The objective of the research was to assess the reproducibility of retinal nerve fiber layer (RNFL) and macular thickness using spectral domain optical coherence tomography and to establish whether the same investigator can get the same or similar results when performing the scan thrice in an hour, without reference to the previous scan and the repeat function.
Materials and Methods
In this prospective observational study, 200 subjects who fulfilled the inclusion and exclusion criteria were scanned 3 times according to predefined guidelines at 0, 30 and 60 minutes on the same day, by the same investigator, using spectral domain optical coherence tomography for measurements of RNFL and macular thickness; observations were statistically analyzed and correlated.
In RNFL thickness, the temporal sector showed the worst reproducibility as compared to other sectors. RNFL was the greatest in the superior quadrant and the thinnest in the temporal quadrant. For macular thickness, the temporal sector (mid zone) showed the worst reproducibility, while in the outer zone, the inferior sector showed the worst reproducibility; macular thickness was the thinnest at the central zone (innermost 1-mm ring), the thickest within the inner 3-mm ring and diminished peripherally.
RNFL and macular thickness measurements using spectral domain optical coherence tomography by the same observer at 0, 30 and 60 minutes were very reproducible, except for the sectors specifically mentioned. The greater the thickness of the RNFL in any sector the better was the reproducibility in that sector. For macular thickness, the temporal sector (mid zone) showed the worst reproducibility.
Keywords:Diagnostic Laboratory Technique, Reproducibility, Macular Thickness, Retinal Nerve Fiber Layer, Spectral Domain Optical Coherence Tomography
Copyright & License
Research Article | Pathology
The authors thank to all hospital staff directly or indirectly involved in the study.
Optical coherence tomography (OCT), introduced by Huang et al. in 1991 for the first time, has become an invaluable tool for early diagnosis and follow-up of cases of neurodegenerative disorders like glaucoma, optic neuritis, multiple sclerosis, etc. [1-3]. It allows for non-invasive micrometer resolution of cross-sectional images of the retina in human beings . Macular spectral domain optical coherence tomography (SD-OCT) system provides highly reproducible scan results for early detection of glaucoma . It also increases the responsibility of ophthalmologists to be confident in minor axonal loss results so as not to initiate undue lifelong treatment in suspected and properly diagnosed cases based on OCT results.
The fundamental principle of OCT is to measure the echo time delay of reflected infrared light with an interferometer and a low-coherence light source. The axial resolution of the stratus time-domain OCT (TD-OCT) system is achieved at around 10 µm, which is insufficient to detect early changes in the retinal nerve fiber layer (RNFL), since peripapillary RNFL thickness is less than 200 µm [5, 6].
SD-OCT, known variously as high-definition OCT or Fourier-domain OCT, is a spatially encoded frequency domain OCT system having advantages and improvements over TD-OCT systems. The depth scan is immediately converted to spectral information by Fourier transformation without movement of the reference arm. Hence, SD-OCT provides fast and detailed structural information as compared to other available ophthalmic equipment. It provides significantly improved scan coverage, image resolution (around 6 µm), imaging speed and retinal segmentation algorithms as compared to conventional TD-OCT system [5, 6]. In addition, it provides three-dimensional (3D) cubic data. When analyzing OCT scans, reproducibility of results is very important for diagnosis and assessing progression, regardless of the imaging instrument used. Though the systems are computerized and programmed to evaluate the scans automatically, yet the role of the investigator/operator is very significant. Retinal tissue segmentation/reference marking is pre-programmed (automatic segmentation), but the investigator needs to check it visually and make manual correction of the segment, if necessary. Single or multiple operators need to be careful in their assessment of the scans. The OCT software can identify previous scan locations (follow-up mode) and guide the OCT system to scan the same locations repeatedly during follow-up visits . Thus, first time OCT scanning is of paramount importance to establish a good baseline confidence. Assessment of reproducibility of RNFL and macular thickness using OCT is of utmost importance as reproducibility affects both accuracy and the ability to monitor disease progression.
In this study on Indian healthy subjects, a relatively large sample size was taken to assess the reproducibility of RNFL and macular thickness using Cirrus SD-OCT under the same conditions (same machine and investigator) without using the repeat function.
The objective of the research was to establish whether the same investigator can get the same or similar results when performing the scan thrice in an hour, without reference to the previous scan.
Materials and Methods
After calculating the minimum sample size required at 80% power and 5% significance level, 200 (400 eyes) subjects of Indian origin, at least 40 subjects of each gender, who were treated at the Ophthalmology Department of Mohan Eye Institute, a tertiary care teaching hospital in New Delhi, India, were included in this prospective observational study. The ophthalmology assessment of study subject was done from December 2014 to December 2015.
Written informed consent was obtained from all subjects before inclusion in the study. The information sheet approved by the Ethics Committee was given to the participants to obtain their consent. It was then duly signed and dated by both the researcher who obtained the consent and the participant. History was taken to rule out previous retinal or choroidal pathology and any other intraocular intervention.
The subjects were assessed regarding:
distance visual acuity;
slit lamp examination;
Goldmann applanation tonometry.
After pharmacological dilation of the pupils with phenylephrine 5% + tropicamide 0.8% ophthalmic solution eye drops and instillation of artificial tears, each subject was scanned 3 times (at 0, 30, and 60 minutes) on the same day, by the same investigator, using the Cirrus SD-OCT 400 machine (Fig. 1). The scans of each individual subject were categorized independently as A (taken at 0 minutes), B (taken at 30 minutes), C (taken at 60 minutes). These scans were then correlated and analyzed for the study. All scans had an image quality factor of 50/100 or greater.
Fig. 2 shows printouts with measurements obtained from one of the patients taken from the optic nerve head (ONH) and RNFL areas, using one of the commercially available SD-OCT instruments, the Cirrus-OCT (Carl Zeiss Meditec Inc, Dublin, CA, USA) system . For RNFL thickness measurements (Fig. 2), the Optic Disc Cube 200x200 scan acquisition protocol was used. In this protocol, a 3.4-mm-diameter circular scan centered on the optic disc is obtained. Cirrus SD-OCT presents RNFL thickness on two circular charts, one with 12 equal sectors, each representing one clock hour, and the other with four equal 90° sectors, each representing one quadrant. The chart displays RNFL thickness in micro-meters (μm) and the average RNFL thickness [9, 10].
Fig. 3 shows printouts with measurements taken from one of the patients obtained for macular thickness measurement, using one of the commercially available SD-OCT instruments, the Cirrus-OCT (Carl Zeiss Meditec Inc, Dublin, CA, USA) system . For macular thickness measurement (Fig. 3), the Macular Cube 512x128 protocol was used. According to the Early Treatment Diabetic Retinopathy Study (ETDRS) grid, the macula is divided into 9 regions with 3 concentric rings with diameters of 1 mm (innermost ring), 3 mm (inner ring) and 6 mm (outer ring) centered at the fovea. The 1-mm innermost ring is the fovea (central zone), while the 3-mm inner ring (mid zone) and 6-mm outer ring (outer zone) are further divided into four equal regions [8, 10]. For this type of acquisition, the patient must fixate on the target for 2.4 seconds. During scanning, the screen shows the operator an external view of the eye, a real-time fundus image, OCT images of the central crosshair, and the top and bottom B-scans. After capture, the “Review” screen provides qualitative information on the scan. If a subject blinks during the scan, the horizontal segments will appear black on the OCT image. If a subject loses fixation, saccades will be present where the blood vessels are not contiguous. If blinks or numerous artifacts are present, the operator clicks the “Try Again” button to return to the “Scan Acquisition” screen [8, 11].
The collected data were entered in a MS EXCEL spreadsheet and statistical analysis was done, using Statistical Package for the Social Sciences (SPSS) version 21.0.
Statistical tests were applied as follows:
Categorical variables were presented in numbers and percentages (%) and continuous variables were presented as mean ± SD and median.
The square root of the mean within subject variance was the common standard deviation of the repeated measurements.
Reliability analysis using a one-way random model was used to determine the intraclass correlation coefficient (ICC).
The coefficient of variation (COV) and test-retest variability were calculated to quantify the repeatability.
A p-value of < 0.05 was considered statistically significant.
There were analyzed 400 eyes of 200 (90 males, 110 females) healthy subjects. Forty eyes were excluded due to low signal strength of images, and 14 eyes were excluded due to blinks during scanning. The mean patients’ age was 40.70 ± 12.53 years (from 16 to 60 years). The first reading was labeled as baseline reading (Table 1, 2).
Here, the COV is the variation in values between the subjects at first reading. Table 3 tabulates change in values of RNFL thickness measurements at 30 and 60 min as compared to the baseline values of 0 min.
Table 4 shows change in values of macular thickness measurements at 30 and 60 min as compared to the baseline values of 0 min.
Table 5 demonstrates that for RNFL thickness, the temporal sector showed the worst reproducibility as compared to other sectors. For macular thickness, the temporal sector (mid zone) showed the worst reproducibility, while in the outer zone, the inferior sector showed the worst reproducibility.
Our sample size of 400 eyes of healthy subjects was reasonably large and statistically acceptable. Cirrus SD-OCT has 840-nm super luminescent diode as an optical source which acquires 27,000 A-scans/sec. Due to the high acquisition speed of Cirrus SD-OCT, image capture is possible at extremely low light exposures. The power incident of SD-OCT scan on the eye is less than 725 µm; therefore, repeated measurements on normal healthy eyes is safe. This power incident is within the maximum permissible exposure limit for continuous exposure at that wavelength . Cirrus SD-OCT shows probability code results for RNFL thickness using a white-green-yellow-red color code. For instance, when the thinnest 1% of normal age-matched population has similar RNFL thickness, the red code (‘outside normal limits’) is indicated. Yellow code represents ‘suspect’ (1% ≤ yellow < 5%); green code represents ‘normal’ (5% ≤ green ≤ 95%); white code represents the thickest 5% of the population (white > 95%). Theoretically, if RNFL thickness measurements are stable, the probability code results will also be stable .
In our study, the baseline values of the mean RNFL thickness at 0 min were as follows: 90.06 µm - the global RNFL thickness; 84.97 µm - the RNFL thickness symmetry; 121.2 µm in the superior quadrant; 115.63 µm in the inferior quadrant; 70.86 µm in the nasal quadrant; 54.72 µm in the temporal quadrant. Table 5 shows the intersession ICC, COV, and test-retest variability of three scans performed on a single day for the RNFL and macular measurements, respectively. For the average RNFL thickness measurements, the ICC was 0.991, the COV was 1.01%, and the test-retest variability was 1.8 µm. For the RNFL symmetry, the ICC was 0.993, the COV was 1.18%, and the test-retest variability was 2.00 µm. For quadrants, the ICC ranged from 0.991 (superior quadrant) to 0.995 (inferior quadrant) and 0.995 (nasal quadrant) to 0.987 (temporal quadrant). The COVs were 1.46% (superior quadrant), 1.25% (inferior quadrant), 1.38% (nasal quadrant) and 2.38% (temporal quadrant). The test-retest variability ranged from 3.49 µm (superior quadrant) to 2.88 µm (inferior quadrant) and 1.90 µm (nasal quadrant) to 2.57 µm (temporal quadrant). Therefore, our study showed that in RNFL thickness, the temporal sector showed the worst reproducibility as compared to other sectors. RNFL thickness measurements using SD-OCT by the same observer on the same day at 0, 30 and 60 minutes were very reproducible in normal healthy eyes; except for the sectors specifically mentioned.
Samin Hong et al. conducted a similar study in 2010 using Cirrus SD-OCT and have found that the reproducibility of RNFL thickness in healthy eyes was excellent (the ICC was 0.970, the COV was 2.38%, the test-retest variability was 4.5 µm) as compared to the previous reports for Stratus TD-OCT, which is in line with our study results. However, the variability was relatively higher in the nasal area, which is inconsistent with our study, and may be attributed to the factors associated with study sample size, geographic and demographic variations .
D. Cabrera DeBuc et al. scanned RNFL thickness using SD-OCT for 3 sessions with a 30-second rest in between to study the inter-session repeatability and identify the pitfalls affecting the reliabilities. They have found that the values of the coefficient of repeatability (CR) < 5μ m and the COV of 5% were revealed in the outer layers only. The values of the COV were not significantly different in the unregistered scanning session. They have concluded that the rotations in the unregistered scanning sessions did not cause significant change in repeatability. In our study, we have found the highest COV value in the temporal sector (2.38 ± 1.55 %); the superior sector had the highest test-retest variability (3.49 ± 1.88 µm); the values of the COV were not significantly different (p < 0.05) in the three scanning sessions, suggesting very good reproducibility and repeatability. These results can play a superior role in arguing the stability of SD-OCT results .
In our study, the ICC values of Cirrus SD-OCT for quadrants ranged from 0.98 to 0.99. In previous studies conducted by Sull AC et al., Chan A et al., using Stratus TD-OCT, they ranged from 0.67 to 0.97. These variations between the two systems were more prominent in the nasal sector as the nasal ICC values of Stratus TD-OCT (from 0.67 to 0.88) were lower than those of Cirrus SD-OCT (0.99), and the nasal test-retest variability of Stratus TD-OCT (16.0 µm) was much larger than that of Cirrus SD-OCT (1.9 µm) [17, 18]. Stratus TD-OCT selects the inner segment/outer segment junction as the outer retinal boundary for macular thickness measurements, whereas the SD-OCT system selects the retinal pigment epithelium as the outer retinal boundary for thickness measurements, thus, leading to an increase in macular thickness reported with these OCT systems as compared to the TD-OCT systems .
Chen X et al. compared the repeatability and reproducibility of the axial and lateral retinal measurements, using handheld OCT systems and a tabletop OCT system. Three OCT systems were used: handheld Leica Envisu, investigational handheld swept-source OCT (UC3), and Heidelberg Spectralis tabletop system. All three OCT systems have been found to have good repeatability and reproducibility with the COV less than 8.5% for the RNFL thickness . Our results have shown the COV < 1.5 % for the average RNFL thickness in the test-retest measurements and high reproducibility for the SD-OCT system.
In our study, for macular thickness central zone, the ICC was 0.989, the COV was 0.84%, the test-retest variability was 3.89 µm. For the mid zone, the ICC ranged from 0.996 (superior sector) to 0.994 (inferior sector) and 0.999 (nasal sector) to 0.991 (temporal sector). The COVs were 0.68% (superior sector), 0.59% (inferior sector), 0.49% (nasal sector) and 0.90% (temporal sector). The test-retest variability ranged from 4.15 µm (superior sector) to 3.70 µm (inferior sector) and 2.93 µm (nasal sector) to 5.32 µm (temporal sector). For the outer zone, the ICC ranged from 0.993 (superior sector) to 0.932 (inferior sector) and 0.984 (nasal sector) to 0.993 (temporal sector). The COVs were 0.40% (superior sector), 1.10% (inferior sector), 0.64% (nasal sector) and 0.57% (temporal sector). The test-retest variability ranged from 2.25 µm (superior sector) to 5.88 µm (inferior sector) and 3.79 µm (nasal sector) to 2.88 µm (temporal sector).
Therefore, our study showed that for macular thickness, in the mid zone, the temporal sector showed the worst reproducibility and in the outer zone, the inferior sector showed the worst reproducibility.
Demographic variations can be important parameters when comparing macular thickness measurements and diagnosing and monitoring macular pathologies. Natung T et al. (2016) studied macular thickness in healthy Indian eyes, using Zeiss SD-OCT. The mean central subfield thickness of all subjects was 240.40 ± 18.26 μ m. Overall, the nasal quadrant was the thickest followed by the superior, inferior, and temporal subfields . We have found the similar results in the outer zone with extremely high level of the ICC (0.932-0.993) in the repeated measurement test.
According to García-Franco R et al., who studied Mexican healthy population at the age of 18-70 years in 2020 using Huvitz OCT, normal macular thickness values in the foveal macular region of the subjects studied were thinner than values reported in other populations. Using the ETDRS grid, the mean central subfield thickness was 227.4± 18.9 µm; macular thickness was thicker in the inner ring than in the outer ring . The results of our study are in an agreement with the data obtained by García-Franco R et al. except for the nasal region where the average macular thickness values were similar in the test-retest measurements. However, the range of values in the nasal outer zone was wider than that in the inner one. These differences were probably caused by study sample size, demographic and geographic factors.
Wang KL et al. evaluated the variability and reproducibility of central foveal thickness measurements using handheld SD-OCT in supine infants versus conventional adult tabletop SD-OCT imaging. The authors have found that handheld SD-OCT was a reproducible instrument to measure foveal thicknesses in supine infants . The results obtained in our study have shown a high level of SD-OCT result reproducibility in adult patients. According to our results, SD-OCT is a high-precision clinical laboratory instrument for diagnosing pathological conditions in adult and pediatric patients.
RNFL and macular thickness measurements using SD-OCT by the same observer on the same day at 0, 30 and 60 minutes were very reproducible in normal healthy eyes; except for the sectors specifically mentioned.
RNFL thickness in the temporal sector had the worst reproducibility as compared to other sectors.
For macular thickness, the temporal sector (mid zone) showed the worst reproducibility, while in the outer zone, the inferior sector showed the worst reproducibility.
Our study has shown that in RNFL thickness evaluation, the temporal sector had the worst reproducibility, which was naturally thin as compared to other sectors. Therefore, the following question arises: “Will the reproducibility be good in diseases such as glaucoma and neurodegenerative disorders in which RNFL thinning occurs?” This suggests that reproducibility in diseased states needs to be evaluated separately.
This research study on human subjects was conducted in accordance with the ethical standards of the institutional and national research committees and according to the WMA Declaration of Helsinki – Ethical Principles for Medical Research Involving Human Subjects. The research study protocol was approved by the Institutional Ethics Committee - Letter No. MEI/IEC/2014/71 Dated 02/12/2014.
The authors thank to all hospital staff directly or indirectly involved in the study.
Conflict of interest: The authors declare that no conflicts exist.
Financial Disclosure: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Huo YJ, Thomas R, Li L, Cao K, Wang HZ, Wang NL. Comparison of peripapillary retinal nerve fiber layer thickness, functional subzones, and macular ganglion cell-inner plexiform layer in differentiating patients with mild, moderate, and severe open-angle glaucoma. Journal of Glaucoma. 2020;29(9):761–766. Available from: https://doi.org/10.1097/IJG.0000000000001598
Karvonen E, Stoor K, Luodonpää M, Hägg P, Lintonen T, Liinamaa J, et al. Diagnostic performance of modern imaging instruments in glaucoma screening. British Journal of Ophthalmology. 2020;104(10):1399–1405. Available from: https://doi.org/10.1136/bjophthalmol-2019-314795
García-Bella J, Martínez De La Casa JM, Talavero González P, Fernández-Vigo JI, Valcarce Rial L, García-Feijóo J. Variations in retinal nerve fiber layer measurements on optical coherence tomography after implantation of trifocal intraocular lens. European Journal of Ophthalmology. 2018;28(1):32–35. Available from: https://doi.org/10.5301/ejo.5001028
Chopra R, Wagner SK, Keane PA. Optical coherence tomography in the 2020s—outside the eye clinic. Eye. 2021;35(1):236–243. Available from: https://doi.org/10.1038/s41433-020-01263-6
Mohammadzadeh V, Fatehi N, Yarmohammadi A, Lee JW, Sharifipour F, Daneshvar R, et al. Macular imaging with optical coherence tomography in glaucoma. Survey of Ophthalmology. 2020;65(6):597–638. Available from: https://doi.org/10.1016/j.survophthal.2020.03.002
Aumann S, Donner S, Fischer J, Müller F. Optical coherence tomography (oct): principle and technical realizationHigh Resolution Imaging in Microscopy and Ophthalmology. 2019. p. 59–85. Available from: https://doi.org/10.1007/978-3-030-16638-0_3
Dong ZM, Wollstein G, Schuman JS. Clinical utility of optical coherence tomography in glaucoma. Investigative Opthalmology & Visual Science. 2016;57(9):OCT556. Available from: https://doi.org/10.1167/iovs.16-19933
Lumbroso B, Rispoli M. Guide to interpreting spectral domain optical coherence tomography. 2011. Available from: https://doi.org/10.5005/jp/books/11328
Bhende M, Shetty S, Parthasarathy M, Ramya S. Optical coherence tomography: a guide to interpretation of common macular diseases. Indian Journal of Ophthalmology. 2018;66(1):20. Available from: https://doi.org/10.4103/ijo.IJO_902_17
Komaratih AE. Early diagnosis glaucoma using spectral domain OCT. Adv Ophthalmol Vis Syst. 2020;10(3):65–68. Available from: https://medcraveonline.com/AOVS/early-diagnosis-glaucoma-using-spectral-domain-oct.html
Xu X, Guo X, Xiao H, Mi L, Chen X, Liu X. Reproducibility of macular ganglion cell–inner plexiform layer thickness in normal eyes determined by two different OCT scanning protocols. BMC Ophthalmology. 2017;17(1):37. Available from: https://doi.org/10.1186/s12886-017-0434-2
Z136 Committee. American national standard for safe use of lasers: ANSI Z136.1-2000. New York: Laser Institute of America; 2017. Available from: https://eliceirilab.org/sites/default/files/2017-04/American%20National%20Standard%20for%20Safe%20Use%20of%20Lasers.pdf
Sampani K, Abdulaal M, Peiris T, Lin MM, Pitoc C, Ledesma M, et al. Comparison of SDOCT scan types for grading disorganization of retinal inner layers and other morphologic features of diabetic macular edema. Translational Vision Science & Technology. 2020;9(8):45. Available from: https://doi.org/10.1167/tvst.9.8.45
Claessens D, Schuster AK, Krüger RV, Liegl M, Singh L, Kirchhof B. Test-retest-reliability of computer-based metamorphopsia measurement in macular diseases. Klinische Monatsblätter für Augenheilkunde. 2021;238(06):703–710. Available from: https://doi.org/10.1055/a-1252-2910
Hong S, Kim CY, Lee WS, Seong GJ. Reproducibility of peripapillary retinal nerve fiber layer thickness with spectral domain cirrus high-definition optical coherence tomography in normal eyes. Japanese Journal of Ophthalmology. 2010;54(1):43–47. Available from: https://doi.org/10.1007/s10384-009-0762-8
Cabrera Debuc D, Tian J, Bates N, Somfai GM. Inter-session repeatability of retinal layer thickness in optical coherence tomography. Medical Imaging 2017: Biomedical Applications in Molecular, Structural, and Functional Imaging. 2017. Available from: https://doi.org/10.1117/12.2254640
Sull AC, Vuong LN, Price LL, Srinivasan VJ, Gorczynska I, Fujimoto JG, et al. Comparison of spectral/fourier domain optical coherence tomography instruments for assessment of normal macular thickness. Retina. 2010;30(2):235–245. Available from: https://doi.org/10.1097/IAE.0b013e3181bd2c3b
Chan A. Normal macular thickness measurements in healthy eyes using stratus optical coherence tomography. Archives of Ophthalmology. 2006;124(2):193. Available from: https://doi.org/10.1001/archopht.124.2.193
Adhi M, Aziz S, Muhammad K, Adhi MI. Macular thickness by age and gender in healthy eyes using spectral domain optical coherence tomography. PLoS ONE. 2012;7(5):e37638. Available from: https://doi.org/10.1371/journal.pone.0037638
Chen X, Tai V, Mcgeehan B, Ying G, Viehland C, Imperio R, et al. Repeatability and reproducibility of axial and lateral measurements on handheld optical coherence tomography systems compared with tabletop system. Translational Vision Science & Technology. 2020;9(11):25. Available from: https://doi.org/10.1167/tvst.9.11.25
Natung T, Keditsu A, Lyngdoh LA, Dkhar B, Prakash G. Normal macular thickness in healthy Indian eyes using spectral domain optical coherence tomography. Asia-Pacific Journal of Ophthalmology. 2016;5(3):176–179. Available from: https://doi.org/10.1097/APO.0000000000000153
García-Franco R, Méndez-Marín D, García-Roa M, Ramirez-Neria P, Valera-Cornejo D, Lansingh VC. Central macular thickness in a healthy Mexican population using huvitz optical coherence tomography>. Clinical Ophthalmology. 2020;Volume 14:3931–3940. Available from: https://doi.org/10.2147/OPTH.S272431
Wang KL, Chen X, Stinnett S, Tai V, Winter KP, Tran-Viet D, et al. Understanding the variability of handheld spectral-domain optical coherence tomography measurements in supine infants. PLOS ONE. 2019;14(12):e0225960. Available from: https://doi.org/10.1371/journal.pone.0225960