At RetinAI, we are always looking for ways to enhance our platform to ease and accelerate R&D initiatives for pharma, clinics and academia - that’s why we are excited to announce the launch of our latest OCT segmentation model, now available in Research Use Only¹  across all our Discovery products. 

With its new automated detection and measurement of seven layers of the retina, the choroid, fluids, and other relevant biomarkers, and the possibility of quantifying the disruption of RPE and photoreceptors, you will be able to better assess retinal diseases. 

In this blog, we will dive into how our new OCT Segmentation model will help you enhance retinal image analysis, making it easier for you to detect and measure key biomarkers, visualize disruptions in retinal layers, and streamline batch analysis. 

5 minutes read

‍

Upgrading your real-time retinal disease analysis

Our platform RetinAI Discovery® is an agnostic and multimodality cloud-based platform that supports several types of images (OCT, FAF, CFP, OCTA, among others) and formats (DICOM-compliant and proprietary formats).

Discovery equips your team with functionalities ranging from data collection to precise quantification of biomarkers with insights from both certified and validated Research-Use-Only AI modules. 

Assoc Prof. Dr. Dr. Marion R. MUNK, MD, PhD, FEBO, FMH, Senior Physician with Augenarzt-Praxisgemeinschaft Gutblick AG states:

We are using RetinAI Discovery® for an open label study and it really helped us understand the mode of action of the investigational drug, so understanding how it works and its efficacy. Thanks to the immediate access to the images and to the individual study visits that Discovery provides, we were able to understand faster how the drug was working and adapt the study design accordingly.‍

Our current research features our most recent and advanced development in the realm of Optical Coherence Tomography (OCT) image analysis in the retina. We are proud to introduce to the community our new OCT Segmentation module!

This model allows for a more detailed understanding of retinal anatomical landmarks across multiple diseases, including a zoom-in to the photoreceptor layers, reducing the time spent on imaging data analysis. Discover below what’s new. 

‍

Detect and measure biomarkers in seconds 

Our Research-Use-Only OCT Segmentation model provides reliable automated detection and measurement of seven layers of the retina, the choroid, as well as fluids and other relevant biomarkers, including (Figure 1):

• Retinal nerve fibre layer (RNFL) 
• Ganglion cell plus inner plexiform layer (GCL+IPL) 
• Inner nuclear plus outer plexiform layer (INL+OPL) 
• Outer nuclear layer plus Henle’s fibre layer (ONL+HFL)
• Myoid Zone (MZ) 
• Ellipsoid zone, outer photoreceptor segment, and interdigitation zone (EZ+OPR+IZ)
• Retinal Pigment Epithelium (RPE) 
• Choriocapillaris and choroidal stroma (CC+CS)
• Sub-retinal fluid (SRF) 
• Intraretinal fluid (IRF) 
• Sub-RPE material
• Hyperreflective Foci (HRF) 
• Amorphous material (AM) 
  ‍
  ‍

‍

‍

With this advanced segmentation and measurements of structural endpoints, we aim to ease the management of complex retinal diseases. We have recently written an article delving into the benefits of using structural endpoints to monitor diseases such as neovascular Age-related Macular Degeneration (nAMD), Geographic Atrophy (GA) and Central Serous Chorioretinopathy (CSC). 

‍

We do believe structural endpoints are transforming the landscape of retinal disease management, offering clinicians precise tools to monitor disease progression and treatment efficacy. 

‍

‍

‍

Visualize and quantify the disruption of RPE and photoreceptor layers 

In addition, our Research-Use-Only OCT Segmentation model features the possibility of visualizing and quantifying the disruption of RPE and photoreceptor layers through en-face projection of layer loss/attenuation (Defect maps, Figure 2). Defect maps allow areas of atrophy to be easily visualized and located at a glance. 

In addition, the thickness of each layer is also computed along a given A-scan. The average thickness over the whole scan and each ETDRS region is computed and displayed in µm in a heatmap. 

These functionalities are particularly relevant for the assessment of Geographic Atrophy (GA), especially after the FDA's confirmation of EZ to RPE attenuation (defined as the percentage area of regions with distance between segmented EZ and RPE lines <20 μm) as a reliable marker of photoreceptor loss in clinical trials².

‍

‍

Moreover, the evaluation of these biomarkers beyond the atrophic lesion in GA³ provides valuable information about lesion spreading and may help identify early biomarkers for disease progression. 

‍

Predicting Geographic Atrophy Progression with our RUO AI Models 

We designed a Geographic Atrophy (GA) view in Discovery to display all relevant biomarkers extracted from OCT scans by our AI models. Our goal is to simplify the analysis of GA and provide a complete, real-time picture of disease activity. 

This view combines outputs from three RUO AI models: 

1. The GA segmentation model 
2. The GA progression prediction model 
3. The OCT Segmentation Model
   
   

Our GA  view provides detailed segmentation and related measurements of Retinal Pigment Epithelium (RPE) and Outer Retinal Atrophy (RORA) on B-scans (GA Segmentation model). Alongside, the FAF image is displayed when available to facilitate the multimodal comparison of the atrophic lesion. 

‍

In addition, it features a quantitative characterization of Photoreceptors & RPE integrity by showing en-face projections of layer loss area (OCT Segmentation model).  

‍

Finally, when multiple acquisitions are available for a patient, our GA Progression Prediction model shows a progression of the atrophic lesion size over time is also displayed, along with a projection of how this lesion would evolve (up to 54 months in the future!). 

‍

This view will streamline your GA analysis by allowing for precise calculation of the affected GA area and accurate predictions of disease progression based on patient visit data. You will also be able to visualize the layer thickness regions around the GA area within RetinAI Discovery®.

These models combined in an easy-to-understand view are particularly useful for identifying early biomarkers of disease progression, helping users monitor the spread of atrophic lesions.

‍

‍

Enhance your research with RetinAI’s OCT Segmentation Model 

‍

With our new RUO OCT Segmentation Model, elevating retinal disease analysis is possible! Our model is designed to make your research and clinical workflows more efficient and insightful. With our platform RetinAI Discovery®, you will have access to powerful tools to visualize, quantify and predict disease progression with precision. In addition, our expert team can help you understand your data with AI-driven technology supported by State-of-The-Art statistical analysis.

Want to explore how our platform can transform your R&D initiatives and enhance patient outcomes? Request a demo, and see our technology in action! We are here to enable the right decisions sooner in Healthcare. 

References

¹ Not intended for use in clinical, diagnostic procedures, or therapeutic applications.
‍
² Yordi S, Cakir Y, Kalra G, et al. Ellipsoid Zone Integrity and Visual Function in Dry Age-Related Macular Degeneration. J Pers Med. 2024;14:543. doi:10.3390/jpm14050543
‍
³ Blair J, Wu Z, Lasagni Vitar R, et al. On the Edge: Quantitative Analysis of Retinal Layer Thickness Surrounding Geographic Atrophy Lesions Using OCT. Investigative Ophthalmology & Visual Science. 2024;65:5474.

‍

This article was written by Romina Lasagni Vitar, PhD

‍