Δ (%)
Patient 1	RE	193	198	+5	+2.6%
	LE	202	206	+4	+2%
Patient 2	RE	237	383	+146	+61.6%
	LE	332	245	-87	-26.2%
Patient 3	RE	229	242	+13	+5.7%
	LE	282	275	-7	-2.5%
Patient 4	RE	244	238	-6	-2.5%
	LE	223	199	-24	-10.8%
Patient 5	RE	66	84	+18	+27.3%
	LE	56	167	+111	+198.2%
Patient 6	RE	131	278	+147	+112.2%
	LE	232	231	-1	-0.4%
Patient 7	RE	148	154	+6	+4.1%
	LE	104	139	+35	+33.7%
Patient 8	RE	278	281	+3	+1.1%
	LE	263	263	+0	+0%
Patient 9	RE	406	836	+430	+105.9%
	LE	511	681	+170	+33.3%
Patient 10	RE	233	235	+2	+0.9%
	LE	167	181	+14	+8.4%

Table 6. OCT – central foveal thickness (µm) mean Δ values pre and post-injection.by patient.

Patient	Eye	Pre (µm)	Post (µm)	Δ (µm)	Δ (%)
Patient 1	RE	193	198	+5	+2.6%
	LE	202	206	+4	+2%
Patient 2	RE	237	383	+146	+61.6%
	LE	332	245	-87	-26.2%
Patient 3	RE	229	242	+13	+5.7%
	LE	282	275	-7	-2.5%
Patient 4	RE	244	238	-6	-2.5%
	LE	223	199	-24	-10.8%
Patient 5	RE	66	84	+18	+27.3%
	LE	56	167	+111	+198.2%
Patient 6	RE	131	278	+147	+112.2%
	LE	232	231	-1	-0.4%
Patient 7	RE	148	154	+6	+4.1%
	LE	104	139	+35	+33.7%
Patient 8	RE	278	281	+3	+1.1%
	LE	263	263	+0	+0%
Patient 9	RE	406	836	+430	+105.9%
	LE	511	681	+170	+33.3%
Patient 10	RE	233	235	+2	+0.9%
	LE	167	181	+14	+8.4%

Table 7. OCT — Central Foveal Thickness (µm) Group statistical summary (n = 20 eyes).

Measure	Pre-application	Post-application	Mean Δ
Mean µm (SD)	226.8 (±108.2)	275.8 (±178.4)	+49

Table 7. OCT — Central Foveal Thickness (µm) Group statistical summary (n = 20 eyes).

Measure	Pre-application	Post-application	Mean Δ
Mean µm (SD)	226.8 (±108.2)	275.8 (±178.4)	+49

Significance tests (paired):

• Paired t-test: t = -1.9838, p = 0.0619 → difference not statistically significant (p ≥ 0.05) (borderline).
• Wilcoxon signed-rank test: W = 41.5, p = 0.0313 → difference statistically significant (p < 0.05).

Interpretation: there was a mean increase in foveal thickness (~49 µm) after application, statistically significant on the Wilcoxon test (p = 0.031) and borderline on the paired t-test (p = 0.062). This increase is strongly influenced by extreme values, notably Patient 9 (RE: 406→836 µm; LE: 511→681 µm), suggesting a possible localized edema/effect that warrants individual investigation, in addition to more modest gains in Patient 3 and Patient 6 (Figure 3 and Figure 4).

3.4. Visual Field (VF) —MD and PSD

The Mean Deviation (MD, in dB) and Pattern Standard Deviation (PSD, in dB) indices were assessed under the 24-2 and 10-2 protocols, for right eye (RE) and left eye (LE). MD values closer to zero indicate less visual field loss; lower PSD values indicate less irregularity of the visual field (Figure 5).

Table 8. Patient 1 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-33.09	-33.03	+0.06	1.44	1.50	+0.06
24-2	LE	-33.05	-33.04	+0.01	1.54	1.88	+0.34
10-2	RE	-33.81	-31.05	+2.76	3.91	6.34	+2.43
10-2	LE	-33.01	-31.43	+1.58	4.28	6.24	+1.96

Patient's mean Δ MD: +1.1 dB| Patient's mean Δ PSD: +1.2 dB.

Table 8. Patient 1 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-33.09	-33.03	+0.06	1.44	1.50	+0.06
24-2	LE	-33.05	-33.04	+0.01	1.54	1.88	+0.34
10-2	RE	-33.81	-31.05	+2.76	3.91	6.34	+2.43
10-2	LE	-33.01	-31.43	+1.58	4.28	6.24	+1.96

Patient's mean Δ MD: +1.1 dB| Patient's mean Δ PSD: +1.2 dB.

Table 9. Patient 2 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-17.67	-9.87	+7.80	11.81	10.89	-0.92
24-2	LE	-16.92	-18.06	-1.14	13.86	13.58	-0.28
10-2	RE	-16.16	-12.59	+3.57	11.77	8.90	-2.87
10-2	LE	-13.65	-15.82	-2.17	10.05	10.41	+0.36

Patient's mean Δ MD: +2.02 dB | Patient's mean Δ PSD: -0.93 dB.

Table 9. Patient 2 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-17.67	-9.87	+7.80	11.81	10.89	-0.92
24-2	LE	-16.92	-18.06	-1.14	13.86	13.58	-0.28
10-2	RE	-16.16	-12.59	+3.57	11.77	8.90	-2.87
10-2	LE	-13.65	-15.82	-2.17	10.05	10.41	+0.36

Patient's mean Δ MD: +2.02 dB | Patient's mean Δ PSD: -0.93 dB.

Table 10. Patient 3 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-22.71	-15.19	+7.52	6.85	11.31	+4.46
24-2	LE	-18.87	-22.84	-3.97	10.04	7.04	-3.00
10-2	RE	-21.71	-21.26	+0.45	13.14	13.54	+0.40
10-2	LE	-20.97	-19.85	+1.12	11.39	12.64	+1.25

Patient's mean Δ MD: +1.28 dB | Patient's mean Δ PSD: +0.78 dB.

Table 10. Patient 3 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-22.71	-15.19	+7.52	6.85	11.31	+4.46
24-2	LE	-18.87	-22.84	-3.97	10.04	7.04	-3.00
10-2	RE	-21.71	-21.26	+0.45	13.14	13.54	+0.40
10-2	LE	-20.97	-19.85	+1.12	11.39	12.64	+1.25

Patient's mean Δ MD: +1.28 dB | Patient's mean Δ PSD: +0.78 dB.

Table 11. Patient 4 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-30.02	—	—	6.62	—	—
24-2	LE	-30.19	—	—	5.18	—	—
10-2	RE	-22.25	-21.52	+0.73	10.06	11.06	+1.00
10-2	LE	-22.39	-24.48	-2.09	8.51	9.88	+1.37

Patient's mean Δ MD: -0.68 dB | Patient's mean Δ PSD: +1.19 dB.

Table 11. Patient 4 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-30.02	—	—	6.62	—	—
24-2	LE	-30.19	—	—	5.18	—	—
10-2	RE	-22.25	-21.52	+0.73	10.06	11.06	+1.00
10-2	LE	-22.39	-24.48	-2.09	8.51	9.88	+1.37

Patient's mean Δ MD: -0.68 dB | Patient's mean Δ PSD: +1.19 dB.

Table 12. Patient 5 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-10.06	-6.90	+3.16	7.30	6.78	-0.52
24-2	LE	-8.96	-6.72	+2.24	6.83	7.96	+1.13
10-2	RE	-15.47	-13.43	+2.04	11.04	11.02	-0.02
10-2	LE	-16.38	-13.83	+2.55	10.11	10.98	+0.87

Patient's mean Δ MD: +2.5 dB | Patient's mean Δ PSD: +0.37 dB.

Table 12. Patient 5 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-10.06	-6.90	+3.16	7.30	6.78	-0.52
24-2	LE	-8.96	-6.72	+2.24	6.83	7.96	+1.13
10-2	RE	-15.47	-13.43	+2.04	11.04	11.02	-0.02
10-2	LE	-16.38	-13.83	+2.55	10.11	10.98	+0.87

Patient's mean Δ MD: +2.5 dB | Patient's mean Δ PSD: +0.37 dB.

Table 13. Patient 6 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-24.72	-28.98	-4.26	6.79	6.27	-0.52
24-2	LE	-21.14	-22.95	-1.81	9.05	8.18	-0.87
10-2	RE	-24.37	-21.25	+3.12	8.81	9.56	+0.75
10-2	LE	-20.56	-18.17	+2.39	9.53	11.28	+1.75

Patient's mean Δ MD: -0.14 dB | Patient's mean Δ PSD: +0.28 dB.

Table 13. Patient 6 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-24.72	-28.98	-4.26	6.79	6.27	-0.52
24-2	LE	-21.14	-22.95	-1.81	9.05	8.18	-0.87
10-2	RE	-24.37	-21.25	+3.12	8.81	9.56	+0.75
10-2	LE	-20.56	-18.17	+2.39	9.53	11.28	+1.75

Patient's mean Δ MD: -0.14 dB | Patient's mean Δ PSD: +0.28 dB.

Table 14. Patient 7 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-30.35	-30.29	+0.06	5.24	5.67	+0.43
24-2	LE	-30.77	-31.18	-0.41	4.35	3.01	-1.34
10-2	RE	-24.08	-23.55	+0.53	7.91	7.99	+0.08
10-2	LE	-24.77	-23.59	+1.18	6.81	7.05	+0.24

Patient's mean Δ MD: +0.34 dB | Patient's mean Δ PSD: -0.15 dB.

Table 14. Patient 7 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-30.35	-30.29	+0.06	5.24	5.67	+0.43
24-2	LE	-30.77	-31.18	-0.41	4.35	3.01	-1.34
10-2	RE	-24.08	-23.55	+0.53	7.91	7.99	+0.08
10-2	LE	-24.77	-23.59	+1.18	6.81	7.05	+0.24

Patient's mean Δ MD: +0.34 dB | Patient's mean Δ PSD: -0.15 dB.

Table 15. Patient 8 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-0.72	-0.79	-0.07	1.40	1.17	-0.23
24-2	LE	-0.96	-1.12	-0.16	1.49	1.68	+0.19
10-2	RE	—	-0.51	—	—	0.98	—
10-2	LE	-0.89	-0.52	+0.37	1.34	1.19	-0.15

Patient's mean Δ MD: +0.05 dB | Patient's mean Δ PSD: -0.06 dB.

Table 15. Patient 8 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-0.72	-0.79	-0.07	1.40	1.17	-0.23
24-2	LE	-0.96	-1.12	-0.16	1.49	1.68	+0.19
10-2	RE	—	-0.51	—	—	0.98	—
10-2	LE	-0.89	-0.52	+0.37	1.34	1.19	-0.15

Patient's mean Δ MD: +0.05 dB | Patient's mean Δ PSD: -0.06 dB.

Table 16. Patient 9 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-22.76	-6.41	+16.35	11.18	8.38	-2.80
24-2	LE	-21.44	-15.78	+5.66	12.08	12.28	+0.20
10-2	RE	—	—	—	—	—	—
10-2	LE	—	—	—	—	—	—

Patient's mean Δ MD: +11.01 dB | Patient's mean Δ PSD: -1.3 dB.

Table 16. Patient 9 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-22.76	-6.41	+16.35	11.18	8.38	-2.80
24-2	LE	-21.44	-15.78	+5.66	12.08	12.28	+0.20
10-2	RE	—	—	—	—	—	—
10-2	LE	—	—	—	—	—	—

Patient's mean Δ MD: +11.01 dB | Patient's mean Δ PSD: -1.3 dB.

Table 17. Patient 10 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-30.66	-28.11	+2.55	6.45	9.61	+3.16
24-2	LE	-30.56	-29.36	+1.20	7.00	7.80	+0.80
10-2	RE	-20.52	-17.04	+3.48	10.77	7.10	-3.67
10-2	LE	-23.73	-18.01	+5.72	10.60	7.36	-3.24

Patient's mean Δ MD: +3.24 dB | Patient's mean Δ PSD: -0.74 dB.

Table 17. Patient 10 pre and post-injection MD and PSD mean Δ values.

Protocol	Eye	MD Pre	MD Post	Δ MD	PSD Pre	PSD Post	Δ PSD
24-2	RE	-30.66	-28.11	+2.55	6.45	9.61	+3.16
24-2	LE	-30.56	-29.36	+1.20	7.00	7.80	+0.80
10-2	RE	-20.52	-17.04	+3.48	10.77	7.10	-3.67
10-2	LE	-23.73	-18.01	+5.72	10.60	7.36	-3.24

Patient's mean Δ MD: +3.24 dB | Patient's mean Δ PSD: -0.74 dB.

Table 18. MD Index (Mean Deviation).

Measure	Pre-injection	Post-injection	Mean Δ
Mean dB (SD) — n=35	-20.86 (±8.87)	-19.09 (±9.47)	+1.77

Table 18. MD Index (Mean Deviation).

Measure	Pre-injection	Post-injection	Mean Δ
Mean dB (SD) — n=35	-20.86 (±8.87)	-19.09 (±9.47)	+1.77

• Paired t-test: t = -2.8098, p = 0.0082 → difference statistically significant (p < 0.05).
• Wilcoxon signed-rank test: W = 140, p = 0.0042 → difference statistically significant (p < 0.05).

Table 19. PSD Index (Pattern Standard Deviation).

Measure	Pre-application	Post-application	Mean Δ
Mean dB (SD) — n=35	7.85 (±3.63)	7.93 (±3.56)	+0.08

Table 19. PSD Index (Pattern Standard Deviation).

Measure	Pre-application	Post-application	Mean Δ
Mean dB (SD) — n=35	7.85 (±3.63)	7.93 (±3.56)	+0.08

Table 20. Comparative Summary Across Parameters.

Parameter	n (eyes)	Mean Δ	Wilcoxon (p)	Paired t (p)	Significance
BCVA (logMAR)	20	-0.109	0.0904	0.1003	No
IOP (mmHg)	20	+0.5	0.3343	0.3092	No
OCT (µm)	20	+49	0.0313	0.0619	Yes (Wilcoxon)
VF — MD (dB)	35	+1.77	0.0042	0.0082	Yes
VF — PSD (dB)	35	+0.08	0.4561	0.7883	No

Table 20. Comparative Summary Across Parameters.

Parameter	n (eyes)	Mean Δ	Wilcoxon (p)	Paired t (p)	Significance
BCVA (logMAR)	20	-0.109	0.0904	0.1003	No
IOP (mmHg)	20	+0.5	0.3343	0.3092	No
OCT (µm)	20	+49	0.0313	0.0619	Yes (Wilcoxon)
VF — MD (dB)	35	+1.77	0.0042	0.0082	Yes
VF — PSD (dB)	35	+0.08	0.4561	0.7883	No

3.4. ERG Statistical Analysis Results

A coefficient of variation (CV) of 30% was assumed for test-retest variability in the same patient over a 30-day interval — meaning a difference between two measurements of up to 30% of the baseline value may occur due to inherent variability of the exam (electrode positioning, adaptation, patient cooperation), even without any real clinical change.

Criterion applied: a pre-to-post change was only classified as a "trend toward improvement" or "trend toward decline" when |Δ%| > 30%. Changes of up to 30% were classified as "stable", since they fall within the expected noise range of the method itself.

Direction: for amplitude parameters (µV), an increase = improvement; for latency parameters (ms), a decrease = improvement.

This analysis is exploratory: the 30% figure was assumed as a premise, not derived from published data for this specific equipment/protocol.

Even without the 30% adjustment, none of the 8 parameters showed a statistically significant difference between pre- and post-treatment at the group level (paired t-test and Wilcoxon, all p > 0.05 for the t-test; Flicker 30Hz – P1 showed p = 0.020 in the Wilcoxon test, not confirmed by the t-test). Applying the 30% criterion, the average magnitude of group change remains small relative to the expected exam variability — reinforcing that there is no evidence of a consistent treatment effect at the group level (Table 21).

Count of parameters (out of 8) classified as trend toward improvement, trend toward decline, or stable within 30% (Table 22).

4. Discussion

This prospective, open-label, single-arm exploratory study evaluated the short-term structural, functional, electrophysiological, and safety outcomes following suprachoroidal administration of Wharton’s jelly-derived mesenchymal stromal cells (WJ-MSCs) in patients with inherited retinal diseases (IRDs). The principal findings were a statistically significant improvement in visual-field mean deviation (MD) in the paired analyses, a statistically significant increase in central foveal thickness (CFT) detected by the Wilcoxon test but not by the paired Student’s t-test, no statistically significant group-level improvement in best-corrected visual acuity (BCVA), stable pattern standard deviation (PSD) and intraocular pressure (IOP), and heterogeneous electrophysiological responses without consistent group-level ERG improvement. These findings should be interpreted as preliminary because of the small sample size, absence of an untreated control group, heterogeneous diagnoses, short follow-up, mixed visual-field protocols, and potential correlation between measurements obtained from both eyes of the same participant [3,14,16,19].

4.1. Visual Acuity

Mean BCVA improved by 0.109 logMAR from baseline to day 45, but this change did not reach statistical significance in either the paired Student’s t-test or the Wilcoxon signed-rank test. The magnitude of the observed change was comparable to the pooled improvement of approximately 0.12 logMAR reported at 6 months in patients with retinitis pigmentosa receiving cell-based therapy. However, direct comparison is limited by substantial differences in follow-up duration, study design, cell source, administration route, patient characteristics, and outcome assessment [3,14,16,19,20].

The absence of statistical significance in the present study may be related to the small sample size and considerable baseline variability. In addition, the conversion of counting-fingers visual acuity to 2.0 logMAR may have influenced the magnitude and distribution of the observed changes. Although this conversion permits exploratory statistical analysis, it does not provide the same measurement properties as standardized ETDRS letter scores. Future controlled studies should therefore use standardized ETDRS-based visual-acuity assessment and validated approaches for very low visual acuity.

The BCVA findings consequently do not provide evidence of a statistically demonstrable treatment-related improvement at 45 days. Nevertheless, the numerical direction of change warrants further investigation in larger studies with longer follow-up and more precise visual-acuity measurement.

4.2. Visual-Field Mean Deviation

The principal functional finding was an improvement in visual-field MD from −20.86 ± 8.87 dB at baseline to −19.09 ± 9.47 dB at day 45, corresponding to a mean change of approximately +1.77 dB. This difference was statistically significant in both the paired Student’s t-test and the Wilcoxon signed-rank test. Visual-field MD may provide complementary information to BCVA in IRDs because it reflects global visual-field sensitivity and may capture functional changes that are not adequately represented by central visual acuity [1,2].

However, the statistical significance of this finding should not be equated with proof of treatment efficacy. Automated perimetry is susceptible to learning effects, fatigue, fixation instability, attention, response bias, and other sources of short-term variability [1,2]. These effects may be particularly relevant when measurements are obtained only at baseline and 45 days without repeated pretreatment examinations.

An additional methodological limitation is that both 24-2 and 10-2 visual-field protocols were included in the dataset. Although both provide quantitative measures of visual sensitivity, they sample different retinal regions and should not be considered fully interchangeable. The use of mixed protocols therefore limits the interpretation of a pooled MD analysis. Future trials should prespecify a single perimetric protocol as the primary functional endpoint, preferably with repeated baseline examinations to establish individual test–retest stability.

The individual response of Patient 9 also requires particular consideration. This participant demonstrated a marked improvement in MD of approximately +11.01 dB but simultaneously exhibited substantial increases in CFT and predominantly declining ERG parameters. Because these findings are discordant, the improvement in MD should not automatically be interpreted as evidence of a biological treatment response. Changes in fixation, perimetric reliability, retinal morphology, or localized pathology may have contributed to the observed change. Patient 9 was retained in the primary analysis because post hoc exclusion of an extreme observation could introduce selection bias. Nevertheless, sensitivity analyses excluding this influential observation would be informative in future analyses.

Importantly, the present study included both eyes of each participant. Measurements from fellow eyes are biologically and statistically correlated, and therefore treating each eye as an independent observation may underestimate uncertainty. Although the paired analyses account for the pre- versus post-treatment structure, they do not fully address the hierarchical nature of two eyes nested within the same participant. Confirmatory studies should therefore use patient-level mixed-effects models, with participant included as a random effect and eye-level observations appropriately modeled.

4.3. Pattern Standard Deviation

PSD remained essentially stable, with a mean change of approximately +0.08 dB. Thus, the observed improvement in MD was not accompanied by a corresponding measurable change in localized visual-field irregularity.

The divergence between MD and PSD may be compatible with a generalized change in visual-field sensitivity rather than restoration of focal visual-field defects. However, this interpretation remains uncertain because both parameters are influenced by perimetric variability and by the characteristics of the visual-field protocol used. Pointwise visual-field analysis, regional sensitivity analysis, and longitudinal progression modeling may provide additional information in future controlled studies.

Accordingly, the stability of PSD should be considered a neutral finding rather than evidence either for or against retinal functional recovery.

4.4. Optical Coherence Tomography

Mean CFT increased from 226.8 ± 108.2 μm at baseline to 275.8 ± 178.4 μm at day 45. The difference reached statistical significance in the Wilcoxon signed-rank analysis (p = 0.0313), whereas the paired Student’s t-test did not demonstrate statistical significance (p = 0.0619). The discrepancy between the two statistical approaches suggests that the result may be sensitive to the distribution of the observations and to individual extreme values.

Patient 9 contributed substantially to this finding, with CFT increasing from 406 to 836 μm in the right eye and from 511 to 681 μm in the left eye. This marked structural change was retained in the primary analysis to avoid post hoc exclusion bias. However, it substantially affects the interpretation of the group-level result and illustrates the importance of patient-level OCT assessment in small exploratory studies.

An increase in CFT should not be interpreted as evidence of retinal regeneration. Increased retinal thickness may result from several mechanisms, including intraretinal fluid, cystoid changes, inflammatory thickening, epiretinal proliferation, vitreomacular traction, or segmentation artifacts. Qualitative assessment of the OCT images, including evaluation of retinal layers, fluid compartments, and segmentation accuracy, is therefore necessary before attributing an increase in CFT to a biological effect of cell therapy.

The discordance between the CFT increase in Patient 9 and the predominantly declining ERG pattern in the same participant further supports a cautious interpretation. In future studies, quantitative CFT should be complemented by predefined qualitative OCT grading and layer-specific structural endpoints, particularly measures of outer retinal integrity and photoreceptor-related biomarkers.

4.5. ERG Findings

At the group level, none of the eight analyzed ERG parameters demonstrated a statistically significant difference in the paired Student’s t-test. The Wilcoxon signed-rank test identified a statistically significant difference only in 30-Hz flicker P1 latency (p = 0.0195), which was not confirmed by the paired t-test [14,16,19,20]. Therefore, the present electrophysiological data do not demonstrate consistent group-level improvement in retinal function at 45 days.

This finding is compatible with the heterogeneous electrophysiological responses reported in the limited clinical literature on cell-based approaches for retinal degeneration. Functional outcomes following cell therapy have varied across studies, and electrophysiological endpoints have not consistently demonstrated improvement. Differences in disease stage, cell source, administration route, dose, follow-up duration, and electrophysiological methodology make direct comparisons difficult [3,14,16,19,20].

The exploratory patient-level analysis based on a 30% change threshold revealed substantial heterogeneity. Patient 10 demonstrated improvement trends in 7 of 8 ERG parameters, whereas Patient 4 showed decline trends in 6 of 8 parameters and Patient 9 showed predominantly declining electrophysiological patterns. These observations indicate that group-level averages may obscure heterogeneous individual trajectories. However, the 30% threshold used in this study was assumed rather than empirically validated for the RETIport system, the specific laboratory, and the testing protocol. It should therefore be regarded exclusively as a hypothesis-generating classification and not as a validated minimally clinically important difference [14,16,19,20].

The interpretation of ERG responses also requires consideration of the directionality of individual parameters. In general, increased amplitude and decreased latency may be directionally compatible with improved retinal electrophysiological function. Nevertheless, a change in an isolated ERG component cannot establish global retinal recovery when other components remain unchanged or deteriorate [2,12].

Furthermore, eight electrophysiological parameters were evaluated, together with several structural and functional endpoints. Because multiple exploratory comparisons were performed without formal adjustment for multiplicity, an isolated nominal p-value should be interpreted cautiously. In particular, the 30-Hz flicker P1 latency result should be considered hypothesis-generating rather than confirmatory.

Repeated baseline ERG examinations, standardized acquisition conditions, and a dedicated test–retest reproducibility study are therefore important methodological priorities for future trials.

4.6. Comparison with Previous WJ-MSC Studies

Preclinical studies provide a biological rationale for investigating WJ-MSCs in retinal degeneration. In an RCS rat model, subretinal administration of human WJ-MSCs was associated with preservation of retinal structure and absence of tumor formation during the study period, with electrophysiological outcomes also evaluated over time. These findings support further investigation of WJ-MSCs but cannot be directly extrapolated to the present human study because of differences in species, disease model, administration route, cell dose, and timing of assessment [13].

Clinical studies using WJ-MSCs administered through the sub-Tenon route have reported preliminary changes in visual function and generally favorable short-term safety profiles. However, differences in administration route, cell preparation, culture conditions, dosing, patient selection, and outcome definitions prevent direct quantitative comparison with the present suprachoroidal approach [14,15,16,17,19,20].

A systematic review and meta-analysis of 21 prospective studies involving 382 patients and 496 eyes reported early visual-acuity improvement after cell-based therapy in RP and Stargardt disease. In RP, approximately 49% of eyes demonstrated visual-acuity improvement at 6 months, decreasing to approximately 30% at 12 months, while the pooled visual-acuity improvement at 6 months was approximately 0.12 logMAR. The analysis also identified suprachoroidal administration as a potentially relevant route for RP, although the evidence was limited and randomized controlled trials were lacking [3].

The present study differs from these previous investigations by specifically evaluating a suprachoroidal WJ-MSC approach and by incorporating visual-field testing, OCT, and ERG in addition to BCVA. These multimodal assessments may provide a broader characterization of short-term functional and structural changes. Nevertheless, the absence of a control group and the short 45-day follow-up prevent determination of whether the observed changes are attributable to the intervention rather than to measurement variability, natural fluctuation, or other confounding factors.

4.7. Safety and Targeted Delivery

No ocular inflammatory or infectious signs were observed during the 45-day follow-up, and no clinically significant treatment-related IOP events were identified. These observations indicate that the procedure was feasible and was not associated with major ocular safety events during the observation period [14,15,16,17,19,20].

However, a 45-day follow-up period is insufficient to characterize delayed complications associated with cellular therapies or the suprachoroidal route. Potential longer-term events may include persistent inflammation, epiretinal membrane formation, proliferative vitreoretinopathy, retinal detachment, choroidal neovascularization, abnormal tissue proliferation, or delayed immune reactions [7,9,19,20]. Accordingly, the present findings should be described as evidence of short-term procedural feasibility rather than definitive evidence of long-term safety.

The suprachoroidal route may represent a biologically attractive alternative between intravitreal and subretinal administration. It potentially permits exposure of the choroid, retinal pigment epithelium, and outer retina while avoiding direct penetration of the retinal tissue [2,6,7,18]. Nevertheless, the actual distribution, persistence, survival, and biological activity of WJ-MSCs following suprachoroidal administration remain incompletely characterized [2,6,7,18,21].

Consequently, the present clinical observations provide an initial basis for further investigation of this delivery route but do not establish the biological mechanism responsible for the observed functional changes.

4.8. Strengths and Limitations

Several aspects strengthen the exploratory value of this study. These include its prospective design, standardized 45-day follow-up, genetic and multimodal confirmation of the retinal diagnoses, use of a defined WJ-MSC product, standardized suprachoroidal administration, multimodal outcome assessment, inclusion of visual-field and ERG endpoints, systematic safety monitoring, and retention of Patient 9 in the primary analysis despite its marked individual structural and functional changes.

The study nevertheless has important limitations. The sample size was small, there was no untreated or sham-controlled comparison group, and the cohort included heterogeneous inherited retinal disorders. Follow-up was limited to 45 days, preventing assessment of durability and delayed adverse events. Both eyes were analyzed, creating potential within-patient correlation that is not fully addressed by conventional paired tests. The visual-field dataset included both 24-2 and 10-2 protocols, limiting direct comparability of MD measurements. Missing visual-field observations further reduced the effective sample size for this endpoint.

Additional limitations include the conversion of counting-fingers acuity to 2.0 logMAR, which may introduce measurement distortion; the absence of repeated pretreatment ERG and visual-field examinations; the lack of a validated ERG test–retest threshold; and the evaluation of multiple exploratory endpoints without correction for multiplicity.

The 30% ERG threshold was not derived from a dedicated reproducibility study of the RETIport device, the laboratory, or the exact acquisition protocol. Its use was therefore exploratory and should not be interpreted as evidence of clinically meaningful electrophysiological improvement.

Patient 9 deserves particular attention because of its substantial CFT increase and marked improvement in MD accompanied by predominantly declining ERG parameters. Retaining this participant in the primary analysis was methodologically appropriate because exclusion based on post-treatment observations could introduce selection bias. Nevertheless, the influence of this participant indicates the need for prespecified sensitivity analyses and patient-level modeling in future studies.

Finally, the absence of a control group is particularly important when interpreting the statistically significant MD finding. Without an untreated comparator and repeated baseline measurements, it is not possible to distinguish a potential treatment-related effect from learning effects, measurement variability, regression to the mean, or other temporal changes.

4.9. Implications for Future Research

The findings of this preliminary study provide several methodological considerations for future investigations. Controlled prospective or randomized studies with larger sample sizes and disease-specific stratification will be necessary to determine whether the observed functional changes are reproducible and treatment-related.

Future protocols should prespecify standardized cell-product characteristics, including dose, viability, manufacturing and release criteria, together with detailed documentation of cell identity and preparation. Repeated baseline assessments of visual fields and ERG should be incorporated to establish individual test–retest variability before treatment.

A single prespecified visual-field protocol should be used as the primary functional endpoint, and analyses should account explicitly for the correlation between fellow eyes using patient-level mixed-effects models. OCT assessment should combine quantitative thickness measurements with predefined qualitative grading of retinal morphology. ERG endpoints should be supported by dedicated reproducibility studies and validated thresholds for clinically meaningful change.

Longer follow-up is also necessary to determine the persistence of functional changes and to monitor delayed structural or inflammatory complications. Patient-reported visual-function outcomes may provide complementary information regarding clinically meaningful changes that are not fully captured by BCVA, visual-field indices, OCT, or ERG.

Finally, future studies should prespecify primary and secondary endpoints and apply appropriate methods for controlling multiplicity when multiple exploratory comparisons are performed.

Overall, the present findings should be interpreted as preliminary evidence regarding the feasibility of suprachoroidal WJ-MSC administration and as a basis for hypothesis generation. The observed improvement in visual-field MD is of interest and warrants confirmation, but it should not be interpreted in isolation as evidence of treatment efficacy. The heterogeneous electrophysiological responses, the CFT increase observed in one participant, and the methodological limitations of the current study underscore the need for adequately powered controlled trials with standardized functional endpoints, patient-level statistical modeling, validated ERG reproducibility criteria, and longer follow-up.

5. Conclusions

These preliminary findings indicate that suprachoroidal administration of WJ-MSCs was feasible and was not associated with major ocular safety events during 45 days of follow-up. The observed improvement in visual-field mean deviation warrants further investigation but should be interpreted cautiously given the single-arm design, small sample size, heterogeneous diagnoses, mixed perimetric protocols, and potential inter-eye correlation. Larger controlled studies with standardized functional endpoints and longer follow-up are required to determine whether these findings represent a treatment-related biological and functional effect. The study support further controlled investigation of the suprachoroidal route as a targeted strategy for inherited retinal diseases.

Appendix A.1. Parameter Classification Matrix (All Patients)

Patient	Scotopic 0.01 – B latency (ms)	Scotopic 0.01 – B amplitude (µV)	Photopic 3.0 – A latency (ms)	Photopic 3.0 – B latency (ms)	Photopic 3.0 – A amplitude (µV)	Photopic 3.0 – B amplitude (µV)	Flicker 30 Hz – P1 (ms)	Flicker 30 Hz – N1-P1 (µV)
Patient 10	Trend toward improvement	Trend toward improvement	Trend toward improvement	Trend toward improvement	Trend toward improvement	Trend toward improvement	Stable (within 30%)	Trend toward improvement
Patient 1	Stable (within 30%)	Stable (within 30%)	Trend toward improvement	Stable (within 30%)	Trend toward improvement	Trend toward improvement	Stable (within 30%)	Trend toward decline
Patient 6	Stable (within 30%)	Trend toward improvement	Trend toward improvement	Trend toward improvement	Trend toward decline	Trend toward decline	Trend toward decline	Trend toward improvement
Patient 8	Trend toward decline	Trend toward improvement	Stable (within 30%)	Stable (within 30%)	Trend toward improvement	Trend toward improvement	Stable (within 30%)	Trend toward improvement
Patient 3	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Trend toward decline	Stable (within 30%)	Stable (within 30%)	Trend toward improvement
Patient 5	Trend toward improvement	Stable (within 30%)	Trend toward decline	Trend toward decline	Trend toward improvement	Trend toward improvement	Stable (within 30%)	Trend toward decline
Patient 7	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)
Patient 4	Trend toward improvement	Trend toward decline	Stable (within 30%)	Trend toward decline	Trend toward decline	Trend toward decline	Trend toward decline	Trend toward decline
Patient 2	Stable (within 30%)	Trend toward decline	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)
Patient 9	Stable (within 30%)	Trend toward decline	Stable (within 30%)	Stable (within 30%)	Stable (within 30%)	Trend toward decline	Stable (within 30%)	Trend toward decline

Appendix B.1. Detailed Pre/Post Values by Patient

Patient 10. (Pre-injection → Post-injection).

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	78.4	29.65	-48.75	-62.2%	Yes
Scotopic 0.01 – B amplitude (µV)	9.8	35.8	26.0	265.3%	Yes
Photopic 3.0 – A latency (ms)	21.6	14.35	-7.25	-33.6%	Yes
Photopic 3.0 – B latency (ms)	48.15	23.35	-24.8	-51.5%	Yes
Photopic 3.0 – A amplitude (µV)	3.16	32.85	29.69	940.4%	Yes
Photopic 3.0 – B amplitude (µV)	19.05	45.1	26.05	136.7%	Yes
Flicker 30 Hz – P1 (ms)	47.25	48.6	1.35	2.9%	No
Flicker 30 Hz – N1-P1 (µV)	7.0	12.0	5.0	71.4%	Yes

Patient 1. (Pre-injection → Post-injection).

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	51.4	62.05	10.65	20.7%	No
Scotopic 0.01 – B amplitude (µV)	10.45	9.7	-0.75	-7.2%	No
Photopic 3.0 – A latency (ms)	19.1	10.45	-8.65	-45.3%	Yes
Photopic 3.0 – B latency (ms)	39.5	49.9	10.4	26.3%	No
Photopic 3.0 – A amplitude (µV)	10.3	14.6	4.3	41.7%	Yes
Photopic 3.0 – B amplitude (µV)	8.45	56.9	48.45	573.4%	Yes
Flicker 30 Hz – P1 (ms)	46.95	56.8	9.85	21.0%	No
Flicker 30 Hz – N1-P1 (µV)	8.05	3.25	-4.8	-59.6%	Yes

Patient 6. (Pre-injection → Post-injection).

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	61.8	79.85	18.05	29.2%	No
Scotopic 0.01 – B amplitude (µV)	13.05	18.65	5.6	42.9%	Yes
Photopic 3.0 – A latency (ms)	17.35	9.1	-8.25	-47.6%	Yes
Photopic 3.0 – B latency (ms)	40.35	16.45	-23.9	-59.2%	Yes
Photopic 3.0 – A amplitude (µV)	13.95	9.15	-4.8	-34.4%	Yes
Photopic 3.0 – B amplitude (µV)	15.2	2.05	-13.15	-86.5%	Yes
Flicker 30 Hz – P1 (ms)	46.8	66.75	19.95	42.6%	Yes
Flicker 30 Hz – N1-P1 (µV)	4.5	13.75	9.25	205.6%	Yes

Patient 8. (Pre-injection → Post-injection).

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	34.35	69.15	34.8	101.3%	Yes
Scotopic 0.01 – B amplitude (µV)	2.44	15.9	13.46	550.9%	Yes
Photopic 3.0 – A latency (ms)	10.45	10.7	0.25	2.4%	No
Photopic 3.0 – B latency (ms)	49.6	49.9	0.3	0.6%	No
Photopic 3.0 – A amplitude (µV)	3.67	11.5	7.83	213.4%	Yes
Photopic 3.0 – B amplitude (µV)	10.75	39.25	28.5	265.1%	Yes
Flicker 30 Hz – P1 (ms)	58.6	66.05	7.45	12.7%	No
Flicker 30 Hz – N1-P1 (µV)	3.35	4.85	1.5	44.8%	Yes

Patient 3. (Pre-injection → Post-injection).

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	84.85	99.65	14.8	17.4%	No
Scotopic 0.01 – B amplitude (µV)	78.15	68.9	-9.25	-11.8%	No
Photopic 3.0 – A latency (ms)	14.5	17.15	2.65	18.3%	No
Photopic 3.0 – B latency (ms)	33.05	32.45	-0.6	-1.8%	No
Photopic 3.0 – A amplitude (µV)	26.4	9.4	-17.0	-64.4%	Yes
Photopic 3.0 – B amplitude (µV)	79.8	88.2	8.4	10.5%	No
Flicker 30 Hz – P1 (ms)	66.9	67.2	0.3	0.4%	No
Flicker 30 Hz – N1-P1 (µV)	53.35	72.45	19.1	35.8%	Yes

Patient 5. (Pre-injection → Post-injection).

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	84.85	35.8	-49.05	-57.8%	Yes
Scotopic 0.01 – B amplitude (µV)	14.6	12.65	-1.95	-13.4%	No
Photopic 3.0 – A latency (ms)	9.25	14.95	5.7	61.6%	Yes
Photopic 3.0 – B latency (ms)	16.0	51.1	35.1	219.4%	Yes
Photopic 3.0 – A amplitude (µV)	4.15	11.55	7.4	178.3%	Yes
Photopic 3.0 – B amplitude (µV)	4.6	21.0	16.4	356.5%	Yes
Flicker 30 Hz – P1 (ms)	59.6	67.1	7.5	12.6%	No
Flicker 30 Hz – N1-P1 (µV)	20.05	3.0	-17.05	-85.0%	Yes

Patient 7 (Pre-injection → Post-injection)

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	99.8	90.55	-9.25	-9.3%	No
Scotopic 0.01 – B amplitude (µV)	50.0	55.65	5.65	11.3%	No
Photopic 3.0 – A latency (ms)	14.85	16.1	1.25	8.4%	No
Photopic 3.0 – B latency (ms)	31.25	31.4	0.15	0.5%	No
Photopic 3.0 – A amplitude (µV)	37.95	30.3	-7.65	-20.2%	No
Photopic 3.0 – B amplitude (µV)	99.0	70.95	-28.05	-28.3%	No
Flicker 30 Hz – P1 (ms)	63.4	63.0	-0.4	-0.6%	No
Flicker 30 Hz – N1-P1 (µV)	78.8	57.0	-21.8	-27.7%	No

Patient 4 (Pre-injection → Post-injection)

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	94.35	49.35	-45.0	-47.7%	Yes
Scotopic 0.01 – B amplitude (µV)	193.0	34.9	-158.1	-81.9%	Yes
Photopic 3.0 – A latency (ms)	14.25	14.85	0.6	4.2%	No
Photopic 3.0 – B latency (ms)	30.2	44.2	14.0	46.4%	Yes
Photopic 3.0 – A amplitude (µV)	42.75	9.8	-32.95	-77.1%	Yes
Photopic 3.0 – B amplitude (µV)	186.5	42.45	-144.05	-77.2%	Yes
Flicker 30 Hz – P1 (ms)	25.8	68.4	42.6	165.1%	Yes
Flicker 30 Hz – N1-P1 (µV)	112.9	19.8	-93.1	-82.5%	Yes

Patient 2 (Pre-injection → Post-injection)

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	77.5	71.6	-5.9	-7.6%	No
Scotopic 0.01 – B amplitude (µV)	140.0	91.15	-48.85	-34.9%	Yes
Photopic 3.0 – A latency (ms)	12.9	15.3	2.4	18.6%	No
Photopic 3.0 – B latency (ms)	30.5	31.25	0.75	2.5%	No
Photopic 3.0 – A amplitude (µV)	23.2	24.9	1.7	7.3%	No
Photopic 3.0 – B amplitude (µV)	93.65	75.5	-18.15	-19.4%	No
Flicker 30 Hz – P1 (ms)	62.95	61.9	-1.05	-1.7%	No
Flicker 30 Hz – N1-P1 (µV)	101.4	83.55	-17.85	-17.6%	No

Patient 9 (Pre-injection→ Post-injection)

Parameter	Pre	Post	Δ	Δ%	Beyond 30%?
Scotopic 0.01 – B latency (ms)	83.2	87.2	4.0	4.8%	No
Scotopic 0.01 – B amplitude (µV)	196.5	96.75	-99.75	-50.8%	Yes
Photopic 3.0 – A latency (ms)	13.05	15.25	2.2	16.9%	No
Photopic 3.0 – B latency (ms)	31.85	32.15	0.3	0.9%	No
Photopic 3.0 – A amplitude (µV)	50.5	46.85	-3.65	-7.2%	No
Photopic 3.0 – B amplitude (µV)	182.9	105.2	-77.7	-42.5%	Yes
Flicker 30 Hz – P1 (ms)	63.1	64.75	1.65	2.6%	No
Flicker 30 Hz – N1-P1 (µV)	115.25	74.75	-40.5	-35.1%	Yes