Comparison of Macular Thickness Measurements between Time Domain and Spectral Domain Optical Coherence Tomography

Abstract

purpose. To compare macular thickness measurements obtained from time domain optical coherence tomography (OCT) and spectral domain OCT and to evaluate their repeatability and agreement.

methods. Thirty-five healthy normal subjects were included. In one randomly selected eye in each subject, three serial macular measurements were obtained from a time domain OCT (Stratus OCT, Carl Zeiss Meditec, Dublin, CA) and a spectral domain OCT (3D OCT; Topcon, Tokyo, Japan) by an experienced technician in random order. Total and regional macular thicknesses obtained by the two OCTs were compared. Their agreement was examined with Bland-Altman plots. Repeatability (2.77 × within subject SD [Sw]), coefficient of variation (CVw; Sw/overall mean), and intraclass correlation coefficient (ICC) were calculated to evaluate repeatability.

results. Low variability for macular thickness measurements was found in both time domain and spectral domain OCTs. The range of the respective CVw and ICC values were 1.6% to 3.2% and 0.85 to 0.91 for Stratus OCT and 0.6% to 2.4% and 0.92 to 0.99 for 3D OCT. 3D OCT demonstrated better repeatability for total and regional macular thicknesses (all with P ≤ 0.014). The foveal and total macular thicknesses measured by 3D OCT were significantly greater than those measured by Stratus OCT (both with P < 0.001). The spans of 95% limits of agreement for foveal and total macular thicknesses were 33.9 and 21.3 μm, respectively.

conclusions. Although both time domain and spectral domain OCTs are reliable for macular thickness measurements, spectral domain OCT has better measurement repeatability compared with time domain OCT. Macular measurements obtained from the two OCT systems may not be used interchangeably.

Accurate and reliable measurement of the macula is of clinical importance. Measurement of macular thickness is not only important in the diagnosis and monitoring of macular diseases, it has also been found to be useful in evaluating glaucomatous change. ^{1 2 3 4}  

Several imaging instruments are commercially available for measurement of macular thickness. These include the retinal thickness analyzer, confocal scanning laser ophthalmoscope, and optical coherence tomography (OCT). OCT is the most commonly used imaging technology for examination of the macula because it is the only instrument that provides direct visualization of the layered retinal structures. The recent introduction of spectral domain OCT allows imaging of the macula with a much faster scan rate and at a higher scan resolution. Compared with the commercially available time domain OCT (Stratus OCT, Carl Zeiss Meditec, Dublin, CA) which collects 400 axial measurements per second with an axial resolution of approximately 10 μm, the scan rate of spectral domain OCT is at least 18,000 axial measurements per second with an axial resolution of 5 μm. The basic working principle of spectral domain OCT is similar to that of time domain OCT. Both systems measure the echo time delay of backscattered light signals via an interferometer. In time domain OCT, the depth information of the retina is collected as a function of time by moving the reference mirror. The reference mirror in spectral domain OCT, in contrast, is stationary. The light spectrum from the interferometer is detected by a spectrometer. The interference spectrum data are then Fourier transformed to generate axial measurements of the retina. Although spectral domain OCT has been shown to improve visualization of the intraretinal structures, ^{5 6 7} quantitative evaluation of macular thickness has not been validated. It is uncertain if the higher sampling rate and the increased scan resolution of spectral domain OCT could increase its measurement reliability. The purpose of this study was to compare macular thickness measurements obtained from time domain OCT and spectral domain OCT and evaluate their repeatability and agreement. 

Methods

Subjects

Thirty-nine healthy normal volunteers were enrolled and 35 subjects were included in the analysis. All volunteers had visual acuity of at least 20/40, spherical error within the range between +3.0 and −6.0 D and with no evidence of ocular diseases. Subjects with clinical evidence of macular diseases, previous refractive surgery, neurologic diseases, or diabetes were excluded. In one randomly selected eye, three serial measurements were obtained from both time domain and spectral domain OCTs in the same visit by an experience technician in a random order. No mydriatic agent was used during the imaging. The study was conducted in accordance with the ethical standards stated in the 1964 Declaration of Helsinki and approved by the Clinical Research Ethics Committee (Kowloon Central/East) with informed consent obtained. 

Time Domain OCT Imaging

Time domain OCT imaging was performed with the Stratus OCT (Carl Zeiss Meditec Inc., Dublin, CA). A fast macular-thickness scan with six 6-mm linear scans oriented 30° apart in a radial spokelike pattern was acquired in a continuous automated sequence. Each of the six linear scans is composed of 128 equally spaced transverse axial scans. The calculation of macular thickness is based on the 6-mm retinal thickness map analysis printout. The map is composed of nine sectorial thickness measurements in three concentric circles with diameters of 1, 3, and 6 mm (Fig. 1) . The area bounded by the outer (6 mm) and middle (3 mm) circles forms the outer ring while the area bounded by the middle (3 mm) and inner circles (1 mm) forms the inner ring. Each ring is divided into four quadrants, superior, nasal, inferior, and temporal. The central 1-mm circular region represents the foveal area. Each of the scans acquired had a signal strength of at least 8. One subject was excluded because of suboptimal image quality. 

Spectral Domain OCT Imaging

Spectral domain OCT imaging was performed with the 3D OCT (Topcon, Tokyo, Japan). The 3D OCT combines a spectral domain OCT and a nonmydriatic fundus camera (TRC-NW200; Topcon, Tokyo, Japan). The details of the principle of spectral domain OCT have been described. ^{8 9} The 3D OCT uses a superluminescent diode laser with a center wavelength of 840 μm and a bandwidth of 50 nm as a light source. The acquisition rate of the 3D OCT is up to 18,000 A-scans per second. The transverse and axial resolutions are 20 and 5 μm, respectively. The 3D scan protocol was used in this study. This is a raster scan composed of 256 × 256 (vertical × horizontal) axial scans covering a 6 × 6-mm macular region. A built-in correlation-based algorithm is used to cancel axial eye motion artifacts. All the images were obtained with image quality score of at least 60, as recommended by the manufacturer. The 3D OCT provides the same circular macular map analysis as in Stratus OCT which is composed of nine sectorial thickness measurements in three concentric circles with diameters of 1, 3, and 6 mm (Fig. 1) . Three subjects were excluded because of image quality scores less than 60. 

The total (6 mm) and regional macular thicknesses including fovea, temporal inner, superior inner, nasal inner, inferior inner, temporal outer, superior outer, nasal outer thickness, and inferior outer macular thicknesses were analyzed in both OCTs. The total macular thickness was calculated based on the proportional contribution of the regional macular thicknesses as previously described. ^{2 10} The three individual measurements in each eye were used to calculate repeatability (see Table 2 ) whereas the average of the three measurements was used in the comparative and agreement analyses (see Table 1 ). For the calculation of average retinal thickness in the 1-mm central area, 1820 and 128 A-scan measurements were used in 3D OCT and Stratus OCT, respectively. For each of the inner/outer ring sectors, 2860/9651 and 128/192 A-scan measurements were respectively included for 3D OCT and Stratus OCT. 

Statistical Analysis

Statistical analyses were performed with commercial software (SPSS ver. 15.0; SPSS Inc., Chicago, IL). By setting the confidence interval as 20% on either side of the estimate of within-subject SD: Sw {n = 1.96²/[2 × 0.2² × (m − 1)]}, where n is the number of subjects and m is the number of observations, a minimum of 25 subjects is necessary to calculate repeatability. ¹² The repeatability (2.77 × within subject SD [Sw]), coefficient of variation CVw (100 × Sw/overall mean) and intraclass correlation coefficient (ICC) were computed. The Sw was calculated as the square root of the within-subject mean square of error (the unbiased estimator of the component of variance due to random error) in a one-way random-effects model. ¹² The ICC is the ratio of the intersubject component of variance to the total variance (intersubject variance + within subject variance). The within-subject variances of total and regional macular thicknesses obtained by the two OCTs were compared with paired t-test on log-transformed data. ¹¹ The average of macular measurements obtained from the OCTs were compared by paired t-tests. Bland and Altman plots were used to assess agreement. ¹³  

Results

The mean ± SD age of the 35 subjects was 36.4 ± 12.6 years with average spherical equivalent of −1.98 ± 2.26 D. The average image quality score was 9.8 ± 0.4 (signal strength) for Stratus OCT and 67.5 ± 3.5 for 3D OCT. The mean foveal and total average macular thicknesses obtained by Stratus OCT and 3D OCT were 195.6 ± 17.2 and 260.0 ± 12.2 μm and 216.4 ± 18.0 and 263.2 ± 12.6 μm, respectively (Table 1) . Significant differences were found between total and regional (except temporal and inferior inner; superior and nasal outer) macular thicknesses measured by the two OCTs (Table 1) . 3D OCT macular thicknesses generally were thicker than those measured by Stratus OCT. The spans of 95% limits of agreement for foveal and total macular thicknesses were 33.9 and 21.3 μm, respectively (Fig. 2) . 

Table 2presents the repeatability, coefficient of variation (CVw) and intraclass correlation coefficient (ICC) of Stratus OCT and 3D OCT macular measurements. The repeatability of Stratus OCT/3D OCT foveal and total macular measurements were 17.4/14.5 and 12.2 μm/6.3 μm, respectively. The ICC for 3D OCT ranged between 0.92 (foveal thickness) and 0.99 (superior inner, nasal inner and nasal outer) whereas they were between 0.85 (temporal inner and nasal inner) and 0.91 (superior outer) for Stratus OCT. Except for foveal thickness, all the 3D OCT measurements had CVw at or below 1%. Significant differences of within-subject variances were found in total and regional macular thicknesses between Stratus OCT and 3D OCT (all with P ≤ 0.014). 

Discussion

Macular thickness measurements with time domain OCT has been demonstrated to be reliable in previous studies. ^{10 14 15 16} Using Stratus OCT, Polito et al. ¹⁴ showed that the coefficient of repeatability (2 SDs of the difference between pairs of measurements obtained in the same subjects divided by the average of the means of each pair of readings) ranged between 1.68% and 7.43% in a study with 10 healthy subjects. In the study by Gurses-Ozden et al. ¹⁵ of 10 normal subjects, the coefficient of variation (SD of the measurement divided by its mean) of Stratus OCT macular measurements ranged between 4.7% and 6.4% for mean foveal thickness and between 0.7% and 1.1% for 6 mm total macular volume. With a similar study design and sample size, Paunescu et al. ¹⁰ reported that the ICC of total and regional macular thickness measurements varied between 0.55 and 0.97. In agreement with the previous studies, we also observed relatively good repeatability for Stratus OCT macular thickness measurements. The coefficient of variation ranged from 1.6% to 3.2% and the ICC ranged from 0.85 to 0.91. Using spectral domain OCT, we found that the measurement repeatability was even better. All the macular measurements had coefficient of variability at or less than 1% except foveal thickness (CVw = 2.42%). The 6-mm total macular thickness had repeatability of 6.3 μm with coefficient of variation of 0.86%. In other words, the difference in total average macular thickness between two separate measurements obtained by 3D OCT would be less than 6.3 μm in 95% of pairs of observations. Despite the fact that 3D OCT generally had greater macular thickness measurements compared with Stratus OCT, the within-subject variances of 3D OCT were significantly lower (P ≤ 0.014). The better repeatability in 3D OCT may be attributable to its higher scan rate and the increased sampling frames. Compared with Stratus OCT in which 6 × 128 axial scans are acquired in 1.92 seconds to construct the 6 mm macular map, 3D OCT takes 3.5 seconds to perform 256 × 256 axial scans over the same retinal area. The faster acquisition rate and the increased sampling frames allow fine and detailed mapping of the macula in a relatively short period. This is in contrast to Stratus OCT in which the 6-mm macular map is generated based on extrapolation from only six radial scans. It is conceivable that the enhanced measurement reliability of 3D OCT could translate to higher sensitivity in detecting changes during serial monitoring of macular diseases and glaucoma. 

The spans of 95% limits of agreement of macular measurements ranged between 21.3 μm (6 mm total macular thickness) and 38.6 μm (nasal inner macular thickness). For foveal thickness measurement, the mean difference between the 2 OCT instruments was 20.8 μm, and the span of 95% limit of agreement ranged between 3.9 and 37.8 μm (Fig. 2a) . In other words, the difference in foveal thickness measured by the two OCT instruments would lie between 3.9 and 37.8 μm in 95% of pairs of observation. The poor agreement between 3D OCT and Stratus OCT may be attributable to the different definitions of the posterior retinal boundary (Fig. 3) . In Stratus OCT, the inner segment/outer segment (IS/OS) interface of the photoreceptor layer is set as the posterior retinal boundary. Having a higher axial resolution, delineation of the IS/OS photoreceptor interface and the retinal pigment epithelium (RPE) is possible in 3D OCT, and the RPE is set as the posterior retinal boundary. The more posteriorly located reference line resulted in the greater values of macular measurement obtained by 3D OCT. Clinicians should be aware of the differences obtained between the two different OCT systems. Measurement obtained from one system may not be used interchangeably with the other. 

All the images taken in this study were obtained by a trained and experienced technician, and the image quality scores were high. The minimum image quality score was at least 8 (average, 9.7 ± 0.6) for Stratus OCT and at least 61 (average, 67.5 ± 3.5) for 3D OCT. Results obtained in this study may not be directly translated to other populations with different clinical settings. Because of the foveal depression, larger variations in retinal thickness along the foveal curvature would be expected. Small changes in scan position secondary to microsaccade could have resulted in the higher measurement variability of the foveal thickness (coefficient of variation = 2.42%/3.21% for 3D OCT/Stratus OCT) observed in this study. We selected the fast macular scan in Stratus OCT because it is the more commonly adopted scanning protocol in clinical practice. For the new 3D OCT, there are five different protocols for macular scan (256 × 256, 256 × 128, 512 × 128, 512 × 64, and 512 × 32). As the primary objective of the study was to compare the agreement and repeatability between spectral domain OCT and time domain OCT, we selected only one protocol from each instrument, to minimize the scan time in each eye. 

In summary, although there are differences in macular thickness measurements between Stratus OCT and 3D OCT, both systems are reliable for macular thickness measurements. 3D OCT demonstrates better measurement repeatability compared with Stratus OCT. In this regard, spectral domain OCT offers the potential for sensitively detecting changes during serial monitoring of macular diseases and glaucoma. 

 Supported by Carl Zeiss Meditec in the form of instruments (CL and RW) and by Topcon, Inc. in the form of instruments (CL).

 Submitted for publication October 16, 2007; revised March 21, 2008; accepted September 11, 2008.

 Disclosure: C.K. Leung, Carl Zeiss Meditec, Inc. (F, R), Topcon, Inc. (R); C.Y. Cheung, None; R.N. Weinreb, Carl Zeiss Meditec, Inc. (F, C); G. Lee, None; D. Lin, None; C.P. Pang, None; D.S.C. Lam, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Circular macular thickness map. The diameters of the three concentric circles are 1, 3, and 6 mm. Area 1 represents the fovea. Areas 2, 3, 4, and 5 form the inner ring. Area 6, 7, 8, and 9 form the outer ring.

Table 1.

 

Comparison of Regional and Total Macular Thicknesses Measured by Spectral Domain and Time Domain OCT

	3D OCT (mean ± SD)	Stratus OCT (mean ± SD)	P *	Mean Difference (Spectral Domain − Time Domain OCT)	95% Limits of Agreement
Foveal minimum, †	—	155.4 ± 15.8	—	—	—
Fovea thickness	216.4 ± 18.0	195.6 ± 17.2	<0.001	20.8	33.9 (3.9–37.8)
Temporal inner thickness	277.1 ± 12.0	275.1 ± 12.8	0.195	2.0	32.4 (−14.2–18.2)
Superior inner thickness	297.2 ± 14.4	292.0 ± 13.2	<0.001	5.2	24.5 (−7.1–17.4)
Nasal inner thickness	294.2 ± 16.5	288.0 ± 12.5	0.001	6.2	38.6 (−13.1–25.5)
Inferior inner thickness	288.4 ± 13.7	286.5 ± 12.9	0.116	2.0	28.1 (−12.1–16.0)
Temporal outer thickness	244.3 ± 13.3	235.1 ± 13.3	<0.001	9.3	28.5 (−5.0–23.5)
Superior outer thickness	257.5 ± 13.2	258.4 ± 14.9	0.516	−0.9	30.6 (−16.2–14.4)
Nasal outer thickness	276.6 ± 15.9	278.4 ± 15.3	0.246	−1.8	35.2 (−19.4–15.8)
Inferior outer thickness	259.5 ± 14.7	247.3 ± 14.0	<0.001	12.2	26.9 (−2.3–24.6)
Total macular thickness	263.2 ± 12.6	260.0 ± 12.2	<0.001	3.2	21.3 (−7.4–13.9)

*  Paired t-test.

†  Foveal minimum was not reported in the 3D OCT analysis printout.

Figure 2.

 

Bland-Altman plots of the foveal thickness (a) and total macular thickness (b) difference between 3D OCT and Stratus OCT against the average of the two.

Table 2.

 

ICC, Repeatability (2.77 × Sw), and CVw of Total and Regional Macular Thicknesses Obtained with Stratus OCT (upper rows) and 3D OCT (lower rows)

	Foveal Minimum*	Fovea Thickness	Temporal Inner Thickness	Superior Inner Thickness	Nasal Inner Thickness	Inferior Inner Thickness	Temporal Outer Thickness	Superior Outer Thickness	Nasal Outer Thickness	Inferior Outer Thickness	Total Macular Thickness
ICC (95% CI)	0.709 (0.555–0.828)	0.877 (0.796–0.931)	0.850 (0.756–0.915)	0.861 (0.772–0.922)	0.845 (0.748–0.913)	0.883 (0.806–0.934)	0.867 (0.781–0.925)	0.906 (0.842–0.948)	0.855 (0.764–0.918)	0.878 (0.798–0.931)	0.8 89 (0.816–0.938)
	—	0.918 (0.862–0.955)	0.962 (0.934–0.979)	0.985 (0.974–0.992)	0.988 (0.978–0.993)	0.964 (0.939–0.981)	0.975 (0.957–0.986)	0.973 (0.953–0.985)	0.988 (0.979–0.994)	0.969 (0.946–0.983)	0.968 (0.944–0.982)
Repeatability (95% CI)	26.29 (21.93–30.64)	17.41 (14.52–20.29)	14.83 (12.37–17.28)	15.25 (12.73–17.78)	14.77 (12.32–17.22)	12.83 (10.71–14.96)	14.30 (11.93–16.67)	13.52 (11.28–15.76)	17.22 (14.37–20.08)	14.63 (12.20–17.05)	12.16 (10.14–14.17)
	—	14.48 (12.08–16.88)	6.59 (5.50–7.69)	4.77 (3.98–5.56)	5.14 (4.29–5.99)	6.85 (5.71–7.98)	5.95 (4.96–6.93)	6.01 (5.02–7.01)	4.76 (3.97–5.55)	6.99 (5.83–8.14)	6.25 (5.22–7.29)
CVw (%) (95% CI)	6.11 (5.09–7.12)	3.21 (2.68–3.74)	1.95 (1.62–2.27)	1.89 (1.57–2.20)	1.85 (1.54–2.16)	1.62 (1.35–1.89)	2.20 (1.83–2.56)	1.89 (1.58–2.20)	2.23 (1.86–2.60)	2.13 (1.78–2.49)	1.66 (1.41–1.97)
	—	2.42 (2.02–2.82)	0.86 (0.72–1.00)	0.58 (0.48–0.67)	0.63 (0.53–0.74)	0.85 (0.71–1.00)	0.88 (0.73–1.03)	0.84 (0.70–0.98)	0.62 (0.52–0.72)	1.00 (0.84–1.17)	0.86 (0.72–1.00)
P , †	—	0.014	<0.001	<0.001	0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001

*  Foveal minimum was not reported in the 3D OCT analysis printout.

†  Comparison of within-subject variances between Stratus OCT and 3D OCT (paired t test on log-transformed data). ¹¹

Figure 3.

 

In Stratus OCT, the inner segment/outer segment (IS/OS) interface of the photoreceptor layer is set as the posterior retinal boundary (a). In contrast, the RPE is set as the posterior retinal boundary in 3D OCT (b). White outlines: retinal boundaries.