Daniela Giannini

Statistics of Optical Coherence Tomography Data From Human Retina

Optical coherence tomography (OCT) has recently become one of the primary methods for noninvasive probing of the human retina. The pseudoimage formed by OCT (the so-called B-scan) varies probabilistically across pixels due to complexities in the measurement technique. Hence, sensitive automatic procedures of diagnosis using OCT may exploit statistical analysis of the spatial distribution of reflectance. In this paper, we perform a statistical study of retinal OCT data. We find that the stretched exponential probability density function can model well the distribution of intensities in OCT pseudoimages. Moreover, we show a small, but significant correlation between neighbor pixels when measuring OCT intensities with pixels of about 5 ¿m. We then develop a simple joint probability model for the OCT data consistent with known retinal features. This model fits well the stretched exponential distribution of intensities and their spatial correlation. In normal retinas, fit parameters of this model are relatively constant along retinal layers, but varies across layers. However, in retinas with diabetic retinopathy, large spikes of parameter modulation interrupt the constancy within layers, exactly where pathologies are visible. We argue that these results give hope for improvement in statistical pathology-detection methods even when the disease is in its early stages.

Investigation of Adaptive Optics Imaging Biomarkers for Detecting Pathological Changes of the Cone Mosaic in Patients with Type 1 Diabetes Mellitus.

Purpose To investigate a set of adaptive optics (AO) imaging biomarkers for the assessment of changes of the cone mosaic spatial arrangement in patients with type 1 diabetes mellitus (DM1). Methods 16 patients with ≥20/20 visual acuity and a diagnosis of DM1 in the past 8 years to 37 years and 20 age-matched healthy volunteers were recruited in this study. Cone density, cone spacing and Voronoi diagrams were calculated on 160x160 μm images of the cone mosaic acquired with an AO flood illumination retinal camera at 1.5 degrees eccentricity from the fovea along all retinal meridians. From the cone spacing measures and Voronoi diagrams, the linear dispersion index (LDi) and the heterogeneity packing index (HPi) were computed respectively. Logistic regression analysis was conducted to discriminate DM1 patients without diabetic retinopathy from controls using the cone metrics as predictors. Results Of the 16 DM1 patients, eight had no signs of diabetic retinopathy (noDR) and eight had mild nonproliferative diabetic retinopathy (NPDR) on fundoscopy. On average, cone density, LDi and HPi values were significantly different (P<0.05) between noDR or NPDR eyes and controls, with these differences increasing with duration of diabetes. However, each cone metric alone was not sufficiently sensitive to discriminate entirely between membership of noDR cases and controls. The complementary use of all the three cone metrics in the logistic regression model gained 100% accuracy to identify noDR cases with respect to controls. Conclusion The present set of AO imaging biomarkers identified reliably abnormalities in the spatial arrangement of the parafoveal cones in DM1 patients, even when no signs of diabetic retinopathy were seen on fundoscopy.

Multimodal Approach to Monitoring and Investigating Cone Structure and Function in an Inherited Macular Dystrophy

PURPOSE To examine a female subject, her father, and a brother harboring a missense mutation of the retinitis pigmentosa 1-like 1 (RP1L1) gene, over 2 years of follow-up. DESIGN Observational case series. METHODS setting: Fondazione G.B. Bietti IRCCS, Rome, Italy. STUDY POPULATION RP1L1 family members and controls. MAIN OUTCOME MEASURES Images of the cone mosaic acquired with an adaptive optics retinal camera, spectral-domain optical coherence tomography (SD OCT), and full-field and multifocal electroretinography (mfERG). RESULTS In the proband, best-corrected visual acuity (≤0.7 logMAR) was stable in both eyes during follow-up, though analysis of adaptive optics images showed decreased cone density in the central 9 degrees from the fovea with respect to controls (P < .05) and cone density loss in the parafoveal area (2 degrees; <12%-16%) during follow-up. Texture analysis of SD OCT images identified abnormalities of the ellipsoid zone in the central 7 degrees, while mfERG response amplitudes were reduced only in the central 5 degrees relative to controls. In the probands father, who had 0.0 logMAR visual acuity, significant cone loss was found in the central 7 degrees from the fovea (P < .05); abnormal SD OCT and mfERG values with respect to controls were found in corresponding retinal areas. No defects in the cone structure and function were found in the probands brother, who had 0.0 logMAR visual acuity. CONCLUSIONS Occult macular dystrophy was diagnosed based on genetic and multimodal ophthalmic findings. The quantitative assessment of photoreceptor survival or loss, based on analysis of adaptive optics retinal images, was valuable to monitor disease progression at a cellular level.

HIGH-RESOLUTION MULTIMODAL IMAGING AFTER IDIOPATHIC EPIRETINAL MEMBRANE SURGERY.

Purpose: To investigate the changes of the vitreomacular interface during a 1-year follow-up after idiopathic epiretinal membrane (iERM) surgery. Methods: Six patients affected by fovea-attached iERM were recruited in this pilot study. Pars plana vitrectomy associated with epiretinal membrane peeling was performed uneventfully in all cases. In four cases, the inner limiting membrane was removed using Brilliant blue G. En face high-resolution adaptive optics and cross-sectional spectral domain optical coherence tomography retinal imaging were performed before and at 1, 3, 6, and 12 months after surgery. The microstructures of vitreomacular interface in high-resolution adaptive optics images were correlated to the cross-sectional spectral domain optical coherence tomography data. Results: Preoperatively, adaptive optics images showed multiple abnormalities of the vitreomacular interface, such as macrofolds, microfolds, and hyperreflective microstructures. We identified two subtypes of iERM according to the distribution of microfolds over the foveal area, which included the radial-type and the grid-type iERM. After surgery, the morphology of the vitreomacular interface changed compared with the preoperative state. The number of both macrofolds and microfolds was reduced in all cases. The hyperreflective structures were still resolvable in all cases, however presenting different shape and morphology than preoperatively. In addition, they showed marked differences between eyes that had internal limiting membrane removal and eyes that did not. Conclusion: Adaptive optics imaging gives new insight into the changes of vitreomacular interface after iERM surgery. Enhanced multimodal imaging of the vitreomacular interface and retinal structures can be valuable to monitor treatment outcome of iERM.

Reliability and agreement between metrics of cone spacing in adaptive optics images of the human retinal photoreceptor mosaic

Purpose To assess reliability and agreement among three metrics used to evaluate the distribution of cell distances in adaptive optics (AO) images of the cone mosaic. Methods Using an AO flood illumination retinal camera, we acquired images of the cone mosaic in 20 healthy subjects and 12 patients with retinal diseases. The three spacing metrics studied were the center-to-center spacing (Scc), the local cone spacing (LCS), and the density recovery profile distance (DRPD). Each metric was calculated in sampling areas of different sizes (64 × 64 μm and 204 × 204 μm) across the parafovea. Results Both Scc and LCS were able to discriminate between healthy subjects and patients with retinal diseases; DRPD did not reliably detect any abnormality in the distribution of cell distances in patients with retinal diseases. The agreement between Scc and LCS was high in healthy subjects (intraclass correlation coefficient [ICC] ≥ 0.79) and moderate in patients with retinal diseases (ICC ≤ 0.51). The DRPD had poor agreement with Scc (ICC ≤ 0.47) and LCS (ICC ≤ 0.37). The correlation between the spacing metrics of the two sampling areas was greater in healthy subjects than in patients with retinal diseases. Conclusions The Scc and LCS provided interchangeable estimates of cone distance in AO retinal images of healthy subjects but could not be used interchangeably when investigating retinal diseases with significant cell reflectivity loss (≥30%). The DRPD was unreliable for describing cell distance in a human retinal cone mosaic and did not correlate with Scc and LCS. Caution is needed when comparing spacing metrics evaluated in sampling areas of different sizes.