Fundus autofluorescence abnormalities can predict fluorescein angiography abnormalities in patients with chronic central serous chorioretinopathy

Central serous chorioretinopathy (CSC) is a chorioretinal disease, characterized by a serous retinal detachment due to accumulation of subretinal fluid (SRF), typically affecting middle-aged male patients [1]. In this disease, choroidal thickening and hyperpermeability have been hypothesized to lead to dysfunction of the retinal pigment epithelium (RPE) and serous fluid leakage into the subretinal space [2].

In most cases of acute CSC, SRF resolves spontaneously within 4 to 6 months, often without residual visual impairment [3]. Chronic CSC (cCSC), characterized by persistent or recurrent accumulation of SRF, may lead to severe and permanent visual symptoms and to decreased quality of life [4, 5]. Despite the fact that indocyanine green angiography (ICGA) reveals important information on CSC pathogenesis by visualizing the underlying widespread choroidal abnormalities, fluorescein angiography (FA) is still the most widely used technique to obtain information on the area and morphology of RPE abnormalities and the leakage through the RPE in CSC [1].

Several studies described using short-wave fundus autofluorescence (FAF) for the evaluation of retinal abnormalities in CSC patients and compared them with other imaging modalities [6‐13]. Next to optical coherence tomography (OCT)—providing accurate information on the accumulation of SRF and intraretinal fluid—FAF was also found to establish the spectrum of pathological retinal changes noninvasively. Using FAF, different patterns of RPE alterations can be distinguished, and this imaging modality can be used to facilitate monitoring disease progression in different phenotypes of CSC [6‐13].

For CSC, several patterns of increased and decreased areas of autofluorescence were described in previous studies. The various autofluorescence intensity changes depend on the type of RPE alterations and the chronicity of disease (generally, the more chronic a case is, the more RPE alterations are observed) [12, 13]. Furthermore, in some cCSC cases with residual SRF, a significant discrepancy is observed in the size of the areas of hyperfluorescent abnormalities on FAF versus FA [14]. The aim of the present study was to assess the possible correlation between the findings on FAF and FA in cCSC patients with FAF and FA imaging discrepancies (i.e., difference in size of hyperfluorescent RPE abnormalities on these modalities) at first visit. This study also assessed the predictive ability of FAF for structural hyperfluorescent RPE abnormalities measured by FA over time.

In the current study, we aimed to find out whether findings on short-wave FAF at baseline could correlate with or predict FA findings in cCSC patients on the long-term. In our retrospective study, we analysed 92 eyes with cCSC with a notable discrepancy in the size of corresponding hyperfluorescent alterations areas on FAF (HF-FAF) and hyperfluorescent alterations on FA (HF-FA) at baseline. We observed that the discrepancy in the size of these areas significantly diminished during the follow-up period, and this decrease independently correlated with the follow-up duration, and the size of HF-FAF at first visit, but not with invasive treatment approaches. Also, no relationship was found with SRF persistence, possibly due to the partial absence of OCT data during the follow-up period in 24% of the cases. Based on current data, we proposed a model which predicted the enlargement of hyperfluorescent RPE abnormalities on FA at final visit in 82.6% of cases, based on HF-FAF abnormalities at first visit.

Different authors previously stated the possibility of using FAF along with other imaging modalities and described correlations between FAF and FA to differentiate between acute CSC and Ccsc [7, 8, 10, 12, 14, 17‐19]. It was noted that during the first screening in 72–96% of acute CSC patients, a hypoautofluorescence was registered that corresponded with the leakage site on FA and to the area of serous retinal detachment, which was visualized as a blockage of fluorescein on FA by SRF [9, 12, 13, 20]. At the same time, cCSC patients with persistent SRF showed hyperautofluorescence on FAF imaging that tended to increase in intensity in time (which we refer to as HF-FAF in the current study) [6, 14, 19, 20]. Framme et al. noted that irregular and increased hyperautofluorescence surrounding the original leakage point was registered in about one-fifth of cCSC patients with long-lasting SRF.¹⁸ These abnormal hyperautofluorescent lesions were also found in the periphery in 57% of (presumably chronic) CSC cases using ultra-widefield FAF, proving the statement that cCSC patients often do not realize when an extramacular serous retinal detachment occurs [20‐22].

A possible mechanism has been reported to describe the phenomenon of increased HF-FAF. It was stated that an increased autofluorescence is caused by a higher metabolic activity of the RPE accompanied by accumulation of photoreceptor debris and macrophages, containing fluorophores such as lipofuscin [6, 14, 19]. This abnormal accumulation at the level of the outer segments of photoreceptors and RPE usually precede their degradation and thus can be of predictive value for detection of RPE atrophy in a later disease state, which can occur within the course of CSC [23]. Framme et al. also noted that a clear demarcation of the leakage area based on hyperfluorescent abnormalities was only seen in those patients with more extensive leakage, indicating larger defects of the RPE. These well-defined areas of increased HF-FAF, mostly apparent inferior to the site of RPE changes, were reported in cCSC cases with persistent SRF for more than 6 weeks [24]. We observed the same phenomenon in our study, which underlines the hypothesis that RPE autofluorescence must be enhanced in those areas, because FAF intensity is significantly stronger. An increased HF-FAF in our study was observed even in cases with residual SRF, which should act as a blockade of autofluorescence. HF-FAF in acute CSC generally recovers in 4 to 6 months after the SRF resolution, which is in contrast with findings in cCSC [18].

It was previously stated that hypoautofluorescent abnormalities are associated with worse BCVA and retinal sensitivity, whereas hyperautofluorescent were associated with less RPE damage and, therefore, presumably with a better visual outcome [25], although reduced retinal sensitivity has been reported in cases with hyperautofluorescent RPE changes [17, 25, 26]. In our study, median LogMAR BCVA decreased, although insignificantly, from 0.3 (− 0.1–1.3) at first visit to 0.4 (− 0.1–1.3) at last visit (p > 0.05). The RPE alterations on final FAF images in our study could not be interpreted due to the relatively frequent occurrence of a mixed (hypo- and hyperautofluorescent abnormalities) pattern. Therefore, a correlation between FAF abnormalities and BCVA was not analysed. However, this BCVA decline is probably due to the foveal involvement of these RPE abnormalities (Spearman correlation analysis, ρ = 0.724, p < 0.001).

Based on our findings, we proposed a model that may predict the disease progression in cCSC based on FAF/FA imaging discrepancy. According to this model, larger hyperfluorescent areas on FAF in cases with FAF/FA discrepancy can predict the enlargement of hyperfluorescent areas on FA during follow-up. We may speculate that FAF imaging could especially be used for predictive evaluation of hyperfluorescent RPE alterations in cCSC patients.

The study has several limitations. First of all, it was retrospective in design, and thus we were unable to unveil all influential factors, such as SRF persistence and number of disease recurrences, which presumably could be of importance in the decrease in FAF/FA imaging discrepancy in cCSC. Conducting a retrospective study implicates heterogeneous data, including cases with different characteristics of disease, CSC history, and follow-up duration. We understand that there might have been occurred a selection bias in our study, because we only included patients from tertiary academic referral centres. Therefore, we might have been dealing with more severe cases with complicated disease course and therefore a prolonged follow-up time. Secondly, the measurement of hyperfluorescent abnormalities on FA and FAF images was performed manually, which could have induced a certain error. The threshold of 0.5 disc area that was used to define FA-FAF discrepancy in hyperfluorescent RPE abnormalities is arbitrary. However, this was chosen based on clinical experience and as a reasonable difference that can be detected by ophthalmologists with certain accuracy. Automated or semi-automated software may improve the accuracy of quantitative estimation. An ultra-widefield FA and OCT can be used for monitoring cCSC patients to visualize possible extramacular disease activity. Larger and prospective studies are warranted to prove the predictive ability of FAF in cCSC patients with FAF/FA imaging discrepancy at baseline, and these studies will also contribute to detect further potential factors associated with this possible phenomenon.

In conclusion, at the last follow-up visit in the current study, we found a statistically significant relationship between the areas of hyperfluorescent alterations on FAF and FA in patients with FAF/FA discrepancy at first presentation. Our data show that these changes were time-dependent, which were proved by the proposed mathematical formula.