ESCRS 2024: OCT (Artefact versus Glaucoma)
Speaker- Marta Pazos
The 5th edition of the European Glaucoma Society guidelines emphasized that Optical Coherence Tomography (OCT) alone was insufficient for clinical diagnosis. OCT was a support tool, but it could not independently diagnose glaucoma. Instead, it offered statistical results, requiring clinicians to understand the instrument and its database to interpret findings properly. OCT data were presented through a colour code system—green, yellow, and red—indicating the probability of measurements being outside normal limits based on the database. For instance, a patient with 100 microns in the green area could lose up to 30% and fall within the normal range. The variability could lead to interpretation errors, resulting in conditions like "green disease," where OCT results appeared normal, yet clinical evidence pointed to glaucoma. Conversely, "red disease" referred to clinically healthy patients, but OCT suggested abnormal findings.
It was stressed that ensuring high-quality OCT scans and understanding their functionality was essential in clinical practice. Poor-quality input could lead to poor results, impacting diagnosis and treatment. Clinicians must be aware of artefacts and check signal strength, which varies by instrument. For example, Cirrus OCT requires a signal strength above 6, while Spectralis needs it above 15. Inconsistent signal quality over time could affect outcomes, even when results seem normal, as signal strength plays a key role in the reliability of OCT readings.
Several factors can lower signal strength, such as age, poor visual acuity, myopia, and cortical or sub-capsular posterior cataracts. While signal strength may still fall within the normal range, its reduction can affect measurements, as lower signal strength underestimates retinal nerve fiber layer (RNFL) thickness. For example, a decrease in signal strength from 10 to 6 resulted in a 25% reduction in RNFL thickness in the superior quadrant. Therefore, monitoring signal strength is essential but is not the only factor. Artefacts, even with good signal strength, can influence results. Artificial intelligence can detect tiny artefacts in nearly 80% of images, though they are clinically significant in about 21% of cases. Thus, relying solely on signal strength may overlook other significant artefacts that affect clinical outcomes.
The most common artefacts in OCT imaging at the optic disc include decentration, occurring in about 28% of cases, and errors linked to posterior vitreous detachment (PVD), which can cast a posterior shadow. Poor signal quality is another frequent issue, along with misidentification of the anterior RNFL, missing sections, segmentation errors, and motion artefacts. A Spectralis OCT study found artefacts in over 40% of cases, though newer instruments have reduced this occurrence. Examples include motion artefacts, where vessels appear cut due to misalignment, and PVD, where floaters cause a shadow that may thin the measured RNFL segment and show zero in the RNFL profile. It's important to note that the RNFL never goes down to zero because of the floor effect, typically at 50-60 microns, depending on the instrument. If a zero or black image appears, it's undoubtedly an artefact. Other artefacts include blinking, OCT lens misalignment affecting segmentation, and incorrect axial alignment, leading to inaccurate patterns of RNFL loss.
Macular OCT scans also face challenges, with artefacts as frequent as 28%. The primary cause of these artefacts is epiretinal membranes, which can distort segmentation even when small. Segmentation errors at the disc level are also common. Verifying how the instrument detects and measures different layers is crucial by scrolling through the image or checking the deviation map. Vessels, peripapillary atrophy, high myopia, and myelinated nerve fibers can all interfere with accurate imaging. Segmentation issues are particularly prominent at the macular level, with myopia being the leading cause of eye alterations without other diseases. A severe segmentation error is seen in some cases where the instrument completely misidentifies layers in the macula.
The basic checklist for ensuring good quality OCT images was as follows: first, signal strength needed to be checked according to the manufacturer's guidelines, as signal reduction impacted measurements. Alignment and contrast should be confirmed by carefully checking centering, motion, and other artefacts using deviation maps or thickness profiles. It ensures accurate OCT results and helps identify potential issues affecting diagnostic accuracy. It was important to note that a zero value was never considered normal. Finally, segmentation had to be verified, especially in myopic eyes.
In a case involving a 54-year-old female suspected of pre-parametric normal-tension glaucoma, the examination appeared normal, and the OCT signal strength was recorded at 8. The majority correctly identified the issue as a PVD artefact, evidenced by the black shadow on the OCT scan, which displayed a zero RNFL, indicating that the observed changes were due to an artefact rather than actual glaucomatous damage.
In a recent case, a 52-year-old female with an intraocular pressure 23 underwent OCT imaging. The analysis focused on the key aspects of image quality, alignment, and segmentation. Initially, the scan displayed a red sector in the superior quadrant, raising concerns about potential glaucoma. However, closer examination determined that the finding was due to an artefact rather than a true disease. The low signal strength, highlighted in red, was caused by incorrect alignment, which led to truncation of the disc and an artificial red appearance in the scan. Once the alignment issue was addressed, the signal strength returned to normal, and the OCT colour code turned green, verifying that the observed anomaly was indeed an artefact.
In a final case, a 72-year-old female with a family history of glaucoma presented with normal visual fields and a consistent OCT profile over several years. However, recent ganglion cell layer (GCL) measurements from different years, specifically from 2011 to 2013, revealed unexpected changes. Despite her normal pressure and absence of treatment, the OCT images displayed anomalies. Upon review, it was determined that these deviations were due to a segmentation error rather than glaucoma. The unusual GCL deviation maps, lacking the typical snail-shaped appearance associated with glaucoma, indicated a segmentation issue. The patient, who had developed macular oedema due to diabetes, showed altered OCT measurements that mimicked RNFL thinning but were ultimately attributed to incorrect segmentation rather than true glaucomatous damage.
OCT alone should not be relied upon exclusively for managing glaucoma patients, as it cannot independently provide a complete clinical diagnosis. The guidelines stress the importance of integrating various diagnostic tools. While OCT is valuable, particularly in early glaucoma detection, it is crucial to continue utilizing visual field tests. The approach is essential because artefacts in OCT scans, which can affect a significant proportion of images, may lead to incorrect assessments if used in isolation. A comprehensive diagnostic strategy becomes even more important in busy clinical environments, where the potential for overlooked artefacts is higher. Therefore, a combined approach that includes both OCT and visual field examinations is necessary to ensure accurate diagnosis and effective management of glaucoma.
42nd Congress of the European Society of Cataract and Refractive Surgeons, 6 – 10 September 2024, Fira de Barcelona, Spain.
