Measurement Reliability and Functional Validity of Bruch's Membrane Calcification in Pseudoxanthoma Elasticum: PROPXE Study Report 2

Abstract

Purpose: The purpose of this study was to assess the inter-reader and inter-visit reliability of en face Bruch's membrane (BrM) calcification measurements in Pseudoxanthoma elasticum (PXE), investigate how the patient's age influences the BrM calcification extent, and evaluate the association between BrM calcification and delayed rod-mediated dark adaptation.

Methods: This prospective natural history study (PROPXE, ClinicalTrials.gov ID: NCT05662085) included 26 patients (14 women and 12 men; median age = 55 years, interquartile range = 43-59 years) diagnosed with PXE. Participants underwent comprehensive ophthalmic evaluations, including widefield infrared reflectance imaging (up to 83 degrees eccentricity) and dark adaptometry at retinal eccentricities of 8 degrees, 15 degrees, 30 degrees, and 46 degrees along the temporal retina. The primary outcomes were inter-visit and inter-reader reliability of the temporal inner peau d'orange boundary extent, its correlation with age, and its relationship with rod-intercept time (RIT) during dark adaptation.

Results: The temporal inner peau d'orange boundary showed excellent inter-reader and inter-visit reliability (inter-reader intraclass correlation coefficient [ICC] = 0.92 and inter-visit ICC = 0.95) and varied significantly with age (+4.35 deg/decade, P = 0.001). A greater temporal inner peau d'orange boundary extent was strongly associated with delayed rod-mediated dark adaptation at 8 degrees (+1.05 min/deg, P = 0.002) and 15 degrees (+0.91 min/deg, P = 0.001) eccentricities. A threshold effect was observed, with delayed dark adaptation at 8 degrees and 15 degrees eccentricity (from the fovea; i.e. 23 degrees and 30 degrees from the optic nerve head) manifesting once the temporal inner boundary exceeded 27.15 degrees and 32.56 degrees eccentricity (from the optic nerve head), respectively.

Conclusions: Patients with PXE demonstrate significant rod-mediated dark adaptation deficits correlating with the temporal inner peau d'orange boundary extent. These findings support the use of BrM calcification extent as an objective marker for disease severity, facilitating earlier therapeutic intervention than currently possible in PXE.

Figure 1.

Wide-field infrared (IR) reflectance imaging montage and image annotation. The 9-field 55 degrees IR images were registered semi-automatically to obtain a wide-field IR montage (A, B). Subsequently, three readers independently annotated the peau d'orange outer and inner boundaries (C). Last, the temporal, superior, and inferior boundary extents were extracted automatically. The temporal meridian was defined as the horizontal line starting at the level of the optic disc center and going through the fovea. The vertical superior and inferior meridians had the exact same origin (i.e. vertical position based on the vertical position of the foveal center and horizontal position based on the horizontal center of the optic disc).

Figure 2.

Inter-reader reliability and inter-visit reliability for the temporal inner peau d'orange boundary extent. (A, B, C) The Bland-Altman plots show the inter-reader reliability, incorporating data from both the baseline and retest visits. The solid gray line and ribbon indicate the bias and 95% confidence interval, whereas the orange dashed lines represent the 95% limits of agreement. (D) The last Bland-Altman plot displays the inter-visit reliability (including averaging of the grader's individual measurements).

Figure 3.

Inter-eye concordance for the peau d'orange boundary extents. The panels display the scatter plots for the peau d'orange boundary extents along the temporal (A), superior (B), and inferior (C) meridians for the right and left eyes. The gray circles indicate outer boundary measurements, whereas the orange circles represent inner boundary measurements.

Figure 4.

Association between age and the temporal inner peau d'orange boundary and relationship to dark adaptation. (A) The first panel shows the association between age and the extent of the temporal inner peau d'orange boundary, with data points from both eyes and visits plotted. (B, C, D) The panels B to D illustrate the relationship between the temporal inner boundary and rod-intercept time (RIT), a measure of dark adaptation, at 8 degrees, 15 degrees, and 30 degrees eccentricity. Regression lines were derived from Bayesian linear or nonlinear mixed models, with step functions applied where appropriate. Thin lines represent posterior draws of the expected value, visualizing the range of plausible fits under the model. Marginal R² values (with 95% credible intervals [CrI]) indicate the proportion of variance explained by the fixed effects alone. (B–D) Are shown without age adjustment, illustrating the unadjusted association between peau d'orange extent and RIT. For reference, the extent of peau d'orange was defined as eccentricity from the optic nerve head, whereas RIT was defined relative to the fovea (i.e. 15 degrees temporal to the optic nerve head). Therefore, the observed breakpoints in panels C and D align very closely with the locations at which RIT was measured.

Figure 5.

Patient examples. The upper row shows a patient with a temporal inner peau d'orange boundary within the central 30 degrees. The rate of dark adaptation at all loci is within normal limits (panels for 8 degrees, 15 degrees, 30 degrees, and 46 degrees eccentricity, blue dots for the 500 nm stimulus, red dots for the 647 nm stimulus, and vertical dashed lines denoting the cone-rod-break). The lower row shows a patient with proximity of the inner and outer peau d'orange boundary. Notably, the dark adaptation at 15 degrees eccentricity is markedly delayed. The red overlays highlight areas with subretinal drusenoid deposits (SDDs), which can be mistaken for peau d'orange.