Pediatric ophthalmologist, Retina imaging expert
6 min read
The first normative database for OCT was created in the early 2000s and were based on small studies of mostly white patients. However, as OCT technology has evolved, so too have the normative databases. Recent databases are larger and more diverse, reflecting the increasing ethnic and racial diversity of the population.
Nowadays, eye care specialists use normative database to compare the characteristics of a patient to a population-wide norm. This allows them to quickly and easily assess whether a patient’s retinal dimensions fall within normal limits. According to our survey, 79% of eye care specialists rely on the normative databases for OCT verdict with every patient.
However, despite the fact that normative databases are very widespread among specialists worldwide, they are not perfect. They can be affected by factors such as age, gender, axial length, and refractive error.
They can be influenced by low image quality due to different eye pathologies. It is essential to be aware of these limitations when interpreting normative data OCT parameters. That is why, in this article, we will discuss the benefits of the collaboration between AI decision-making tools and normative databases to improve patient outcomes.
What is a normative database, and is there a difference between normative databases for different devices?
Before diving into the subject of the benefits and limitations of normative databases, we would like to remind you what a normative database is. From the moment of its invention, the OCT exam has rapidly gained widespread adoption and has become indispensable in the eye care practice. Critical to this success has been the ability of software to automatically produce important measurements, such as the thickness of the peripapillary retinal nerve fiber layer (RNFL) in tracking glaucoma progression or the total retinal thickness in the assessment of macular diseases.
In order to accurately interpret OCT scans, normative databases were created. These databases are now built into almost all commercial OCT devices, allowing eye care specialists to view colored reports and progression maps that assist in the rapid recognition and tracking of pathology.
Summing up, a normative database for OCT is a set of data that provides references for OCT thickness measurements in a healthy population. These databases are used to compare the OCT measurements of your patient to a population-wide norm.
Here are some of the OCT parameters that are commonly measured and compared to normative databases:
- Retinal nerve fiber layer (RNFL) thickness: the RNFL is a retinal layer that is measured around the optic nerve. This measurement is important for diagnosing optic nerve atrophy.
- Macular thickness: the macula is responsible for sharp central vision.
- Ganglion cell complex thickness: the ganglion cell complex is a group of cells in the retina that are responsible for transmitting visual information to the brain.
- Cup-to-disc ratio, neuroretinal rim, and other optic nerve parameters: are very important for diagnosing glaucoma and other optic nerve pathologies
These are just a few of the OCT parameters that are commonly measured in normative databases. The specific parameters that are measured can vary depending on the type of OCT device and the clinical application.
In addition, different OCT devices can have different measurement capabilities and resolutions. For example, a device that uses time-domain OCT (TD-OCT) technology may have a lower resolution than a device that uses spectral-domain or swept-source OCT (SD or SS-OCT) technology. This means that the normative database for a TD-OCT device may not be as accurate as the normative database for an SD or SS-OCT device.
What is more, the normative database for a particular device may be based on a specific population of patients. What are the benefits and limitations of normative databases?
Now that we have highlighted different aspects of the normative database definition let us discuss the benefits and limitations of this tool. Normative databases can sometimes be very helpful for eye care specialists in diagnosis, decision-making, and creating a treatment strategy for eye diseases such as glaucoma and macular degeneration.
- The measurement provided by the normative database can be used as a baseline for tracking a patient’s response to medication or other treatment. Eye care specialists can track changes between a few visits and determine the impact on the patient.
- Normative databases show deviations from the norm, which may be a reason for a more comprehensive examination.
- Eye care specialists can also use normative databases to compare the results of different OCT devices. This can help to ensure that they are using the most accurate device for their patients.
There are still challenges that must be overcome to develop normative databases sufficient for use in clinical trials. That is why current normative databases also have a lot of limitations.
Does not detect pathology
The normative database works only with the thickness of the retina and does not detect what is inside the retina. Therefore, it cannot detect all pathologies where there is no change in retinal thickness. In the early stages, these are absolutely all diseases. We can see deviations from the normative base only when the disease progresses to a later and more severe stage when the retinal thickness decreases or increases.
Normative databases can be limited by factors like age, gender, and ethnicity of the population used to create them. This can result in reduced accuracy for patients who are not well-represented in the database.
Even healthy patients can have some anatomical variations that fall within the range of normal. These variations may be falsely flagged as abnormalities when compared to the database.
How Altris AI platform can complement the information provided by the normative database
Normative databases in OCT play a crucial role in aiding diagnosis and treatment planning, but they also have limitations related to representation, disease progression, and data quality. Eye care specialists need to interpret the results in the context of the patient’s individual characteristics and other clinical information, using additional tools for scan interpretations.
Sometimes, low-quality OCT scans can be inaccurately interpreted by the eye care specialist, and the normative database can showcase inaccurate measurements. Altris AI platform detects low-quality scans automatically and warns about the possibility of inaccurate results. In addition, the platform automates the detection of 70+ pathologies and pathological signs. Once the user uploads the scan, they can see visualized and highlighted pathological areas and pathology classification that the algorithm has detected. The user can also calculate the area and volume of detected biomarkers.
Artificial intelligence-based tools for OCT interpretation used along with normative databases can play a crucial role in clinical eye care. Altris AI, for example, can provide eye care specialists with additional and more precise information about separate retinal layer thickness. The system analyzes the thickness of each retina layer or several layers combined.
While normative databases provide information only about the thickness, AI tools equipped with deep learning models can detect pathological signs in OCT scans that might be missed by the normative database or the human eye, enhancing diagnostic accuracy. Altris AI algorithm classifies the OCT scans based on the degree of pathology found. It can distinguish green concern, which indicates normal retina, yellow – moderate with slight deviations, and red concern, which means high severity level.
Despite their limitations, normative databases are an essential tool for the clinical use of OCT. They provide a valuable reference point for assessing patients and can help to identify some diseases. However, the normative database measures only the thickness, which is not enough to accurately diagnose the patient and create a treatment plan.
That is why incorporating AI into OCT interpretation streamlines the decision-making process. By automating the initial analysis of OCT scans, specialists can focus their attention on more complex cases, making the best use of their skills and experience. Moreover, embracing AI technologies empowers eye care specialists to personalize patient care with greater precision.