Comparative analysis of morphometric parameters of the retina and optic nerve head, obtained with different types of optical coherence tomographs

Full Text

INTRODUCTION

Optical coherence tomography is the most sought after techniques nowadays for the retinal and optic nerve pathology diagnosis. However, the practical implementation on a large scale of various types of optical coherent tomographs (OCTs) hinders the patients’ follow-up when repeated studies are performed in different institutions using different devices. Se­veral studies have compared the results obtained from various typesof OCTs [1–10]. The findings of these studies have highlighted significant data variability due to differences in the operating principles of modern tomographs and scanning protocols. However, most studies included only a small number of patients, and a comparative thickness analysis was conducted only in the central subfield (1.0 mm zone) [1,  2,  5, 11]. When comparing the morphometric parameters of the optic nerve head (ONH), some researchers concluded that it is inaccurate to compare these figures directly [12]. Some researchers did not find any statistically significant differences in the ONH parameters between the time-domain and spectral-domain tomographs [13]. Significant discrepancies in data obtained using different OCTs [4], multicenter studies, and scientific publications that have not considered the differences between the different types of OCTs, while reporting findings of various OCTs, could lead to misinterpretation and unfounded conclusions. Owing to the variation in published findings, confirmatory studies are warranted.

Therefore, this study aimed to comparatively analyze the measurement results of the thickness of the central zone of the retina, macular volume, and thickness of the retinal nerve fiber layer (RNFL) obtained using three different optical coherence tomography devices.

MATERIALS AND METHODS

To determine the retinal thickness and the macular volume, 90 eyes of 50 healthy volunteers (39 wo­men and 11 men) aged 21–79 years (average age, 55.06 ± 21.28 years; median, 63 years) were examined. Participants with posterior segment diseases, opacification of the optical media, which impedes the visualization of the fundus and interferes with OCT, and high-degree myopia (spherical equivalent of refractive error ≥ –6.25 diopters) were excluded.

In addition to the complete standard ophthalmologic examination, all patients underwent optical coherence tomography of the macular area using three devices: the Stratus OCT 3000 (Carl Zeiss Meditec, USA) using the Fast Macula scan protocol; the Cirrus HDOCT 4000 (Carl Zeiss Meditec, USA) using the Macular cube 128 × 512 scan protocol, and the Spectralis OCT (Heidelberg Engineering, Germany) using the Fast Retina scan protocol. As a result of scanning, a macular map was formed, which consisted of nine sectors. The retinal thickness and the macular volume in each of the nine sectors were compared.

To determine the RNFL thickness of the ONH, 90 eyes of 50 healthy volunteers (39 women and 11 men) aged 21–79 years (average age, 52.91 ± 21.47 years; median, 61 years) were examined. The ONH was evaluated on the Stratus OCT 3000 using the Fast Optic disc scan protocol for tomography; the Cirrus HDOCT 4000 device using the Optic Disc Cube 200 × 200 scan protocol for tomography; and the Spectralis OCT using the RNFL scan protocol for tomography. The RNFL thickness in four sectors: superior (S), inferior (I), nasal (N), and temporal (T) was compared.

Statistical data analysis was performed using the SAS statistical program (version 9.4). Data of thickness and volume parameters measured on the three devices were compared using the mixed-design analysis of variance model. General pairwise comparisons of the individual devices were performed using the Tukey–Kramer test. Descriptive statistics of quantitative variables are presented in the form of means, standard deviations, and minimum and maximum values of indicators.

RESULTS AND DISCUSSION

The mean values of the differences in the measurements of the macular retinal thickness for each of the nine sectors between the devices Spectralis OCT and Cirrus HD-OCT 4000 (Table 1), Cirrus HD-OCT4000 and Stratus OCT 3000 (Table 2), and Spectralis OCT and Stratus OCT 3000 (Table 3), which measured the thickness of the retinal slices, showed significant differences. Since the retinal thickness is measured from the internal limiting membrane (ILM) to the layer of the outer segments of photoreceptors, the smallest values were obtained using the Stratus OCT 3000. The spectral tomograph Cirrus HD-OCT 4000 device measured the retinal thickness from the ILM to the outer border of the pigment epithelium, and the Spectralis OCT measured the retinal thickness from the ILM to the Bruch’s membrane.

 

Table 1 / Таблица 1

Indices of the difference in measurements of retinal thickness in the macular area for each of the 9 zones between Spectralis OCT and Cirrus HD-OCT 4000
Показатели разницы в измерениях толщины сетчатки в макулярной области для каждой из девяти зон между приборами Spectralis OCT и Cirrus HD-OCT 4000

Scope  of research	Difference in indices (µm)	Minimum value (µm)	Maximum value (µm)	Adj P (Tukey-Kramer)
1	16.5556	14.0812	19.0299	< .0001
2	21.1778	18.3146	24.0409	< .0001
3	19.3778	17.2182	21.5374	< .0001
4	18.7667	16.2275	21.3058	< .0001
5	21.7222	17.5172	25.9272	< .0001
6	16.8222	14.7349	18.9096	< .0001
7	15.8444	13.6566	18.0323	< .0001
8	18.4444	16.6003	20.2886	< .0001
9	20.3333	17.884	22.7826	< .0001

 

Table 2 / Таблица 2

Indices of the difference in measurements of retinal thickness in the macular area for each of the 9 zones between Cirrus HD-OCT 4000 and Stratus OCT 3000
Показатели разницы в измерениях толщины сетчатки в макулярной области для каждой из девяти зон между приборами Cirrus HD-OCT 4000 и Stratus OCT 3000

Scope  of research	Difference in indices (µm)	Minimum value (µm)	Maximum value (µm)	Adj P (Tukey-Kramer)
1	39.9	37.4256	42.3744	< .0001
2	34.1111	31.248	36.9743	< .0001
3	35.1222	32.9626	37.2818	< .0001
4	30.9222	28.3831	33.4614	< .0001
5	31.7444	27.5394	35.9494	< .0001
6	28.0889	26.0015	30.1762	< .0001
7	29.0556	26.8677	31.2434	< .0001
8	28.5222	26.6781	30.3664	< .0001
9	31.6333	29.184	34.0826	< .0001

 

Table 3 / Таблица 3

Indices of the difference in measurements of retinal thickness in the macular area for each of the 9 zones between Spectralis OCT and Stratus OCT 3000
Показатели разницы в измерениях толщины сетчатки в макулярной области для каждой из девяти зон между приборами Spectralis OCT и Stratus OCT 3000

Scope  of research	Difference in indices (µm)	Minimum value (µm)	Maximum value (µm)	Adj P (Tukey-Kramer)
1	56.4556	53.9812	58.9299	< .0001
2	55.2889	52.4257	58.152	< .0001
3	54.5	52.3404	56.6596	< .0001
4	49.6889	47.1497	52.228	< .0001
5	53.4667	49.2617	57.6717	< .0001
6	44.9111	42.8238	46.9985	< .0001
7	44.9	42.7122	47.0878	< .0001
8	46.9667	45.1225	48.8108	< .0001
9	51.9667	49.5174	54.416	< .0001

 

The average difference in the thickness of the central subfield between the Cirrus HD-OCT 4000 and the Stratus OCT 3000 was 39.9 ± 2.48 µm, that between the Spectralis OCT and the Cirrus HD-OCT 4000 was 16.55 ± 2.47 µm, and that between the Spectralis OCT and the Stratus OCT 3000 was 56.45 ± 2.05 μm (p < 0001). The overall average difference in the central retinal zone thickness between the Cirrus HD-OCT 4000 and the Stratus OCT 3000 was 32.12 ± 3.77 µm, that between the Spectralis OCT and the Cirrus HD-OCT 4000 was 18.78 ± 2.08 μm, and that between the Spectralis OCT and the Stratus OCT 3000 was 50.9 ± 4.47 μm (p < 0001). These differences must be taken into account when comparing data obtained using diffe­rent OCTs.

It is noteworthy that with low visual acuity and in the absence of central fixation of gaze in a patient, the subsequent repeated scanning of the macular area in the same area is possible only with the same tomography device, having a function that allows for automatic positioning of the scan in the place same as that at the first visit. Moreover, in the advanced stages of certain diseases, due to gross changes in the architecture of the retina, tomographs cannot accurately determine the boundaries of the layers. In such cases, it is inaccurate to compare the results obtained using different devices, and case follow-up of patient should be performed using the same device.

The macular retina volumes obtained using all the three devices were compared. The average retinal volume measured using the Stratus OCT 3000 was 7.1 mm³ (max–min, 7.19–7.01 mm³), that using the Cirrus HD-OCT 4000 was 9.86 mm³ (max–min, 9.95–9.77 mm³), and that using the Spectralis OCT was 8.46 mm³ (max–min, 8.55–8.37 mm³), detailed data is presented in Table 4. The retinal volume obtained using the Cirrus HD-OCT 4000 was larger compared with those obtained using the Stratus OCT 3000 and the Spectralis OCT. This may be attri­buted to the screening protocol of scanning on the Cirrus HD-OCT 4000 that provided for a more detailed macular area map, which leads to an increase in the macular volume, due to the greater number of scans than in other two tomographs. Conversely, the Spectralis OCT provided the largest retinal thickness reading. This discrepancy may be associated not only with the aspects of retinal thickness measurement but also with the difference in the number of slices used for each device, which form the macular area map.

 

Table 4 / Таблица 4

Indices of the difference in measurements of macular volume between Spectralis OCT, Stratus OCT 3000 and Cirrus HD-OCT 4000
Показатели разницы в измерениях объёма сетчатки в макулярной области между приборами Spectralis OCT, Stratus OCT 3000 и Cirrus HD-OCT 4000

Devices	Difference in indices (µm)	Minimum value (µm)	Maximum value (µm)	Adj P (Tukey-Kramer)
Cirrus HD-OCT 4000 vs. Spectralis OCT	1.396	1.3434	1.4486	< .0001
Cirrus HD-OCT 4000 vs. Stratus OCT 3000	2.7638	2.7111	2.8164	< .0001
Spectralis OCT vs. Stratus OCT 3000	1.3678	1.3151	1.4204	< .0001

 

Unlike various principles of determining the retinal thickness on different OCTs, the algorithms for estimating the RNFL thickness are equal. The RNFL thickness is determined on all tomographs on a circular section with a diameter of 3.46 mm. The RNFL thickness between devices was compared in the four sectors, S, I, N, and T. Tables 5–7 represent the average differences in the measurements of the RNFL thickness for each of the four sectors between the Spectralis OCT and the Cirrus HD-OCT 4000, the Cirrus HD-OCT 4000 and the Stratus OCT 3000, the Spectralis OCT and the Stratus OCT 3000, respectively.

 

Table 5 / Таблица 5

Indices of the difference in measurements of retinal nerve fibre layer thickness in 4 zones of peripapillary area between Spectralis OCT and Cirrus HD-OCT 4000
Показатели разницы в измерениях толщины слоя нервных волокон сетчатки в четырёх секторах перипапиллярной зоны между приборами Spectralis OCT и Cirrus HD-OCT 4000

Scope of research	Difference in indices (µm)	Standard deviation	Adj P (Tukey-Kramer)
Superior sector	2.2667	0.9484	0.0468
Nasal sector	–0.5222	1.1573	0.8939
Inferior sector	2.4222	0.9977	0.0426
Temporal sector	4.6556	0.8253	< .0001

 

Table 6 / Таблица 6

Indices of the difference in measurements of retinal nerve fibre layer thickness in 4 zones of peripapillary area between Stratus OCT 3000 and Cirrus HD-OCT 4000
Показатели разницы в измерениях толщины слоя нервных волокон сетчатки в четырёх секторах перипапиллярной зоны между приборами Stratus OCT 3000 и Cirrus HD-OCT 4000

Scope of research	Difference in indices (µm)	Standard deviation	Adj P (Tukey-Kramer)
Superior sector	2.5889	0.9484	0,019
Nasal sector	5.6333	1.1573	< .0001
Inferior sector	7.1667	0.9977	< .0001
Temporal sector	4.8111	0.8253	< .0001

 

Table 7 / Таблица 7

Indices of the difference in measurements of retinal nerve fibre layer thickness in 4 zones of peripapillary area between Stratus OCT 3000 and Spectralis OCT
Показатели разницы в измерениях толщины слоя нервных волокон сетчатки в четырёх секторах перипапиллярной зоны между приборами Stratus OCT 3000 и Spectralis OCT

Scope of research	Difference in indices (µm)	Standard deviation	Adj P (Tukey-Kramer)
Superior sector	0.3222	0.9484	0.9384
Nasal sector	6.1556	1.1573	< .0001
Inferior sector	4.7444	0.9977	< .0001
Temporal sector	0.1556	0.8253	0.9806

 

Differences in the RNFL thickness measured with different devices were assessed, assuming that the indicators of the devices should be the same or their difference should be minimal. The differences in the indicators varied from 0.15 to 7.1 µm. It is plausible that the scatter of the data obtained is due to the inability to position the scan area in the same place as in another tomograph. In the Stratus OCT 3000 and the Spectralis OCT devices, there is no function of automatic detection of the ONH center, and in some cases it is necessary to manually position the scanning area. Consequently, a circular cut of the retina can be performed in the mismatched areas. The absence of a consistent pattern in the identified differences in measuring the RNFL thickness does not allow the creation of a mathematical recalculation algorithm for comparing the results obtained using various OCTs. Thus, case follow-up should only be conducted on the same device.

CONCLUSIONS

Significant differences in the indicators of the retinal thickness and macular volume were established. The difference in the retinal thickness measurement between the Cirrus HD-OCT 4000 and the Stratus OCT 3000 was 32.12 ± 3.77 µm, that between the Spectralis OCT and the Cirrus HD-OCT 4000 was 18.78 ± 2.08 μm, and that between the Spectralis OCT and the Stratus OCT 3000 was 50.9 ± 4.47 μm. These differencesmust be taken into account when comparing data obtained using different OCTs.

The average difference in the measurements of the macular volume between the Cirrus HD-OCT 4000 and the Stratus OCT 3000 was 2.77 mm³, that between the Cirrus HD-OCT 4000 and the Spectralis OCT was 1.40 mm³, and that between the Spectralis OCT and the Stratus OCT 3000 was 1.37 mm³. There was an absence of a consistent pattern of differences in the measurements of the RNFL thickness using various OCTs. In this regard, for individual case follow-up, it is inaccurate to compare the results obtained using different devices.

PRACTICAL RECOMMENDATIONS

1. If it is necessary to compare the results of the retinal thickness and macular volume of the retina in the macular region obtained using various types of OCTs, it should be remembered that the measurements taken may focus on different anatomical structures. The correction factors proposed in this study can be used for an approximate comparison of the data from these studies over time. However, it is noteworthy that with low visual acuity and the absence of central fixation of gaze in a patient, as well as with gross retinal changes, the correction coefficients determined in this study have very limited application because the significance of the comparison between the results in such cases would be extremely low.
2. For case follow-up or multicenter clinical studies, comparison of the results of the measurement of the RNFL thickness obtained using different types of OCTs is unnecessary.

Authors’ contribution:

Yu. S. Astakhov was involved in the concept and design of the study;

S.G. Belekhova performed collection and proces­sing of material, analysis of data obtained, and drafting of the manuscript.

About the authors

Svetlana G. Belekhova

Academician I.P. Pavlov First Saint Petersburg State Medical University

Author for correspondence.
Email: beleksv@yandex.ru

Assistant, Ophthalmology Department

Russian Federation, Saint Petersburg

Yury S. Astakhov

Academician I.P. Pavlov First Saint Petersburg State Medical University

Email: astakhov73@mail.ru

MD, PhD, DMedSc, Professor, Ophthalmology Department

Russian Federation, Saint Petersburg

References

Supplementary files