- Top of page
- GRANTS AND FINANCIAL ASSISTANCE
Purpose: The aim was to evaluate the intra-observer repeatability of the ocular optical quality measurements provided by a double-pass system in healthy eyes.
Methods: A total of 24 healthy eyes of 24 patients with ages ranging from 20 to 60 years were included in the study. Three consecutive measurements were obtained by an experienced examiner with the Optical Quality Assessment System (Visiometrics) of the following parameters defined by the manufacturer as follows: cut-off spatial frequency for the modulation transfer function (COMTF), Strehl ratio, width of the point spread function (PSF) at 10 per cent of its maximal height (PSF10) and width of the PSF at 50 per cent of its maximal height (PSF50). Intra-observer repeatability for 3.0 mm pupil measurements was evaluated by the within-subject standard deviation (Sw) and intra-class correlation coefficient (ICC).
Results: The Sw values for the different parameters evaluated were 4.34 cycles per degree for COMTF, 0.03 for the Strehl ratio, 1.14 arcmin for PSF10 and 0.36 arcmin for PSF50. The ICC values for these parameters were 0.746, 0.627, 0.783 and 0.814 for COMTF, Strehl ratio, PSF10 and PSF50, respectively. Statistically significant correlations were found between COMTF and the Sw for PSF50 (r = -0.45, p = 0.03), and between the Sw and the mean value for PSF50 (r = 0.42, p = 0.04). The significance of these correlations would vanish when considering the Bonferroni correction.
Conclusions: Measurements provided by the Optical Quality Assessment System should be considered and interpreted with caution because their consistency seems to be limited, especially in eyes with poor optical quality. The limitation in the validity of measurements due to the use of infrared light instead of middle-wavelength light should also be considered.
In the human eye, scattering, diffraction and optical aberrations are the phenomena leading to the degradation of optical quality of the visual system.1 Analysis of the optical quality of the eye can provide information about the contribution of these factors to the deterioration of the retinal image and allows the clinician to have a better understanding of the visual complaints of some patients. Numerous systems and devices have been developed as objective tools for the clinical assessment of the optical performance of the eye, such as ocular wavefront analysers or aberrometers,2,3 or double-pass (DP) systems.4,5 Wavefront sensors allow the clinician to characterise only one of the limiting factors for optical quality of the retinal image, wavefront aberrations, considering the combined optical performance of the cornea and crystalline lens.2,3
Different optical bases have been developed for ocular aberrometry, such as the Hartmann–Shack method,6 the Tscherning principle7 or ray tracing.8 There are other optical effects generated by refractive index variations or inhomogeneities on a microscopic scale, specifically of the order of the wavelength of light. This kind of inhomogeneity causes light rays to spread over much larger angles than wavefront aberrations. The effect of this kind of inhomogeneity cannot be detected with the wavefront sensors currently available and it is essential for proper understanding of visual function that they are differentiated from those induced by classical low- and higher-order wavefront aberrations.9
The DP technique has been demonstrated to be a valid method for evaluating retinal image quality for half a century.10–17 This technique is based on the recording of the retinal image after DP through the ocular media and retinal reflection.10–17 From the images obtained with this system, the point spread function (PSF) and the ocular modulation transfer function (MTF) can be calculated. The PSF and MTF derived using this technology are affected not only by ocular wavefront aberrations but also by diffraction and scattering, although it should be noted that DP analysis is limited to relatively small angles.2,14 In addition, the use of middle wavelength (green-yellow) is mandatory for proper assessment of DP images.10,11,18,19 A DP technique provides more information about the real optical performance of the eye than the analysis of wavefront aberrations. A device based on the DP concept and designed for clinical use was developed several years ago, the Optical Quality Assessment System (OQAS; Visiometrics SL, Terrassa, Spain).4 Regrettably, the instrument uses infrared light causing 90 per cent of the recorded light to be an artifact.20 Several studies21–24 evaluated ocular optical performance using the OQAS device after different types of ophthalmological surgical procedures. There are few studies evaluating the consistency and validity of the measurements provided by this device,25,26 with some limitations. The aim of the current study was to evaluate the intra-observer repeatability of the measurements of the ocular PSF and MTF provided by the DP system OQAS in healthy eyes.
- Top of page
- GRANTS AND FINANCIAL ASSISTANCE
The age of the subjects ranged from 20 to 60 years, with only four subjects (16.7 per cent) older than 40 years. A total of 16 women (66.7 per cent) and eight men were evaluated. There were five right eyes (20.8 per cent) and 19 left eyes (79.2 per cent). The mean pupil size at the time of measurement was 4.08 mm, ranging from 3.01 to 7.67 mm. The mean spherical equivalent of the examined subjects was -1.03 D, ranging from -7.08 to +0.50 D.
Table 1 summarises the intra-observer repeatability outcomes for the optical quality parameters obtained for a 3.0 mm pupil with the OQAS. As shown, the Sw values for the different parameters evaluated were around a 10th of its corresponding average values: 12.6 per cent for COMTF; 9.89 per cent for PSF10; 9.89 per cent for PSF50; and 15.79 per cent for the Strehl ratio. The Sw corresponding to the PSF50 was significantly lower than that corresponding to the PSF10 (p < 0.01). The ICC was below 0.82 for all OQAS parameters. The lowest ICC (0.627) was found for the Strehl ratio. In addition, the CV was around 10 per cent for PSF10 and PSF50 and near to 14 per cent for COMTF and the Strehl ratio.
Table 1. Summary of intra-observer repeatability for the optical quality parameters obtained with the OQAS: the within-subject standard deviation (SW), precision, repeatability and intra-class correlation coefficient (ICC) values are shown
|Parameter||Overall mean (range)||Sw||PR||CV (%)||ICC (range)|
|COMTF (cycles/degree)||34.56 (13.73 to 48.81)||4.34||8.51||13.52||0.746 (0.579 to 0.866)|
|PSF10 (arcmin)||11.53 (7.88 to 19.32)||1.14||2.23||9.63||0.783 (0.633 to 0.887)|
|PSF50 (arcmin)||3.64 (2.43 to 6.92)||0.36||0.70||9.43||0.814 (0.679 to 0.904)|
|Strehl ratio||0.19 (0.10 to 0.28)||0.03||0.05||13.98||0.627 (0.417 to 0.794)|
Poor and not statistically significant correlations were found among spherical equivalent and mean Sw values associated with COMTF (r = -0.09, p = 0.69), PSF10 (r = -0.21, p = 0.33), PSF50 (r = -0.13, p = 0.55) and the Strehl ratio (r = -0.14, p = 0.52). The mean value of COMTF was found to be significantly correlated only with the Sw for PSF50 (r = -0.45, p = 0.03) (Figure 1); however, a poor and not statistically significant correlation was found among the Sw for PSF10 and the mean value of COMTF (r = -0.39, p = 0.06), although with the same sign. Furthermore, a statistically significant correlation was found between the Sw and the mean value for PSF50 (r = 0.42, p = 0.04) (Figure 2). Regarding the Strehl ratio, no statistically significant correlations were found among the mean value of this parameter and the Sw values for all of the evaluated OQAS parameters (r ≤ 0.31, p ≥ 0.15). After Bonferroni correction all significant results would vanish.
Figure 1. Scattergram showing the relationship between the within-subject standard deviation (Sw) for the width of the point spread function in minutes of arc at 50 per cent of its maximal height (PSF50) and the modulation transfer function cut-off point (COMTF). The adjusting line to the data obtained by means of the least-squares fit is shown in the graph: Sw for PSF50 = -0.02 × COMTF + 0.88 (R2= 0.20)
Download figure to PowerPoint
Figure 2. Scattergram showing the relationship between the within-subject standard deviation (Sw) for the width of the point spread function in minutes of arc at 50 per cent of its maximal height (PSF50) and the mean value of this parameter. The adjusting line to the data obtained by means of the least-squares fit is shown in the graph: Sw for PSF50 = 0.12 × PSF50 - 0.09 (R2= 0.18)
Download figure to PowerPoint
- Top of page
- GRANTS AND FINANCIAL ASSISTANCE
Several new diagnostic instruments have been developed and introduced into the routine of clinical practice as a consequence of the technological advance in ophthalmology and optics. One of these diagnostic advances is the introduction of the DP technique into clinical practice for the evaluation of optical quality of the eye. Devices based on the DP principle allow the clinician to evaluate the real retinal image and to derive the PSF and MTF corresponding to the evaluated eye. Retinal images evaluated with this procedure are not only affected by wavefront aberrations but also by diffraction and scattering. Therefore, it provides a more complete and realistic evaluation of ocular optical quality. The OQAS from Visiometrics is the only DP device for clinical use that is commercially available. The most important drawback of this technology is the use of infrared light because light of middle wavelength (green-yellow) is mandatory for a proper measurement procedure with a DP system.10,11,18,19 The artifact due to this issue dominates the recording grossly outside the central peak area, particularly limiting estimation of the scattering effect.20 Independent of this limitation, there are previous experiences that attempted to evaluate the precision of some of the measurements provided by this specific intrument in different groups (healthy subjects,25 post-refractive surgery eyes and cataract eyes26). Recently, Vilaseca and colleagues25 reported a good intra-observer repeatability in 20 eyes of 10 healthy subjects, without considering the potential correlation of the values between the two eyes of the same patient. Saad, Saab and Gatinel26 also reported a good intra-observer repeatability of the OQAS parameters in young healthy (15 eyes), post-refractive surgery (seven eyes) and cataract eyes (six eyes) from the outcomes obtained in a relatively limited population and assuming a tolerance of 50 per cent for the repeatability limit for drawing their conclusions. The aim of the current study was to evaluate the intra-observer repeatability of the measurements of ocular PSF and MTF provided by the DP system OQAS in a sufficiently large sample of healthy eyes from different individuals to assess its potential for clinical purposes and to check the consistency with previous reports.
First we evaluated the Sw of three consecutive measurements obtained for each OQAS parameter. Mean Sw values were around one-tenth of its corresponding means, with the lowest value for PSF50. The Sw for the COMTF and Strehl ratio obtained in the present study were quite similar to those reported by Vilaseca and colleagues25 and Saad, Saab and Gatinel,26 which confirms a relative limitation of the consistency of these parameters. In addition, precision coefficients were derived from Sw after confirming that data followed a normal distribution. The intra-observer precision (±1.96 × Sw) indicates the large range of error of the repeated measurements for 95 per cent of observations.28,29 In the present study, the variability of almost all observations was within ±12.0 cycles per degree for the COMTF and within ±0.07 for the Strehl ratio. This magnitude of intra-observer variability for COMTF and the Strehl ratio was in the range of the mean values reported for these parameters in an older population for a 5.0 mm pupil.31 In addition, the coefficients of variation of the repeated measurements for these two parameters were calculated and reached a mean value close to 14 per cent. This indicates that the OQAS device has a relative limitation for the determination of COMTF and the Strehl ratio, which becomes more significant when evaluating the optical quality of ageing patients. Regarding the parameters evaluating the width of the PSF, the Sw was small compared with the means. Specifically, the variability of all observations was within ±3.15 minutes of arc for the PSF10 and within ±0.99 minutes of arc for the PSF50. The CV associated with the repeated measurements of these parameters was below 10 per cent. Therefore, it seems that a better consistency of PSF measurements is achieved with this instrument.
After analysing the variability of OQAS measurements by means of Sw and CV, calculation of the ICC was performed to confirm the repeatability outcomes. The use of ICC is an alternative way of evaluating the relative homogeneity within groups (between the repeated measurements) with respect to the total variation.30 It should be remembered that the range of values for the ICC is between 0 and 1, with the following grading system: ICC less than 0.75 for low intra-observer repeatability; between 0.75 and 0.90 for moderate intra-observer repeatability; and greater than 0.90 for a high rate of intra-observer repeatability.32 By analysing the ICC, it was confirmed that the repeatability of measurements provided by the OQAS in our sample for the Strehl ratio was low and moderate for COMTF, PSF10 and PSF50. Therefore, the relative inconsistency of optical quality measurements provided by the OQAS was confirmed.
Two factors could have accounted for this inconsistency, namely, the tear film and centration of the system. Several recent investigations have shown the role of the dynamics of the tear film in the optical quality of the eye.33–39 After a blink, a gradual increase in optical aberrations occurs in association with the increasingly irregular tear film.33–37 This effect causes a progressive reduction in the optical quality of the eye. Indeed, Ferrer-Blasco and colleagues33 demonstrated that the air-tear film interface MTF profile varies as a function of time post-blink, showing the highest values at six to seven seconds. It should be remembered that the OQAS takes several seconds to obtain the measurement of PSF and ocular MTF and during this period tear film changes might induce some variability in the optical quality measurements. Future studies on the potential limitation of OQAS as a consequence of tear film dynamics should be performed to ascertain the extent of this effect. It seems clear that in patients with dry-eye syndrome, the limitation in optical quality due to tear film problems is more significant40 and the use of artificial tears prior to the measurement should be considered as mandatory. In our series, no patients with dry-eye syndrome were included.
A substantial decrement in ocular MTF can be observed with decentring even with small pupils and appropriate spherocylindrical refractive correction.41 As the measurement with the OQAS takes several seconds, some minor decentration can occur during the process of measurement, which can introduce additional variability. This effect might be minimised by the inclusion of an eye-tracking system controlling the eye position and avoiding measurements taken under eye position instability.
There are doubts about the procedure for calculation of the parameter PSF10, which might be a source of variability. It was not clear how the width of PSF10 was calculated by the software of the OQAS, because at that level the PSF is no longer strictly Gaussian. The PSF at 10 per cent of its maximal height might resemble a non-circular feature that has multiple peaks and therefore it would be difficult to reconstruct and analyse its shape using a simple common algorithm. More information about this issue should be provided by the manufacturer to indicate the potential use of this parameter.
We also investigated the correlation among the optical quality parameters of the eye evaluated using the OQAS and the within-subject standard deviation of the repeated measurements of such parameters. Only two statistically significant correlations, although weak, were found: a negative correlation between COMTF and the Sw for the PSF50 and a positive correlation between the Sw and the mean value for PSF50. This means that more consistent measurements of PSF50 were obtained in those eyes presenting with good optical quality (large COMTF and low PSF50). This suggests that measurements of optical quality with the OQAS in eyes with an associated poor retinal image are less consistent and subject to larger levels of variability. The variability induced by tear film dynamics and decentration in those eyes with poor visual quality could have a more significant impact. This is something that should be addressed in future studies. Furthermore, whether improvements can be made in the computational and mathematical reconstruction of the PSF obtained by this DP device in eyes with poor optical quality should be evaluated.
In conclusion, the measurements provided by the OQAS should be considered and interpreted with caution because their consistency is limited, especially in those eyes with poor optical quality. It is recommended to always perform repeated measurements of ocular optical quality with this instrument, when the results are intended to be used for clinical or research purposes to avoid the inherent variability in the measurements. In addition, the validity of measurements with this device is questionable because of the use of infrared instead of middle wavelength light. Therefore, several improvements are required to optimise the DP system for clincial use. The limiting effects of tear film dynamics and system centration should be addressed and investigated in the future. Furthermore, the trends observed in the present study should be corroborated with larger samples of eyes including subgroups of patients of different ages.