This retrospective study aimed to investigate how changes in the macular capillary morphology and the macular vessel density affect recurrence of ME associated with BRVO. Using en face OCTA images after resolution of ME, the recurrence group showed a less intact perifoveal capillary ring and a broader range of ring loss than the no ME recurrence group. Moreover, the recurrence group also showed a larger hemi-VD disparity between the unaffected and affected areas in the SCP and DCP and a lower hemi-VD in the affected area in the DCP, when compared with the no ME recurrence group.
The most common causes of visual loss in BRVO cases are known to be ME and macular ischemia [19, 20], however, the definition of macular ischemia remains unclear, but the occurrence of ME and macular ischemia are closely related [21,22]. Sim et al. [21] defined macular ischemia as an expansion of the FAZ and non-perfusion in perimacular capillaries. Meanwhile, Finkelstein [22] reported that the macular ischemia is associated with the destruction of the perifoveal capillary ring, while Wakabayashi et al. [23] defined capillary ring loss of ≥ 1/4 as the destruction of the FAZ. In our study, although a definition of macular ischemia was not provided, we analyzed whether the changes in the capillary network (the FAZ area, capillary ring morphology) and macular vascular density (VD in 3 x 3 mm macular region, the hemi-VD disparity) affect ME recurrence.
The mean superficial and deep FAZ areas in healthy people is 0.2–0.4 mm2 and 0.3–0.6 mm2, respectively, with the deep FAZ areas being wider than superficial FAZ areas [24–26]. Changes in FAZ area and VD have been reported as important biomarkers for the progression of diabetic retinopathy or retinal vascular diseases and prognosis of visual acuity [27–29]. These factors are more useful to OCT angiography than fluorescein angiography [30–31].
Wakabayashi et al. [23] reported that smaller FAZ areas resulted in better visual acuity after resolution of ME, and Parodi et al. [26] reported that the association between ME and visual loss was weak, but an increase in the FAZ area was closely associated with visual loss. As shown, most of the previous studies reported associations between the FAZ and the prognosis of visual acuity in retinal vascular disease [17, 23, 26–29]. However, studies on the association between the FAZ area and the ME recurrence are lacking. Our study found that the recurrence group had wider superficial and deep FAZ areas after resolution of ME when compared with the no ME recurrence group, with greater statistical significance in the superficial FAZ area values. Unfortunately, we did not analyze whether the FAZ areas changes over time after ME resolution, therefore, additional studies are needed to establish the changes that occur in FAZ areas following ME resolution.
In our study, the extent of perifoveal capillary ring loss in the SCP and DCP was significantly higher in the recurrence group when compared with the no ME recurrence group, indicating that ring destruction was more severe at DCP than SCP level in the recurrence group. Destruction of the capillary ring actually had a greater effect on ME recurrence than FAZ enlargement. Identification of DCP status by OCTA is very important for identifying macular ischemia [32]. Deep capillaries play a role as a watershed zone that supplies blood to the outer plexiform and inner nuclear layers [24], and supplies the necessary oxygen to the inner segments of visual cells under scotopic conditions [33]. Destruction of a capillary ring is an indicator of macular ischemia, where destruction of superficial and deep perifoveal capillary ring promotes secretion of VEGFs, which may cause ME to occur more readily due to increased vascular permeability.
The VD of capillary network in 3 x 3 mm macular region and the hemi-VD disparity were calculated, which indirectly reflects the degree of macular capillary perfusion. Hasegawa and colleagues [34] reported that patients with severe macular VD reduction in en face OCTA SCP images had less edema recurrence and need less intravitreal anti-VEGF injections. Sakimoto and colleagues [35] divided the macular perfusion into the three grades using FAG; full perfusion area, partial perfusion area, and nonperfusion area. They found that partial perfusion area with the dilated and irregular capillary net was a source of macular edema.
In this study, we checked the macular VD using both SCP and DCP images in en face OCTA. Our results showed no difference in the macular VD between the recurrence and no ME recurrence groups, but the recurrence group showed a greater hemi-VD disparity between the hemi-superior areas and hemi-inferior areas in the SCP and DCP than the no ME recurrence group. Moreover, macular capillary loss in the affected areas with BRVO in the DCP was especially higher in the ME recurrence group. How can we get the opposite results about between the capillary loss and edema recurrence? We think that the hemi-VD disparity does not mean non-perfusion area. Rehak and colleagues [36] found that alteration of retinal VEGF gene expression and dysfunction of water homeostasis of Müller cells are not restricted in vein-occluded area but also involved in neighboring non-occluded area in rat animal model of BRVO. We consider that macular edema occurs often until restoration of damaged retinal vessels after vein occlusion. So, the hemi-VD disparity beween hemi-superior and hemi-inferior macular area might reflect the disturbances of edema control in entire macular area. and contributes the development of macular edema.
This study had some limitations. Because the images were acquired from just a single round of OCTA, excluding images with poor image qualities from the analysis therefore, a relatively small number of samples were retrospectively analyzed. The study did not identify the range of retinal nonperfusion area by fluorescein fundus angiography, while VD measured by OCTA used only a 3 x 3 mm macular region. So, future studies are needed to examine the change of macular capillary network morphology over time after ME resolution using wide-viewing OCT angiography image.