For detecting structural damage in glaucoma, OCTRNFL thickness measurement and red-free photography have been standard methods 2)3). However, the measurement accuracy of OCT has limitations, and it cannot completely distinguish glaucoma from normal eyes 1). ROTA has the potential to overcome the limitations of these conventional methods.
The advantage of ROTA is that it can reveal the course of arcuate nerve fiber bundles, papillomacular nerve fiber bundles, and nasal radial nerve fiber bundles, as well as a wide area including optical texture. Furthermore, the ROTA algorithm can distinguish between retinal blood vessels and RNFL.
Since it can be analyzed using measurements obtained from commercially available OCT devices, it can be easily applied to existing clinical settings.
QHow is ROTA different from conventional OCT analysis?
A
Conventional OCT analysis focuses on quantitative measurement of RNFL thickness, comparing values obtained from peripapillary circumferential scans with a normative database 2)3). ROTA integrates reflectivity (optical density) in addition to RNFL thickness and calculates it as an optical texture signature. This allows visualization of the course patterns of a wide range of axon fiber bundles, including arcuate fiber bundles and papillomacular fiber bundles, enabling detection of subtle RNFL defects that are difficult to detect with conventional methods.
ROTA analysis is performed according to the following steps.
Acquire raster scans covering the optic nerve head and macula with OCT
Segment the anterior and posterior boundaries of the RNFL
Calculate the optical texture signature (Sxy) from RNFL reflectivity and thickness measurements
Generate a texture analysis map, and a machine learning algorithm detects RNFL abnormalities
The optical texture signature is calculated using measurements of optical density at retinal positions (x, y) and depth (z). Parameters such as gamma transformation function, gamma correction function, and a preset constant proportional to RNFL tissue thickness are used.
Wide-range assessment: Can comprehensively evaluate the course and optical texture of arcuate fiber bundles, papillomacular fiber bundles, and nasal radial fiber bundles
Differentiation from blood vessels: Retinal blood vessels and axon fiber bundles have unique optical texture signatures, so ROTA distinguishes between them.
Axial length correction: The algorithm corrects for axial length, resulting in fewer false positives in myopic eyes.
Feasibility: It can analyze data from commercially available OCT devices, eliminating the need for new equipment.
Limitations of ROTA
Media opacities: In eyes with media opacities such as cataracts, image quality is reduced.
Motion artifacts: Eye movement during OCT imaging affects analysis accuracy.
Subjective interpretation: Interpretation of RNFL defects involves subjective elements, similar to red-free photography.
Still under clinical implementation: Implementation in commercial software is not yet widespread.
Currently, SD-OCT and SS-OCT are widely used for RNFL assessment1)2)3). Three parameter groups are measured: peripapillary RNFL thickness, optic nerve head, and macular inner layers2)3).
OCT results are classified into three categories: “within normal limits,” “borderline,” and “outside normal limits”3). However, a result outside normal limits does not necessarily indicate glaucoma; interpretation in a clinical context is essential2)3). Artifacts and segmentation errors are more common in highly myopic eyes and tilted optic discs2)3)5).
Since the agreement between structural assessment and visual field testing is only partial, glaucoma diagnosis should not be based on a single test alone2)3)4)5).
QIs ROTA widely used in clinical practice?
A
ROTA is currently a research-stage technology and has not been widely implemented in commercial OCT software. However, because ROTA’s algorithm works with standard data from commercial OCT devices, it is technically feasible to introduce into existing clinical settings. Future commercialization of software and accumulation of validation studies will be key to its clinical adoption.
In glaucoma, retinal ganglion cell damage leads to loss of their axons, the RNFL1). Approximately 50% of all retinal ganglion cells are concentrated in the central 20° region of the macula, and even in early glaucoma, about 50% of retinal ganglion cells are lost.
The layers evaluated for RNFL assessment include the RNFL, ganglion cell layer (GCL), and inner plexiform layer (IPL), collectively called the ganglion cell complex (GCC) 6). Some devices use the combined GCL+IPL (GCIPL) as a diagnostic parameter.
Principles and Limitations of RNFL Assessment by OCT
cpRNFL thickness: RNFL thickness is measured by a circular scan around the optic disc. It is displayed on a TSNIT graph, showing a bimodal pattern in the superior and inferior directions in normal eyes 1).
Thickness map: RNFL thickness around the optic disc is displayed as a map from raster scans. It has the highest sensitivity for detecting localized RNFL thinning.
Significance map: Abnormal areas are color-coded by comparison with a normal database. Caution is needed for false positives (red disease).
BMO-MRW: Evaluation of rim width based on Bruch’s membrane opening. It has excellent reproducibility.
Measurement Considerations
Floor effect: In advanced glaucoma, OCT measurements no longer change. Macular parameters show a later floor effect than RNFL thickness 3)6).
Age-related changes: RNFL thickness decreases by approximately 0.5 μm/year with age. Most commercial software does not correct for age 2)3).
Inter-device compatibility: Measurement values are not interchangeable between different OCT devices 1)2)3).
Effect of myopia: In high myopia, RNFL thickness is underestimated due to magnification effects, leading to false positives 1).
ROTA may complement the limitations of these conventional methods by integrating not only RNFL thickness but also reflectance information. In particular, the axial length correction function reduces false positives in myopic eyes, and visualization of a wide range of fiber bundle trajectories is expected to provide analysis less affected by the floor effect.
In a multi-eye study, ROTA, OCT, and red-free photography were compared in 363 patients (531 eyes) with RNFL defects and 177 healthy subjects (315 eyes). The sensitivity of ROTA for glaucoma detection was 98.9% (95% CI: 95.4–100.0%), significantly higher than the 79.3% of red-free photography. The specificity of ROTA was 94.3% (95% CI: 91.3–97.2%), higher than the 87.9% of peripapillary RNFL thickness and 78.1% of GC-IPL.
In a study by Su et al., 600 eyes of patients with ocular hypertension were evaluated. Clinical optic disc examination and OCT analysis showed no RNFL defects in any case, but ROTA analysis detected RNFL defects in 10.8% of eyes. The most common defect location was the superior arcuate fiber bundle. Older age and higher pattern standard deviation were significantly associated with RNFL defects on ROTA.
Involvement of Papillomacular Bundle in Early Glaucoma
In another study by Leung et al., 204 eyes with early glaucoma (MD ≥ −6 dB) were examined. RNFL defects involving the papillomacular bundle were found in 71.6%, and those involving the papillo-foveal bundle in 17.2%. RNFL defects were not limited to the hemiretina but also involved the fovea and macula, and ROTA revealed this extensive neural damage.
Differential diagnosis from non-glaucomatous optic neuropathy
ROTA can depict the loss of RNFL fiber bundles corresponding to areas of optic disc rim pallor. Since non-arteritic anterior ischemic optic neuropathy (NAION) and optic neuritis have characteristic patterns of rim pallor, ROTA is expected to help differentiate them from glaucoma. Even in cases with optic disc drusen or RNFL edema, ROTA may be able to identify RNFL defects.
External validation of ROTA diagnostic performance through large-scale multicenter studies
Implementation of ROTA algorithm in commercial OCT software
Evaluation of glaucoma progression detection ability through longitudinal studies
Integration with AI-based automated diagnosis
QIs ROTA effective in myopic eyes?
A
The ROTA algorithm has a function to correct for axial length, reducing false positives in myopic eyes, which is a problem with conventional OCT analysis. Conventional RNFL thickness measurements are prone to false positives due to magnification effects in high myopia1), but ROTA with axial length correction has been reported to reduce false positive detection of RNFL defects and GC-IPL abnormalities. However, the presence of media opacities can affect accuracy due to reduced image quality.
