Subjects and image identification
This was an observational study conducted at a single tertiary eye institute in South India. Institutional ethical committee approval was obtained and the study adhered to tenets of declaration of Helsinki. From a MacTel registry of 745 patients, we searched for patients with at least one eye with a clinical diagnosis of stage-3 MacTel based on Gass and Blodi classification for whom multimodal imaging was done. We excluded eyes with MacTel stages-4 and 5, those with coexistent macular conditions and those with incomplete imaging and poor imaging quality across any modality as well as image artefact from vessel shadowing on OCT and motion artefacts on OCT-A and where artefacts persisted despite using the projection artefact removal tool. A total of 43 eyes of 38 patients were recruited for the study. Of the patients analyzed, none had record of history of drug intake, solar exposure, trauma or previous vitreoretinal surgery. Electronic medical records of complete ophthalmologic examination, best-corrected visual acuity (BCVA), fundus examination, CFP and SD-OCT were performed in all cases. Other multimodal imaging such as confocal blue reflectance (CBR) 20 and OCT-A 21 were evaluated, if available.
- Image acquisition and analysis
In all patients, CFP was taken using Carl Zeiss (FF 450 plus IR) fundus camera. CBR images were acquired using the Heidelberg confocal system, followed by SD-OCT [volume scans of 15 mm×10 mm (high-resolution detail-scan mode, 97 scans) Spectralis; Heidelberg Engineering, Heidelberg, Germany], and OCT-A [10° x 10° (3 mm × 3 mm) high resolution mode; lateral resolution of 5.7 µm/pixel (512 A-scans x 512 B-scans), Heidelberg]. These images were captured by senior fellows of the retina department in our institute and centered on the posterior pole covering 30-degrees scanning field.
Image analysis
Outer retinal layer microstructural alterations on SD-OCT were defined as follows:
- IDZ attenuation: Faint hypo-reflective IDZ band/remnant IDZ fragments with intact overlying EZ band (Fig. 1(b)-C, 2(a)-D, 2(b)-D).
- IDZ loss: Absence of IDZ band, as confirmed by discontinuity in IDZ (Fig. 2(a)-E, 2(b)-E, 2(b)-F, 3(a)-C, 3(b)-C, 3(b)-D).
- EZ attenuation: Faint but present hypo-reflective EZ band (Fig. 2(b)-F, 3(a)-C1).
- EZ loss: Absent EZ/slit in band between ELM and IDZ (Fig. 3(a)-C2, 3(a)-D, 3(b)-D).
Although not a defect, “at-risk” IDZ was defined as photoreceptor areas with a ragged (irregular) appearance that precede an IDZ and EZ defect (Fig. 2(a)-C and 2(b)-C). 2 The SD-OCT B-scans were carefully analyzed in all 97 detail scan sections. Corresponding SD-OCT radial distance measurement of IDZ attenuation, IDZ loss, EZ attenuation and EZ loss was traced manually using the built-in overlay “caliper measurement tool” of Heidelberg eye explorer of Heidelberg machine. The length of IDZ was defined as ends of the continuous and highly reflective line between EZ and retinal pigment epithelium-Bruch’s membrane (RPE-BM) and the radial distance measure of IDZ defect was any discontinuity in IDZ band. The length of EZ was defined as ends of the continuous line between external limiting membrane (ELM) and IDZ and radial distance measure of the EZ defect was any discontinuity involving EZ band. The measurements were done independently by two authors (KC, AG) in a masked fashion. In the event of any discrepancy, the two graders evaluated the images together and reached a final consensus.
Our definition of RAV is in line with published data - on CFP as one or several visible, blunted, slightly dilated vessels, mainly located in the temporal parafovea, and that seem not to narrow towards the foveola, but suddenly dive at a right angle into deeper retinal layers. 1, 18, 19 On CBR and OCT-A (especially whole retina slab), vessels demonstrating the same characteristics that course the superficial and deep retinal layers with an apparent discontinuation were similarly defined as RAV. 18
RAV was determined initially on CFP and further corroborated using CBR and OCT-A, if available. 19
OCT-A segmentation was done to study the changes in the macular capillary beds at the level of superficial vascular complex (SVC) and DCP. SVC was segmented from internal limiting membrane (ILM) to 55µm above inner plexiform layer (IPL) while DCP was segmented from 6µm above IPL to 50µm below IPL. Telangiectatic vessels in DCP were defined as dilated capillaries with non-tapering ends. 2 Artefacts from SVC were removed using the in-built projection removal tool of the Heidelberg OCT-A machine. The confirmation of DCP telangiectasia as opposed to superficial vessel artifact was done by 2 authors (KC and AG).
Follow-up was available in a subset of eyes (15 eyes) for which we determined whether IDZ defect was a precursor to EZ defect in eyes with early MacTel. We also calculated the predictive value which described the likelihood of a preceding IDZ defect to predict future EZ defect.
For statistical analyses, the annual progression in IDZ loss [(a) baseline to follow-up 1 and (b) follow-up 1 to final follow-up] and EZ attenuation (follow-up 1 to final follow-up) were determined (all patients with IDZ attenuation at presentation had IDZ loss at follow-up). This annual progression was defined as the values calculated by dividing the difference between baseline and follow-up 1 values and between follow-up 1 and final follow-up values by the number of years between the examinations. Comparison in changes of IDZ loss and EZ attenuation values was done during the observation period calculated from baseline.
- Statistical analysis
Statistical analysis was performed using SPSS Statistics software version 26 (IBM Corp., Armonk, NY). Paired t-test and Wilcoxon signed rank test were used to determine the significance of changes in the IDZ (IDZ loss) and EZ (EZ attenuation) values, respectively, during the observation period. The association between baseline IDZ defect and follow-up EZ defect was determined using generalized estimating equations (GEE) regression analysis. A p-value ≤ 0.05 was accepted as statistically significant.