Spectral-domain optical coherence tomography (SD-OCT) is an imaging technique that visualizes the layered structure of the retina by analyzing interference patterns of reflected laser light. First reported in 1991, time-domain OCT (TD-OCT) became commercially available in 2002 and gained widespread use. SD-OCT, introduced after 2006, is a next-generation technology that significantly improves upon TD-OCT.
SD-OCT has improved depth resolution and dramatically increased scan speed. It enables shape analysis not only in cross-section but also in surface and volume. An automatic segmentation algorithm precisely delineates the retinal nerve fiber layer (RNFL) 1).
Representative commercially available SD-OCT devices include the following:
In recent years, swept-source OCT (SS-OCT) with greater penetration depth has also been developed and is applied to the analysis of the lamina cribrosa of the optic nerve head and the choroid1).
In glaucoma diagnosis, the high utility of SD-OCT-based assessment is recognized 1). However, measurement accuracy has limitations, and there is overlap in values between glaucomatous and normal eyes; therefore, final judgment must be made by integrating clinical findings 1)2).
QWhat is the difference between SD-OCT and TD-OCT?
A
TD-OCT obtains retinal cross-sectional images by superimposing A-scans in one axial direction, requiring time for examination. SD-OCT adopts the Fourier domain method, increasing scan speed to 26,000 A-scans/second or more. Axial resolution has also improved to approximately 5 µm, enabling high-speed analysis of RNFL thickness, optic nerve head, and macular ganglion cell complex. Shape analysis by surface and volume is also achieved 1).
SD-OCT evaluates glaucomatous changes using the following three parameters. All values are compared with a normal database and displayed in color codes: white, green, yellow, and red 2). Yellow indicates a probability of less than 5%, and red indicates less than 1%.
RNFL Thickness
Measurement principle: Quantifies the thickness between the internal limiting membrane (ILM) and the RNFL boundary.
TSNIT map: Displays RNFL thickness on a 3.4 mm circle centered on the optic nerve head in the order T (temporal) → S (superior) → N (nasal) → I (inferior) → T (temporal).
Normal pattern: Shows a bimodal peak in the superior and inferior directions (reflecting the anatomical distribution of arcuate nerve fibers) 1).
Quadrant and clock-hour display: Displays RNFL thickness by quadrant and clock-hour sectors.
ONH Parameters
Optic nerve head analysis: Automatically delineates the optic disc, cup, and rim.
Bruch’s membrane reference: Defines the disc margin at the Bruch’s membrane opening and calculates the shortest distance to the ILM.
High diagnostic value indicators: Vertical rim thickness, rim area, and vertical cup-to-disc ratio have the highest diagnostic performance 2).
BMO-MRW: Rim width assessment based on Bruch’s membrane opening, with excellent reproducibility 1).
Ganglion cell analysis (GCA): Measures the combined thickness of the ganglion cell layer (GCL) and inner plexiform layer (IPL) around the macula. Cirrus evaluates GCL+IPL (GCIPL), while Optovue evaluates the ganglion cell complex (GCC) which includes the RNFL1)2). The minimum, inferotemporal sector, and average values are the most diagnostically useful parameters.
Key findings regarding the glaucoma detection ability of SD-OCT are as follows:
Detection by average RNFL thickness: SD-OCT sensitivity 83%, specificity 88% (at 5% level). At the 1% level, specificity 100%, sensitivity 65%.
ONH parameters have diagnostic ability equivalent to RNFL thickness parameters 2).
GCA parameters also have diagnostic ability comparable to ONH and RNFL parameters.
In preperimetric glaucoma, SD-OCTRNFL measurement is particularly useful for detecting structural changes before the onset of visual field defects 1)3). An increasing number of glaucoma cases are being diagnosed for the first time using OCT1).
Myopic eyes: In high myopia, RNFL thickness is underestimated, leading to false positives. Due to the temporal shift of the RNFL bundle, even normal eyes may be judged as “thinning” 1).
Media opacities: Cataracts cause underestimation of RNFL thickness. Reports indicate that RNFL thickness increases by 4.8–9.3% after cataract surgery.
Axial length: The longer the axial length, the thinner the RNFL, and the smaller the optic disc area and rim area measured. Cirrus does not perform axial length correction.
Measurement-related factors
Segmentation errors: These are more likely to occur in cases of tilted optic disc, scleral staphyloma, peripapillary atrophy, or epiretinal membrane. SD-OCT has a lower frequency than TD-OCT.
Eye movements and blinking: These disrupt the alignment of A-scans, leading to inaccurate RNFL thickness measurements. This is improved by eye-tracking functions.
Signal strength: Scans with a signal strength below 6 should be repeated. Defocus causes falsely thin RNFL measurements.
Limitations of the normative database should also be noted 2). The Cirrus normative database consists of 284 individuals (ages 18–84) with refractive errors ranging from −12.00 D to +8.00 D. In patients with characteristics not included in the database, caution is needed for “red disease” (displayed in red even though not actually diseased).
QHow should SD-OCT evaluation be performed in high myopia?
A
In high myopia, comparison with the normative database has limitations. Because the RNFL bundle shifts temporally, even normal eyes may be judged as “thinning.” In such cases, longitudinal comparison using each patient as their own baseline is effective. Evaluate progressive thinning on a series of SD-OCT scans. However, note that RNFL thickness decreases by approximately 0.52 µm per year due to aging even in healthy individuals, so this natural decrease must be considered.
In glaucoma, damage to retinal ganglion cells (RGCs) leads to loss of their axons, the retinal nerve fiber layer (RNFL) 1). Approximately 50% of all RGCs are concentrated in the central 20° area of the macula. Even in early glaucoma, about 50% of RGCs may have been lost 1).
RGC cell bodies and axons at the optic nerve head (ONH) experience different levels of stress 4). Stress from intraocular pressure (IOP) is significantly greater at the ONH than in the retina. Mechanical stress at the lamina cribrosa consists of hoop stress from the peripapillary sclera and trans-lamina cribrosa pressure difference due to the pressure gradient between IOP and the pressure of the myelinated optic nerve tissue 4).
The upstream mechanisms of RGC death are multifactorial and involve the following 4):
SD-OCT evaluates RGC axon loss by measuring RNFL thickness and thinning of the inner layers including cell bodies by GCA (GCIPL) 1)2). Macular parameters have a later floor effect than RNFL thickness, making them useful for evaluating advanced stages 1).
There are two approaches for determining glaucoma progression: event analysis and trend analysis.
Event analysis: progression is determined when follow-up measurements exceed a threshold from baseline.
Trend analysis: progression is determined by calculating the rate of change over time (µm/year) using regression analysis.
Cirrus GPA (Guided Progression Analysis) integrates both approaches 2). It compares baseline and follow-up RNFL thickness maps pixel by pixel to detect changes exceeding test-retest variability. Two baseline scans and three follow-up scans are required to generate an overall trend plot.
The test-retest limit for average RNFL thickness is 3.89 µm, and a reproducible decrease of 4 µm or more suggests a statistically significant change.
In advanced glaucoma, RNFL thickness plateaus, and because non-neural tissues such as glial tissue and blood vessels remain, it rarely falls below 50 µm1)2). This “floor effect” reduces the clinical utility of SD-OCT in the end stage, making visual field testing the primary method for progression assessment. Macular parameters show a later onset of the floor effect than RNFL thickness1).
Detailed analysis of the lamina cribrosa and choroid using SS-OCT (swept-source OCT)1)
Clinical application of ultrahigh-resolution OCT, polarization-sensitive OCT, and adaptive opticsOCT
Development of AI-based automated diagnosis and progression detection algorithms
Standardization of measurements across different OCT devices1)2)
Simultaneous assessment of structure and blood flow through integration with OCT-A
QWhat is the floor effect in SD-OCT?
A
The floor effect is a phenomenon in which RNFL thickness no longer decreases in advanced glaucoma. Even when nerve fibers are severely lost, non-neural tissues such as glial tissue and blood vessels remain, so RNFL thickness usually does not fall below 50 µm. At this stage, progression detection by SD-OCT becomes difficult, and assessment by visual field testing becomes primary1)2). Macular parameters (GCIPL) show a later onset of the floor effect than RNFL thickness, thus maintaining some utility even in advanced stages.
