Lernziele
- Die Inhalte der Einzelvorlesung „Retinal Revolution: Biomarkers and Artificial Intelligence in Ophthalmic Imaging“ strukturiert wiedergeben.
- Diagnostische Befunde und Differenzialdiagnosen fachärztlich einordnen.
- Therapie, Verlauf und Warnzeichen sicher beurteilen.
Ausführliche Vorlesung
Retinal Revolution: Biomarkers and Artificial Intelligence in Ophthalmic Imaging
Executive Summary
This document synthesizes the latest insights into the management of retinal diseases, specifically Diabetic Macular Edema (DMÖ) and Macular Edema secondary to Retinal Vein Occlusion (RVV). The core findings reveal that inflammation is a central, unifying pathophysiological mechanism in these and other retinal conditions. The analysis of morphological and biochemical biomarkers, primarily through Optical Coherence Tomography (OCT), has become critical for diagnosis, prognosis, and therapeutic decision-making. Prognostic biomarkers help predict the natural course of the disease, while predictive biomarkers indicate the likely response to specific treatments, enabling a more individualized approach. The emergence of Artificial Intelligence (AI) marks a significant technological shift, offering standardized, rapid, and comprehensive analysis of vast imaging datasets. Currently available and developing AI systems can automatically segment and quantify key biomarkers like intraretinal fluid, support therapy planning, and facilitate remote patient monitoring. While AI presents transformative potential for improving patient care and alleviating clinical workload, its integration is accompanied by challenges, including device compatibility, training data bias, and the evolving role of the physician, who remains central to patient communication and final therapeutic judgment.
1. The Pathophysiological Foundation: Inflammation in Retinal Disease
Macular edema resulting from diabetic retinopathy (DMÖ) or retinal vein occlusion (RVV) is a leading cause of significant vision loss, impacting patients' quality of life and ability to work. DMÖ is the primary cause of severe visual impairment and blindness in the working-age population, affecting approximately 5.5% of people with diabetes worldwide. RVV is the second most common cause of vascular-related vision reduction, affecting around 28 million people globally, including about 300,000 in Germany.
Although multifactorial, both DMÖ and RVV share a common pathophysiological pathway centered on inflammation. The typical trigger is ischemia—caused by microvascular changes in DMÖ and vessel blockage in RVV. The resulting hypoxia stimulates a cascade of inflammatory responses:
- Cytokine and Growth Factor Release: Hypoxia leads to an increase in pro-inflammatory cytokines and growth factors, including Interleukin-6 (IL-6), IL-8, IL-10, Monocyte Chemotactic Protein-1 (MCP-1), Tumor Necrosis Factor-alpha (TNF-α), and Vascular Endothelial Growth Factor (VEGF).
- Leukocyte Activation and Migration: Changes in the vascular surface promote the adhesion and migration of leukocytes (hypoxia-activated macrophages) into the surrounding retinal tissue. This process is mediated by proteins like ICAM-1.
- Amplification of Inflammation: Once in the retinal tissue, these leukocytes, along with local glial and microglial cells, release more pro-inflammatory mediators, further fueling the inflammatory cascade.
- Blood-Retina Barrier Breakdown: This sustained inflammation ultimately leads to increased vascular permeability and a breakdown of the tight junctions that form the blood-retina barrier. The disruption of the finely tuned retinal fluid homeostasis results in fluid leakage and the formation of macular edema.
This inflammatory cascade is not unique to DMÖ and RVV; it plays a central role in a variety of ophthalmological conditions, including age-related macular degeneration (AMD). The key mechanisms consistently observed include increased blood flow and vascular permeability, elevated expression of inflammatory mediators, macrophage infiltration, neurodegeneration, neovascularization, and microglial activation.
2. The Clinical Utility of Biomarkers
A biomarker is an objectively measurable characteristic that serves as an indicator of normal biological processes, pathological processes, or a pharmacological response to therapy. In retinal disease management, the differentiation between various types of biomarkers is crucial for clinical practice.
| Biomarker Type | Definition & Purpose | Clinical Application |
|---|
| Diagnostic | Identifies or confirms the presence of a disease or condition. | Establishing a diagnosis of DMÖ or RVV based on findings like retinal thickening. |
| Monitoring | Can be repeatedly measured to assess the status of a disease over time. | Tracking changes in central retinal thickness or fluid volume to evaluate treatment success. |
| Prognostic | Indicates the likely course of a disease (e.g., recurrence, progression) *independent* of a specific treatment. | Using features like DRIL to estimate the patient's long-term visual outcome regardless of the therapy chosen. |
| Predictive | Identifies individuals who are more likely to experience a positive or adverse effect from a particular treatment. | Using the presence of HRF to suggest that a patient may respond better to anti-inflammatory steroids than to anti-VEGF therapy. |
The effective use of biomarkers enhances therapy planning and improves patient communication. Explaining the presence of specific biomarkers can help patients understand their diagnosis, prognosis, and the rationale for treatment, thereby improving therapeutic adherence.
3. Key Biomarkers in Optical Coherence Tomography (OCT)
OCT has become the standard imaging modality for retinal diseases, providing high-resolution cross-sectional images that allow for the detailed visualization and assessment of morphological biomarkers. A comprehensive evaluation of multiple biomarkers is recommended to form a complete picture of an individual's disease status.
3.1. Central Retinal Thickness and Intraretinal Fluid (IRF)
Central retinal thickness is a foundational biomarker used for diagnosis and monitoring treatment response. It is often correlated with the presence of large intraretinal cystoid spaces (IRF), which are fluid-filled cavities within the macular area.
- Prognostic Value: The presence of large cystoid spaces is correlated with poorer visual prognosis and structural damage to the outer retina.
- Predictive Value: The size of these cysts can indicate disease duration. Larger cysts (>250 µm) in the outer nuclear layer are associated with chronic DMÖ, which often responds better to intravitreal steroid treatments than to anti-VEGF therapies.
3.2. Hyperreflective Foci (HRF)
HRF are small (<30 µm), discrete, punctate lesions visible on OCT that have a reflectivity equal to or greater than the retinal pigment epithelium (RPE).
- Pathophysiology: Some studies suggest HRF represent activated microglial cells or macrophage infiltrates, making them a direct marker of inflammation.
- Prognostic & Predictive Value: The presence of numerous HRF is significantly associated with poorer visual acuity under therapy. A decrease in their number and size indicates a positive response to treatment. Because HRF signal a heightened inflammatory state, their presence suggests that a broader anti-inflammatory therapy with steroids may be more advantageous than a targeted anti-VEGF therapy.
- Challenge: Manual quantification of HRF is extremely time-consuming and difficult in a clinical setting, highlighting a key area where automated AI systems can provide value.
3.3. Disorganization of the Retinal Inner Layers (DRIL)
DRIL refers to the loss of discernible boundaries between the inner retinal layers (inner nuclear layer, outer plexiform layer, and the ganglion cell layer-inner plexiform layer complex) on an OCT scan.
- Pathophysiology: DRIL visualizes structural damage to the inner retina, likely resulting from chronic or recurrent edema, and may indicate a disruption of the neural signaling pathways from photoreceptors to ganglion cells.
- Prognostic Value: DRIL is a powerful prognostic biomarker that strongly correlates with visual acuity in both active and resolved DMÖ. An extensive DRIL is a sign of poor prognosis.
- Predictive Value: In cases with extensive DRIL, particularly when persisting under anti-VEGF treatment, a switch to intravitreal steroids should be considered, as their multiple anti-inflammatory mechanisms may yield better structural improvement.
4. Biochemical Biomarkers: The Role of Cytokines
Beyond morphological changes, the concentration of cytokines in the aqueous humor and vitreous serves as a direct biochemical biomarker of inflammation.
- Elevated Levels: Two meta-analyses identified significantly elevated concentrations of IL-6, IL-8, MCP-1, and VEGF in the aqueous humor and/or vitreous of patients with DMÖ and RVV.
- Correlation with Severity: One study showed that levels of inflammatory mediators like IL-6 and IL-8 increase with the severity of diabetic retinopathy, while VEGF levels do not change significantly across different severity grades.
- Differential Treatment Effects: A study comparing off-label triamcinolone (a steroid) and bevacizumab (an anti-VEGF agent) in bilateral DMÖ demonstrated that the steroid down-regulated a wide range of cytokines as well as VEGF, whereas the anti-VEGF therapy primarily lowered VEGF levels.
- Predicting Treatment Response: A Japanese study on treatment-naïve DMÖ patients found that those with higher baseline aqueous VEGF concentrations responded better and more quickly to ranibizumab (anti-VEGF) injections, achieving better long-term visual outcomes.
While direct measurement of aqueous cytokines is not yet routine, it points toward a future of personalized medicine where such profiles could guide the selection of the most appropriate therapeutic agent.
5. The Emergence of Artificial Intelligence in Retinology
AI is rapidly entering the field of ophthalmology, driven by the large volume of imaging data generated in routine practice. The number of publications on AI in ophthalmology has grown exponentially since 2014. A EURETINA survey revealed that 75% of ophthalmologists would use AI applications in the future, with 72% believing AI will significantly support diagnosis and monitoring.
AI's primary capabilities in retinopathies include:
- Classification: Differentiating between diseases (e.g., AMD vs. DMÖ) and classifying disease severity.
- Segmentation: Automatically identifying and outlining biomarkers, such as fluid compartments, and quantifying their volume.
- Prediction: Forecasting disease progression, therapy response, and even predicting the onset of disease based on complex data patterns.
5.1. Currently Available and Developing AI Systems
Several AI-powered "clinical decision support systems" are already certified or in development:
| System | Developer / Origin | Primary Function & Disease Focus | Status |
|---|
| RetinAI Discovery® | Spin-off, University of Bern | Quantifies fluid biomarkers in nAMD, GA, DR/DME, and CME. Provides a "Fluids Report." | CE Certified |
| Vienna Fluid Monitor | RetinSight (Spin-off, University of Vienna) | Segments and quantifies intraretinal fluid, subretinal fluid, and pigment epithelial detachment in AMD. Reports volume in nanoliters. | CE Certified (MDR) |
| deepeye TPS® | deepeye Medical | Therapy planning support for nAMD. Annotates fluid, predicts disease activity, and estimates injection needs over 12 months. | CE Certification expected mid-2025 |
| Scanly® Home OCT | Notal Vision, Inc. | A portable, patient-operated home OCT for remote monitoring of nAMD, DMÖ, and RVV. AI quantifies fluid and alerts physicians to activity. | FDA Approved; No CE cert. |
| RetinSight GA Monitor | RetinSight | Segments and measures areas of photoreceptor and RPE loss in Geographic Atrophy (GA) to monitor progression. | CE Certified (MDR) |
| LumineticsCore™ | Digital Diagnostics | Fundus-based screening for diabetic retinopathy. | FDA Approved; EU certification not renewed due to regulatory costs (EU-MDR). |
| Optomed Aurora AEYE | AEYE Health | Handheld, fundus-based screening for diabetic retinopathy, usable by other disciplines like diabetologists. | FDA Approved; No CE cert. |
5.2. Opportunities and Limitations of AI
AI offers significant opportunities to enhance patient care by providing objective, standardized, and rapid data analysis, which can alleviate the workload in busy clinics. Studies have shown that diagnoses are more effective when made by a physician in combination with an AI system than by either alone.
However, several limitations exist:
- Device Compatibility: Many AI algorithms are trained on data from specific OCT manufacturers (e.g., Heidelberg) and are not yet compatible with all devices.
- Training Data Bias: The databases used to train AI are often not globally representative. A significant portion comes from Asian populations, with a lack of data from regions like Africa, leading to concerns about a "white AI" that may perform less accurately on underrepresented groups.
- Regulatory Hurdles: The stringent European Union Medical Device Regulation (EU-MDR) has increased costs and complexity, leading some companies (e.g., Digital Diagnostics) to withdraw from the EU market.
- The Physician's Role: While AI can automate routine tasks, the physician's expertise remains irreplaceable for interpreting results in the context of the whole patient, ensuring empathetic communication, and making the final, nuanced therapeutic decision. The future role of the ophthalmologist will be to leverage AI as a powerful supplementary tool while retaining ultimate clinical oversight.
Prüfungsorientierter Überblick
Retinal Revolution: Biomarkers and Artificial Intelligence in Ophthalmic Imaging
This document synthesizes the latest insights into the management of retinal diseases, specifically Diabetic Macular Edema (DMÖ) and Macular Edema secondary to Retinal Vein Occlusion (RVV). The core findings reveal that inflammation is a central, unifying pathophysiological mechanism in these and other retinal conditions. The analysis of morphological and biochemical biomarkers, primarily through Optical Coherence Tomography (OCT), has become critical for diagnosis, prognosis, and therapeutic decision-making. Prognostic biomarkers help predict the natural course of the disease, while predictive biomarkers indicate the likely response to specific treatments, enabling a more individualized approach. The emergence of Artificial Intelligence (AI) marks a significant technological shift, offering standardized, rapid, and comprehensive analysis of vast imaging datasets. Currently available and developing AI systems can automatically segment and quantify key biomarkers like intraretinal fluid, support therapy planning, and facilitate remote patient monitoring. While AI presents transformative potential for improving patient care and alleviating clinical workload, its integration is accompanied by challenges, including device compatibility, training data bias, and the evolving role of the physician, who remains central to patient communication and final therapeutic judgment.
Macular edema resulting from diabetic retinopathy (DMÖ) or retinal vein occlusion (RVV) is a leading cause of significant vision loss, impacting patients' quality of life and ability to work. DMÖ is the primary cause of severe visual impairment and blindness in the working-age population, affecting approximately 5.5% of people with diabetes worldwide. RVV is the second most common cause of vascular-related vision reduction, affecting around 28 million people globally, including about 300,000 in Germany.
Although multifactorial, both DMÖ and RVV share a common pathophysiological pathway centered on inflammation. The typical trigger is ischemia—caused by microvascular changes in DMÖ and vessel blockage in RVV. The resulting hypoxia stimulates a cascade of inflammatory responses:
Cytokine and Growth Factor Release: Hypoxia leads to an increase in pro-inflammatory cytokines and growth factors, including Interleukin-6 (IL-6), IL-8, IL-10, Monocyte Chemotactic Protein-1 (MCP-1), Tumor Necrosis Factor-alpha (TNF-α), and Vascular Endothelial Growth Factor (VEGF). Leukocyte Activation and Migration: Changes in the vascular surface promote the adhesion and migration of leukocytes (hypoxia-activated macrophages) into the surrounding retinal tissue. This process is mediated by proteins like ICAM-1. Amplification of Inflammation: Once in the retinal tissue, these leukocytes, along with local glial and microglial cells, release more pro-inflammatory mediators, further fueling the inflammatory cascade. Blood-Retina Barrier Breakdown: This sustained inflammation ultimately leads to increased vascular permeability and a breakdown of the tight junctions that form the blood-retina barrier. The disruption of the finely tuned retinal fluid homeostasis results in fluid leakage and the formation of macular edema.
This inflammatory cascade is not unique to DMÖ and RVV; it plays a central role in a variety of ophthalmological conditions, including age-related macular degeneration (AMD). The key mechanisms consistently observed include increased blood flow and vascular permeability, elevated expression of inflammatory mediators, macrophage infiltration, neurodegeneration, neovascularization, and microglial activation.
A biomarker is an objectively measurable characteristic that serves as an indicator of normal biological processes, pathological processes, or a pharmacological response to therapy. In retinal disease management, the differentiation between various types of biomarkers is crucial for clinical practice.
Biomarker Type Definition & Purpose Clinical Application :--- :--- :--- Diagnostic Identifies or confirms the presence of a disease or condition. Establishing a diagnosis of DMÖ or RVV based on findings like retinal thickening. Monitoring Can be repeatedly measured to assess the status of a disease over time. Tracking changes in central retinal thickness or fluid volume to evaluate treatment success. Prognostic Indicates the likely course of a disease (e.g., recurrence, progression) independent of a specific treatment. Using features like DRIL to estimate the patient's long-term visual outcome regardless of the therapy chosen. Predictive Identifies individuals who are more likely to experience a positive or adverse effect from a particular treatment. Using the presence of HRF to suggest that a patient may respond better to anti-inflammatory steroids than to anti-VEGF therapy.
The effective use of biomarkers enhances therapy planning and improves patient communication. Explaining the presence of specific biomarkers can help patients understand their diagnosis, prognosis, and the rationale for treatment, thereby improving therapeutic adherence.
OCT has become the standard imaging modality for retinal diseases, providing high-resolution cross-sectional images that allow for the detailed visualization and assessment of morphological biomarkers. A comprehensive evaluation of multiple biomarkers is recommended to form a complete picture of an individual's disease status.
Diagnostik
- This document synthesizes the latest insights into the management of retinal diseases, specifically Diabetic Macular Edema (DMÖ) and Macular Edema secondary to Retinal Vein Occlusion (RVV). The core findings reveal that inflammation is a central, unifying pathophysiological mechanism in these and other retinal conditions. The analysis of morphological and biochemical biomarkers, primarily through Optical Coherence Tomography (OCT), has become critical for diagnosis, prognosis, and therapeutic decision-making. Prognostic biomarkers help predict the natural course of the disease, while predictive biomarkers indicate the likely response to specific treatments, enabling a more individualized approach. The emergence of Artificial Intelligence (AI) marks a significant technological shift, offering standardized, rapid, and comprehensive analysis of vast imaging datasets. Currently available and developing AI systems can automatically segment and quantify key biomarkers like intraretinal fluid, support therapy planning, and facilitate remote patient monitoring. While AI presents transformative potential for improving patient care and alleviating clinical workload, its integration is accompanied by challenges, including device compatibility, training data bias, and the evolving role of the physician, who remains central to patient communication and final therapeutic judgment.
- Biomarker Type Definition & Purpose Clinical Application :--- :--- :--- Diagnostic Identifies or confirms the presence of a disease or condition. Establishing a diagnosis of DMÖ or RVV based on findings like retinal thickening. Monitoring Can be repeatedly measured to assess the status of a disease over time. Tracking changes in central retinal thickness or fluid volume to evaluate treatment success. Prognostic Indicates the likely course of a disease (e.g., recurrence, progression) independent of a specific treatment. Using features like DRIL to estimate the patient's long-term visual outcome regardless of the therapy chosen. Predictive Identifies individuals who are more likely to experience a positive or adverse effect from a particular treatment. Using the presence of HRF to suggest that a patient may respond better to anti-inflammatory steroids than to anti-VEGF therapy.
- The effective use of biomarkers enhances therapy planning and improves patient communication. Explaining the presence of specific biomarkers can help patients understand their diagnosis, prognosis, and the rationale for treatment, thereby improving therapeutic adherence.
- OCT has become the standard imaging modality for retinal diseases, providing high-resolution cross-sectional images that allow for the detailed visualization and assessment of morphological biomarkers. A comprehensive evaluation of multiple biomarkers is recommended to form a complete picture of an individual's disease status.
- 3.1. Central Retinal Thickness and Intraretinal Fluid (IRF) Central retinal thickness is a foundational biomarker used for diagnosis and monitoring treatment response. It is often correlated with the presence of large intraretinal cystoid spaces (IRF), which are fluid-filled cavities within the macular area. Prognostic Value: The presence of large cystoid spaces is correlated with poorer visual prognosis and structural damage to the outer retina. Predictive Value: The size of these cysts can indicate disease duration. Larger cysts ( 250 µm) in the outer nuclear layer are associated with chronic DMÖ, which often responds better to intravitreal steroid treatments than to anti-VEGF therapies.
- 3.2. Hyperreflective Foci (HRF) HRF are small (<30 µm), discrete, punctate lesions visible on OCT that have a reflectivity equal to or greater than the retinal pigment epithelium (RPE). Pathophysiology: Some studies suggest HRF represent activated microglial cells or macrophage infiltrates, making them a direct marker of inflammation. Prognostic & Predictive Value: The presence of numerous HRF is significantly associated with poorer visual acuity under therapy. A decrease in their number and size indicates a positive response to treatment. Because HRF signal a heightened inflammatory state, their presence suggests that a broader anti-inflammatory therapy with steroids may be more advantageous than a targeted anti-VEGF therapy. Challenge: Manual quantification of HRF is extremely time-consuming and difficult in a clinical setting, highlighting a key area where automated AI systems can provide value.
- 3.3. Disorganization of the Retinal Inner Layers (DRIL) DRIL refers to the loss of discernible boundaries between the inner retinal layers (inner nuclear layer, outer plexiform layer, and the ganglion cell layer-inner plexiform layer complex) on an OCT scan. Pathophysiology: DRIL visualizes structural damage to the inner retina, likely resulting from chronic or recurrent edema, and may indicate a disruption of the neural signaling pathways from photoreceptors to ganglion cells. Prognostic Value: DRIL is a powerful prognostic biomarker that strongly correlates with visual acuity in both active and resolved DMÖ. An extensive DRIL is a sign of poor prognosis. Predictive Value: In cases with extensive DRIL, particularly when persisting under anti-VEGF treatment, a switch to intravitreal steroids should be considered, as their multiple anti-inflammatory mechanisms may yield better structural improvement.
- AI is rapidly entering the field of ophthalmology, driven by the large volume of imaging data generated in routine practice. The number of publications on AI in ophthalmology has grown exponentially since 2014. A EURETINA survey revealed that 75% of ophthalmologists would use AI applications in the future, with 72% believing AI will significantly support diagnosis and monitoring.
Differenzialdiagnosen
- Beyond morphological changes, the concentration of cytokines in the aqueous humor and vitreous serves as a direct biochemical biomarker of inflammation. Elevated Levels: Two meta-analyses identified significantly elevated concentrations of IL-6, IL-8, MCP-1, and VEGF in the aqueous humor and/or vitreous of patients with DMÖ and RVV. Correlation with Severity: One study showed that levels of inflammatory mediators like IL-6 and IL-8 increase with the severity of diabetic retinopathy, while VEGF levels do not change significantly across different severity grades. Differential Treatment Effects: A study comparing off-label triamcinolone (a steroid) and bevacizumab (an anti-VEGF agent) in bilateral DMÖ demonstrated that the steroid down-regulated a wide range of cytokines as well as VEGF, whereas the anti-VEGF therapy primarily lowered VEGF levels. Predicting Treatment Response: A Japanese study on treatment-naïve DMÖ patients found that those with higher baseline aqueous VEGF concentrations responded better and more quickly to ranibizumab (anti-VEGF) injections, achieving better long-term visual outcomes.
Therapieprinzipien
- This document synthesizes the latest insights into the management of retinal diseases, specifically Diabetic Macular Edema (DMÖ) and Macular Edema secondary to Retinal Vein Occlusion (RVV). The core findings reveal that inflammation is a central, unifying pathophysiological mechanism in these and other retinal conditions. The analysis of morphological and biochemical biomarkers, primarily through Optical Coherence Tomography (OCT), has become critical for diagnosis, prognosis, and therapeutic decision-making. Prognostic biomarkers help predict the natural course of the disease, while predictive biomarkers indicate the likely response to specific treatments, enabling a more individualized approach. The emergence of Artificial Intelligence (AI) marks a significant technological shift, offering standardized, rapid, and comprehensive analysis of vast imaging datasets. Currently available and developing AI systems can automatically segment and quantify key biomarkers like intraretinal fluid, support therapy planning, and facilitate remote patient monitoring. While AI presents transformative potential for improving patient care and alleviating clinical workload, its integration is accompanied by challenges, including device compatibility, training data bias, and the evolving role of the physician, who remains central to patient communication and final therapeutic judgment.
- A biomarker is an objectively measurable characteristic that serves as an indicator of normal biological processes, pathological processes, or a pharmacological response to therapy. In retinal disease management, the differentiation between various types of biomarkers is crucial for clinical practice.
- Biomarker Type Definition & Purpose Clinical Application :--- :--- :--- Diagnostic Identifies or confirms the presence of a disease or condition. Establishing a diagnosis of DMÖ or RVV based on findings like retinal thickening. Monitoring Can be repeatedly measured to assess the status of a disease over time. Tracking changes in central retinal thickness or fluid volume to evaluate treatment success. Prognostic Indicates the likely course of a disease (e.g., recurrence, progression) independent of a specific treatment. Using features like DRIL to estimate the patient's long-term visual outcome regardless of the therapy chosen. Predictive Identifies individuals who are more likely to experience a positive or adverse effect from a particular treatment. Using the presence of HRF to suggest that a patient may respond better to anti-inflammatory steroids than to anti-VEGF therapy.
- The effective use of biomarkers enhances therapy planning and improves patient communication. Explaining the presence of specific biomarkers can help patients understand their diagnosis, prognosis, and the rationale for treatment, thereby improving therapeutic adherence.
- 3.1. Central Retinal Thickness and Intraretinal Fluid (IRF) Central retinal thickness is a foundational biomarker used for diagnosis and monitoring treatment response. It is often correlated with the presence of large intraretinal cystoid spaces (IRF), which are fluid-filled cavities within the macular area. Prognostic Value: The presence of large cystoid spaces is correlated with poorer visual prognosis and structural damage to the outer retina. Predictive Value: The size of these cysts can indicate disease duration. Larger cysts ( 250 µm) in the outer nuclear layer are associated with chronic DMÖ, which often responds better to intravitreal steroid treatments than to anti-VEGF therapies.
- 3.2. Hyperreflective Foci (HRF) HRF are small (<30 µm), discrete, punctate lesions visible on OCT that have a reflectivity equal to or greater than the retinal pigment epithelium (RPE). Pathophysiology: Some studies suggest HRF represent activated microglial cells or macrophage infiltrates, making them a direct marker of inflammation. Prognostic & Predictive Value: The presence of numerous HRF is significantly associated with poorer visual acuity under therapy. A decrease in their number and size indicates a positive response to treatment. Because HRF signal a heightened inflammatory state, their presence suggests that a broader anti-inflammatory therapy with steroids may be more advantageous than a targeted anti-VEGF therapy. Challenge: Manual quantification of HRF is extremely time-consuming and difficult in a clinical setting, highlighting a key area where automated AI systems can provide value.
- 3.3. Disorganization of the Retinal Inner Layers (DRIL) DRIL refers to the loss of discernible boundaries between the inner retinal layers (inner nuclear layer, outer plexiform layer, and the ganglion cell layer-inner plexiform layer complex) on an OCT scan. Pathophysiology: DRIL visualizes structural damage to the inner retina, likely resulting from chronic or recurrent edema, and may indicate a disruption of the neural signaling pathways from photoreceptors to ganglion cells. Prognostic Value: DRIL is a powerful prognostic biomarker that strongly correlates with visual acuity in both active and resolved DMÖ. An extensive DRIL is a sign of poor prognosis. Predictive Value: In cases with extensive DRIL, particularly when persisting under anti-VEGF treatment, a switch to intravitreal steroids should be considered, as their multiple anti-inflammatory mechanisms may yield better structural improvement.
- Beyond morphological changes, the concentration of cytokines in the aqueous humor and vitreous serves as a direct biochemical biomarker of inflammation. Elevated Levels: Two meta-analyses identified significantly elevated concentrations of IL-6, IL-8, MCP-1, and VEGF in the aqueous humor and/or vitreous of patients with DMÖ and RVV. Correlation with Severity: One study showed that levels of inflammatory mediators like IL-6 and IL-8 increase with the severity of diabetic retinopathy, while VEGF levels do not change significantly across different severity grades. Differential Treatment Effects: A study comparing off-label triamcinolone (a steroid) and bevacizumab (an anti-VEGF agent) in bilateral DMÖ demonstrated that the steroid down-regulated a wide range of cytokines as well as VEGF, whereas the anti-VEGF therapy primarily lowered VEGF levels. Predicting Treatment Response: A Japanese study on treatment-naïve DMÖ patients found that those with higher baseline aqueous VEGF concentrations responded better and more quickly to ranibizumab (anti-VEGF) injections, achieving better long-term visual outcomes.
Red Flags
- Warnzeichen, Komplikationen und Notfallindikationen fachärztlich validieren.
Prüfungsfragen
1. Was sind die wichtigsten Lernpunkte der Vorlesung „Retinal Revolution: Biomarkers and Artificial Intelligence in Ophthalmic Imaging“?
Die Antwort ergibt sich aus den strukturierten Abschnitten und muss am individuellen klinischen Befund überprüft werden.
2. Welche Befunde erfordern eine dringliche Abklärung?
Rasche Sehverschlechterung, starke Schmerzen, ausgeprägte Entzündung, Druckanstieg oder Verdacht auf eine infektiöse oder neurologische Ursache.