optical coherence tomography angiography ...

OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY ASSESSMENT OF VASCULAR EFFECTS OCCURRING AFTER AFLIBERCEPT INTRAVITREAL INJECTIONS IN TREATMENT-NAIVE PATIENTS WITH WET AGE-RELATED MACULAR DEGENERATION LEONARDO MASTROPASQUA, MD,* LISA TOTO, MD, PHD,* ENRICO BORRELLI, MD,* PAOLO CARPINETO, MD,* LUCA DI ANTONIO, MD, PHD,* RODOLFO MASTROPASQUA, MD† Purpose: To investigate vessel changes occurring after aflibercept injections in treatment-naive exudative age-related macular degeneration patients. Methods: Fifteen eyes of 15 patients affected by wet age-related macular degeneration were enrolled in the study. All the patients had a diagnosis of Type 1 choroidal neovascularization and were treated with 3 monthly aflibercept intravitreal injections (IVI). Subjects were evaluated by means of optical coherence tomography angiography at baseline, the day after the first injection and one month after both the first and the second IVI. At last, all the patients were followed up to 2 months after the third IVI. Results: Foveal superficial vascular plexus flow density was 29.01% (21.13–37.32%) at baseline and was significantly reduced as soon as 1 month after the first IVI (median: 20.78%; interquartile range: 14.75–23.13%; P = 0.017). Parafoveal superficial vascular plexus flow density was 47.09% (44.91–51.72%) at baseline and significantly decreased as soon as 1 month after the second IVI (median: 44.40%; interquartile range: 41.59–49.29%; P = 0.034). Choroidal neovascularization lesion area remained stable throughout the follow-up. Nevertheless, interestingly, choroidal neovascularization flow area was significantly reduced as soon as the next day the first IVI (median: 0.37 mm2 and interquartile range: 0.27–0.72 mm2 at baseline; median: 0.30 mm2 and interquartile range: 0.24–0.64 mm2 at 1 day after the first IVI; P = 0.047). Conclusion: Intravitreal aflibercept injections are associated with a significant change in native retinal and choroidal vasculature. Moreover, the treatment did not cause a reduction in lesion area, but rather reduced the flow in the choroidal neovascularization. RETINA 37:247–256, 2017

A

There are several types of commercially available VEGF antagonists, including receptor fragments (aflibercept); monoclonal antibodies (ranibizumab and bevacizumab); and other molecules (pegaptanib). Aflibercept (Eylea; Regeneron Pharmaceuticals, Tarrytown, NY) was approved in 2011 for the treatment of wet AMD with a recommended dose of 2 mg administered by intravitreal injection every 4 weeks (monthly) for the first 3 injections, followed by 1 IVI every 8 weeks. Thereafter, several studies showed aflibercept being effective as much as ranibizumab for the wet AMD

ge-related macular degeneration (AMD) is the leading cause of irreversible central vision loss in the Western world.1 Two clinically recognized subtypes of AMD exist: nonneovascular (dry) AMD and neovascular exudative (wet) AMD.2 Wet AMD is caused by proliferation of choroidal neovascularization (CNV), the latter causing loss of photoreceptors through fibrovascular scarring. Over the last years, the intravitreal injections (IVI) of vascular endothelial growth factor (VEGF) antagonists have become the first-line treatment of wet AMD. 247

248

RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ! 2017 ! VOLUME 37 ! NUMBER 2

treatment, in terms of visual acuity outcome and morphologic changes.3,4 Despite the potential and rare systemic and/or ocular side effects, the importance of the VEGF antagonist in the treatment of wet AMD is well recognized.3,5 Nevertheless, the effects on retinal and choroidal vasculature, as well as on CNV, after IVI, have been much less studied. Spectral domain optical coherence tomography (SD-OCT) shows bands that seem to correspond to the anatomical layers of the human retina, and allow seeing also the choroidal layers. Both SD-OCT and fluorescein angiography are essential diagnostic tools in CNV diagnosis and follow-up. Moreover, a new diagnostic tool, the optical coherence tomography angiography (OCTA), has recently been developed to study retinal and choroidal vasculature without needing for the dye injection.6 In the current prospective study, we used an OCTA system based on an SD-OCT device using a splitspectrum amplitude decorrelation angiography algorithm, to investigate changes in CNV, as well as in superficial vascular plexus flow density and in choroidal thickness (CT), in eyes undergoing 3 monthly consecutive aflibercept IVI for wet AMD, over a 4-month period. Methods Study Participants A total of 15 eyes of 15 patients with wet AMD with AMD (10 males and 5 females; mean age 69.77 ± 9.62 years, range 51–83 years) were enrolled in the study. All patients were intravitreal treatment-naive and were affected by a decreased best-corrected visual acuity (BCVA) secondary to a recent diagnosis of wet AMD. Patients consecutively presented at the University Gabriele D’Annunzio Department of Ophthalmology, between January 2015 and May 2015. This study received a prospective international review board approval and patients’ signed informed consent to the use of their data were collected. The study adhered to the tenets of the Declaration of Helsinki. Criteria for inclusion were 1) age .50 years old; 2) fluorescein angiography evidence of active occult (Type 1) CNV secondary to AMD not determining an edema From the *Department of Medicine and Science of Ageing, Ophthalmology Clinic, University G. D’Annunzio Chieti-Pescara, Chieti, Italy; and †Ophthalmology Unit, Department of Neurological, Neuropsychological, Morphological and Movement Sciences, University of Verona, Verona, Italy. None of the authors have any financial/conflicting interests to disclose. Reprint requests: Lisa Toto, MD, PhD, Department of Medicine and Science of Ageing, Ophthalmology Clinic, University G. D’Annunzio Chieti-Pescara, Chieti 66100, Italy; e-mail: [email protected]

of the innermost retinal layers; 3) BCVA of at least 0.7 LogMAR (Snellen visual acuity of 20/100) in the study eye at baseline examination (to ensure proper execution of examination). Exclusion criteria were 1) presence of geographic atrophy; 2) any previous ocular surgery (including antiVEGF and photodynamic therapies); 3) any maculopathy secondary to causes other than AMD (including presence of an epiretinal membrane or vitreomacular traction syndrome); 4) any history of vascular diseases, including diabetic retinopathy or previous retinal vein occlusion in the study and fellow eyes; 5) any optic neuropathy, including glaucoma, or any condition increasing the risk of secondary glaucoma (e.g., pigment dispersion syndrome or pseudoexfoliation syndrome); 6) any neurodegenerative diseases known to influence retinal nerve fiber layer thickness (RNFL) thickness7,8 (e.g., Alzheimer or Parkinson diseases); 7) intraocular pressure .21 mmHg; 8) refractive error greater than 3 diopters; and 9) significant media opacities. Study Protocol At baseline, all patients underwent a complete ophthalmic evaluation, including assessment of BCVA using Early Treatment Diabetic Retinopathy Study charts, tonometry, slit-lamp biomicroscopy, indirect fundus ophthalmoscopy. Furthermore, all patients were tested by means of XR Avanti AngioVue OCTA (Optovue Inc, Fremont, CA) and fluorescein angiography with Heidelberg Retina Angiograph 2 (HRA + OCT Spectralis: Heidelberg Engineering, Heidelberg, Germany). All patients were treated with aflibercept IVI within 3 days from the baseline evaluation. Then, aflibercept was administered monthly for a total of 3 injections. All injections were performed in the operating theater following the standard aseptic intravitreal technique detailed in the package insert.9 After each treatment, patients were treated with topical netilmicin. Subjects were evaluated the day after the first injection and one month after both the first and the second IVI. At last, all the patients were followed up to 2 months after the third IVI. At each follow-up visit, patients were evaluated both with a complete ophthalmologic evaluation, to evaluate changes in BCVA and incidence of side effects, and by means of OCTA. Outcome measures included the change in BCVA, macular thickness, superficial vascular plexus flow density, subfoveal CT, and RNFL values from baseline to each evaluation. Moreover, we evaluated the whole CNV lesion area and the CNV flow area throughout the follow-up.

AFLIBERCEPT VASCULAR EFFECTS IN WET AMD ! MASTROPASQUA ET AL

Procedures Spectral Domain Optical Coherence Tomography Angiography With XR Avanti XR Avanti AngioVue OCTA (Optovue Inc) is a device using a system based on split-spectrum amplitude decorrelation angiography algorithm (Version: 2015.100.0.35). The latter algorithm let the device to detect the flow as a variation over time in the speckle pattern formed by interference of light scattered from RBC and adjacent tissue structure.10 In addition, the SD-OCT tool was skilled of acquiring the standard structural OCT typically used by commercially available devices. The examination was performed as previously described.11 In brief, each subject’s pupil was dilated with a combination of 0.5% tropicamide and 10% phenylephrine, then study participants underwent SD-OCT imaging following a protocol that included AngioVue OCT 3D volume set of 3 · 3 mm, consisting of 304 · 304 pixels in the transverse dimension. Low-quality scans (i.e., if the subject blinked or if there were much motion artifacts in the data set) were excluded and repeated until a good quality was achieved. Three scans for each patient were captured, then the best one for quality (without significant motion artifacts and with a signal strength index .60) was considered for the analysis. Vascular layer Segmentation and flow density analysis. Vascular retinal layers were visualized and segmented as previously described.10–14 To evaluate the superficial vascular retinal plexus, we used a layer thickness of 60 micron from the inner limiting membrane, to include all the vessels of this plexus. Two investigators checked the segmentation quality before testing flow density. Objective quantification of flow density was evaluated for each eye using the split-spectrum amplitude decorrelation angiography software. Quantitative analysis was performed on the OCTA en face image using the AngioVue software. The flow density was defined as the percentage area occupied by vessels in a circle region of interest (ROI) centered on the center of the foveal avascular zone (FAZ) and with a diameter of 2.5 mm (Figure 1). The AngioVue software automatically splatted the ROI into 2 fields 1) the foveal area, a central circle with a diameter of 1 mm; and 2) the parafoveal area that constitutes the remaining part inside the ROI (Figure 1). The AngioVue software automatically outputs the flow density percentage inside the foveal area and the parafoveal area, as previously described15,16 (Figure 1).

249

Moreover, the software outputted the mean macular thickness in the foveal (foveal macular thickness) and in the parafoveal (parafoveal macular thickness) areas. Choroidal neovascularization analysis. The same method showed by de Carlo et al17 was used to determinate the best segmentation for the CNV evaluation (Figure 1). In brief, the OCTA software was used to determinate the best segmentation for each patient, to visualize the whole CNV at the en face image. Initially, a retinal layer, with the inner border at the level of the outer aspect of the inner nuclear layer and the outer border at the level of Bruch’s membrane, was set. In this manner, minimal choroidal vasculature and deep retinal vasculature would be included in the region being imaged by OCTA. Therefore, the inner and outer border levels were changed finely to include all the area suspicious for CNV, as visualized on the corresponding OCT B-scan images. A software artifact removal toggle function removes retinal vessel shadowing from the en face flow image. Finally, by means of a dedicated tool, on the en face image, we drew an ROI following the CNV perimeter. After determining the ROI, the software automatically outputted the CNV lesion area (the entire area inside the drawn ROI) and the CNV flow area (the area of flow in the user-defined CNV lesion area) (Figure 2). Choroidal Thickness Cross-sectional SD-OCT (10 mm scan length, 3 mm scan depth) of macular region was performed using the XR Avanti SD-OCT. The obtained scan provides visualization of structures from deep choroid well into the vitreous, in a single B-scan. The images were shown and measured with the XR Avanti software. The choroid was measured in masked fashion from the outer portion of the hyper-reflective line corresponding to the retinal pigment epithelium to the inner surface of the sclera. Macular measurements of the choroid thickness were made in the subfoveal (SF) location. Retinal Nerve Fiber Layer The RNFL thickness measurement was generated from a 3.45-mm diameter circle and is calculated as the distance between the internal limiting membrane and the outer edge of the inner plexiform layer.18 The RNFL parameters provided by the software include the average RNFL thickness and eight sectorial measurements. The eight sectors are temporal upper, superior temporal, superior nasal, nasal upper, nasal lower, inferior nasal, inferior temporal, and temporal lower. The average and the sectorial RNFL thickness values were considered for the analysis.

250


Fig. 1. Representative OCTA macula 3 · 3 scan from a wet patient with AMD at baseline evaluation. To evaluate the superficial retinal plexus (right figure), a layer thickness of 60 micron from the inner limiting membrane was set. The superficial vascular plexus flow density (left figure) was defined as the percentage area occupied by vessels in a circle ROI centered on the center of the FAZ and with a diameter of 2.5 mm. The AngioVue software automatically splatted the ROI and the superficial vascular plexus flow density evaluation into 2 fields: 1) the foveal area (F), a central circle with a diameter of 1 mm; and 2) the parafoveal area (PF) that constitutes the remaining part inside the ROI.

Statistical Analysis Statistical calculations were performed using Statistical Package for Social Sciences (version 20.0. SPSS Inc, Chicago, IL). To detect departures from normality distribution, Shapiro–Wilk’s test was performed for all variables. All quantitative variables were presented as median in the results and as median and interquartile range (IQR) in the tables. Wilcoxon signed-rank test with correction for multiple comparison was applied to compare baseline values with each follow-up controls.

Fig. 2. (bottom image) Spectral domain optical coherence tomography B-scan image from a wet patient with AMD at baseline evaluation. To evaluate the Type 1 CNV, a retinal layer, with the inner border at the level of the outer aspect of the inner nuclear layer and the outer border at the level of Bruch membrane, was set. Then, the inner and outer border levels were manually changed finely to include all the area suspicious for CNV. (top left image) OCT angiography en face projection image of the neovascular complex. (top right image) Colorcoded flow analysis with manually drawn lesion borders. The CNV borders define the ROI of which the software automatically outputted the CNV lesion area (the entire area inside the drawn ROI) and the CNV flow area (the area of flow in the user-defined CNV lesion area) values.

Moreover, we compared the mean change of vascular parameters from baseline to 2 months after the last injection, expressed as a percentage, between patients experiencing a BCVA improvement and patients without BCVA progression, using Wilcoxon signed-rank test. The chosen level of statistical significance was P , 0.05. Results The BCVA was 0.45 (0.17–0.70) LogMAR (approximately Snellen acuity of 20/50) at baseline

Wet AMD Patients (n = 15) Baseline

1 Day

BCVA (LogMAR) 0.45 (0.17–0.70) ("20/50) 0.45 (0.20–0.70) (Snellen ("20/50), P = 1.0 equivalent) Superficial vascular plexus flow density (%) Foveal 29.01 (21.13–37.32) 28.00 (19.90–32.03), P = 0.646 Parafoveal 47.09 (44.91–51.72) 45.52 (41.82–46.89), P = 0.241 Choroidal SF 185.50 (140.25–254.25) 177.50 (132.00–230.50), thickness (mm) P = 0.107 CNV Lesion area 0.89 (0.47–1.71) 0.79 (0.46–1.69), (mm2) P = 0.249 Flow area 0.37 (0.27–0.72) 0.30 (0.24–0.64), (mm2) P = 0.047 Macular thickness (mm) Foveal 372.00 (251.25–443.25) 347.00 (253.50–420.50), P = 0.304 Parafoveal 378.50 (351.00–410.00) 366.00 (341.00–411.50), P = 0.635 FAZ area (mm2) 0.20 (0.11–0.22) 0.20 (0.11–0.25), P = 1.0 FAZ flow density 0.03 (0.01–0.04) 0.03 (0.01–0.04), P = 1.0 (%)

1 Month

2 Months

4 Months

0.35 (0.10–0.55) ("20/40), P = 0.155

0.25 (0.10–0.52) ("20/32), P = 0.042

0.20 (0.10–0.50) ("20/32), P = 0.035

20.78 (14.75–23.13), P = 0.017 45.09 (43.91–50.72), P = 0.074 147.50 (125.50–192.75), P = 0.036

21.14 (16.09–24.51), P = 0.043 44.40 (41.59–49.29), P = 0.034 155.00 (117.00–199.75), P = 0.045

20.04 (14.77–23.63), P = 0.035 43.97 (41.58–46.79), P = 0.006 160.00 (111.25–196.75), P = 0.025

0.65 (0.38–1.55), P = 0.249

0.67 (0.36–1.15), P = 0.116

0.60 (0.32–1.36), P = 0.055

0.31 (0.18–0.52), P = 0.046

0.28 (0.18–0.52), P = 0.046

0.29 (0.14–0.50), P = 0.031

235.00 (216.50–256.00), P = 0.008 304.00 (287.00–323.50), P = 0.005 0.21 (0.12–0.22), P = 1.0 0.03 (0.01–0.05), P = 1.0

237.00 (218.00–275.00), P = 0.015 301.00 (291.50–317.00), P = 0.008 0.20 (0.10–0.22), P = 1.0 0.02 (0.01–0.04), P = 1.0

240.00 (218.00–259.00), P = 0.043 307.00 (291.50–318.00), P = 0.011 0.21 (0.14–0.23), P = 0.857 0.03 (0.02–0.04), P = 1.0

Data are presented as median (IQR). The Wilcoxon signed-rank test was performed to obtain P-values. LogMAR, logarithm of the minimum angle of resolution; SF, subfoveal.


Table 1. Parameters Evaluated at Each Time-Point of the Follow-up

251

252


and 0.45 (0.20–0.70) LogMAR (approximately Snellen acuity of 20/50) the next day after the first IVI (P = 1.0). The BCVA improved at 1 month after the first IVI, without reaching the statistical significance (median: 0.35 LogMAR and IQR: 0.10–0.55 LogMAR [approximately Snellen acuity of 20/40]; P = 0.155). Finally, the BCVA significantly improved as soon as 1 month after the second IVI (median: 0.25 LogMAR and IQR: 0.10–0.52 LogMAR [approximately Snellen acuity of 20/32]; P = 0.042) and was still significantly enhanced at 2 months after the third IVI (median: 0.20 LogMAR and IQR: 0.10–0.50 LogMAR [approximately Snellen acuity of 20/32]; P = 0.035) (Table 1). At 2 months after the last IVI, 9 of 15 eyes completed the follow-up experiencing a BCVA improvement (of at least 5 letters on the Early Treatment Diabetic Retinopathy Study chart), after the comparison with baseline values. Superficial Vascular Plexus Flow Density Analysis Foveal superficial vascular plexus flow density was 29.0% (21.1–37.3%) in patients affected by wet AMD, at baseline. One day after the first IVI, the foveal superficial vascular plexus flow density was stable (median: 28.0%; IQR: 19.9–32.0%; P = 0.646). Finally, compared with baseline values, the foveal superficial vascular plexus flow density was significantly reduced both at 1 month after the first IVI (median: 20.8%; IQR: 14.7–23.1%; P = 0.017) and at 1 month after the second IVI (median: 21.1%; IQR: 16.1–24.5%; P = 0.043) and at 2 months after

the third IVI (median: 20.0%; IQR: 14.8–23.6%; P = 0.035) (Table 1, Figure 3). Parafoveal superficial vascular plexus flow density was 47.1% (44.9–51.7%) in pretreatment patients affected by wet AMD. One day after the first IVI, the parafoveal superficial vascular plexus flow density was not significantly changed, compared with baseline values (median: 45.5%; IQR: 41.8–46.9%; P = 0.241). Moreover, parafoveal superficial vascular plexus flow density was 45.1% (43.9–50.7%) at 1 month after the first IVI (P = 0.074), 44.4% (41.6–49.3%) at 1 month after the second IVI (P = 0.034), and finally, 43.97% (41.6–46.8%) at 2 months after the third IVI (P = 0.006) (Table 1, Figure 3). Noise and Foveal Avascular Zone Analysis Decorrelation can also be generated by bulk motion, determining noise. To be sure the analysis was not influenced by noise, as already shown by Wei E et al,19 we tested the flow density in the FAZ area. Indeed, because there is no retinal circulation in the FAZ, this method is an assessment of background motion noise. Table 1 shows these results. Furthermore, because variations in FAZ area throughout the follow-up could influence the results, we compared FAZ area values between baseline and each follow-up visit, as shown in the Table 1. Macular Thickness Analysis Foveal macular thickness was 372.0 mm (251.2– 443.2 mm) in wet AMD patients, at baseline. Moreover, one day after the first IVI, the foveal macular

Fig. 3. (top images) Optical coherence tomography angiography scans from an enrolled patient affected by Type 1 CNV in AMD and treated with 3 monthly aflibercept IVI. Scans were acquired at baseline visit and 1 day, 1 month, 2 months, and 4 months after the first aflibercept injection. The figure shows the superficial vascular plexus flow density changes throughout the follow-up. (bottom images) Color-coded flow density maps of the superficial vascular plexus flow density (the warmer is the color the greater is the flow).


253

Fig. 4. Optical coherence tomography angiography scans, from an enrolled patient affected by Type 1 CNV in AMD and treated with 3 monthly aflibercept IVI. Scans were acquired at baseline visit and 1 day, 1 month, 2 months, and 4 months after the first aflibercept injection. The figure shows the CNV changes throughout the follow-up.

thickness was stable (median: 347.0 mm; IQR: 253.5– 420.5 mm; P = 0.304). Finally, compared with baseline values, the foveal macular thickness was significantly decreased both at 1 month after the first IVI (median: 235.0 mm; IQR: 216.5–256.0 mm; P = 0.008) and at 1 month after the second IVI (median: 237.0 mm; IQR: 218.0–275.0 mm; P = 0.015) and at 2 months after the third IVI (median: 240.0 mm; IQR: 218.0–259.0 mm; P = 0.043) (Table 1). Parafoveal macular thickness was 378.5 mm (351.0– 410.0 mm) in pretreatment patients affected by wet AMD. Therefore, the next day the first IVI, the parafoveal macular thickness was not significantly changed, compared with baseline values (median: 366.0 mm; IQR: 341.0–411.5 mm; P = 0.635). Moreover, parafoveal macular thickness was 235.0 mm (216.5–256.0 mm) at 1 month after the first IVI (P = 0.008); 237.00 (218.0–275.0 mm) at 1 month after the second IVI (P = 0.015), and finally, 240.00 mm (218.0–259.0 mm) at 2 months after the third IVI (P = 0.011) (Table 1). Choroidal Thickness Analysis Choroidal subfoveal thickness was 185.5 mm (140.2–254.2 mm) at baseline and 177.5 mm (132.0– 230.5 mm) the next day after the first IVI (P = 0.107). The subfoveal CT significantly decreased as soon as 1 month after the first aflibercept injection (median: 147.5 mm; IQR: 125.5–192.7 mm; P = 0.036). Moreover, its value remained reduced both at 1 month after the second injection (median: 155.0 mm; IQR:

117.0–199.7 mm) and at 2 months after the third IVI (median: 160.0 mm; IQR: 111.2–196.7 mm) (Table 1). Choroidal Neovascularization Analysis The lesion area did not significantly decrease throughout all the follow-up. Nevertheless, the flow area was significantly reduced as soon as 1 day after the first IVI (median: 0.37 mm2 and IQR: 0.27– 0.72 mm2 at baseline; median: 0.30 mm2 and IQR: 0.24–0.64 mm2 at 1 day after the first IVI; P = 0.047). Moreover, the CNV flow area remained reduced compared with baseline values for all the follow-up (Table 1, Figure 4). Furthermore, the CNV flow area reduction was significantly higher in patients experiencing a BCVA improvement, after the comparison with patients without BCVA progression (Table 2). Retinal Nerve Fiber Layer Thickness Analysis Average RNFL thickness was 100 mm (97–109 mm) at baseline and its value was stable the day after the first IVI (median: 101 mm; IQR: 100–112 mm; P = 0.248 mm) and at one month after the first IVI (median: 101 mm; IQR: 96–103 mm; P = 0.176). However, average RNFL thickness was significantly reduced both at 1 month after the second IVI (median: 97; IQR: 96– 101; P = 0.042) and at 2 months after the third IVI (median: 96; IQR: 92–101; P = 0.030), after the comparison with the baseline values (Table 3).

Table 2. Comparison Between Patients With and Without Visual Acuity Improvement Patients With BCVA Improvement (n = 9) CNV lesion area CNV flow area Foveal superficial vessel density Parafoveal superficial vessel density

233.3 240.5 232.0 28.5

(244.7 (255.0 (236.4 (240.7

to to to to

232.9) 243.0) 222.7) 27.0)

Data are delta percentages (median and IQR) after 2 months from the last treatment. The Mann–Whitney U test was performed to obtain P-values. n, number of patients.

Patients Without BCVA Improvement (n = 6) 230.8 230.0 233.6 28.7

(235.2 (238.2 (237.0 (235.0

to to to to

228.0) 225.5) 221.0) 28.0)

P 0.275 0.040 0.656 0.557

Data are presented as median (IQR). The Wilcoxon signed-rank test was performed to obtain P-values. IN, inferior nasal; IT, inferior temporal; NL, nasal lower; NU, nasal upper; RNFL, retinal nerve fiber layer; SN, superior nasal; ST, superior temporal; TL, temporal lower; TU, temporal upper.

96 (92–101), P = 0.030 130 (123–133), P = 0.043 87 (77–97), P = 0.310 72 (69–77), P = 0.028 125 (100–142), P = 0.043 110 (102–117), P = 0.237 75 (71–81), P = 0.068 85 (76–89), P = 0.061 95 (88–107), P = 0.293 97 (96–101), P = 0.042 126 (122–140), P = 0.047 86 (77–97), P = 0.075 78 (74–84), P = 0.043 132 (93–143), P = 0.063 105 (99–120), P = 0.093 79 (71–82), P = 0.833 91 (83–92), P = 0.528 100 (95–109), P = 1.0 101 (96–103), P = 0.176 127 (122–141), P = 0.093 83 (78–94), P = 0.075 76 (74–87), P = 0.028 134 (103–149), P = 0.093 113 (81–129), P = 0.499 77 (72–87), P = 0.465 84 (81–89), P = 1.0 98 (93–102), P = 0.528 101 (100–112), P = 0.248 130 (122–144), P = 0.172 87 (82–103), P = 0.395 92 (81–105), P = 0.753 137 (99–141), P = 0.172 123 (90–132), P = 0.753 80 (73–103), P = 0.075 88 (85–92), P = 0.075 105 (100–106), P = 0.244 100 (97–109) 133 (129–141) 95 (83–103) 87 (76–97) 136 (111–142) 124 (88–144) 77 (67–98) 86 (81–89) 100 (93–106)

2 Months 1 Month 1 Day Baseline

Wet AMD Patients (n = 15)

Table 3. Retinal Nerve Fiber Layer Evaluated at Each Time-Point of the Follow-up

4 Months


RNFL thickness (mm) Average ST TU TL IT IN NL NU SN

254

In the evaluation of the individual sectors, the RNFL thickness reduction was still present in the superior-temporal sector (as soon as one month after the second IVI), in the TL sector (1 month after the first injection), and in the IT sector (only at the last assessment) (Table 3). Discussion In this prospective study, we investigated vascular effects of 3 consecutive monthly aflibercept IVI, in treatment-naive patients with wet AMD with AMD affected by Type 1 CNV. In the last years, several authors have studied vessels’ changes after VEGF antagonist treatment. Bonnin et al,20,21 by means of an ultrasound approach, showed that both ranibizumab and bevacizumab induced a reversible decrease in blood flow velocity in retrobulbar arteries. Moreover, the VEGF antagonist effect on the native vessels was also studied by means of the retinal vessel analyzer; the latter showing a retinal arteriolar vasoconstriction after ranibizumab treatment, as soon as one month after the first intravitreal injection.22,23 Nevertheless, at the best of our knowledge, no study exists evaluating retinal vascular changes after aflibercept IVI. We showed, by means of a new imaging approach, a reduction in superficial vascular plexus flow density occurring after the treatment both in the foveal area and in the parafoveal area. Then, our results demonstrated that aflibercept treatment is associated with changes in retinal vessels, as well as already shown for ranibizumab and bevacizumab. Furthermore, many authors showed aflibercept is associated with a reduction in CT in eyes with neovascular AMD.24,25 Interestingly, Mazaraki et al24 showed aflibercept reducing CT with a more pronounced effect in treatment-naive eyes. Our data confirmed these results, showing for the first time that changes after aflibercept injection affect both the retinal and the choroidal native vasculature. Possible explanations for a decrease in superficial vascular plexus flow density, as well as in CT, after aflibercept IVI, could be 1) the induced decreased of nitric oxide, a potent vasodilator, causing vessel vasoconstriction26; 2) the capillary “rarefaction,” the latter described in experimental models of mice treated with a VEGF signaling inhibitor.27 Moreover, the CT decrease could be also secondary to a reduction of choriocapillaris endothelial cell fenestrations after VEGF antagonist treatment.28,29 Note that the latter experimental studies have assessed the effect of ranibizumab and bevacizumab, but not aflibercept; nevertheless, we


can suppose that the effect is relatively similar for all the VEGF antagonist drugs. Furthermore, OCTA allows the clinician to obtain depth-resolved information and detailed images of CNV in neovascular AMD.17,30 Recently, Kuehlewein et al31 demonstrated that in most cases of Type 1 CNV, the lesion area remains unchanged, even after the VEGF antagonist therapy. Our data confirmed the latter results; indeed the CNV lesion area did not change throughout the follow-up. Nevertheless, interestingly, we showed a reduction in CNV flow area, as soon as 1 day after the first IVI, and this reduction remained stable up to the end of the follow-up. In accord with our results, Huang et al32 recently showed a similar pattern as our of decreasing in CNV flow area in a single case of exudative CNV closely followed by means of OCTA. However, at the best of our knowledge, we first showed the CNV flow area reduction pattern in a prospective study enrolling as many as 15 treatment-naive patients with AMD. Our results give credit to the theory that, in most cases, the treatment does not lead to a reduction in the size of CNV, but rather reduces the number and the perfusion of the smallest pathological vessels. Indeed, the larger trunks remained well-perfused after the treatment.33 Finally, we demonstrated that the reduction in CNV flow area was higher in patients experiencing a BCVA improvement after the 3 IVI. Moreover, retinal nerve fiber layer thickness changes occurring after IVI VEGF antagonist treatment have been much debated in the last years. Several authors showed a reduction in RNFL thickness after IVI.34,35 On the contrary, other authors showed that long-term treatment with anti-VEGF drugs did not lead to changes in RNFL thickness.36,37 We first showed a reduction in RNFL thickness after aflibercept IVI as soon as 2 months after the first injection. In theory, RNFL modifications could be induced by several mechanisms 1) the intraocular volume increase produced by the injection could lead to intraocular pressure spikes38; 2) a second hypothesis, although less plausible, would be a toxic effect on the nerve fiber layer, by reducing VEGF free quantities, the latter having a protective effect on ganglion cells.39 Further studies should evaluate the relationship between the reduction in superficial vascular plexus flow density and in RNFL thickness, to examinee a fascinating theory of ischemia conditioning death of retinal nerve fibers. Our study has several limitations. The series presented here is relatively small. The latter aspect, in addition to the fact that this study lacks a control group, implies that differences during the follow-up could have been occurred even if the treatment was not applied. However, one should look at the current series in

255

consideration of the strict inclusion criteria for patients with wet AMD with AMD. In particular, we did not enroll patients already treated with intravitreal injection, to evaluate as better as possible the effect of aflibercept IVI on the native vasculature. Furthermore, we did not evaluate the deep vascular plexus because of the altered macular architecture, secondary to the Type 1 CNV, did not allow a proper segmentation of the deep vascular layer. Another major limitation is that we followed up the patients up to 2 months after the third IVI. The latter limit did not allow us to assess whether the changes occurring in retinal vessels are reversible or not. In conclusion, we provide the first fully integrated study of retinal and choroidal blood supply in treatment-naive patients with wet AMD with AMD undergoing 3 monthly aflibercept treatments. We showed that both choroid and retinal vessels modified throughout the follow-up. Moreover, we showed that the CNV flow reduced as soon as 1 day after the IVI. This study raises many questions and prompts further investigations including the evaluation of the relationship between these changes and the retinal atrophy sometimes associated with the VEGF treatment.40 However, these effects had no impact on the beneficial effect of aflibercept on the CNV, indeed the treatment with VEGF antagonists remain the standard of care in exudative AMD. Key words: AMD, wet AMD, aflibercept, intravitreal injection, optical coherence tomography angiography, OCTA, flow density, superficial plexus. References 1. Friedman DS, O’Colmain BJ, Muñoz B, et al. Prevalence of age-related macular degeneration in the United States. Arch Ophthalmol 2004;122:564–572. 2. Coleman HR, Chan CC, Ferris FL, Chew EY. Age-related macular degeneration. Lancet 2008;372:1835–1845. 3. Schmid MK, Bachmann LM, Fäs L, et al. Efficacy and adverse events of aflibercept, ranibizumab and bevacizumab in agerelated macular degeneration: a trade-off analysis. Br J Ophthalmol 2015;99:141–146. 4. Yuzawa M, Fujita K, Wittrup-Jensen KU, et al. Improvement in vision-related function with intravitreal aflibercept: data from phase 3 studies in wet age-related macular degeneration. Ophthalmology 2015;122:571–578. 5. Singh RP, Srivastava SK, Ehlers JP, et al. A single-arm, investigator-initiated study of the efficacy, safety, and tolerability of intravitreal aflibercept injection in subjects with exudative age-related macular degeneration previously treated with ranibizumab or bevacizumab (ASSESS study): 12-mo. Clin Ophthalmol 2015;9:1759–1766. 6. Spaide RF, Klancnik JM, Cooney MJ. Retinal vascular layers imaged by fluorescein angiography and optical coherence tomography angiography. JAMA Ophthalmol 2015; 133:45–50. 7. Bambo MP, Garcia-Martin E, Otin S, et al. Visual function and retinal nerve fibre layer degeneration in patients with

256

8. 9. 10. 11.

12.

13. 14. 15. 16. 17. 18. 19. 20.

21.

22.

23.

24.


Alzheimer disease: correlations with severity of dementia. Acta Ophthalmol 2015;93:e507–508. Yu JG, Feng YF, Xiang Y, et al. Retinal nerve fiber layer thickness changes in Parkinson disease: a meta-analysis. PLoS One 2014;9:e85718. Eylea [package Insert]. Tarrytown, NY: Regeneron Pharmaceuticals, Inc. Available at: http://www.regeneron.com/Eylea/ eylea-Fpi.pdf. Accessed February 15, 2016. Lumbroso B, Huang D, Jia Y, et al. Clinical Guide to AngioOct Non Invasive, Dyeless OCT Angiography. 1st ed. New Delhi, India: Jaypee Brothers Medical Publisher (P) Ltd; 2015. Toto L, Borrelli E, Di Antonio L, et al. Retinal vascular plexuses’ changes in dry age-related macular degeneration, evaluated by means of optical coherence tomography angiography. Retina 2016 Jan 19. Epub ahead of print. doi: 10.1097/IAE. 0000000000000962. Huang Y, Zhang Q, Thorell MR, et al. Swept-source OCT angiography of the retinal vasculature using intensity differentiation-based optical microangiography algorithms. Ophthalmic Surg Lasers Imaging Retina 2014;45:382–389. Moult E, Choi W, Waheed NK, et al. Ultrahigh-speed sweptsource OCT angiography in exudative AMD. Ophthalmic Surg Lasers Imaging Retina 2014;45:496–505. Savastano MC, Lumbroso B, Rispoli M. In vivo characterization of retinal vascularization morphology using optical coherence tomography angiography. Retina 2015;35:2196–2203. Jia Y, Tan O, Tokayer J, et al. Split-spectrum amplitudedecorrelation angiography with optical coherence tomography. Opt Express 2012;20:4710–4725. Jia Y, Morrison JC, Tokayer J, et al. Quantitative OCT angiography of optic nerve head blood flow. Biomed Opt Express 2012;3:3127. de Carlo TE, Bonini Filho MA, Chin AT, et al. Spectraldomain optical coherence tomography angiography of choroidal neovascularization. Ophthalmology 2015;122:1228–1238. Sehi M, Grewal DS, Sheets CW, Greenfield DS. Diagnostic ability of Fourier-domain vs time-domain optical coherence tomography for glaucoma detection. Am J Ophthalmol 2009;148:597–605. Wei E, Jia Y, Tan O, et al. Parafoveal retinal vascular response to pattern visual stimulation assessed with OCT angiography. PLoS One 2013;8:e81343. Bonnin P, Pournaras JAC, Lazrak Z, et al. Ultrasound assessment of short-term ocular vascular effects of intravitreal injection of bevacizumab (Avastin) in neovascular age-related macular degeneration. Acta Ophthalmol 2010;88:641–645. Bonnin P, Pournaras JAC, Makowiecka K, et al. Ultrasound assessment of ocular vascular effects of repeated intravitreal injections of ranibizumab for wet age-related macular degeneration. Acta Ophthalmol 2014;92:e382–e387. Papadopoulou DN, Mendrinos E, Mangioris G, et al. Intravitreal ranibizumab may induce retinal arteriolar vasoconstriction in patients with neovascular age-related macular degeneration. Ophthalmology 2009;116:1755–1761. Mendrinos E, Mangioris G, Papadopoulou DN, et al. Long-term results of the effect of intravitreal ranibizumab on the retinal arteriolar diameter in patients with neovascular age-related macular degeneration. Acta Ophthalmol 2013;91:e184–e190. Mazaraki K, Fassnacht-Riederle H, Blum R, et al. Change in choroidal thickness after intravitreal aflibercept in pretreated and treatment-naive eyes for neovascular age-related macular degeneration. Br J Ophthalmol 2015;99:1341–1344.

25. Koizumi H, Kano M, Yamamoto A, et al. Short-term changes in choroidal thickness after aflibercept therapy for neovascular age-related macular degeneration. Am J Ophthalmol 2015;159: 627–633. 26. Tilton RG, Chang KC, LeJeune WS, et al. Role for nitric oxide in the hyperpermeability and hemodynamic changes induced by intravenous VEGF. Invest Ophthalmol Vis Sci 1999;40: 689–696. 27. Kamba T, McDonald DM. Mechanisms of adverse effects of anti-VEGF therapy for cancer. Br J Cancer 2007;96:1788– 1795. 28. Peters S, Heiduschka P, Julien S, et al. Ultrastructural findings in the primate eye after intravitreal injection of bevacizumab. Am J Ophthalmol 2007;143:995–1002. 29. Shimomura Y, Hirata A, Ishikawa S, Okinami S. Changes in choriocapillaris fenestration of rat eyes after intravitreal bevacizumab injection. Graefes Arch Clin Exp Ophthalmol 2009; 247:1089–1094. 30. Jia Y, Bailey ST, Wilson DJ, et al. Quantitative optical coherence tomography angiography of choroidal neovascularization in age-related macular degeneration. Ophthalmology 2014; 121:1435–1444. 31. Kuehlewein L, Bansal M, Lenis TL, et al. Optical coherence tomography angiography of type 1 neovascularization in agerelated macular degeneration. Am J Ophthalmol 2015;160: 739–748.e2. 32. Huang D, Jia Y, Rispoli M, et al. Optical coherence tomography angiography of time course of choroidal neovascularization in response to anti-angiogenic treatment. Retina 2015;35: 2260–2264. 33. Coscas G, Lupidi M, Coscas F, et al. Optical coherence tomography angiography during follow-up: qualitative and quantitative analysis of mixed type I and ii choroidal neovascularization after vascular endothelial growth factor trap therapy. Ophthalmic Res 2015;54:57–63. 34. Martinez-de-la-Casa JM, Ruiz-Calvo A, Saenz-Frances F, et al. Retinal nerve fiber layer thickness changes in patients with age-related macular degeneration treated with intravitreal ranibizumab. Invest Ophthalmol Vis Sci 2012;53:6214–6218. 35. Parlak M, Oner FH, Saatci AO. The long-term effect of intravitreal ranibizumab on retinal nerve fiber layer thickness in exudative age-related macular degeneration. Int Ophthalmol 2015;35:473–480. 36. Horsley MB, Mandava N, Maycotte MA, Kahook MY. Retinal nerve fiber layer thickness in patients receiving chronic antivascular endothelial growth factor therapy. Am J Ophthalmol 2010;150:558–561.e1. 37. Demirel S, Batio˘glu F, Özmert E, Erenler F. The effect of multiple injections of ranibizumab on retinal nerve fiber layer thickness in patients with age-related macular degeneration. Curr Eye Res 2015;40:87–92. 38. Falkenstein IA, Cheng L, Freeman WR. Changes of intraocular pressure after intravitreal injection of bevacizumab (avastin). Retina 2007;27:1044–1047. 39. Brar VS, Sharma RK, Murthy RK, Chalam KV. Bevacizumab neutralizes the protective effect of vascular endothelial growth factor on retinal ganglion cells. Mol Vis 2010;16:1848–1853. 40. Channa R, Sophie R, Bagheri S, et al. Regression of choroidal neovascularization results in macular atrophy in anti-vascular endothelial growth factor-treated eyes. Am J Ophthalmol 2015; 159:9–19.e1–e12.