Introduction

Diabetic retinal disease, which includes diabetic retinopathy (DR) and diabetic macular edema (DME), remains a leading cause of preventable vision loss among working-age adults worldwide. The advent of dual therapy—combining anti-vascular endothelial growth factor (anti-VEGF) agents with corticosteroids—has provided a powerful option for patients who respond inadequately to monotherapy. Accurately monitoring the efficacy of this combination strategy demands imaging techniques that capture structural, vascular, and inflammatory changes with high sensitivity and specificity. Over the past decade, advances in retinal imaging have profoundly transformed clinical decision-making, enabling clinicians to tailor therapy, reduce treatment burden, and preserve vision. This article reviews the latest imaging modalities and how they are applied to track dual therapy response in diabetic retinal disease.

Key Imaging Modalities for Monitoring Therapy Response

Spectral-Domain and Swept-Source Optical Coherence Tomography

Optical coherence tomography (OCT) remains the cornerstone of retinal imaging in diabetic eye disease. Spectral-domain OCT (SD-OCT) and swept-source OCT (SS-OCT) provide axial resolution in the range of 5–7 µm, allowing precise measurement of retinal thickness and identification of intraretinal and subretinal fluid. In the context of dual therapy, high-resolution OCT enables clinicians to quantify fluid resolution after anti-VEGF-corticosteroid injections and detect early signs of disease reactivation. SS-OCT, with its longer wavelength (1050 nm) and faster scan speed, offers deeper penetration through the choroid and better visualization of vitreoretinal interface abnormalities. These improvements are critical for assessing the chronic effects of corticosteroids, which can alter vitreous structure and choroidal thickness over time.

Recent studies have consistently shown that changes in central subfield thickness (CST) and cube volume correlate strongly with visual acuity outcomes in patients receiving dual therapy. Automated segmentation algorithms now provide layer-by-layer analysis, identifying biomarkers such as hyperreflective retinal foci (indicative of lipid exudation or inflammatory cells) and disorganization of retinal inner layers (DRIL). The presence and resolution of hyperreflective foci have been associated with inflammatory-driven DME, a subtype that may respond preferentially to the corticosteroid component of dual therapy. Moreover, advanced OCT platforms allow for three-dimensional volumetric analysis, giving clinicians a comprehensive view of disease activity across the macula rather than relying solely on a single cross-sectional line.

Optical Coherence Tomography Angiography

OCT angiography (OCTA) is a non-invasive, dye-free technique that generates depth-resolved maps of retinal and choroidal microvasculature. Unlike fluorescein angiography, OCTA can separately visualize the superficial and deep capillary plexuses, the intermediate capillary plexus, and the choriocapillaris. This level of detail is essential for monitoring dual therapy response because anti-VEGF agents primarily target vascular leakage and neovascularization, while corticosteroids reduce vascular hyperpermeability and modulate inflammatory cytokines. OCTA has become indispensable for evaluating the microvascular effects of combination treatment.

Key metrics derived from OCTA include vessel density, perfusion density, and the area of the foveal avascular zone (FAZ). Enlargement of the FAZ is a hallmark of capillary dropout in diabetic retinal disease, and serial OCTA can track whether dual therapy halts or reverses this progression. Studies have demonstrated that corticosteroid therapy is associated with a greater reduction in macular capillary non-perfusion compared with anti-VEGF alone, suggesting a synergistic effect on the microcirculation. OCTA also enables identification of choroidal neovascularization (CNV) in eyes with DME, which may require a modified dual therapy approach. With wider field-of-view OCTA (12×12 mm and beyond), clinicians can now detect peripheral neovascularization and monitor regression after treatment, providing a more complete assessment of disease burden.

Fundus Autofluorescence

Fundus autofluorescence (FAF) imaging captures the natural fluorescence of lipofuscin in the retinal pigment epithelium (RPE), providing a metabolic map of retinal health. In diabetic retinal disease, FAF patterns can indicate RPE stress, subretinal fluid, and the presence of hard exudates. For patients undergoing dual therapy, FAF may reveal hyperautofluorescent or hypoautofluorescent areas that correlate with ongoing inflammation or photoreceptor damage. The addition of a corticosteroid to anti-VEGF therapy often leads to a reduction in RPE hyperautofluorescence, reflecting decreased metabolic overload. Serial FAF imaging can also help differentiate between transient macular edema and irreversible photoreceptor injury, guiding the decision to continue or adjust treatment. When combined with SD-OCT, FAF adds a metabolic dimension that enhances the understanding of tissue health beyond simple thickness measurements.

Multimodal Imaging Approaches

No single imaging modality provides a complete picture of treatment response. Multimodal imaging—combining SD-OCT, OCTA, FAF, and sometimes fluorescein angiography—allows clinicians to cross-reference structural, vascular, and metabolic data. For example, a patient may show complete resolution of fluid on OCT but have persistent capillary non-perfusion on OCTA, indicating ongoing risk of vision loss despite apparent anatomical improvement. Conversely, a patient with stable OCT findings but new hyperreflective foci on FAF may warrant a corticosteroid boost. Incorporating multimodal protocols into routine practice has been shown to improve the detection of disease progression by up to 30% compared with OCT alone. Several clinical trials now mandate multimodal imaging at each follow-up visit to capture the full spectrum of dual therapy effects, and this approach is increasingly adopted in academic and high-volume retina practices.

Biomarkers of Treatment Response

The proliferation of advanced imaging has led to the identification of specific biomarkers that predict and track dual therapy response. Among the most robust are:

  • Disorganization of retinal inner layers (DRIL): Present in intermediate and advanced stages of DME, DRIL indicates damage to bipolar and Müller cells. Resolution of DRIL after dual therapy is associated with better visual outcomes, whereas persistent DRIL suggests a need for alternative treatment strategies.
  • Hyperreflective retinal foci (HRF): These small, discrete spots on OCT are thought to represent activated microglia or lipoprotein extravasation. A rapid decrease in HRF count after corticosteroid injection aligns with the anti-inflammatory mechanism of dual therapy.
  • Subretinal fluid (SRF): The presence of SRF in DME is less common than intraretinal fluid but often signals a more inflammatory phenotype. Dual therapy leads to faster SRF resolution compared with anti-VEGF alone.
  • Choroidal thickness: Swept-source OCT enables accurate choroidal thickness measurements. A thinning of the subfoveal choroid after treatment may indicate a reduction in choroidal vascular hyperpermeability, a target of corticosteroid therapy.
  • Peripapillary vessel density: OCTA of the optic nerve head region can reveal microvascular dropout that correlates with diabetic neuropathy. Improvement in peripapillary perfusion has been reported after sustained-release corticosteroid implants.
  • Foveal avascular zone (FAZ) area: Enlargement of the FAZ on OCTA is a sign of ischemia. Serial measurements can show whether dual therapy stabilizes or reduces the rate of capillary dropout.

Clinicians must integrate these biomarkers with functional measures such as best-corrected visual acuity and microperimetry to avoid over-relying on imaging alone. However, the objectivity and reproducibility of imaging biomarkers make them invaluable for serial monitoring and for guiding treatment adjustments in real time.

Timing and Frequency of Imaging in Dual Therapy

Optimal timing of imaging is crucial for accurately capturing treatment response. In patients receiving dual therapy with a corticosteroid implant (e.g., dexamethasone or fluocinolone acetonide) combined with monthly anti-VEGF injections, imaging is typically performed at each visit—usually every 4 to 8 weeks. The peak effect of corticosteroids occurs 2–4 weeks after injection, so OCT and OCTA at this time point can identify maximal fluid reduction. Late follow-up (12–16 weeks) may reveal the waning effect, guiding the interval for re-treatment. For eyes on fixed-interval dual therapy (e.g., bimonthly ranibizumab plus an implant), imaging at trough and peak drug levels helps differentiate between insufficient dosing and drug refractory disease. Standardized imaging protocols (e.g., consistent scan patterns, illumination, and segmentation algorithms) are essential to minimize variability and ensure that changes are true biological responses rather than artifacts. Many retina specialists now adopt a treat-and-extend approach using imaging biomarkers to personalize the interval between injections, a strategy that can reduce the number of visits while maintaining disease control.

Artificial Intelligence and Automated Analysis

The increasing volume of imaging data in diabetic retinal disease has spurred the development of artificial intelligence (AI) algorithms for automated analysis. Deep learning models can now segment retinal layers, detect fluid, quantify vessel density, and predict treatment outcomes with accuracy comparable to human experts. In the context of dual therapy monitoring, AI offers several advantages:

  • Quantitative trend analysis: AI tools automatically generate trend plots of CST, vessel density, and HRF counts, highlighting clinically significant changes that might be missed by manual review.
  • Early warning systems: Machine learning classifiers can flag eyes at risk of disease recurrence before fluid reaccumulates visibly on OCT, potentially enabling proactive treatment adjustments.
  • Personalized dosing algorithms: Some platforms use reinforcement learning to recommend the optimal timing and combination of anti-VEGF and corticosteroid treatments based on imaging biomarkers and patient history.

Several AI-based platforms have received regulatory clearance for diabetic eye disease screening and quantification, and their integration into electronic health records is accelerating. For instance, the American Academy of Ophthalmology highlights the growing role of AI in diabetic retinopathy screening. However, validation in real-world populations with diverse ethnicities and disease histories remains an ongoing priority. Clinicians should critically evaluate the performance of AI tools in their own practice settings and use them as decision support rather than autonomous diagnostic agents. As AI continues to evolve, it will likely become a standard component of monitoring dual therapy response.

Imaging as an Endpoint in Clinical Trials

Advanced imaging modalities have also become essential endpoints in clinical trials investigating dual therapy. Regulatory agencies now accept CST measured by SD-OCT as a primary anatomical endpoint, and OCTA metrics such as vessel density and FAZ area are increasingly included as secondary or exploratory endpoints. The use of multimodal imaging in trials has enabled researchers to detect subtle differences between treatment arms that might not be apparent with visual acuity alone. For example, recent studies have shown that dual therapy leads to a greater reduction in HRF and DRIL compared with anti-VEGF monotherapy, suggesting a more pronounced anti-inflammatory effect. A 2022 meta-analysis published in Ophthalmology confirmed that combination therapy produces superior anatomical outcomes in DME, especially in eyes with subretinal fluid. Moving forward, the inclusion of OCTA and FAF in trial protocols will provide a more comprehensive understanding of how dual therapy modifies the natural history of diabetic retinal disease.

Future Perspectives

Emerging imaging technologies promise to further refine the monitoring of dual therapy. Adaptive optics OCT can resolve individual photoreceptor and RPE cells, enabling detection of microscopic damage that precedes visible clinical signs. Ultra-widefield OCTA will allow comprehensive assessment of the entire retina and choroid in a single scan, capturing peripheral ischemia and neovascularization that may influence treatment efficacy. Hyperspectral imaging, which measures light reflectance across many wavelengths, can non-invasively measure retinal oxygen saturation and blood flow, providing a direct functional correlate of therapy response. These innovations are still in early clinical development but show great potential for refining patient selection and treatment monitoring.

Another frontier is the integration of imaging with systemic biomarkers. Combining retinal imaging data with serum levels of inflammatory cytokines (e.g., interleukin-6, interleukin-8) may help identify patients who would benefit most from corticosteroid augmentation. Ongoing longitudinal studies such as the DRCR Retina Network are expected to produce robust evidence linking specific imaging features to dual therapy outcomes. Additionally, advances in portable OCT devices could expand access to high-quality imaging in low-resource settings, helping to reduce disparities in diabetic eye care.

Despite these advances, challenges remain. Cost and access limit the widespread adoption of advanced imaging in many regions. Moreover, interpretation requires specialized training, and inter-reader variability can still affect clinical decisions. Standardized reporting frameworks and automated quality control will be essential to ensure that imaging delivers its full potential in guiding dual therapy for diabetic retinal disease. The retina community is actively working on consensus guidelines for imaging protocols in DME management.

Conclusion

Advances in imaging—from high-resolution OCT and OCTA to multimodal and AI-enhanced approaches—have greatly improved our ability to monitor dual therapy response in diabetic retinal disease. By harnessing these tools, clinicians can optimize treatment individualization, reduce the burden of unnecessary injections, and ultimately preserve long-term vision. The ability to track biomarkers such as DRIL, HRF, and FAZ area allows for more targeted and responsive care. Continued innovation and validation will further solidify imaging’s central role in the era of combination therapy. As technology becomes more accessible and AI integration matures, the dream of truly personalized, imaging-guided treatment for diabetic retinal disease moves closer to reality.