To determine the frequency of AMD among patients with ODD and to compare the clinical and imaging features, we performed a retrospective analysis from a dataset of 650 prospectively recruited patients with various optic neuropathies, including ODD, at Byers Eye Institute, Stanford University, USA. The Stanford Institutional Review Board approval was obtained and electronic medical records were reviewed for each patient in accordance with the Declaration of Helsinki and the Health Insurance Portability and Accountability Act.

We identified 94 consecutive patients with ODD and a cohort of 100 consecutive neuro-ophthalmology patients without ODD presenting for evaluation of optic neuropathy (July 2022 and June 2024). The non-ODD group included patients presenting for evaluation of optic neuropathy. Among these, the most frequent diagnosis was ischemic optic neuropathy in 38%. Other diagnoses included papilledema, optic atrophy, Leber’s hereditary optic neuropathy, mitochondrial deletion syndrome, nonspecific visual disturbances, and others. An additional patient with ODD-AMD met our inclusion criteria through chart review, leading to a total of 10 patients (20 eyes) with ODD-AMD. All patients had comprehensive ophthalmic examination and multimodal imaging. AMD was diagnosed and classified into early, intermediate, or late (dry form with geographic atrophy or exudative form with neovascularization) stages based on the Beckmann Classification for AMD (10). The diagnosis for ODD was established based on the EDI-OCT images, the gold standard for diagnosis according to The Optic Disc Drusen Studies Consortium. Furthermore, the drusen were categorized into superficial and buried in relation to Bruch’s membrane (11). Patients with nonspecific retinal deposits, retinal dystrophies, or neuro-ophthalmic diseases that can impact analysis were excluded.

We performed multimodal imaging using FAF and OCT. FAF was acquired using an ultra-wide field Optos system (Optos PLC, Dunfermline, Scotland). Spectral-domain OCT images were acquired for the optic disc and macula using Cirrus HD-OCT (Carl Zeiss Meditec Inc., Jena, Germany). We acquired an Optic Disc Cube (200 × 200) with 200 horizontal scan lines and a Macular Cube (512 × 128) with 128 horizontal scan lines. We also acquired the 97-line spectral-domain EDI-OCT scans of the optic nerve (Spectralis HRA; Heidelberg Engineering, Heidelberg, Germany) (11).

We identified 26 likely calcified macular drusen, 34 noncalcified macular drusen, and 19 ODD from the study eyes based on published criteria (3). Then, we measured the near-infrared light reflectivity of the drusen core, the drusen cap, and the difference between them using the image processing software, FIJI (An enhanced version of ImageJ). We used a previously described method with some modifications tailored to our parameters (12–15). Figure 1 illustrates the regions defined as drusen core and cap in ODD and macular drusen. Reflectivity was defined as the average gray value of the selected pixels, and a value of 0 corresponds to pure black and a value of 255 corresponds to pure white. To eliminate the signal variation between the OCT images, average gray values from quadrangular areas of the vitreous and retinal nerve fiber layer (RNFL) were used for standardization. We calculated the average reflectivity of the drusen core (a region of low to no reflectivity). The drusen cap (a region of high reflectivity) signal was calculated as the average of the top five pixels above the core. We calculated the difference between the drusen cap and drusen core for each macular drusen using the formula:

To calculate macular drusen disease burden, defined as the total volume of macular drusen per eye, we manually outlined all macular drusen on the SD-OCT Macular Cube 512 × 128 scans (128 OCT B-scan sections/eye) and calculated drusen volume using ImageJ. To minimize variability, this process was done by a single investigator.

Statistical analysis was performed using R 2023.12.1 + 402 (https://www.r-project.org/). Quantitative variables were reported as mean and standard deviation and qualitative variables were reported as a percentage. The odds ratio was used to calculate the risk of AMD among ODD and non-ODD cohorts. A logistic regression model was used to account for possible confounding variables between the two cohorts. The Cap-Core scores in each group were tested for normality of distribution using the Shapiro–Wilk test, and the mean and standard deviation were reported. GraphPad Prism (Prism 10.2.2; GraphPad Software, Inc., La Jolla, CA) was then used to compare the means of the three groups using one-way ANOVA and post hoc Tukey test. p-values of less than 0.05 were considered statistically significant.

To assess whether ODD is associated with increased risk of AMD, we performed a multivariable logistic regression analysis adjusting for age and family history of AMD. Among 94 patients with ODD, 9 (9.6%) had AMD, compared to 3 out of 100 patients without ODD (3.0%) ( Table 1 ). In the adjusted model, ODD was associated with higher odds of AMD (AOR = 3.93, 95% CI: 0.89–21.85, p = 0.084), although this association did not reach statistical significance. Age and family history were both significant predictors of AMD (age: AOR = 1.12 per year, 95% CI: 1.06–1.22, p = 0.001; family history: AOR = 11.95, 95% CI: 2.44–68.64, p = 0.003) ( Supplementary Table 1 ). The ODD group was significantly younger than the non-ODD group (median age 44 vs. 60 years, p = 0.003) ( Table 1 ), and AMD cases were primarily observed in older patients (median age: 78, IQR: 73–81). Therefore, the small sample size and the younger age in the ODD group likely limited statistical power and contributed to the wide confidence interval and lack of statistical significance.

Analysis of the 10 patients with ODD-AMD showed that all 10 patients were Caucasians and older than 55 years old ( Table 2 ). Detailed clinical information about each participant is available online ( Supplementary Table 2 ). All patients had bilateral features of ODD. The AMD ranged from early dry AMD to late dry and wet AMD. Extramacular drusen were present in one patient. In addition, seven patients (70%) had a family history of AMD, six of whom were from first-degree relatives. The patients had an adult-onset presentation, with no syndromic features. History of smoking was reported in three patients only.

The severity of vision loss was diverse and occurred as central vision loss or peripheral field constriction. The visual acuities and field defects are described in Supplementary Table 2 . The visual acuity was worse than 20/40 in 35% of the eyes. The visual field defects (defined as mean deviation worse than −2 dB) were present in 10 eyes (55.6%).

We investigated the different ophthalmic imaging modalities to assess ODD-AMD. Typical features of ODD and AMD were identified bilaterally in the 10 patients. The macular drusen were mainly in association with non-exudative AMD. Four patients showed early AMD with small to medium-sized drusen. One of these also showed scattered peripheral extramacular retinal drusen ( Figure 2A ). Three patients had large (>125 μm) drusen ( Figure 2B ), and two developed geographic atrophy ( Figure 2C ). Signs of neovascularization were apparent in one patient only.

ODD was detectable in eight patients on the en face color fundus photographs (refractile bodies) or FAF imaging (hyper-autofluorescent signal). They were observed in the inferior sector, superonasal sector, and diffusely in different patients ( Figure 2 ). ODD was present in at least two quadrants of the optic disc in 70% of the eyes. EDI-OCT was necessary for the detection of ODD in two cases, which showed mild features.

Typical calcified macular drusen were noted in 40% of patients. Figure 2 shows an example of ectopic calcification of the optic disc and the macula. The elevated RPE with a hyporeflective core and a hyperreflective cap ( Figure 1A ) represents calcified macular drusen in patients with AMD. A similar pattern in ODD is shown with hyperreflective horizontal lines and signal-poor core ( Figure 1B ). A grayscale plot of the different patterns of drusen appearance on OCT imaging is demonstrated in Figure 3 . An increased grayscale signal followed by a decreased signal is observed in the calcified AMD and ODD images ( Figures 3A, B, E, F ). This pattern was not apparent in non-calcified AMD drusen ( Figures 3C, D ).

When quantifying the difference in reflectivity between the hyperreflective cap and the hyporeflective core, ODD showed the highest mean difference while non-calcified macular drusen had the least difference ( Figure 4A ). The measured means were 0.76 (SD = 0.14) in ODD, 0.5 (SD = 0.16) in the likely calcified macular drusen, and 0.16 (SD = 0.14) in the non-calcified macular drusen. A statistical significance was observed between all three groups (P< 0.001). To correlate the macular Cap-Core score with disease burden (based on the number of discrete macular drusen and total area of macular drusen), the drusen with the highest Cap-Core scores were plotted against the total drusen volume and drusen count in each eye ( Figure 4B ). This revealed that macular drusen with a higher Cap-Core score tend to be in eyes with higher drusen count and total drusen volume. Finally, the drusen size was not correlated with its corresponding Cap-Core score in ODD and macular drusen ( Supplementary Figure 5 ).