Background
Polypoidal choroidal vasculopathy (PCV) is a subtype of neovascular age-related macular degeneration (AMD) that accounts for 22.3–61.6% of neovascular AMD patients in Asia [
1,
2]. PCV consists of a branching vascular network (BVN) and its characteristic terminal polyps, both of which are located between the retinal pigment epithelium (RPE) and Bruch’s membrane [
3‐
5]. The rupture of polyps in PCV can lead to massive subretinal hemorrhage and cause sudden and severe vision deterioration [
6]. Another devastating nature of polyps even after successful treatment of PCV is their high recurrence rate and eventual severe vision loss [
2]. Many treatment modalities for PCV have been evaluated in terms of their polyp closure rate. Indocyamine green angiography (IA) has been essential for accurately monitoring the regression of polyps and BVN since it employs a longer wavelength than does fluorescein angiography (FA) to provide more fluorescence through the melanin pigments of the RPE and more clearly depict the polyps and BVN underneath.
The major treatments for PCV are anti-vascular endothelial growth factor (VEGF) agents, photodynamic therapy (PDT), and a combination of both [
1,
2]. All 3 treatment modalities could provide relatively high polyp closure rate, while BVN persists in follow-up FA/IA usually performed 3 months or later after the initiation of treatments. [
2,
7‐
13]
Optical coherence tomography angiography (OCT-A) is a modern technique that depicts retinal and choroidal vessels by detecting flow signals. Consequently, it is considered a non-invasive form of angiography not requiring intravenous injection of fluorescent dye. Inoue et al. reported that en face images from OCT-A provided anatomical information about the BVN that was comparable to that from IA. Polyps were less clearly depicted in the en face OCT-A images than on IA but were clearly defined in cross-sectional OCT-A images with flow signals [
14,
15]. Since OCT-A is non-invasive and requires only a few seconds for acquiring retino-choroidal vascular images, more frequent and intensive longitudinal follow-up of the BVN and polyps is now possible. Indeed, anaphylactic shock induced by IA is extremely rare but possible [
16], resulting in more conservative use by clinicians.
The present study analyzed the early changes of BVN and polyps that had been resistant to multiple anti-VEGF agents following combination therapy of IVR and PDT. From as early as 1 week after treatment, OCT-A was performed at monthly visits and the findings at 3 months were compared with those of IA.
Discussion and conclusions
In PCV, exudative changes including subretinal hemorrhage are usually derived from active polyps [
6,
7,
18‐
21]. Consequently, polyp closure is important for its successful treatment. Anti-VEGF monotherapy is currently the first line treatment for AMD, including PCV [
1,
19]. However, anti-VEGF therapy using ranibizumab or bevacizumab has a limited effect on PCV, with polyp regression rates ranging from 26 to 33% over a period of 1 year [
12,
13]. Moreover, BVN size increased in most cases [
22,
23]. Intravitreal injection of aflibercept (IVA) has recently been reported to be effective for PCV, producing a 51.8–72.5% polyp regression rate in cohorts that included patients responding poorly to ranibizumab, although the BVN persisted in all eyes [
7,
18‐
21].
PDT could achieve a 95% polyp regression rate over 1 year, but BVN regression remained minimal. Persistent BVN often serves as the origin of the recurring or newly developed polyps associated with subretinal hemorrhage and pigment epithelium tear over years of follow-up [
2,
7‐
11]. Consequently, the long-term visual outcome of PDT is negative despite a high polyp closure rate.
In the EVEREST II study [
17], combination therapy of intravitreal ranibizumab injection (IVR) and PDT achieved superior vision outcomes than did IVR monotherapy: higher polyp closure rates were obtained with less frequent IVRs and complete polyp regression rates at months 3, 6, and 12 were consistently higher for combination therapy (71.4%, 71.3%, and 69.7%, respectively) than for ranibizumab monotherapy (23.3%, 28.0%, and 33.8%, respectively). However, not even combination therapy could induce complete regression of the BVN, leaving a risk of recurrent active polyp development [
8].
There have been detailed reports about early FA/IA changes of choroidal neovascularization (CNV), which showed regression in 5 h and became inapparent 1 day after PDT in both FA/IA. At 3 months, however, the CNV size was consistently larger than at baseline [
24,
25]. On the other hand, there have been no reports on the early morphological changes of the BVN and polyps in PCV soon after intravitreal injection of anti-VEGF drugs and/or PDT.
We herein present sequential early OCT-A evidence on how combination therapy of IVR and PDT induces transient, but complete, regression of BVN and polyps that had persisted after multiple intravitreal injections of anti-VEGF agents. OCT-A is non-invasive since it does not require intravenous dye injection to visualize retino[24]-choroidal vessels, which enabled frequent angiographic examination as early as 1 week after combination therapy. In contrast, the first IA was performed 3 months after treatment in many earlier studies to minimize the potential risk of anaphylactic shock.
OCT-A examination of the retino-choroidal circulation was possible at every outpatient clinic visit. In case 1, 42 anti-VEGF injections had been given to maintain adequate vision. After we started to use OCT-A in our outpatient clinic,9 IVRs were performed but OCT-A revealed no signs of BVN or polyp regression. As subretinal fluid persisted, we performed additional PDT 3 days after the 43rd IVR based on the EVEREST II study [
17]. Surprisingly, OCT-A revealed complete regression of the BVN and polyp 1 week later. BVN gradually displayed reperfusion and restored trunk vessels in OCT-A at 3 months, almost perfectly resembling the pre-PDT findings. Conventional IA at that time confirmed the hyperfluorescence indicative of the BVN. We previously evaluated the treatment of PCV by IA only and generally believed that anti-VEGF agents, PDT, or their combination had no effect on the BVN while inducing polyp closure. However, more frequent OCT-A examination indicated that PDT induced transient complete regression of the BVN. Early cross-sectional OCT-A depicted both PED flattening and loss of flow signals, suggesting collapse of the BVN lumen and associated polyp. Although the BVN had restored its original network in 3 months, neither OCT-A nor IA showed the recurrence of polyps.
In case 2, OCT-A was performed 2 weeks after PDT. Regression of the BVN was remarkable but traces of BVN trunk vessels could already be distinguished. Cross-sectional OCT-A showed the disappearance of flow signals without PER flattening. These findings indicated that blood flow was reduced in the BVN but that vessel lumens were still present.
The above cases show OCT-A to be a useful tool for investigating the effects of anti-VEGF agents and PDT in the PCV treatment. Through OCT-A, we were able to characterize the changes in BVN and polyp status much more frequently than by IA alone.
Differently from previous observations by IA, PDT in conjunction with anti-VEGF agents were found by OCT-A to induce complete regression of the BVN, which then reappeared within 3 months. Since polyps are usually formed at the terminal portions of the BVN, such early-stage regression may explain the superior effects of combination therapy in polyp closure and visual outcomes in the EVEREST II study.
The limitations of this study are very low number of observed patients and their short follow up period. This study included only two PCV patients and one of them failed to visit our outpatient clinic on month 3, when the findings of OCT-A and IA were compared. Larger prospective trials including OCT-A monitoring are warranted to better characterize the nature of BVN and PCV, and merits of combination IVR and PDT treatment.