Optical coherence tomography angiography in assessment of response to therapy in retinal capillary hemangioblastoma and diffuse choroidal hemangioma.

In this series, we discuss the role of optical coherence tomography angiography (OCTA) in assessing response to treatment in intraocular vascular tumors. This is a series of two cases: Multiple retinal capillary hemangioblastoma (RCH) treated with laser photocoagulation and diffuse choroidal hemangioma (DCH) with radiotherapy. In large RCH and DCH, optical coherence tomography (OCT) showed significant reduction of subretinal and intraretinal fluid. But post-treatment mean tumor vascular density (MTVD) was slightly reduced. In one small RCH, vascular loop was seen suggesting minimal residual disease. So, OCTA helps in identifying treatment inadequacy and understanding alternate mechanism involved in treatment response in vascular tumors.

Vascular tumors of retina include cavernous hemangioma of retina, peripheral, and juxtapapillary retinal capillary hemangioblastoma (RCH). Larger RCH needs multiple sessions of treatment and it is preferable to treat these tumors at an earlier stage when they are small.[1] Vascular tumors of choroid include solitary and diffuse choroidal hemangioma (DCH), and needs treatment if associated with exudative retinal detachment (RD). Optical coherence tomography angiography (OCTA) can identify the intrinsic vascularity of RCH[2] and choroidal hemangioma.[3] Hence, OCTA could help in identifying post-treatment changes in intrinsic vascularity of these tumors.

This is a retrospective series of two cases. Pretreatment and post-treatment color photograph, OCT, and OCTA of tumors were captured using swept source OCTA (Topcon DRI OCT Triton, Topcon Corporation, Japan). The available software in the swept source OCTA can measure the capillary density only in the superficial capillary plexus slab. Hence, the segmentation lines in the superficial capillary plexus slab were modified manually to include the entire thickness of tumor that could be captured. Capillary density was measured by repositioning the capillary density grid on the tumor repeatedly to include entire tumor area. Mean tumor vascular density (MTVD) was calculated as average of these readings [Fig. 1].

Figure 1

Measurement of mean tumor vascular density (MTVD). The central circle of the grid is positioned on different areas of tumor to include entire tumor area. MTVD is calculated as average of these readings

Case Reports

Case 1

A boy with von Hippel-Lindau disease presented with blurred vision in left eye. On examination, a large RCH was seen in superior mid-peripheral retina with sub foveal fluid. Six additional small tumors were detected on fundus fluorescein angiography. Five small tumors were identified on OCTA. The large tumor had rich intrinsic vascularity on OCTA. Pretreatment MTVD was 80.36%. Green laser photocoagulation (LP) was used to treat all the tumors. Feeder vessel treatment was performed for the large tumor. Two months later, complete resolution of subretinal fluid was noted. Five months later, best corrected visual acuity (BCVA) was 6/6 with normal foveal contour. But no significant change in size of the large tumor (that resulted in macular edema) was noted. Minimal change in color was seen [Fig. 2]. OCTA showed minimal decrease in vascular density with increase in intervening signal void areas. Larger vessels within the tumor were less tortuous [Fig. 3]. Post-treatment MTVD was 78.4%. The smaller tumors were replaced by chorio-retinal scars. Only two small tumors could be captured by OCTA, as the patient felt it to be exhausting. On OCTA, abnormal vascular signals were not seen in the area corresponding to one treated tumor, indicating complete treatment. In the other, vascular loop was seen suggesting residual disease [Fig. 4].

Figure 2

Case 1: (a) Pretreatment color photograph showing tortuous feeder and draining vessels. (b) Post-treatment color photograph showing minimal whitening of tumor and decrease in the vascular tortuosity. (c) Pretreatment OCT through the fovea showing presence of intraretinal and subretinal fluid. (d) Post-treatment OCT showing resolution of subretinal and intraretinal fluid. Disruption of the ellipsoid zone is seen

Figure 3

Case 1: (a) Pretreatment optical coherence tomography angiography (OCTA). The pre-treatment and post treatment OCTA were segmented manually. (b) Post-treatment OCTA showing minimal increase in signal void areas in the superior half of tumor. (c) Pretreatment OCTA vascular density map (the vascular density is represented as a color map. Violet corresponds to lesser density and red to high density). (d) Post-treatment OCTA vascular density map showing minimal enlargement of central blue zone

Figure 4

Case 1: (a) Color photo showing chorioretinal scar at the site of laser photocoagulation. (b) Absence of abnormal vasculature corresponding to the lesion (within the green circle) indicating complete treatment. Vascular lesion with feeder and draining vessel corresponding to the lesion (within white circle) indicating residual disease

Case 2

A gentleman with right-sided facial port-wine stain presented with blurred vision in right eye. On examination, he was detected to have DCH with exudative RD. The tumor was treated with external beam radiotherapy (20 Gy). Two months later, resolution of exudative RD was noted [Fig. 5]. Pretreatment and post-treatment comparative OCTA images showed very minimal reduction in bright areas with increase in intervening signal void areas [Fig. 6]. The bright signal is considered as vasculature in the subsequent discussion. The pretreatment MTVD was 56.70%. The post-treatment MTVD was 55.73%. Projection artifacts from the retinal vessels were seen obscuring the underlying choroidal vessels in both pre- and post-treatment OCTA.

Figure 5

Case 2: (a) Color photograph of diffuse choroidal hemangioma showing inferior exudative retinal detachment (white arrow). (b) Post-treatment color photograph showing decrease in exudative retinal detachment. (c) OCT through diffuse choroidal hemangioma showing undulating choroid with presence of subretinal fluid. (d) Post-treatment OCT showing decrease in choroidal undulation and resolution of subretinal fluid

Figure 6

Case 2: (a): Pretreatment OCTA of choroid. (b) Post-treatment OCTA showing increase in signal void areas. (c) Pretreatment OCTA vascular density map. (d): Post-treatment OCTA vascular density map showing increase in blue zones indicating decrease in vascular density

Discussion

In large RCH, decrease in capillary leak, as demonstrated by resolution of macular edema, was evident after treatment as early as 2 months. But no significant changes in tumor morphology could be appreciated even at 5 months. OCTA could detect decrease in MTVD. But the reduction in vascular density was minimal, suggesting that LP would decrease capillary leak by mechanisms other than capillary closure. Exudation from RCH is due to the fenestrated nature of endothelial cells.[4] It would be possible that LP would alter the endothelial fenestrations or would induce changes in the basement membrane. In an experimental study, increase in number and length of choroidal endothelial cell processes was noted 5 days after retinal LP. This was associated with increase in thickness of endothelial basement membrane.[5] Hence, it is possible that thickening of basement membrane can decrease leakage from the tumor and it appears inessential to achieve closure of tumor vessels. Decrease in the tortuosity of larger vessels within the lesion was noted in our case, similar to the case reported in literature.[6]

In our case, complete resolution of abnormal vessels was appreciated on OCTA in one small RCH. Hence, it is likely that LP can lead to capillary closure. Laser may not penetrate adequately in thicker tumor to achieve capillary closure, while it is possible in smaller lesions. It is possible that milder burns lead to remodeling of capillary architecture and severe burns lead to capillary closure. Vascular loop was identified post-treatment in one small RCH, suggesting that OCTA can identify incomplete treatment.

In DCH, reduction in exudative RD and decrease in choroidal thickness were noted 2 months after treatment. Radiotherapy leads to apoptosis of endothelial cells and sloughing in early phase. Smooth muscle and endothelial proliferation occurs in late phase and leads to perivascular fibrosis.[7] The reduction in intrinsic vascularity would parallel the reduction in tumor thickness. But there was minimal reduction in vascular density on OCTA. Hence, there is a possibility that radiation induced vascular remodelling would decrease leak.

Changes in OCT characteristics, such as resolution of exudation, serve as early markers of response to therapy in vascular intraocular tumors. But decrease in vascular density precedes morphological changes and also helps in assessment of response to treatment. The measurement of MTVD by the technique described above would not be accurate due to overlap between the circles in each frame. But it would be an approximate measure. Patient co-operation is essential and it would be a limiting factor in imaging. Presence of projection artifacts may limit assessment of vascular density in choroidal tumors. Short comings of our study include retrospective nature, small sample size, and short follow up. Study of OCTA changes in early post-treatment period (prior to OCT changes) would help in better understanding.

Conclusion

OCTA helps in identifying treatment inadequacy and understanding alternate mechanism involved in treatment response in vascular tumors.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.