Predictive Relationship of OCT Characteristics for Efficacy in an Intravitreal Injection of Allogeneic Human Retinal Progenitor Cells (hRPC) for the Treatment of Retinitis Pigmentosa

Abstract

Purpose : The recession of the ellipsoid zone (EZ) and more generally the thinning of the outer zone have become potential structural surrogates for disease severity in retinitis pigmentosa (RP). During a Phase 2b trial conducted to evaluate the intravitreal injection of allogeneic hRPC for treatment of RP, spectral-domain optical coherence tomography (SD-OCT) baseline characteristics of subjects were explored to identify predictors of efficacy in a more responsive target population.

Methods : In a post hoc analysis, a target subgroup of patients with less measurement variability was identified (n=37). From this subgroup, readable SD-OCT volumetric data (n=29; sham, n=10, low dose, n=9, high dose, n=10) was analyzed. Frame by frame macular cube scans were processed using automated segmentation followed by EZ mapping with manual correction for segmentation errors by masked graders. A three-dimensional volume map of the EZ-RPE for the entire macular cube generated measures of mean foveal thickness within the subfield or mid-subfield, mean central foveal thickness (CFT), EZ-RPE subfield or mid-subfield volume and thickness (with or without fluid). Correlational analysis was performed between each OCT parameter and change in each of the phase 2b trial endpoints from baseline to 12 months (Best Corrected Visual Acuity, Contrast Sensitivity, Kinetic Visual Fields, Low Luminance Mobility Test, and a VFQ).

Results : Two parameters from the OCT analysis demonstrated moderate to strong correlations in all trial endpoints in the high dose (single injection of 6 x 10⁶ cells) target group with most relationships being statistically significant despite the small sample size. The parameters were mean CFT (p<0.05 for BCVA, VF, CS, LLMT, VFQ) and mid-subfield mean EZ thickness (p<0.05 for VF, CS, VFQ) with greater values in each parameter corresponding to greater improvement in endpoints. The low dose (3 x 10⁶ cells) did not demonstrate a significant relationship.

Conclusions : Given the small number of subjects in the target population subset and the post hoc nature of the analyses, these results should be viewed cautiously. However, we believe that having a certain minimum EZ or mean CFT thickness will facilitate selection of a population with the greatest chance for response in a phase 3 study from the neutrophic effects of hRPCs

This is a 2021 ARVO Annual Meeting abstract.