Influence of Clinical Factors and Magnification Correction on Normal Thickness Profiles of Macular Retinal Layers Using Optical Coherence Tomography

Abstract

Purpose: To identify the factors which significantly contribute to the thickness variabilities in macular retinal layers measured by optical coherence tomography with or without magnification correction of analytical areas in normal subjects.

Methods: The thickness of retinal layers {retinal nerve fiber layer (RNFL), ganglion cell layer plus inner plexiform layer (GCLIPL), RNFL plus GCLIPL (ganglion cell complex, GCC), total retina, total retina minus GCC (outer retina)} were measured by macular scans (RS-3000, NIDEK) in 202 eyes of 202 normal Asian subjects aged 20 to 60 years. The analytical areas were defined by three concentric circles (1-, 3- and 6-mm nominal diameters) with or without magnification correction. For each layer thickness, a semipartial correlation (sr) was calculated for explanatory variables including age, gender, axial length, corneal curvature, and signal strength index.

Results: Outer retinal thickness was significantly thinner in females than in males (sr2, 0.07 to 0.13) regardless of analytical areas or magnification correction. Without magnification correction, axial length had a significant positive sr with RNFL (sr2, 0.12 to 0.33) and a negative sr with GCLIPL (sr2, 0.22 to 0.31), GCC (sr2, 0.03 to 0.17), total retina (sr2, 0.07 to 0.17) and outer retina (sr2, 0.16 to 0.29) in multiple analytical areas. The significant sr in RNFL, GCLIPL and GCC became mostly insignificant following magnification correction.

Conclusions: The strong correlation between the thickness of inner retinal layers and axial length appeared to result from magnification effects. Outer retinal thickness may differ by gender and axial length independently of magnification correction.

Conflict of interest statement

Competing Interests: The authors of this manuscript have the following competing interests: F, NIDEK (MH, HT), Canon (HT, NY), C, Santen (SO), JAPAN FOCUS (SO), NIDEK (SO, MH, KS, NY), Canon (NY), R, Carl Zeiss (HT), NIDEK (HT, KL, NY), Canon (NY). This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.

Figures

Fig 1. Thickness maps of the ganglion cell complex showing the effects of magnification correction on the analytical area.

A highly myopic eye with an axial length of 28.41 mm (A, B) and a hyperopic eye with an axial length of 21.71 mm (C, D). The 30×30°square scan area is supposed to be 9×9 mm square area for any eye when magnification correction is not considered (A, C). In this situation, the apparent sizes of the analytical area on the recorded images, which is defined by three concentric circles (1-, 3- and 6-mm nominal diameters), are identical regardless of the axial length (A, C). However, the scan area corresponded to a 10.5x10.5 mm (B) or 7.9x7.9 mm (D) square area after magnification correction in eyes with a long or short axial length, respectively. The sizes of three concentric circles were rescaled according to the dimension of the ‘magnification-corrected’ scan area, and defined as the ‘magnification-corrected’ analytical area (B, D). The analytical area became relatively smaller or larger compared to the scan area by magnification correction in eyes with a long or short axial length, respectively (B, D).

Fig 2. The relationship between axial length and thickness changes of retinal layers in the total analytical area after magnification correction.

The thickness changes (μm; A) and relative thickness changes compared to the magnification-uncorrected thickness (%; B) after magnification correction were plotted against axial length. Green squares: retinal nerve fiber layer, Red circles: ganglion cell layer plus inner plexiform layer, Blue circles: ganglion cell complex, Light blue diamonds: total retina, Black triangles: outer retina.

Fig 3. The relationship between axial length and thickness changes of retinal layers in the inner ring of analytical area after magnification correction.

The thickness changes (μm; A) and relative thickness changes compared to the magnification-uncorrected thickness (%; B) after magnification correction were plotted against axial length. Green squares: retinal nerve fiber layer, Red circles: ganglion cell layer plus inner plexiform layer, Blue circles: ganglion cell complex, Light blue diamonds: total retina, Black triangles: outer retina.

Fig 4. The relationship between axial length and thickness changes of retinal layers in the outer ring of analytical area after magnification correction.

The thickness changes (μm; A) and relative thickness changes compared to the magnification-uncorrected thickness (%; B) after magnification correction were plotted against axial length. Green squares: retinal nerve fiber layer, Red circles: ganglion cell layer plus inner plexiform layer, Blue circles: ganglion cell complex, Light blue diamonds: total retina, Black triangles: outer retina.

Fig 5. Semipartial correlation of gender with the thickness of retinal layers in each analytical area with or without magnification correction.

The significance of the semipartial correlation coefficient (sr) and magnitude of the sr squared (sr2) are color coded in each analytical area. S = superior, N = nasal, I = inferior, T = temporal. A, F: retinal nerve fiber layer, B, G: ganglion cell layer plus inner plexiform layer, C, H: ganglion cell complex, D, I: total retina, E, J: outer retina. A—E: ‘magnification-uncorrected’ analytical area, F—J: ‘magnification-corrected’ analytical area. A significant sr was found in the inner ring of inner retinal layers and in all the analytical areas of outer and total retina. The sr2 was relatively large in total and outer retina compared to inner retinal layers. The magnification correction of analytical areas did not cause remarkable changes in the magnitude of sr2 values compared to the sr of axial length with the thickness of retinal layers (Fig 6).

Fig 6. Semipartial correlation of axial length with the thickness of retinal layers in each analytical area with or without magnification correction.

The significance of the semipartial correlation coefficient (sr) and magnitude of the sr squared (sr2) are color coded in each analytical area. S = superior, N = nasal, I = inferior, T = temporal. A, F: retinal nerve fiber layer (RNFL), B, G: ganglion cell layer plus inner plexiform layer (GCLIPL), C, H: ganglion cell complex (GCC), D, I: total retina, E, J: outer retina. A—E: ‘magnification-uncorrected’ analytical area, F—J: ‘magnification-corrected’ analytical area. A large sr2 (≥0.2) was observed in RNFL, GCLIPL and outer retina in ‘magnification-uncorrected’ analytical areas. After magnification correction, most of the sr in the inner retinal layers, especially in GCC, became insignificant. In contrast, the sr were still significant in most of the analytical areas of the outer and total retina after magnification correction.

Fig 7. Partial residual plots showing the relationship between axial length and the average thickness of each retinal layer in total analytical area controlling for the effects of other five explanatory variables.

Component plus residual was plotted against axial length for each retinal layer. A, F: retinal nerve fiber layer, B, G: ganglion cell layer plus inner plexiform layer, C, H: ganglion cell complex, D, I: total retina, E, J: outer retina. A—E: magnification uncorrected analytical area, F—J: magnification corrected analytical area. The other five explanatory variables = age, gender, eye laterality, corneal curvature, and signal strength index.