Optical Coherence Tomography Biomarkers as Potential Predictors for Visual Function and Response to Intravitreal Therapy in Diabetic Macular Edema

Abstract

Purpose: To synthesize a broad range of optical coherence tomography (OCT) biomarkers in diabetic macular edema (DME) and their associations with visual acuity (VA) and response to intravitreal antivascular endothelial growth factor (anti-VEGF) and corticosteroid therapy.Methods: A literature review was conducted.Results: Disorganization of the inner retinal layers (DRIL), presence of epiretinal membrane (ERM), ellipsoid zone (EZ) integrity, and external limiting membrane (ELM) continuity were the most commonly reported biomarkers. In most studies, absence of baseline DRIL/ERM, as well as preserved ELM/EZ continuity, was associated with a better final VA and improved response to intravitreal injections (IVI). Lower choroidal vascularity index was associated with a favorable response to anti-VEGF therapy and a poorer response to corticosteroids. Studies reporting hyperreflective dots showed mixed results. Degradation of the retinal pigment epithelium (RPE), increased central retinal thickness (CRT), and foveal eversion were associated with poor visual outcomes and poor response to both anti-VEGF and corticosteroid IVI. Conclusions: Although many studies reported similar trends in biomarker levels after anti-VEGF and corticosteroid IVI, certain biomarkers (DRIL, hyperreflective dots, and CRT) were found to be more responsive to global cytokine suppression using corticosteroids than anti-VEGF agents. The absence of DRIL and preserved EZ/ELM correlated with improved visual outcomes. However, data concerning the choroidal vascularity index, RPE continuity, CRT, and hyperreflective dots are relatively limited. Future prospective studies that focus on these biomarkers, as well as the ongoing advancements in OCT imaging and the integration of artificial intelligence, will help to individualize treatment strategies for patients with DME.

Keywords

diabetic macular edema optical coherence tomography biomarkers intravitreal injections

Introduction

Diabetic macular edema (DME) is characterized by the pathologic accumulation of extracellular fluid within the macula, arising secondary to disruption of the blood–retinal barrier and often involving the inner nuclear layer, outer plexiform layer, Henle fiber layer, and the subretinal space.¹ DME is recognized as one of the most common causes of visual impairment in patients with diabetes, with a global prevalence ranging from 4.2% to 14.3%. It is expected that nearly 25% of patients with type 2 diabetes will develop DME, posing a significant risk of losing at least 2 lines on the Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity (VA) chart if left untreated.^2,3 Moreover, nearly 69% of patients with insulin-dependent diabetes will have significant progression of their retinopathy, and 24% will progress to proliferative retinopathy in 10 years.⁴

The introduction of optical coherence tomography (OCT) revolutionized the assessment of DME, offering rapid and precise evaluation. Comprehensive multi-layer segmentation is now possible with recent technologic advancements in OCT; thus the accuracy of assessing individual retinal layers and identifying potential pathologies is enhanced.⁵ OCT biomarkers, commonly categorized into structural differences within the retina or the presence of fluid in the retinal and choroidal layers, represent visually evident morphologic variations in retinal anatomy. Common structural OCT biomarkers include disorganization of the retinal inner layers (DRIL), epiretinal membrane (ERM), continuity of the external limiting membrane (ELM) or ellipsoid zone (EZ), and thickness-based metrics such as central retinal thickness (CRT) and central subfield thickness (CST). Notably, CST is defined as the mean retinal thickness within the central 1-mm subfield centered on the fovea. Its standardized and reproducible assessment makes it one of the most frequently used OCT biomarkers for diagnosing DME and quantitatively monitoring disease progression in both clinical practice and randomized controlled trials.^6–12 Common fluid-based biomarkers include subretinal fluid (SRF), subretinal pigment epithelium (RPE) fluid, and intraretinal fluid.^13,14 Of note, assessment of the integrity of inner and outer segments, EZ, and ELM, along with the presence of SRF, has shown promise in predicting VA and treatment response in DME.^15,16

Although several OCT biomarkers have demonstrated value as prognosticators of VA, markers such as hyperreflective dots and the choroidal vascularity index remain underexplored with less available data.¹⁷ This review aimed to comprehensively summarize the existing knowledge regarding macular OCT biomarkers’ predictive capabilities for VA and response to intravitreal therapy (both corticosteroids and antivascular endothelial growth factor [anti-VEGF]) in DME. As such, studies reporting on the predictive value of OCT biomarkers for VA and response to intravitreal therapy in DME were reviewed using a narrative synthesis approach, with effect estimates extracted directly from individual studies, where available. These insights hold the potential to facilitate enhanced personalization of treatment and the development of targeted therapeutic approaches for individuals affected by DME.

Ovid MEDLINE, Embase, the Cochrane Library, and Web of Science were searched for relevant literature from their inception through August 2024. The general inclusion and exclusion criteria are detailed in Supplemental Table 1.

OCT Biomarkers

Epiretinal Membrane

Several studies have shown that the absence of ERM is not correlated with improvements in best-corrected visual acuity (BCVA) after treatment.^18,19 Nevertheless, Lai et al²⁰ demonstrated in a multivariate analysis that patients with ERM could still achieve substantial improvement in BCVA from baseline. However, the presence of ERM was associated with a worse final CRT after 1 year of a treat-and-extend regimen with ranibizumab, in which a mean of 7.67 ± 2.09 injections were administered.²⁰ Other studies have shown that both good and poor responders to anti-VEGF, with respect to functional and anatomic outcomes, present with similar frequencies of ERM, thus indicating its presence may not predict anti-VEGF response.²¹

Foveal Eversion

Foveal eversion represents a full convex appearance of the central fovea. Arrigo et al²² reported that patients with and without foveal eversion had similar baseline BCVA. However, only eyes without foveal eversion demonstrated a significant improvement in BCVA when treated with either anti-VEGF or corticosteroids, whereas eyes with foveal eversion had relatively static BCVA values until the final follow-up. Furthermore, everted eyes required significantly greater retreatment with corticosteroids.

Disorganization of Retinal Inner Layers

Multiple studies have found that the presence of DRIL at baseline confers reduced VA after treatment.^17,21,23 Mazloumi et al²³ found that DRIL was one of the only significant predictors of worse BCVA at 3 months after treatment. Dou et al¹⁷ found that DRIL presents the highest risk (odds ratio [OR], 6.71; P = .001) when predicting final BCVA compared with multiple other biomarkers, including SRF, choroidal vascularity index, and EZ discontinuity. In addition, the absence of DRIL was associated with a gain in BCVA of more than 5 letters. Zur et al¹⁵ demonstrated that DRIL significantly reduced BCVA gains at all time points after treatment (2, 4, 6, and 12 months). Moreover, Oliverio et al²⁴ found that the presence of DRIL at baseline was associated with significantly lower odds of achieving an improvement in BCVA of more than 10 letters at 12 months after the administration of a mean of 1.6 ± 0.5 intravitreal dexamethasone implants.

Koc et al²⁵ demonstrated that extensive DRIL (>500 µm) serves as a significant predictor of both functional and morphologic resistance to anti-VEGF therapy (3 monthly doses of aflibercept or ranibizumab). Similar findings were reported by Santos et al,²¹ who showed that EZ disruption (OR, 10.96; P < .001) poses a slightly greater risk of poor functional response to anti-VEGF treatment compared with DRIL (OR, 7.05; P = .034) after 3 monthly loading doses of ranibizumab and as-needed treatment up to month 12. Additionally, Oliverio et al²⁴ found that both the presence of EZ disruption and DRIL at baseline were independently associated with lower odds of achieving an improvement in BCVA of more than 10 letters at 12 months after the administration of a mean 1.6 ± 0.5 intravitreal dexamethasone implants.

Lopes et al²⁶ observed reduced DRIL (25% in the early switch group vs 90% in the late switch group) and greater BCVA gains when a long-term fluocinolone acetonide implant was administered to patients early in their disease process (before receiving a dexamethasone implant or 6 IVI of bevacizumab, aflibercept, ranibizumab, or triamcinolone acetonide). This further reinforces the deleterious effects of chronic DRIL and the benefit of an early switch to long-term corticosteroids.

Müller Cell Cone Integrity

Patients with good response to anti-VEGF at 1 month (defined as CST ≤250 µm) were more likely to have an intact Müller cell cone (P < .001).²⁷ Both univariate and multivariate analyses showed a significant association of integrity with BCVA. However, there was no significant association with BCVA at 12 months. Given very few follow-up patients in the present study, the temporal effect on BCVA in patients without Müller cell cone integrity requires further analysis.

Hyperreflective Dots and Hyperreflective Walls

Hyperreflective dots are any reflective foci within the retina, thought to represent inflammatory cells (ie, microglia) in DME.²⁸ Choovuthanayakorn et al²⁹ found that the baseline presence of hyperreflective dots on OCT is associated with a higher improvement in VA when treated with anti-VEGF compared with control patients (P = .010). Zur et al¹⁵ have shown that the absence of hyperreflective dots significantly predicts improved visual outcomes after treatment with dexamethasone. Similarly, Korkmaz et al³⁰ found that patients with hyperreflective dots had 2.21 more times visual recovery compared with those without baseline hyperreflective dots after 12 months of intravitreal therapy with a ranibizumab, aflibercept, or dexamethasone implant. Moreover, Narnaware et al³¹ demonstrated that the presence of hyperreflective dots predicted a more significant anatomic response to IVI dexamethasone in patients who had previously received anti-VEGF, with respect to reducing central macular thickness (CMT). Existing evidence suggests that hyperreflective dots may serve as a stronger predictor of response to corticosteroids; however, further research is warranted to directly compare the predictive value of hyperreflective dots in eyes treated with anti-VEGF agents vs corticosteroids.

Sardana et al³² described the presence of hyperreflective walls of foveal cystoid spaces in patients with DME. In their 1-year prospective study, the presence of hyperreflective walls at baseline was predictive of worse BCVA, albeit this was not statistically significant (P = .08).³² Rana et al³³ also found that the hyperreflectivity of cysts at baseline was associated with a poor anatomic response to 3 monthly injections of ranibizumab, defined as a reduction in CMT of less than 10%; however, the number of hyperreflective walls was not associated with this poor response.

Intraretinal Biomarkers: Cystoid Spaces and Fluid

Nagai et al³⁴ showed a significant correlation (P < .001) between both the number and area of intraretinal cystoid spaces and poorer BCVA at baseline and 12 months after anti-VEGF treatment. These findings were further supported by another study at the 24-month time point.³⁵ Pessoa et al³⁶ found that the absence of inner nuclear layer cysts at baseline was associated with a good early functional and anatomic response to fewer than 7 injections of ranibizumab. Limited data suggest that large or expanding intraretinal cysts may disrupt Müller cell and ELM or EZ integrity, thereby reducing VA gains. Consequently, patients presenting with relatively good BCVA at baseline but with large or multiple intraretinal cystoid spaces may benefit from prompt therapy due to the heightened risk of long-term ELM/EZ disruption.³⁴ Gerendas et al³⁷ revealed that patients with intraretinal cystoid fluid ≤380 µm receiving intravitreal ranibizumab experienced significantly better gains in VA at all time points compared with patients with intraretinal cystoid fluid >380 µm (mean letter gains 64.8 vs 61.7). Similarly, because the disparity in gains persisted consistently over time, it appears that prolonged treatment may not substantially benefit patients with increased intraretinal cystoid fluid.³⁷ Hence, early intervention may offer the best approach for these patients. Nonetheless, Vujosevic et al³⁸ noted that more significant reductions were seen in several inner retinal biomarkers, including intraretinal cysts and DRIL, after an intravitreal dexamethasone implant, compared with 3 monthly intravitreal ranibizumab injections. Moreover, Chang et al³⁹ reported that large outer nuclear layer cysts were associated with reductions in CRT of >50 µm from baseline, but not with improvements in BCVA. Additionally, Pessoa et al⁴⁰ found no significant differences in the presence of inner nuclear layer cysts and outer nuclear layer cysts at baseline, irrespective of achieving a functional response to a 0.19 mg fluocinolone acetonide at 12 months of follow-up.

Central Retinal Thickness, Central Subfield Thickness, Foveal/Parafoveal Inner/Outer Retinal Thicknesses

CRT is the mean thickness measured at the intersection of 6 radial scans on OCT, whereas CST is the thickness in a 1 mm diameter at the central retina.¹⁶ Some studies characterize foveal thickness as that directly within the 1 mm diameter from the fovea and parafoveal thicknesses from 1 to 6 mm diameter circles around the center. Manufacturer-supplied OCT software is able to stratify between inner (nerve fiber and outer plexiform layer) and outer (outer plexiform and outer nuclear layer) retinal thicknesses. Ceravolo et al⁴¹ showed that CRT negatively predicts the ability to gain more than 10 letters of final BCVA when patients are treated with either anti-VEGF or corticosteroids (P < .001). In contrast, Gerendas et al³⁷ showed that CRT is only weakly correlated with baseline BCVA (r = -0.344), which decreased over a follow-up period of 36 months (r = -0.259).

Of note, in a multivariate analysis of a treat-and-extend regimen of ranibizumab over 12 months, greater CRT predicted the need for more than 6 anti-VEGF injections, unlike ERM.²⁰ Given that CRT demonstrated significant reduction with both anti-VEGF and corticosteroid IVI, it does not seem to impede either treatment response.⁴¹ In keeping with this, Narnaware et al³¹ found that there is a significant decrease in CRT even with increasing amounts of anti-VEGF injections. Moreover, a lower CST (<405 µm) was associated with nearly 90% of patients gaining at least 1 ETDRS line by 12 months of therapy when treated with anti-VEGF (P < .001).⁴² Finally, while regression analysis showed that parafoveal/foveal outer retinal thickness were significant predictors of a greater increase in BCVA, inner retinal thicknesses were not, indicating that outer retinal damage may play a larger role in DME.¹⁹ Borrelli et al⁴³ found that parafoveal outer retinal thickness was associated with the change in BCVA from baseline, as well as the final BCVA after 5 years of follow-up with bevacizumab, ranibizumab, or aflibercept, administered using an as-needed regimen.

Ellipsoid Zone, External Limiting Membrane Continuity, and Disorganization of Retinal Outer Layers

Both the continuity of the EZ and ELM, which serve as biomarkers for photoreceptor integrity, exhibited similar outcomes. Specifically, disruption of the EZ and ELM was associated with poor BCVA both at baseline and at the 24-month follow-up and was linked to lower overall gains in BCVA.^{19,29,35,44,45} Particularly noteworthy, significantly worse long-term outcomes in VA were found in individuals experiencing ELM disruption concurrent with inner segment-outer segment (IS-OS) disruption after the initial IVI.³⁵ It is essential to acknowledge, however, that assessing both biomarkers can be challenging in patients with severe edema, thereby diminishing their reliability at baseline.

A multivariate analysis by Lai et al²⁰ demonstrated that patients with EZ disruption could still achieve considerable improvements in BCVA. Multivariate regression showed that EZ integrity at baseline was a significant predictor of a good response to dexamethasone implant as measured by improvement in BCVA (OR, 1.3; P < .001).⁴⁶ Moreover, Santos et al²¹ showed that good responders (≥10 ETDRS letter gain) to anti-VEGF had 2.3 and 5.4 times lower areas of EZ and ELM disruption at baseline, respectively, than poor responders (<5 letter gain). Contrary to the previously mentioned studies, Koc et al found that ELM continuity was the strongest predictor of functional and anatomic responses to anti-VEGF therapy in patients receiving 3 monthly injections of ranibizumab or aflibercept. In receiver operating characteristics analysis, Santos et al^21,25 showed that EZ disruption is the stronger of the 2 (EZ area under the curve = 0.71, ELM area under the curve = 0.64). Another study found that DRIL was equally significant to the extent of ELM disruption as a predictor for worse BCVA.²³ Conversely, Pessoa et al⁴⁰ found no significant differences with respect to the presence of baseline EZ and ELM disruption between functional responders and nonresponders to a 0.19 mg fluocinolone acetonide at 12 months of follow-up.

Sardana et al³² measured the disruption of retinal outer layers because the horizontal extent for which any boundaries between the ELM, EZ, and interdigitation zone could not be identified in the 1 mm foveal area. The extent of baseline disruption of retinal outer layers was weakly correlated with logMAR BCVA 12 weeks after treatment with ranibizumab, although this was not statistically significant.³²

Cone Outer Segment Tip and Photoreceptor Outer Segment Length

Szeto et al⁴⁵ showed, in both univariable and multivariable regression, that disruption of the cone outer segment tip at baseline is associated with significantly poorer baseline BCVA (P < .001) and at 6, 12, and 24 months (P < .001) after initial anti-VEGF therapy with ranibizumab or aflibercept. The presence of a cone outer segment tip was also correlated with ELM/EZ disruption and DRIL, albeit these associations were not robust. Interestingly, visibility of a cone outer segment tip was significantly associated with an improvement in VA at 6 and 12 months only with anti-VEGF therapy. Visibility of a cone outer segment tip was also not significantly associated with a specific gain of 5 or more ETDRS letters at 12 or 24 months.⁴⁵ Although more data are required, cone outer segment tip integrity seems to predict poor baseline BCVA but cannot predict significant gains in BCVA. Additionally, in patients with DME refractory to bevacizumab, Salimi et al⁴⁷ found an association between improvements in VA and decreases in disruption of the cone outer segment tip after a 1-year course of treatment with aflibercept. Another study explored the predictive value of photoreceptor outer segment length, measured from the inner surface of the IS-OS band to the inner surface of the RPE.³² Sardana et al³² found that baseline photoreceptor outer segment length was weakly correlated with BCVA at 12 weeks after ranibizumab treatment, but this was not statistically significant.

Retinal Pigment Epithelium Integrity

Borrelli et al⁴³ showed that RPE discontinuity is significantly associated with poorer vision at baseline (P = .019). Regression analysis found that baseline RPE integrity was one of the biomarkers with the strongest association with overall change in BCVA and greater final BCVA (β coefficient = 0.370, P = .30; β coefficient = 0.440, P < .001). Similarly, Wang et al⁴⁸ showed that, compared with eyes with gains in BCVA, eyes without gains demonstrate EZ-RPE disruption (P = .037). Although further studies are required, trends in RPE integrity similar to EZ/ELM integrity are seen and may complement these biomarkers. Furthermore, the appearance of the RPE was strongly linked to changes in BCVA from baseline, as well as the BCVA after 5 years of follow-up with bevacizumab, ranibizumab, or aflibercept, administered using an as-needed regimen.⁴³

Subretinal Fluid and Serous Retinal Detachment

In DME, SRF is seen as the accumulation of exudate containing lipoprotein and serous fluid between the neurosensory retina and the RPE.⁴⁹ Gerendas et al³⁷ observed that changes in VA were negatively influenced by baseline SRF in patients receiving a combination treatment of laser and 6.8 ± 2.95 anti-VEGF injections, albeit this was not significant (P = .1038). However, in patients treated only with ranibizumab, gains in BCVA were associated with SRF resolution after a mean of 7.0 ± 2.81 injections (P = .0563). In contrast, Zur et al¹⁵ demonstrated that reduced SRF is associated with improved BCVA after dexamethasone IVI (P = .01). Dou et al¹⁷ noted that patients with SRF were more likely to achieve a gain of more than 5 letters of VA when treated with anti-VEGF (P = .008), whereas Pessoa et al⁴⁰ showed a gain of more than 15 letters (P = .036) when treated with corticosteroids. Similarly, Schwarzer et al showed that presence of SRF is associated with better gains in BCVA with multiple anti-VEGF agents (P = .003). Despite these significant findings, considerable variability was found in treatment response associated with SRF across the literature.⁵⁰ The presence of persistent SRF has been previously associated with RPE and photoreceptor damage, which may confer poor visual prognosis.⁵¹

Rana et al³³ found no association between the presence of subfoveal serous retinal detachment at baseline and anatomic responses to ranibizumab, as all cases showed resolution after a single injection. Korkmaz et al³⁰ also reported that patients with serous macular detachment achieved more visual gains with ranibizumab (at months 1 and 4) and aflibercept (at months 1, 4, and 6 months) compared with dexamethasone implant.

Koc et al²⁵ noted that SRF did not predict response to anti-VEGF treatment with ranibizumab or aflibercept, wherein the response criteria included achieving a CST of ≤300 µm and an increase of 3 or more lines in VA. Moreover, Chang et al³⁹ found that baseline SRF was associated with reductions in CRT of >50 µm after treat-and-extend ranibizumab therapy but not with improvements in BCVA. Another study found that baseline SRF was correlated with CRT improvement of more than 100 µm over 12 months of follow-up in patients with DME receiving intravitreal dexamethasone.⁵²

Notably, Pessoa et al⁴⁰ demonstrated that the need for additional corticosteroid therapy is influenced by the presence of baseline SRF (P = .005), given this biomarker is important for assessing anatomic resolution and may suggest more chronic and severe DME. These findings may suggest that the presence of SRF indicates more severe disease with greater leakage and may warrant the use of corticosteroid therapy if the response to anti-VEGF treatment is suboptimal.

Choroidal Vascularity Index

Dou et al¹⁷ found that patients with gains of more than 5 ETDRS letters have a significantly higher baseline choroidal vascularity index compared with those with fewer gains (0.69 vs 0.66). Compared with other choroidal markers, such as subfoveal choroidal thickness, choroidal vascularity index demonstrated less variability due to factors such as age and intraocular pressure, thereby increasing reliability. Pessoa et al⁵³ also reported that higher baseline choroidal vascularity index was the only choroidal marker significantly correlated with greater gains in BCVA at 6 months (r = 0.450, P = .041).

In a multivariate analysis, higher choroidal vascularity index significantly predicted gains in BCVA when treatment involved anti-VEGF (OR, 0.17; P = .006).¹⁷ Interestingly, Arrigo et al⁵⁴ reported that lower values correlated with improved response to corticosteroid therapy. However, good responders were categorized as those with worse BCVA at baseline with consecutive gains, in contrast to poor responders, whose BCVA remained static. Although there are limited studies and further investigations are warranted, choroidal vascularity index is a promising biomarker compared with choroidal thickness and subfoveal choroidal thickness because physiologic factors make it less variable.

Choroidal Thickness and Subfoveal Choroidal Thickness

Generally, choroidal thickness is the mean measurement of the distance between Bruch membrane and the choroidal-scleral junction at 3 points: the fovea, 1500 µm nasally, and 1500 µm temporally.⁵⁵ Subfoveal choroidal thickness is the specific thickness at the fovea only. In a linear regression analysis, Liu et al⁵⁶ showed that greater subfoveal choroidal thickness significantly predicts a better final BCVA. Moon et al⁵⁷ found that short-term reductions also accompanied functional treatment response to dexamethasone implant. Additionally, Rayess et al⁵⁸ found that eyes with a thicker baseline subfoveal choroidal thickness had better short-term functional and anatomic responses to 3 monthly intravitreal injections of ranibizumab or bevacizumab. However, Savur et al⁵⁹ reported that eyes with a thinner baseline subfoveal choroidal thickness (ie, ≤220 µm) achieved better improvements in BCVA over 2 years of follow-up and required fewer injections of aflibercept and ranibizumab compared with eyes with greater subfoveal choroidal thickness (ie, >220 to ≤270 µm and >270 µm). In contrast, a separate multivariate analysis showed that choroidal thickness does not significantly predict final BCVA.⁴¹ Likewise, Vujosevic et al³⁸ observed that choroidal thickness remains largely unchanged after either anti-VEGF or corticosteroid therapy. This finding was similarly reported by Ceravolo et al.⁴¹

Conclusions

Although numerous OCT biomarkers have demonstrated their value as prognosticators of VA and treatment response, others remain underexplored. Hyperreflective dots and choroidal vascularity index, for instance, show promise but lack sufficient data for comprehensive analysis. Notably, the choroidal vascularity index exhibits potential due to its lower physiologic variability compared with other choroidal markers. Conversely, ERM is generally not associated with gains in VA, and the variability in visual outcome data on SRF underscores the complexity of its role in DME management. Furthermore, both SRF and foveal eversion have shown an increased need for repeat injections when treated with corticosteroids. CRT generally has not affected functional responses to anti-VEGF agents, and only outer retinal thickness significantly predicted VA outcomes. Additionally, foveal eversion and Müller cell cone integrity are relatively new and understudied areas in DME research.

Our review aims to bridge this gap by summarizing existing knowledge regarding OCT biomarkers’ predictive capabilities for VA and response to intravitreal therapy in DME. Among the biomarkers evaluated, several stand out for their predictive value. SRF has shown varied associations with treatment response, highlighting its multifaceted role in DME pathophysiology. Conversely, DRIL has emerged as a robust predictor of visual outcomes, with its absence correlating with better BCVA gains and its presence associated with increased resistance to VA gains with anti-VEGF therapy. EZ and ELM continuity serve as reliable biomarkers for photoreceptor integrity, with disruption of either linked to worse BCVA outcomes. Interestingly, the absence of ERM does not significantly impact gains in BCVA after treatment, suggesting its limited prognostic value compared with other biomarkers.

Although our review highlights the predictive power of certain OCT biomarkers, many challenges remain. This review was conducted using a narrative synthesis approach. As such, no quantitative pooling of data was performed, precluding the formal assessment of statistical heterogeneity through methods such as the I² statistic. Nevertheless, we acknowledge the potential for heterogeneity in the reported findings, as the studies reviewed varied in design, population characteristics, and definitions of specific OCT biomarkers. Moreover, it is important to emphasize that the narrative design of our review did not involve quantitative synthesis of data or formal comparison of effect sizes across studies. Variability in treatment response and the need for further exploration of less studied biomarkers underscore the complexity of DME management. Nevertheless, our findings emphasize the importance of several, well-established OCT biomarkers in guiding clinical decision-making and advancing personalized care for patients with DME.

Further research is needed to determine whether the prognostic value of OCT biomarkers is consistent across studies in which OCT scans are manually annotated by independent graders, evaluated at a reading center, or interpreted using artificial intelligence algorithms. Deep learning models have demonstrated strong potential for automated segmentation and quantitative fluid-volume analysis in DME on OCT,⁶⁰ showing high concordance with expert manual assessment and generating measurements that correlate with VA.^61,62 Given the expanding literature on the use of large language models for interpreting clinical retinal images,^63–66 future research should evaluate the ability of these algorithms to identify and quantify OCT biomarkers.

Supplemental Material

sj-docx-1-vrd-10.1177_24741264261457974 – Supplemental material for Optical Coherence Tomography Biomarkers as Potential Predictors for Visual Function and Response to Intravitreal Therapy in Diabetic Macular Edema

Supplemental material, sj-docx-1-vrd-10.1177_24741264261457974 for Optical Coherence Tomography Biomarkers as Potential Predictors for Visual Function and Response to Intravitreal Therapy in Diabetic Macular Edema by Andrew Mihalache, Bhadra U. Pandya, Aaditeya Jhaveri, Austin Pereira, Safwan Tayeb, Bernard Hurley, Amin Kherani, Yuyi You and Peng Yan in Journal of VitreoRetinal Diseases

Footnotes

ORCID iDs

Andrew Mihalache

Peng Yan

Ethical Considerations

Not required.

Consent to Participate

Not required.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data availability requests can be made to the corresponding author.

Supplemental Material

Supplemental material is available online with this article.

References

Zhang

Zeng

Bao

Wang

Gillies

MC.

Diabetic macular edema: new concepts in patho-physiology and treatment. Cell Biosci. 2014;4:27.

Zhang

, et al. Diabetic macular edema: current understanding, molecular mechanisms and therapeutic implications. Cells. 2022;11:3362.

Bandello

Cunha-Vaz

Chong

, et al. New approaches for the treatment of diabetic macular oedema: recommendations by an expert panel. Eye (Lond). 2012;26:485-493.

Klein

BEK

Moss

Cruickshanks

KJ.

The Wisconsin epidemiologic study of diabetic retinopathy: XIV. ten-year incidence and progression of diabetic retinopathy. Arch Ophthalmol. 1994;112:1217-1228.

Baskin

DE.

Optical coherence tomography in diabetic macular edema. Curr Opin Ophthalmol. 2010;21:172-177.

You

Tsuboi

Guo

, et al. Comparison of central macular fluid volume with central subfield thickness in patients with diabetic macular edema using optical coherence tomography angiography. JAMA Ophthalmol. 2021;139:734-741.

Browning

Glassman

Aiello

, et al. Optical coherence tomography measurements and analysis methods in optical coherence tomography studies of diabetic macular edema. Ophthalmology. 2008;115:1366-1371, 1371.e1.

Schmidt-Erfurth

Garcia-Arumi

Bandello

, et al. Guidelines for the management of diabetic macular edema by the european society of retina specialists (EURETINA). Ophthalmologica. 2017;237:185-222.

Wells

Glassman

Ayala

, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema: two-year results from a comparative effectiveness randomized clinical trial. Ophthalmology. 2016;123:1351-1359.

10.

Elman

Bressler

Qin

, et al. Expanded 2-year follow-up of ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema. Ophthalmology. 2011;118:609-614.

11.

Wykoff

Abreu

Adamis

, et al. Efficacy, durability, and safety of intravitreal faricimab with extended dosing up to every 16 weeks in patients with diabetic macular oedema (YOSEMITE and RHINE): two randomised, double-masked, phase 3 trials. Lancet. 2022;399:741-755.

12.

Wong

Haskova

Asik

, et al. Faricimab treat-and-extend for diabetic macular edema: two-year results from the randomized phase 3 YOSEMITE and RHINE trials. Ophthalmology. 2024;131:708-723.

13.

Markan

Agarwal

Arora

, et al. Novel imaging biomarkers in diabetic retinopathy and diabetic macular edema. Ther Adv Ophthalmol. 2020;12:2515841420950513.

14.

Kwan

Fawzi

AA.

Imaging and biomarkers in diabetic macular edema and diabetic retinopathy. Curr Diab Rep. 2019;19:95.

15.

Zur

Iglicki

Busch

, et al. OCT biomarkers as functional outcome predictors in diabetic macular edema treated with dexamethasone implant. Ophthalmology. 2018;125:267-275.

16.

Bressler

Odia

Maguire

, et al. Factors associated with visual acuity and central subfield thickness changes when treating diabetic macular edema with anti–vascular endothelial growth factor therapy: an exploratory analysis of the protocol T randomized clinical trial. JAMA Ophthalmology. 2019;137:382-390.

17.

Dou

Tsui

C-K

, et al. Choroidal vascularity index as a biomarker for visual response to antivascular endothelial growth factor treatment in diabetic macular edema. Journal of Diabetes Research. 2021;2021:e3033219.

18.

Chou

H-D

C-H

Chiang

W-Y

, et al. Optical coherence tomography and imaging biomarkers as outcome predictors in diabetic macular edema treated with dexamethasone implant. Sci Rep. 2022;12:3872.

19.

Borrelli

Zuccaro

Zucchiatti

, et al. Optical coherence tomography parameters as predictors of treatment response to eplerenone in central serous chorioretinopathy. J Clin Med. 2019;8:1271.

20.

Lai

C-T

Hsieh

Y-T

Lin

C-J

, et al. Age, initial central retinal thickness, and OCT biomarkers have an influence on the outcome of diabetic macular edema treated with ranibizumab–tri-center 12-month treat-and-extend study. Front Med (Lausanne). 2021;8:668107.

21.

Santos

Costa

MÂ

Schwartz

, et al. Optical coherence tomography baseline predictors for initial best-corrected visual acuity response to intravitreal anti-vascular endothelial growth factor treatment in eyes with diabetic macular edema: the CHARTRES study. Retina. 2018;38:1110.

22.

Arrigo

Aragona

Capone

, et al. Foveal eversion: a possible biomarker of persistent diabetic macular edema. Ophthalmol Ther. 2021;10:115–126.

23.

Mazloumi

Entezari

Samadikhadem

, et al. Spectral domain optical coherence tomography biomarkers of retinal hyperpermeability and choroidal inflammation as predictors of short-term functional and anatomical outcomes in eyes with diabetic macular edema treated with intravitreal bevacizumab. Retina. 2022;42:760.

24.

Oliverio

Meduri

Brancati

, et al. Clinical and optical coherence tomography biomarkers as prognostic factors in dexamethasone intravitreal implant for diabetic macular edema. Eur J Ophthalmol. 2024;34(6):1810-1818.

25.

Koc

Güven

Egrilmez

Aydın

Optical coherence tomography biomarkers in bilateral diabetic macular edema patients with asymmetric anti-VEGF response. Semin Ophthalmol. 2021;36:444-451.

26.

Lopes

Brizido

Aerts

, et al. Prognostic biomarkers of chronic diabetic macular edema treated with a fluocinolone acetonide intravitreal implant. Med Hypothesis Discov Innov Ophthalmol. 2021;10:50-58.

27.

Choi

Yun

J-H

Kim

S-W.

Foveal müller cell cone as a prognostic optical coherence tomography biomarker for initial response to antivascular endothelial growth factor treatment in cystoid diabetic macular edema. Retina. 2022;42:129.

28.

Qin

Zhang

Qin

, et al. Hyperreflective foci and subretinal fluid are potential imaging biomarkers to evaluate anti-VEGF effect in diabetic macular edema. Front Physiol. 2021;12:791442.

29.

Choovuthayakorn

Tantraworasin

Phinyo

, et al. Factors associated with 1-year visual response following intravitreal bevacizumab treatment for diabetic macular edema: a retrospective single center study. Int J Retina Vitreous. 2021;7:17.

30.

Korkmaz

Karti

Zengin

, et al. Real-life outcomes of intravitreal ranibizumab, aflibercept, and dexamethasone implant administrations in patients with treatment-naïve diabetic macular edema. Saudi J Ophthalmol. 2022;36:327-334.

31.

Narnaware

Bawankule

Raje

Short-term outcomes of intravitreal dexamethasone in relation to biomarkers in diabetic macular edema. Eur J Ophthalmol. 2021;31:1185-1191.

32.

Sardana

Singh

VK.

Optical coherence tomography biomarkers DROL, PROS, SND, hyperreflective walls of foveal cystoid spaces as predictors of central macular thickness and visual acuity in diabetic macular edema treated with intravitreal ranibizumab. Indian J Ophthalmol. 2024;72:722-727.

33.

Rana

Agarwal

Arora

, et al. Predicting visual outcomes following anti-vascular endothelial growth factor treatment for diabetic macular edema. Indian J Ophthalmol. 2024;72:S16-S21.

34.

Nagai

Suzuki

Uchida

, et al. The area and number of intraretinal cystoid spaces predict the visual outcome after ranibizumab monotherapy in diabetic macular edema. J Clin Med. 2020;9:1391.

35.

Giannakaki-Zimmermann

Behrndt

Hoffmann

, et al. Predictors for 2-year functional and morphological outcomes of a treat-and-extend regimen with ranibizumab in patients with diabetic macular edema. Ophthalmic Res. 2021;64:465-475.

36.

Pessoa

Leite

Ferreira

, et al. Oct biomarkers for early prognosis in diabetic macular edema treatment with ranibizumab. Eur J Ophthalmol. 2024;34:1141-1148.

37.

Gerendas

Prager

Deak

, et al. Predictive imaging biomarkers relevant for functional and anatomical outcomes during ranibizumab therapy of diabetic macular oedema. Br J Ophthalmol. 2018;102:195-203.

38.

Vujosevic

Toma

Villani

, et al. Diabetic macular edema with neuroretinal detachment: OCT and OCT-angiography biomarkers of treatment response to anti-VEGF and steroids. Acta Diabetol. 2020;57:287-296.

39.

Chang

Y-C

Huang

Y-T

Hsu

, et al. Optical coherence tomography biomarkers in predicting treatment outcomes of diabetic macular edema after ranibizumab injections. Medicina (Kaunas). 2023;59:629.

40.

Pessoa

Ferreira

Leite

, et al. Optical coherence tomography biomarkers: vitreous status influence in outcomes for diabetic macular edema therapy with 0.19-mg fluocinolone acetonide implant. Ophthalmic Res. 2021;64:639-647.

41.

Ceravolo

Oliverio

Alibrandi

, et al. The application of structural retinal biomarkers to evaluate the effect of intravitreal ranibizumab and dexamethasone intravitreal implant on treatment of diabetic macular edema. Diagnostics (Basel). 2020;10:413.

42.

Chatziralli

Sergentanis

Sivaprasad

Hyperreflective foci as an independent visual outcome predictor in macular edema due to retinal vascular diseases treated with intravitreal dexamethasone or ranibizumab. Retina. 2016;36:2319.

43.

Borrelli

Grosso

Barresi

, et al. Long-term visual outcomes and morphologic biomarkers of vision loss in eyes with diabetic macular edema treated with anti-VEGF therapy. Am J Ophthalmol. 2022;235:80-89.

44.

Panozzo

Cicinelli

Dalla Mura

, et al. Enhancing diabetic macular edema treatment outcomes: exploring the ESASO classification and structural OCT biomarkers. Ophthalmol Ther. 2024;13:1383-1398.

45.

Szeto

Hui

VWK

Tang

, et al. OCT-based biomarkers for predicting treatment response in eyes with centre-involved diabetic macular oedema treated with anti-VEGF injections: a real-life retina clinic-based study. Br J Ophthalmol. 2023;107:525-533.

46.

Meduri

Oliverio

Trombetta

, et al. Optical coherence tomography predictors of favorable functional response in naïve diabetic macular edema eyes treated with dexamethasone implants as a first-line agent. J Ophthalmol. 2021;2021:e6639418.

47.

Salimi

Vila

Modabber

Kapusta

One-year outcomes of aflibercept for refractory diabetic macular edema in bevacizumab nonresponders. Indian J Ophthalmol. 2021;69:360-367.

48.

Wang

Srivastava

Vasanji

, et al. Quantitative ultra-widefield fluorescein angiography and volumetric optical coherence tomography analysis in the REACT study: a prospective randomized comparative dosage trial evaluating ranibizumab in bevacizumab-resistant diabetic macular edema. Invest Ophthalmol Vis Sci. 2017;58:936.

49.

Marques

Silva

Subretinal fluid. In: Schmidt-Erfurth

Kohnen

, eds. Encyclopedia of Ophthalmology. Springer; 2016:1-3.

50.

Schwarzer

Ebneter

Munk

, et al. One-year results of using a treat-and-extend regimen without a loading phase with anti-VEGF agents in patients with treatment-naive diabetic macular edema. Ophthalmologica. 2019;241:220-225.

51.

Tsai

M-J

Hsieh

Y-T

Shen

Peng

Y-J.

Systemic associations with residual subretinal fluid after ranibizumab in diabetic macular edema. J Ophthalmol. 2017;2017:4834201.

52.

Huang

Y-T

Chang

Y-C

Meng

P-P

, et al. Optical coherence tomography biomarkers in predicting treatment outcomes of diabetic macular edema after dexamethasone implants. Front Med (Lausanne). 2022;9:852022.

53.

Pessoa

Marques

Leite

, et al. Choroidal blood flow after intravitreal ranibizumab in vitrectomized and non-vitrectomized eyes with diabetic macular edema. Clin Ophthalmol. 2021;15:4081-4090.

54.

Arrigo

Capone

Lattanzio

, et al. Optical coherence tomography biomarkers of inflammation in diabetic macular edema treated by fluocinolone acetonide intravitreal drug-delivery system implant. Ophthalmol Ther. 2020;9:971-980.

55.

Al Marshood

Rubio Caso

AlSaedi

, et al. Optical coherence tomography based choroidal thickness and its determinants in healthy Saudi population: a cross-sectional study. Cureus. 2023;15:e34152.

56.

Liu

Hou

Cong

Choroidal structural changes of posterior subtenon triamcinolone acetonide injection in eyes with refractory diabetic macular edema. J Ophthalmol. 2022;2022:e6882607.

57.

Moon

Choi

Song

JH.

Changes in subfoveal choroidal thickness after intravitreal dexamethasone implant therapy for diabetic macular edema. Retina. 2021;41:1283-1292.

58.

Rayess

Rahimy

Ying

G-S

, et al. Baseline choroidal thickness as a predictor for response to anti-vascular endothelial growth factor therapy in diabetic macular edema. Am J Ophthalmol. 2015;159:85-91.e1-3.

59.

Savur

Kaldırım

Atalay

, et al. Treatment results of diabetic macular edema with different choroidal thickness with intravitreal antivascular endothelial growth factor. BMC Ophthalmol. 2022;22:508.

60.

Schlegl

Waldstein

Bogunovic

, et al. Fully automated detection and quantification of macular fluid in OCT using deep learning. Ophthalmology. 2018;125:549-558.

61.

Tsuboi

You

Guo

, et al. Association between fluid volume in inner nuclear layer and visual acuity in diabetic macular edema. Am J Ophthalmol. 2022;237:164-172.

62.

Roberts

Vogl

W-D

Gerendas

, et al. Quantification of fluid resolution and visual acuity gain in patients with diabetic macular edema using deep learning: a post hoc analysis of a randomized clinical trial. JAMA Ophthalmol. 2020;138:945-953.

63.

Mihalache

Huang

Popovic

, et al. Accuracy of an artificial intelligence chatbot’s interpretation of clinical ophthalmic images. JAMA Ophthalmol. 2024;142:321-326.

64.

Mihalache

Huang

Mikhail

, et al. Interpretation of clinical retinal images using an artificial intelligence chatbot. Ophthalmol Sci. 2024;4:100556.

65.

Aftab

Khan

VanderBeek

, et al. Evaluation of ChatGPT-4 in detecting referable diabetic retinopathy using single fundus images. AJO Int. 2025;2:100111.

66.

Mihalache

Huang

Popovic

, et al. Fundus photograph interpretation of common retinal disorders by artificial intelligence chatbots. AJO Int. 2025;2:100154.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.02 MB