Correlation between fundus autofluorescence and visual function in patients with cone-rod dystrophy

This study aimed to investigate the relationship between autofluorescence (AF) signal measured with ultra-wide field imaging and visual functions in patients with cone-rod dystrophy (CORD). A retrospective chart review was performed for CORD patients. We performed the visual field test and fundus autofluorescence (FAF) measurement and visualized retinal structures with optical coherence tomography (OCT) on the same day. Using binarised FAF images, we identified a low FAF area ratio (LFAR: low FAF/30°). Relationships between age and logMAR visual acuity (VA), central retinal thickness (CRT), central choroidal thickness (CCT), mean deviation (MD) value, and LFAR were investigated. Thirty-seven eyes of 21 CORD patients (8 men and 13 women) were enrolled. The mean patient age was 49.8 years. LogMAR VA and MD were 0.52 ± 0.47 and − 17.91 ± 10.59 dB, respectively. There was a significant relationship between logMAR VA and MD (p = 0.001). LogMAR VA significantly correlated with CRT (p = 0.006) but not with other parameters. Conversely, univariate analysis suggested a significant relationship between MD and LFAR (p = 0.001). In the multivariate analysis, LFAR was significantly associated with MD (p = 0.002). In conclusion, it is useful to measure the low FAF area in patients with CORD. The AF measurement reflects the visual field deterioration but not VA in CORD.


Results
Characteristics of the subjects are shown in Table 1. Thirty-seven eyes ( Table 2).

Discussion
In the current study, the visual field and OCT measurements were performed along with wide-field FAF in patients with CORD. We found that CRT significantly correlated with logMAR VA, suggesting that CRT was the most useful parameter to predict logMAR VA. LFAR correlated with MD, but CRT did not correlate with MD. This implied that the FAF measurement was more useful compared to OCT parameters when analysing visual field deterioration in CORD.
In daily practice with CORD, VA is the most frequently used method to evaluate visual function. The current result that VA was significantly associated with CRT supported the usefulness of this approach. However, VA mainly reflected the visual function around the macula, and more detailed evaluation of retinal sensitivity could not be performed without measuring the visual field. The current results suggested that MD of the visual field correlated with LFAR but not with CRT or CCT. This result was in agreement with that of a previous study by Oishi et al., who suggested that there was a correlation between the area of abnormal FAF and the central scotoma size measured with Goldmann perimetry but not between the area of abnormal FAF and logMAR VA 30 . These findings suggested that the disease status of CORD could not be completely explained simply based on the thicknesses of the retina and choroid and were instead useful to evaluate retinal atrophy, particularly in the www.nature.com/scientificreports/ RPE and photoreceptor cells, with the FAF measurement. However, we did not evaluate the outer retinal structure, such as ellipsoid zone and interdigitation zone, which were reported to be affected in eyes with CORD 31 . It appears that the outer retinal structure rather than CRT might be associated with visual field deterioration.
There are previous reports that suggested the usefulness of measuring the FAF pattern in eyes with CORD. In contrast to low AF resulting from retinal atrophy, particularly in the RPE and photoreceptor cells, the hyper-AF area suggested a degenerative process. The hyperfluorescent AF signal was also important when assessing the disease status of CORD 32,33 . For instance, peripherin-2 (PRPH2) is one of the causative genes of CORD, and the mutation of PRPH2 has been shown to be associated with a speckled pattern of FAF 34 . Other studies reported that FAF imaging in eyes with X-linked RPGR-associated CORD demonstrated parafoveal hyperfluorescent rings, and the size of the rings gradually increased over time 16,35 . Furthermore, hyper-AF was observed in the abnormal area in eyes with RP 36 . Leber congenital amaurosis caused by the Crumb1 gene demonstrates a unique AF pattern in which the AF signal is preserved in the para-arteriolar region 37 . In the current study, to highlight the usefulness of the hypo-AF signal, we calculated LFAR in eyes with CORD even when both hyper-and hypo-AF signals were observed in the abnormal AF area (Fig. 1). It is of our further interest to examine whether similar results are obtained when the abnormal hyperfluorescent area are also analyzed. Furthermore, genetic backgrounds Using the binarised image, the low FAF signal was measured as the sum of black pixels in the presumable abnormal autofluorescence area (yellow circle). Next, LFAR was calculated as the ratio of black pixels in the abnormal FAF area to total pixels within 30° around the macula. www.nature.com/scientificreports/ of CORD were not investigated in the current study, so further studies are required to clarify the relationship between an abnormal FAF area and VF deterioration considering genetic information.
A limitation of the current study was the accuracy of the visual field test. We used HFA to evaluate the visual function in eyes with CORD. However, severe loss of central visual function might hamper accurate visual field tests performed with HFA because of poor fixation. It may be valid to measure the visual field using a microperimeter, in which the position of the retina is tracked, and the same location is stimulated at each target   www.nature.com/scientificreports/ presentation even with the loss of central visual function. Moreover, the sample size was relatively small. A further study involving a larger number of eyes would be required.

Conclusion
In conclusion, it is useful to measure the low FAF area in patients with CORD, as suggested by the significant relationship between LFAR and MD.