This was a retrospective, multicenter, clinical cohort study evaluating eyes with a diagnosis of CSCR with SHRM. All cases were gathered through a collaborative effort by the Macula Society members to create a cohort of CSCR patients with multimodal imaging data. Local institutional review board clearance was obtained from individual institutions, and the study adhered to the tenets of the Declaration of Helsinki. The project has previously published details on data-sharing agreements among collaborators, as well as procedures for collecting and analyzing demographic, clinical, and imaging data. All cases were collected from January 2010 to December 2023. The inclusion criteria were: (a) patients older than 18 years, (b) a confirmed diagnosis of CSCR (c) presence of SHRM between the neurosensory retina and the RPE (in patients with both eyes meeting this criterion, only the right eye was analyzed). The exclusion criteria were: (a) eyes with inadequate or incomplete records, (b) coexisting inflammatory ocular conditions (like VKH) that can present with SHRM, and (c) eyes showing macular neovascularization (MNV) at baseline, diagnosed using OCT angiography or indocyanine green angiography, depending on availability. A control group was also selected from the multicentric cohort, consisting of eyes with CSCR but without SHRM. Information regarding the baseline demographics, best recorded visual acuity (BRVA), systemic co-morbidities, history of psychiatric or sleep disorders, medication use, smoking status, symptom duration, and prior treatment for CSCR was also obtained.

Imaging data were collected at both baseline and final visits, including fundus autofluorescence (FAF), optical coherence tomography (OCT), and fluorescein angiography (FFA). FAF assessment included: (a) the extent of retinal pigment epithelium (RPE) changes quantified in disc areas [including both hypo- and hyperautofluorescent lesions, in disc areas (DA)]; (b) focal vs multifocal RPE alterations; (c) presence of gravitational tracts; and (d) peripapillary RPE changes. CSCR cases were further classified based on the extent of RPE involvement and the presence of multifocal lesions into simple or complex CSCR. Simple CSCR referred to disease limited to 2 or fewer disc areas of RPE alteration, whereas complex CSCR included eyes with multifocal involvement or changes extending more than 2 DA.

Other OCT parameters evaluated included CMT and the height and width of the neurosensory retinal detachment (NSRD). CMT was obtained automatically from the macular cube scan generated by the OCT device. The height of the NSRD was measured from the innermost surface of the elevated neurosensory retina at its highest point to the inner border of the RPE at the fovea. Choroidal measurements included SFCT, Haller vessel diameter, and inner choroidal thickness. SFCT was determined within a 500 µm radius of the fovea using the OCT caliper, extending from the outer surface of the RPE to the choroid–sclera junction. Haller vessel thickness was measured by identifying the largest choroidal vessel lumen within 750 µm of the foveal center. The inner choroid, comprising the choriocapillaris and Sattler’s layer, was measured from the inner margin of the largest Haller vessel to the outer surface of the RPE. Pigment epithelial detachments (PEDs) were identified on OCT as dome-shaped elevations of the RPE. Their number was confirmed using fundus fluorescein angiography (FFA), infrared imaging, or en-face OCT. The tallest PED within the scanned volume was selected for height measurement, defined as the vertical distance from Bruch’s membrane to the apex of the RPE elevation. PED width was measured horizontally between the points where the RPE contour returned to normal, using Bruch’s membrane as a reference. Fovea-involving RPE atrophy was also recorded, defined as focal or diffuse areas of hyper-transmission on OCT, with any portion located within 200 µm of the foveal center. Loss of the ellipsoid zone (EZ) was defined as a discontinuity in the EZ extending over an area of at least 100 microns. EZ integrity was analyzed only in eyes with SRF resolution at the final visit. Assessment (as a categorical variable) was performed by reviewing the entire macular cube to identify EZ disruptions corresponding to regions of prior NSRD. The double-layer sign (DLS) was recorded when shallow, irregular PEDs with internal hyperreflective, hypo-reflective, or mixed signals were observed. Hyperreflective dots (HRDs) in the choroidal stroma were also noted. Photoreceptor outer segment (PROS) thickening was defined as an increased distance between the inner boundary of the ellipsoid zone and the outer surface of the neurosensory retina in the presence of SRF. Any irregularities in the PROS layer in any of the OCT B-scans of the macular cube were also recorded. Detection of macular neovascularization (MNV) was based on findings from OCT angiography or indocyanine green angiography (ICG-A), when available.

OCT images were acquired using Heidelberg Spectralis HRA and OCT (Heidelberg Engineering, Heidelberg, Germany), Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA), and Triton SS-OCT device (Topcon Corporation, Tokyo, Japan). FFA/ICG/FAF images were acquired using Heidelberg Spectralis HRA and OCT (Heidelberg Engineering, Heidelberg, Germany), Zeiss FF 450 plus IR (Carl Zeiss Meditec AG, Jena, Germany), and Optos California device (Optos plc, Dunfermline, UK).

SHRM was identified on OCT and was defined by the presence of medium to high reflectivity material in the sub-retinal space. The SHRM group was further classified into 3 grades using Fiji software (ImageJ, version 2.16.0) based on the methods described in a previous study by Pu et al. The areas of SHRM were manually delineated by an experienced grader (NH). The proportion of SHRM area relative to the total subretinal fluid (SRF) area was subsequently calculated. An initial attempt to quantitatively assess SHRM by binarizing the images was deferred due to suboptimal detection of finer SHRM reflectivities. Grade 1 consisted of occupying less than 20 % of the subretinal space ( Figure 1 B); Grade 2 represented SHRM occupying 20–50 % of the subretinal space ( Figure 1 C); and Grade 3 indicated extensive SHRM involving more than 50 % of the subretinal area, often showing layered or onion-like reflectivity on OCT ( Figure 1 C). Few modifications were made to the classification: the grades were assigned based on the area occupied, irrespective of the configuration [adherent to the neurosensory retina, suspended within subretinal fluid, or appearing as filamentous strands beneath the photoreceptor outer segments (PROS)]. The mean SHRM area was calculated as the average of 3 horizontal OCT scans within the macular cube. The 3 horizontal OCT scans for evaluation were selected as follows: one mandatory scan through the foveal center, and 2 additional non-consecutive scans positioned superior and inferior to the fovea, ensuring that at least one scan captured the region of maximum SHRM accumulation based on review of the entire macular cube. This standardized protocol enabled consistent evaluation across all cases, regardless of leakage-point identification or angiographic availability. This approach also minimized overestimation that could occur when SHRM appeared extensive on a single scan through a shallow detachment segment.

Example of (A) Central serous chorioretinopathy (CSCR) with no subretinal hyper-reflective material (SHRM), (B) CSCR with Grade-1 SHRM, showing continuity of the material with the elongated photoreceptor outer segment (PROS)( white arrow ), (C) CSCR with Grade-2 SHRM with broad attachments to the PROS, and (D) CSCR with Grade-3 SHRM showing layering.

For all measurements, including SHRM grading, data were collected from each center and evaluated collectively by a senior grader (NH) who was masked for the clinical outcomes.

The primary outcome measures were factors affecting the change in BRVA from baseline to the final visit, and the factors associated with persistence of subretinal fluid at the final follow-up. Similarly, the secondary outcomes were changes in CMT, NSRD height, and SFCT; disease course (resolution, recurrence, or persistence); development of MNV; EZ loss in resolved cases; subretinal scar formation; and change in SHRM grade.

All analyses were conducted using R Studio (version 2025.09.1 + 40), R Foundation for Statistical Computing, Vienna, Austria. The cohort consisted of 2 groups according to the presence or absence of SHRM. Continuous variables were summarized as mean ± SD or median (IQR), while categorical variables were expressed as counts and percentages. The normality of data distribution was examined using the Shapiro-Wilk test. Comparisons between groups were made using the independent-samples t-test or Mann-Whitney U test for continuous variables, and the chi-square or Fisher’s exact test for categorical data. To explore determinants of persistent SRF, logistic regression analysis was performed; and to analyze the factors affecting change in BRVA [expressed in logarithm of minimum angle of resolution (logMAR)], linear regression analysis was performed. Each baseline factor was first evaluated in a univariate model, and variables showing a p -value <.10 were subsequently entered into a multivariable logistic regression model. A manual stepwise approach was then applied, where variables were iteratively added or removed. Results were presented as odds ratios (OR) or Regression coefficients with 95% CIs (CI). A 2-sided p <.05 was considered statistically significant.