Highlights
-
•
Microperimetry increases eligibility in RPGR gene therapy trials.
-
•
LLVA underestimates treatment response across disease stages.
-
•
Microperimetry detects more responders than LLVA at Month 12.
-
•
Central 16-point grid performs comparably to full 68-point testing but must be utilized with caution.
-
•
Endpoint choice substantially alters detection of trial efficacy.
Purpose
To evaluate how primary endpoint selection influences patient eligibility, enrolment, and detection of treatment efficacy in RPGR gene therapy trials, comparing low luminance visual acuity (LLVA) with microperimetry-based outcome measures.
Design
Retrospective analysis of a phase 1/2 to 3 interventional clinical trial.
Subjects
Fifteen patients were included in this study. Each was administered with gene therapy in one with the other eye serving as control.
Methods
Month 12 data from patients with RPGR -associated retinopathy enrolled at a single center in the XIRIUS trial (NCT03116113) were retrospectively analyzed. FDA-aligned low-luminance visual acuity (LLVA) responder criteria (≥15-letter gain) and EMA-aligned significant change from baseline in microperimetry mean sensitivity, were applied. Microperimetry outcomes were evaluated across the full 68-point MAIA grid and within a central 16-point subset. Minimal baseline inclusion thresholds were used. The proportion of gene therapy–treated patients meeting FDA- and EMA-aligned (≥2.5 dB improvement in mean sensitivity) responder criteria was determined.
Main Outcome Measures
Proportion of patients meeting the responder criteria in either microperimetry or LLVA.
Results
At baseline, LLVA excluded more patients from study inclusion due to floor effects than microperimetry. At month 12, 2 of 10 patients (20%) met LLVA responder criteria. In contrast, 5 of 11 patients (45%) met EMA-aligned whole-grid microperimetry responder criteria, including all LLVA responders and 3 additional patients with clear functional improvement not captured by LLVA. Restricting analysis to the central 16 microperimetry points further increased responder detection, identifying 7 of 11 patients (64%). Improvements in microperimetry mean sensitivity of at least 2.5 dB, exceeding expected test–retest variability. No responders were identified for any endpoint in untreated fellow eyes.
Conclusions
Primary endpoint choice substantially affects efficacy detection and patient inclusion in RPGR gene therapy trials. LLVA demonstrated limited sensitivity and restricted eligibility to a narrow disease window, risking underestimation of treatment benefit. Microperimetry mean sensitivity, particularly when spatially aligned with the treated retinal area, detected functional improvement in a substantially larger proportion of patients and supports broader enrolment. Microperimetry-based endpoints provide a more sensitive and inclusive primary outcome measure for RPGR gene therapy trials.
INTRODUCTION
R etinitis pigmentosa (RP) encompasses a broad spectrum of genetically heterogeneous disorders characterized by progressive retinal degeneration primarily affecting rod and cone photoreceptors. The disease typically manifests with night blindness and peripheral visual field loss, eventually progressing to severe central vision impairment or complete blindness. , X-linked RP, primarily caused by mutations in the Retinitis Pigmentosa GTPase Regulator ( RPGR ) gene, represents one of the most severe and common forms. RPGR encodes a protein localized to the photoreceptor connecting cilium, where it plays a critical role in protein trafficking between the inner and outer segments of photoreceptors and helps support the delivery of phototransduction-related cargo that is essential for outer segment maintenance and normal photoreceptor survival. ,
Two recent phase 3 gene therapy clinical trials have now concluded, and both have failed to meet their primary endpoint. MeiraGTx/Janssen R&D (NCT04671433) failed to meet its primary endpoint on demonstrating improvement on vision-guided mobility of patients compared with an uninjected control group. Similarly, cotoretigene toliparvovec (XIRIUS, NCT03116113) from Biogen also failed to meet its primary endpoint in the number of participants meeting the predefined microperimetry responder criteria (7 dB improvement at 5 of 16 central loci), largely due to COVID-related recruitment issues. A third clinical trial (VISTA, NCT04850118) delivering a codon-optimized full length RPGR ORF15 sequence is currently underway at sites within the USA, UK and Australia, with 12-month outcomes expected by the end of 2026. There is discrepancy in the agreement of an acceptable primary outcome measures across worldwide regulatory agencies. For the current VISTA trial, the primary outcome measure approved in the United States by the Food and Drug Administration (FDA) is the proportion of participants with a ≥ 15 letter improvement at month 12 in low luminance visual acuity (LLVA) compared to an untreated control group, whereas, in Europe, the European Medicines Agency (EMA) has approved a significant change, compared to controls, from baseline to month 12 in mean sensitivity across the whole grid, as measured by MAIA microperimetry.
Microperimetry has emerged as a valuable tool for quantifying macular retinal function through precise threshold sensitivity measurements correlated with specific retinal locations. , Unlike conventional perimetry, microperimetry incorporates fundus-tracking technology, enabling precise spatial correlation between anatomical features and functional sensitivity, making it particularly valuable for monitoring macular function in progressive retinal conditions. Previous work by the authors have demonstrated the relative sensitivity of both LLVA and microperimetry mean sensitivity (and their correlation) as outcome measures, with LLVA being sensitive only over a narrow range in those patients where LLVA is impaired by central foveal encroachment of the disease, but not yet extinguished. ,, This narrow subset of patients that could demonstrate a functional improvement limits patient recruitment from the already small pool of available patients and risks excluding patients due to insensitivity of the endpoint.
Here, we explore the relative sensitivity and suitability of endpoints by applying the VISTA trial’s EMA-accepted microperimetry-based endpoint criteria and the FDA-accepted LLVA endpoint criteria in a post hoc analysis of month 12 results from the phase 2/3 XIRIUS trial. This focus was chosen because, in gene therapy trials, endpoint selection is not only methodological but also regulatory, directly influencing patient eligibility, efficacy assessment, and ultimately whether a therapy may gain approval. Selection of appropriate outcome measures was identified as a priority research topic by the Monaciano Symposium, a global gathering of scientists and physicians focused on accelerating research and developing treatments for Inherited Retinal Dystrophies. Comparing LLVA to microperimetry based outcome measures therefore addresses a clinically and translationally important question: whether currently relevant regulatory frameworks capture meaningful functional benefit in RPGR-associated retinopathy. We further explore the potential of a central 16-point microperimetry endpoint, in keeping with previous FDA-mandated microperimetry outcomes based on a 7-dB improvement at 5 of 16 central loci, and reflecting the need to align endpoint design with the spatial distribution of treatment effect.
METHODS
This retrospective analysis included data from patients diagnosed with RPGR -related retinopathy enrolled on XIRIUS (Clinicaltrials.gov NCT03116113); part 1 (phase 1 dose escalation) and part 2 (phase 2/3 masked dose expansion) who were administered with a single subretinal injection of cotoretigene toliparvovec. Whilst the XIRIUS trial was a multicenter study, this post-hoc analysis only includes patients enrolled at the Oxford Eye Hospital. The study adhered to the tenets of the Declaration of Helsinki and written informed consent was obtained from all participants. Ethics Committee approval was obtained prior to study initiation.
Low luminance VA was assessed, following a refraction, using a retro-illuminated early treatment diabetic retinopathy study (ETDRS) chart, placed at 4 meters, with all room lights switched off and a 2.0 neutral density filter added in front of the testing eye (background luminance 1.6 cd/m2).
Retinal sensitivity was measured using the MAIA microperimeter (CenterVue, Padova, Italy). The standardized testing protocol employed a 4-2 threshold strategy with a Goldmann III stimulus size, presented for 200 ms against a background luminance of 1.27 cd/m². A standard 68-point grid pattern was utilized, henceforth termed “whole-grid,” covering approximately a 10° radius (central 20°) around the central retina with a consistent 2° spacing between stimulus loci. Tests with poor reliability as defined in the protocol (fixation losses > 30%) were excluded from the analysis as per standard protocol.
Two key metrics were then extracted and considered for post-hoc analysis, specifically evaluating which patients would have achieved the LLVA and microperimetry-based outcome measures. These are:
-
1)
FDA-aligned endpoint: proportion of participants with a ≥ 15 letter improvement from baseline to month 12 in low luminance visual acuity (LLVA).
-
2)
EMA-aligned endpoint: proportion of participants with a significant change from baseline to month 12 in mean sensitivity across the whole grid, as measured by MAIA microperimetry.
-
3)
Exploratory endpoint: proportion of participants with a significant change from baseline to month 12 in mean sensitivity across only the central 16 points, as measured by MAIA microperimetry.
Here, a significant change in microperimetry mean sensitivity is defined in this study as ≥2.5 decibels, which is approximately double the test-retest repeatability and considered functionally significant in the context of RPGR-associated retinopathy . This may differ in other studies with different entrance and testing parameters.
A comprehensive description of methods, demographics and results have previously been published. A set of post-hoc inclusion criteria were applied to ensure the inclusion of the broadest range of patients but with minimal compromise on potential detection of treatment efficacy. As such the inclusion criteria at baseline for those included in this analysis were:
-
1)
LLVA > 0 letters and;
-
2)
Microperimetry (whole grid) mean sensitivity (MP) > 0 dB.
The lower limits of LLVA and microperimetry were set to ensure elimination of patients whose visual function was too poor to be registered at baseline reaching the floor effects of each test. No other bounds were applied ensuring the widest possible inclusion with the above exceptions. Analysis was performed on both the treated and untreated fellow eyes of each patient.
STATISTICAL ANALYSIS
All analyses and figure plotting were performed using R software version 4.1.0. Due to the small sample sizes and research question revolving around proportion of responders using different endpoints, no hypothesis testing was employed in this work and instead we have provided descriptive analyses only.
RESULTS
Fifteen male patients with RPGR -associated retinopathy who received a single subretinal injection of cotoretigene toliparvovec in 1 eye as part of the XIRIUS (part 1 and 2) clinical trial were included in this retrospective analysis. Application of the predefined post-hoc baseline inclusion criteria resulted in the exclusion of 5 patients from the LLVA analysis due to zero baseline LLVA letter scores and 4 patients from the microperimetry analysis due to zero baseline mean sensitivity. Consequently, baseline and month 12 LLVA data were available for 10 patients, while baseline and month 12 microperimetry data were available for 11 patients.
All included microperimetry tests met predefined reliability criteria, with fixation losses ≤30%. A summary of demographic and baseline clinical characteristics has been previously reported.
LOW LUMINANCE VISUAL ACUITY OUTCOMES
Individual LLVA letter scores at baseline and month 12 for all included patients are shown in Figure 1 . At month 12, two of the 10 patients (20%) demonstrated an improvement of ≥15 letters from baseline, meeting the FDA-aligned threshold for clinically meaningful change. The remaining 8 patients demonstrated either smaller improvements, no change, or a decline in LLVA that did not meet responder criteria. Of the 3 patients that showed evidence of LLVA declines from baseline to M12 (pt 8, 9 and 10 with 10, 5 and 6 letter decrease, respectively), all were less than the 15 letter change criteria which would be considered clinically significant.
LLVA responders results from baseline to month 12 for all 10 analysis eligible patients. Of these, only 2 patients (20%) achieved an LLVA gain of 15 or more letters that would be considered clinically significant according to guidance from the FDA.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
