Observation of Natural Course Characteristics of Bietti Crystalline Dystrophy by Fundus Fluorescein Angiography

Background: Bietti crystalline dystrophy (BCD) is an autosomal recessive genetic disorder that causes progressive vision loss. Here, 12 patients were followed up for 1–5 years with fundus uorescein angiography (FFA) to clarify BCD disease development and its classication. Methods: FFA images were collected for 12 patients with BCD who visited our clinic twice or more in 5 years. Peripheral venous blood was collected to identify a pathogenic gene related to the clinical phenotype. Results: FFA images identied two BCD types. Type 1 showed retinal pigment epithelium (RPE) atrophy at the macular area, followed by choriocapillaris atrophy, then RPE atrophy appeared at the peripheral retina. Type 2 showed RPE atrophy at the posterior pole and peripheral retina, followed by choriocapillaris atrophy around the macula and along the superior and inferior vascular arcades and optic disc nasal side. Then the posterior and peripheral lesions of type 1 and type 2 were extended to the mid-periphery; at last, all the RPE and choriocapillars atrophied, exposed choroid great vessels, but the macular RPE atrophy of type 2 can existed for a long time. Conclusions: The two different BCD development types provide understanding


Background
Bietti crystalline dystrophy (BCD) is an autosomal recessive retinal dystrophy characterized by numerous tiny sparkling yellow-white spots at the posterior pole of the fundus. The causative gene has been identi ed as CYP4V2 [1,2]. In 1937, Bietti rst reported three patients with BCD [3,4]. Since then, research on the disease has continued, but the natural progression and pathogenesis of BCD remains poorly understood. Most previous work has been cross-sectional studies [5,6,7,8], gene research [1,2,9,10,11,12], and case reports [13,14,15,16]. Very few of these studies have used fundus uorescein angiography (FFA) for BCD and the FFA images mostly involve the posterior pole of fundus. Few articles have reported long-term observations of BCD [17,18,19], and most of these articles are only case reports and lack comparisons of the before and after examinations. To our knowledge, no studies have used FFA to follow the progression of BCD in patients.
The aim of the present study was to outline the phenotype of BCD more clearly and to obtain a better understanding of the natural course of BCD. Here, we present 12 cases of BCD focusing on the FFA progression with the goal of identifying the features of BCD development.

Subjects
The implementation of all research methods in this study followed the provisions of the Declaration of Helsinki, the Ethics Committee of Beijing Tongren Hospital, Capital Medical University, and the Ethics Committee of Hebei Provincial Eye Hospital. This study included 12 unrelated Chinese patients who were followed up at our hospital clinic in last 5 years. The clinical characteristics and images of these patients were retrospectively analyzed. For convenience, patients 1-12 were designated P1-12.
This was a retrospective analysis of 24 eyes of 12 unrelated Chinese patients who had visited our hospital clinic twice or more between January 2013 and December 2018. All the patients had consented for mutation screening in our laboratory. Written informed consent was obtained from each patient before drawing peripheral venous blood for genomic DNA extraction and mutation screening of the CYP4V2 gene by direct sequencing, as previously described [20].

Procedures
All patients underwent complete ophthalmic examinations, including best corrected visual acuity (BCVA), slit-lamp microscopy, Goldman tonometry, indirect dilatation fundus examination, and fundus photography (Kowa, Nonmyd 7, Kowa, Japan). The BCVA measurements were converted to the logarithm of the minimum angle of resolution (logMAR) values [21]. FFA images of all the patients were obtained over a 55 × 55° eld with a confocal scanning laser ophthalmoscope (Heidelberg Spectralis, Heidelberg Engineering, Heidelberg, Germany). An exception was P6, whose rst visit FFA was done over a 50° eld with a Topcon retina camera (TRC-50DX, Topcon Corporation, Tokyo, Japan). The FFA images were used to create jigsaws every time and the jigsaws were compared to establish the developmental features of the BCD fundus lesions.

Clinical Presentation and Genetic Diagnosis
The general conditions and genetic diagnosis results for P1-12 are shown in Tables 1 and 2. The mean patient age was 39.75 ± 12.81 years. The mean BCVA of the 24 eyes was 0.95 ± 1.06 logMAR. P5 and P7 had a family history; none of the other patients had a family history. The lefteye of P6 had a choroidal neovascularization (CNV) at the macular area; no CNVs were found in the eyes of the other patients.

Fundus Color Photography
The total number of crystalline deposits decreased over time in all eyes. At the rst visit, the posterior pole retina showed dirty bluish gray. At the last visit, the dirty bluish gray of retina was diminished gradually.
And the choroidal great vessels were clearer than before. More pigment clumps were apparent (P1-5) (Fig. 1A, B). In P6, P11, P12, the condition of the macular area showed no signi cant change, but the RPE-choriocapillaris complex of the area around it was more atrophied than before (Fig. 1C, D). A scar cause by choroidal neovascularization was observed at the macular area in the left eye of P6 (Fig. 1F); this had not been evident at the rst visit ( Fig. 1E). At the last visits of P7-10, both eyes showed more pigment clumping and the choroidal great vessels were more apparent than at the rst visit (Fig. 1G, H).

Fundus Fluorescein Angiography (ffa)
The uorescence of both eyes in all patients was similar, except the left eye of P5 had a branch retinal vein occlusion and the left eye of P6 had a CNV. In this study, the FFA images at theearly phases revealed hypo-uorescence due to a lling defect of the choriocapillaris. The hypo-uorescent lesions showed lling of the choroidal great vessels. The FFA images at the late stage indicated hyper-uorescence staining of the hypo-uorescent lesions margin. The mottled uorescence in this study indicated RPE atrophy, which Mataftsi et al 5 had called a "salt and pepper appearance." At their rst visits, P1-P4 showed patchy hypo-uorescence at the macular region and mottled uorescence around it. The peripheral retina also showed mottled uorescence, while the mid-peripheral retina showed normal retinal uorescence ( Fig. 2A, Fig. 4A). Mottled uorescence was observed at the nasal side of P1, while P2 had mottled uorescence and patchy hypo-uorescence at the nasal retina and mottled uorescence at the superior and inferior mid-periphery; however the degree of disorder of the mottled uorescence was slighter in the superior and inferior mid-periphery than in the posterior pole and periphery, normal retinal uorescence was only observed at the temporal mid-periphery. At the last visits by P1-P4, the macular and peripheral lesions had extended to the mid-periphery and the RPE atrophy had progressed to choriocapillaris atrophy. This RPE-choriocapillaris complex atrophy was observed throughout the fundus images from P3 and P4 (Fig. 2B, Fig. 4C).
At the rst visit by P5, patchy hypo-uorescence was observed at the macular region, with mottled uorescence around it, whereas the mid-peripheral and peripheral retinal regions showed normal uorescence. The left eye of P5 had a branch retinal vein occlusion (Fig. 2C). At the last visit of P5, the lesion of the posterior pole had extended, as had the hypo-uorescence, and patchy mottled uorescence was now apparent at the peripheral retina. Points of laser photocoagulation were observed at the branch retinal vein occlusion area of the left eye (Fig. 2D).
At the rst visit by P6, the posterior pole and the peripheral and optic disc nasal side retina showed mottled uorescence, and a few patchy areas of hypo-uorescence were observed at the nasal and superior nasal areas of the optic disc. The mid-periphery, except for the nasal side, showed normal retinal uorescence (Fig. 3A, C). A CNV uorescence was observed at the macular area of the left eye (Fig. 3C), but no treatment was given. At the last visit, the posterior pole and peripheral lesion had extended to the mid-periphery and choriocapillaris atrophy had appeared around the macular area, equivalent to the superior and inferior vascular arcade areas and the optic disc nasal area. At the same time, RPE atrophy had progressed to choriocapillaris atrophy, but the macular RPE atrophy showed no signi cant changes (Fig. 3B). The CNV uorescence of the left eye progressed to uorescence staining of the scar, with annular hypo-uorescence observed around it (Fig. 3D).
At the rst visits by P7 and P8, only few areas of mottled uorescence remained in the peripheral retina and the superior temporal retina of the optic disc. Most of the fundus was hypo-uorescent, and the uorescence of choroidal great vessels were exposed (Fig. 4D). At the last visit, the fundus hypouorescence was extended, the choroidal great vessels were clearer, and the mottled uorescence area was reduced (Fig. 4F).
At the rst visits by P9 and P10, large areas of hypo-uorescence were seen in the posterior poles of both eyes, with the rest of the area showing mottled uorescence (Fig. 2E). Patchy hypo-uorescence was observed at the nasal retina in both eyes of P10. At the last visit, the hypo-uorescence had expanded and the area of mottled uorescence was reduced (Fig. 2F).
At the rst visits by P11 and P12, the macular region showed mottled uorescence, withpatchy hypouorescence observed around the macular area (Fig. 3E). The hypo-uorescent area was larger in P11 than in P12. At the last visit, no macular area changes were evident, and the rest of the retina showed expanded hypo-uorescence and a more disordered mottled uorescence (Fig. 3F).

Discussion
The natural progression of BCD remains poorly understood, with few follow-up cases appearing in the published literature. The previous staging methods for BCD have included the Yuzawa staging [22], fundus uorescein angiography staging [5], and electrophysiological staging [20]. The Yuzawa staging is widely used [5,20,23,24,25,26], but most of these studies have been cross-sectional ones or case reports with only small numbers of observed cases. We followed up 12 patients with BCD to determine the natural progression of the disease, and our discussion focuses on the breadth and depth of the fundus lesion of BCD.
Our rst focus is on the sequence of BCD lesion occurrences. We found that the posterior pole retina was the rst to be attacked, followed by the peripheral retina. Both the posterior and peripheral lesions extended to the mid-periphery, nally all the retina was attacked. The disease progresses more rapidly on the nasal side retina of optic disc, whereas the retina in the mid-peripheral part of the temporal side of the macular area is involved last.
The Yuzawa staging is used widely and divides the disease into three stages. However, these stages were based on three patients, with only one patient followed up for ve years, and lesions have occurred throughout the fundus of this patient at the initial visit. The two other patients were cross-sectional observations. They described the rst stage of the lesion con ned to the macular area, the second stage beyond the posterior pole, and the third stage involving the totalretina. This description tends to give the impression of a centrifugal expansion.
Many subsequent studies have stated that "Yuzawa described a centrifugal expansion of the RPEchoriocapillaris complex atrophy, from the macular area towards the periphery, occurring in three stages" [5,20,24,27], and they agreed with this description. By contrast, our observations were different. In our opinion, the reasons for this difference were the small case numbers in the previous studies, with even fewer cases of total retinal observation, and a focus of attention on the changes in the posterior pole retina.
Halford et al [27] reported that atrophy of areas of the RPE and choroid tended to develop at the posterior pole, became con uent, and expanded centrifugally to involve peripheral retina, but they only observed a 55° auto uorescence in the fundus images, not a total retinal involvement. Consequently, the conclusion that BCD is a centrifugal expansion disease is questionable. Mataftsi et al [5] conducted jigsaw observations in six patients, but all were advanced cases. According to the Yuzawa staging, all patients were stage 3 and had complete retinal involvement, so the conclusion that BCD is a centrifugal expansion was not justi ed. However, Mataftsi et al [5] found one patient with mild disease that showed a signi cant difference, as atrophic changes in the choriocapillaris were evident not only in the posterior pole but also at the equator level, at the eccentricity of the vortex veins. This nding was consistent with our observation: the atrophy in the posterior pole and peripheral choroid appeared before the midperipheral atrophy. Mataftsi et al [5] also performed an FFA staging of BCD, but the number of cases was small and all cases were advanced. Therefore, this FFA staging was not comprehensive and will not be discussed further in this article.
Some reports have suggested a centrifugal expansion of the visual defect based on the centralscotoma seen with the 30° visual eld test [1,20,28]. However, Liu et al [29] con rmed the visual eld features using the 85° visual eld test. They found that the peripheral and central scotomas initially appear, but as the disease progresses, these expand and combine, ultimately resulting in visual islands only in the midperiphery, which are not found centrally. This is consistent with our observation of lesions occurring rst in the center and periphery and then eventually extending to the mid-periphery.
In this study, we also used the Yuzawa staging, regarding it as a cross-sectional staging method according to the width and depth of the BCD lesions. We simpli ed the Yuzawa staging as follows: stage 1 consisted of a lesion con ned to the posterior pole with only RPE atrophy; stage 2 consisted of a lesion beyond the posterior pole or with choroid atrophy; and stage 3 consisted of total retinal involvement. Since the Yuzawa staging has been used widely for a long time, the application of Yuzawa staging in this study was intended as a convenience for readers to understand the condition of the eyes of our patients. It was not meant to indicate the natural progression of the disease.
We also examined the atrophy sequence of the BCD lesions and found two patterns of BCD atrophy. Type 1 showed RPE atrophy at the macular area, followed by choriocapillaris atrophy at the macula, then RPE atrophy at the peripheral retina, and subsequent extension of the macular and peripheral lesion to the mid-periphery; at the same time, the RPE atrophy extended to choriocapillaris atrophy and eventually all the RPE and choriocapillaris atrophied, exposing the choroid great vessels. Type 2 showed RPE atrophy at the posterior pole and peripheral retina, followed by choriocapillaris atrophy around the macular area, along the superior and inferior vascular arcades, and at optic disc nasal side. Subsequently, the posterior lesion and peripherallesion extended to the mid-periphery, with simultaneous transition of the RPE atrophy to choriocapillaris atrophy and, ultimately, full RPE and choriocapillaris atrophy, with exposure of the choroid great vessels, but the RPE atrophy of macular can persist for a long time. However, we found that the last area of atrophy at the most advanced stage of BCD was in the temporal peripheral retina, rather than the retinal area last attacked in the mid-periphery.
Mataftsi et al [5] found that peripheral retinal atrophy was relatively mild in the nal stage of the disease, which is consistent with our ndings. They hypothesized that the posterior ciliary arteries are selectively affected, whereas the anterior uveal circulation (ciliary body and anterior choroid) is preserved until late in the disease, although nevertheless insu cient for supplying the retro-equatorial choroid. We agree with this hypothesis, but it needs further con rmation.
Immunohistochemistry analyses have revealed that CYP4V2 is highly expressed in the choroid and the RPE, while relatively less expressed in the retinal outer and inner nuclear layers, retinal ganglion cells, and corneal epithelial cells, in accordance with the BCD phenotype [30]. The FFA images revealed changes mainly in the RPE and choroid, so RPE dysfunction has been considered the primary change in BCD [20,27,31]. One view holds that vascular endothelial growth factor (VEGF) is produced by the RPE and is necessary for choroidal maintenance [32]; therefore, a lack of VEGF caused by an RPE disorder may play a role in choroidal thinning. Our FFA results showed that RPE atrophy occurs rst, followed by choroidal vessels atrophy, in agreement with the previous research, including the research by Yuzawa et al [22].
We re-examined the FFA images in the previous literature, and we found that those images can be divided into the two atrophy types we have mentioned here. Type 1 shows choriocapillaris atrophy rst appearing at the macula [7,11,27,33,34], and type 2 shows choriocapillaris atrophy rstappearing around the macular area and along the superior and inferior vascular arcades and optic disc nasal side [7,15]. The case numbers are signi cantly smaller for type 2 than for type 1; for example, Wang et al [7] reported that, of the 4 patients examined, 3 were type 1 and 1 was type 2. In the present study, only 3 (P6, 11, and 12) of the 12 cases were type 2. Apart from the macular area changes caused by CNV of P6 left eye, the macular area of the other 5 eyes of the 3 cases showed slow changes and the RPE and choroid atrophy of the mid-periphery and periphery was signi cantly aggravated. This suggests that the type 2 patients can preserve better vision for a longer time, so these pattern differences may aid in the evaluation of a patient's prognosis.
The reason for these two different atrophy patterns is unknown. We looked at the gene mutation sites of P6, P11, and P12 and found that the P6 and P11 had homozygous mutations of c.802-8_810del17insGC, while P12 was heterozygous for a mutation of c.802-8_810del17insGC and c.332T > C; p. l111T. In this study, P1 and P7 also had homozygous mutations of c.802-8_810del17insGC, but their phenotypes differed from those of P6 and P11 (Table 2). Therefore, the speci c causes of these differences need further observation and research.
In the early stage of BCD, distinguishing the progression type is not possible in eyes with disordered pigment epithelium and no choroid atrophy. Only patients with choriocapillaris atrophy can be typed. The type of progression also cannot be determined in patients at the end stage of the disease because the choriocapillaris and RPE are atrophied and no longer visible, leaving only the image of the choroidal great vessels. We propose adding the progression types to the staging criteria to provide a clearer understanding of how the disease progresses, without creating the misconception that the disease develops in a centrifugal way. The progression type cannot be distinguished in the early and the last stages, but the addition of progression typing, when possible, can help to clarify the characteristics of disease development, while providing a more accurate prognosis and a better understanding of the pathogenesis of BCD.
This study had several limitations, including the small number of included eyes and the lack of primary patient observations. However, considering the rarity of the disease and our review of the previous literature, our study on the progression of BCD using FFA picture jigsaws provides one of the largest collections of images and the largest number of patients. Finding BCD patients with early stage disease is di cult because the visual acuity of patients at this stage is not substantially damaged, so they seldom come to the hospital.

Conclusions
The natural progression of BCD shows two patterns. The reasons for these different types of development need further study, but this study provides a better understanding of the phenotype and the development of the disease. The ndings presented here will be helpful for future pathogenesis research and for prognosis assessment of patients with BCD.