Assessing Deep Retinal Capillary Ischemia in Paracentral Acute Middle Maculopathy by Optical Coherence Tomography Angiography JULIA NEMIROFF, LAURA KUEHLEWEIN, EHSAN RAHIMY, IRENA TSUI, RISHI DOSHI, ALAIN GAUDRIC, MICHAEL B. GORIN, SRINIVAS SADDA, AND DAVID SARRAF PURPOSE:
To assess microvascular blood flow of the deep retinal capillary plexus in eyes with paracentral acute middle maculopathy using optical coherence tomography (OCT) angiography. DESIGN: Retrospective, multicenter observational case series. METHODS: Clinical and multimodal imaging findings from 8 patients with paracentral acute middle maculopathy were reviewed and analyzed. OCT angiography scans were analyzed and processed, and vessel density was calculated. RESULTS: Eight patients (7 male, 1 female, aged 9–82 years) were included. OCT angiography was obtained at either the acute (4 cases) or old stage (4 cases). Scans of the deep capillary plexus showed preservation of perfusion in acute lesions and capillary attenuation in old cases. Cases of central retinal artery occlusion showed marked loss of the deep capillary plexus. The mean vessel density of the superficial capillary plexus in normal fellow eyes was 12.8 ± 1.8 mmL1 vs 12.1 ± 1.9 mmL1 in eyes with paracentral acute middle maculopathy (reduction L6.0%, P [ .08). The mean vessel density of the deep capillary plexus in normal fellow eyes was 17.5 ± 1.4 mmL1 vs 14.7 ± 3.5 mmL1 in eyes with paracentral acute middle maculopathy (reduction L19.4%, P [ .04). This significant difference was representative of the eyes with old lesions. CONCLUSION: Paracentral acute middle maculopathy lesions correspond to preservation of perfusion in focal acute lesions and to pruning of the plexus in old cases. Cases of central retinal artery occlusion demonstrate marked hypoperfusion of the deep capillary plexus. Our
Accepted for publication Oct 29, 2015. From the Stein Eye Institute, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, California (J.N., I.T., M.B.G., D.S.); Institute for Ophthalmic Research, Center for Ophthalmology, Eberhard Karls University Tu¨bingen, Tu¨bingen, Germany (L.K., S.S.); Retina Service, Wills Eye Hospital, Philadelphia, Pennsylvania (E.R.); Kaiser Permanante, Tustin, California (R.D.); Ophtalmologie, Hopital Lariboisiere, AP-HP, Universite´ Paris 7 – Sorbonne Paris Cite´, Paris, France (A.G.); and Greater Los Angeles Veterans Affairs Healthcare Center, Los Angeles, California (D.S.). Inquiries to David Sarraf, Retinal Disorders and Ophthalmic Genetics Division, Stein Eye Institute, David Geffen School of Medicine at UCLA, 100 Stein Plaza, Los Angeles, California 90095; e-mail: dsarraf@ ucla.edu 0002-9394/$36.00 http://dx.doi.org/10.1016/j.ajo.2015.10.026
Ó
2015 BY
study further supports an ischemic pathogenesis of this retinal vasculopathy. (Am J Ophthalmol 2015;-: -–-. Ó 2015 by Elsevier Inc. All rights reserved.)
P
ARACENTRAL ACUTE MIDDLE MACULOPATHY IS A
recently described entity in patients presenting with an acute-onset paracentral scotoma. Spectraldomain optical coherence tomography (OCT) reveals hyperreflective band-like lesions at the level of the inner nuclear layer. Although these lesions resolve, patients are left with atrophy of the inner nuclear layer, resulting in a permanent paracentral visual field defect. Paracentral acute middle maculopathy can be idiopathic, or it can be secondary to local retinal vascular or systemic disease. Deep retinal capillary ischemia has been proposed as a causative factor in the development of these lesions, as the intermediate and deep retinal capillary plexuses flank the inner and outer boundaries of the inner nuclear layer, respectively.1–3 Fluorescein angiography, the reference standard for visualizing the retinal vasculature, has limited depth resolution.4 With OCT angiography, it is possible to obtain high-resolution, depth-resolved en face images of the retinal microvasculature by calculating motion contrast in OCT B-scans acquired repeatedly at the same location. OCT angiograms are co-registered with OCT B-scans from the same location, allowing simultaneous visualization of structure and blood flow.5–8 We used OCT angiography to study the retinal capillary microvasculature at different levels of the retina in 8 cases of paracentral acute middle maculopathy occurring in various clinical scenarios.
METHODS APPROVAL FOR THIS STUDY WAS OBTAINED FROM THE UNI-
versity of California Los Angeles (UCLA) Institutional Review Board committee. Research adhered to the tenets of the Declaration of Helsinki and was conducted in accord with regulations set forth by the Health Insurance Portability and Accountability Act. We retrospectively identified 8 patients from 3 tertiary referral centers. All were evaluated between August 2014
ELSEVIER INC. ALL
RIGHTS RESERVED.
1
TABLE 1. Clinical Characteristics and Associated Pathology in Patients Presenting With Paracentral Acute Middle Maculopathy Case
Sex
Age (y)
Affected Eye
1 2 3
Male Male Male
66 72 31
OD OD OS
4 5
Male Female
63 57
6 7 8
Male Male Male
82 16 9
BCVA
Ophthalmologic Diagnosis
Systemic Disease
20/50 20/20-1 20/20-1
CRAO BRAO BRAO
OS OS
20/30 20/300
CRAO CRAO
OS OU OD
CF at 4 ft 20/20 20/20-1
HTN; type II DM; hypothyroidism HTN; atrial fibrillation; mitral valve repair; pacemaker Transposition of the great vessels (status post repair); head trauma 2 days prior Left middle cerebral artery infarction; HTN Internal carotid artery dissection; cerebral vascular occlusion coinciding with ocular event; meningioma; hyperlipidemia HTN; type II DM Sickle cell disease (hemoglobin type SC) Alpha-thalassemia trait
CRAO, NPDR SCR Trauma
BCVA ¼ best-corrected visual acuity; BRAO ¼ branch retinal artery occlusion; CF ¼ count fingers; CRAO ¼ central retinal artery occlusion; DM ¼ diabetes mellitus; HTN ¼ hypertension; NPDR ¼ nonproliferative diabetic retinopathy; SCR ¼ sickle cell retinopathy.
and July 2015 and diagnosed with paracentral acute middle maculopathy. Written informed consent was obtained for each patient. We reviewed the clinical and multimodal imaging data for each patient and obtained OCT angiography analysis. The diagnostic criteria of paracentral acute middle maculopathy included a history of acute-onset paracentral scotoma with or without decline in visual acuity and a nonprogressive course. All patients demonstrated characteristic abnormalities with spectral-domain OCT, including either the acute finding of hyperreflective, plaque-like bands within the inner nuclear layer or old lesions demonstrating thinning of the inner nuclear layer. High-resolution digital color imaging, red-free photography, spectral-domain OCT, and OCT angiography were performed at the time of presentation for each patient. Fluorescein angiography was obtained in 7 of the 8 cases. The AngioVue OCT angiography system (Optovue, Inc, Fremont, California, USA) operates at 70, 000 A-scans per second to acquire OCT angiography volumes consisting of 304 3 304 A-scans in approximately 2.6 seconds. It uses a split-spectrum amplitude decorrelation angiography software algorithm and orthogonal registration and merging of 2 consecutive scan volumes to obtain 3 3 3 mm and 6 3 6 mm OCT angiography volumes of both eyes of each patient. OCT angiograms were co-registered with OCT B-scans, to allow visualization of both retinal vasculature and structure, and were performed in both the normal and affected eyes of each patient. The OCT angiography software was used to segment the superficial capillary plexus and deep capillary plexus in 3 3 3 mm scans. Consistently in each patient, the superficial capillary plexus slab was taken from the internal limiting membrane (offset 3 mm) to the inner plexiform layer (offset 15 mm). The deep capillary plexus slab was taken from the inner plexiform layer (offset 15 mm) to the outer plexiform layer (offset 70 mm). In cases where thinning of the inner nuclear layer caused failure of automatic segmentation, a 2
thinner 30 mm band was manually adjusted to include the deep capillary plexus. OCT angiography analysis of the superficial and deep capillary plexus was performed independently by a trained reader to assess for capillary attenuation and pruning involving either plexus. Quantitative analyses were performed using the publicly available software Fiji ImageJ 2.0.0-rc-29/1.49q (http://fiji. sc/Fiji)9 and GNU Image Manipulation Program GIMP 2.8.14 (http://gimp.org). Fiji was used to binarize and skeletonize the en face image of the superficial and deep retinal capillary plexus, showing the blood vessels as a 1-pixelwide line. GIMP was used to count the number of black pixels and total pixels. Vessel density was then calculated as [(pixels of vessels) (3/304)]/(area in mm2) in mm1.10,11 In cases of partial or complete projection artifact from the superficial capillary plexus, Fiji was used to binarize the superficial and deep capillary plexus images. GIMP was then used to subtract the black pixels of the superficial capillary plexus from the deep capillary plexus image. The resulting deep capillary plexus was skeletonized using Fiji software, and pixels were calculated using GIMP. Vessel density calculation was then repeated in the manner described above.
RESULTS NINE EYES WITH PARACENTRAL ACUTE MIDDLE MACULOP-
athy (8 patients: 7 male, 1 female) were identified and enrolled in this study. Clinical characteristics are summarized in Table 1. Patient ages ranged from 9 to 82 years (mean 49.5 years, median 60). Isolated band-like hyperreflective lesions in the middle retinal layers consistent with acute paracentral middle maculopathy lesions were observed with spectral-domain OCT imaging in 5 patients at baseline. Three patients presented with patchy
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
thinning of the inner nuclear layer on spectral-domain OCT, consistent with old paracentral acute middle maculopathy. Retinal vascular etiologies leading to paracentral acute middle maculopathy included central retinal artery occlusion (CRAO, n ¼ 4), branch retinal artery occlusion (BRAO, n ¼ 2), suction eye injury (n ¼ 1), and sickle cell retinopathy (n ¼ 1). OCT angiography was obtained at the acute stage of the lesion in 4 cases and at the old stage in 4 cases. The superficial capillary plexus demonstrated minimally attenuated perfusion in affected eyes. Focal acute cases (n ¼ 2, Cases 2 and 3) demonstrated patent perfusion of the deep capillary plexus within the lesions, whereas focal old cases (n ¼ 2, Cases 7 and 8) showed pruning of the deep capillary plexus within the lesions. One month follow-up analysis was obtained for Case 2 and demonstrated normal vascular density of the deep capillary plexus. Four cases of diffuse paracentral acute middle maculopathy associated with CRAO (Cases 1 and 4 with acute lesions, and Cases 5 and 6 with old lesions) demonstrated profound to complete absence of the deep capillary plexus with significant projection artifact from the superficial capillary plexus. Quantitative analysis was carried out in all cases. Because the patient in Case 7 was affected bilaterally, and thus had no control fellow eye, Case 7 was excluded from statistical analysis. Table 2 includes the vessel density measurements for each eye (n ¼ 9), and Table 3 shows a summary of the statistical analysis that was carried out comparing the patients with unilateral disease (n ¼ 7). The mean vessel density of the superficial capillary plexus in normal fellow eyes was 12.8 6 1.8 mm1 vs a mean of 12.1 6 1.9 mm1 in eyes with paracentral acute middle maculopathy (reduction 6.0%, P ¼ .33). The mean vessel density of the deep capillary plexus in normal fellow eyes was 17.5 6 1.4 mm1 vs a mean of 14.7 6 3.5 mm1 in eyes with paracentral acute middle maculopathy (reduction 19.4%, P ¼ .04). Cases 1 through 4 were scanned at the acute stage of the paracentral acute middle maculopathy lesion (n ¼ 4), and Cases 5 through 8 were scanned at the old stage of the lesion (n ¼ 4). Focal acute lesions (n ¼ 2, Cases 2 and 3) demonstrated preservation of perfusion of the deep capillary plexus (18.4 mm1) compared with the normal fellow eyes (18.2 mm1), whereas eyes with focal old lesions (n ¼ 1, Case 8) demonstrated attenuation of the deep capillary plexus (16.7 mm1) compared with the normal fellow eye (19.3 mm1) (Table 3). Localized vascular density analysis was also performed and compared between normal eyes and eyes with focal acute lesions. In Case 2, 33% of the 3 3 3 mm cube was involved, and the deep capillary density measured 6.3 mm1 in the affected eye vs 6.5 mm1 for the corresponding area in the unaffected eye. In Case 3, 25.6% of the 3 3 3 mm cube was involved, and capillary density in the affected area measured 4.8 mm1 vs 5.2 mm1 in the corresponding area of the unaffected, contralateral eye. This analysis confirmed the preservation of perfusion VOL. -, NO. -
TABLE 2. Retinal Capillary Density as Measured on Optical Coherence Tomography Angiography images in Paracentral Acute Middle Maculopathy Vessel Density (mm1) Superficial Capillary Plexus
Case
Deep Capillary Plexus
Affected Eye
Acute vs Chronic PAMM
Normal Eye
PAMM Eye
Normal Eye
PAMM Eye
OD OD OS OS OS OS OD OS OD
Acute Acute Acute Acute Chronic Chronic Chronic Chronic Chronic
12.2 12.9 14.9 9.7 13.1 12.2 N/A N/A 14.9
11.8 12.5 14.4 11.7 10.9 9.1 15.0 13.1 14.5
17.1 17.6 19.3 16.4 16.8 15.8 N/A N/A 19.3
15.7 17.8 18.7 13.9 11.0 9.2 21.3 19.4 16.7
1 2 3 4 5 6 7 8
N/A ¼ not applicable; PAMM ¼ paracentral acute middle maculopathy.
within the deep retinal capillary plexus for acute focal lesions. The 4 cases of CRAO with both acute (n ¼ 2) and old (n ¼ 2) paracentral acute middle maculopathy lesions showed partial or complete loss of the deep capillary plexus with overlying projection artifact from the superficial capillary plexus (Cases 1, 4, 5, and 6). The degree of projection artifact was obtained by recalculating the capillary density and subtracting the density of the superficial capillary plexus from that of the deep capillary plexus. A summary of calculated vessel density of the resulting deep capillary plexus is shown in Table 4. The mean vessel density of the deep plexus after subtraction in normal fellow eyes was 11.4 mm1 vs 7.5 mm1 in eyes with paracentral acute middle maculopathy and CRAO (reduction 41.2%) despite a relatively similar superficial vessel density in both groups. Representative cases (2, 3, 6, and 8) are described in detail below with corresponding figures. CASE 2:
A 72-year-old man (Figure 1) presented with a 3day history of seeing a ‘‘red line’’ across his right eye. Past medical history was notable for hypertension, atrial fibrillation, mitral valve repair, and pacemaker placement. Bestcorrected visual acuity (BCVA) was 20/20-1 in the right eye and 20/20-2 in the left eye. Fundus examination showed a patch of retinal whitening along the superior arcade and a normal retinal examination of the left eye. Spectral-domain OCT (Figure 1, Bottom row) showed numerous dense hyperreflective foci of the inner nuclear layer in band-like patches throughout the superior macula, consistent with paracentral acute middle maculopathy. An occult BRAO was suspected.
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
3
TABLE 3. Mean Parafoveal Vessel Density of the Superficial Capillary Plexus and Deep Capillary Plexus in Eyes Affected by Paracentral Acute Middle Maculopathy Compared With the Normal Fellow Eye and Separated Into Acute Versus Chronic Lesions
Variable 1
Superficial capillary plexus normal eye (mm ) Superficial capillary plexus PAMM eye (mm1) Superficial capillary plexus percent reduction (%) Deep capillary plexus normal eye (mm1) Deep capillary plexus PAMM eye (mm1) Deep capillary plexus percent reduction (%)
Mean 6 SD Vessel Density
Mean 6 SD Vessel Density in Acute Lesions
Mean 6 SD Vessel Density in Chronic Lesions
N¼7
N¼4
N¼3
12.8 6 1.8 12.1 6 1.9 6.0 6 14.9 (P ¼ .33a) 17.5 6 1.4 14.7 6 3.5 19.4 6 20.1 (P ¼ .04a)
12.4 6 2.1 12.6 6 1.2 þ2.1 6 10.9 17.6 6 1.2 16.5 6 2.1 6.8 6 7.4
13.4 6 1.4 11.5 6 2.7 16.9 6 13.4 17.3 6 1.8 12.3 6 3.9 36.3 6 19.7
PAMM ¼ paracentral acute middle maculopathy. P values from paired t test.
a
TABLE 4. Vascular Density of the Deep Capillary Plexus After Subtraction of Projection Artifact From the Superficial Capillary Plexus Measured on Optical Coherence Tomography Angiography in Eyes Affected by Paracentral Acute Middle Maculopathy and Central Retinal Artery Occlusion and in Normal Fellow Eyes Vessel Density of the Deep Capillary Plexus Normal Eye
PAMM Eye
Case
PAMM Eye
Acute vs Chronic PAMM
Original (mm-1)
After Subtractiona (mm-1)
PR (%)
Original (mm-1)
After Subtractiona (mm-1)
PR (%)
1 4 5 6
OD OS OS OS
Acute Acute Chronic Chronic
17.1 16.4 16.7 15.8
11.8 11.8 9.2 12.8
36.6 32.2 58.0 20.5
15.7 13.9 11.0 9.2
10.0 8.0 4.4 7.7
44.5 54.4 86.1 17.2
PAMM ¼ paracentral acute middle maculopathy; PR ¼ percent reduction (as calculated between original vessel density and vessel density after subtraction). a Subtraction of the superficial capillary plexus projection artifact from the deep capillary plexus.
On en face OCT angiography, the superficial capillary plexus of the affected right eye appeared to be within normal limits. The deep capillary plexus (Figure 1, First row, second from left) demonstrated patent structure with careful analysis of the en face images, and patent flow and perfusion with careful study of the OCT angiograms, corresponding to the hyperreflective paracentral acute middle maculopathy lesions with adjacent areas of hyporeflectivity, likely due to a contrast effect of the adjacent bright hyperreflective plaques. Vessel density was calculated for the superficial and deep retinal capillary plexus. There was minimal difference in vessel density between each eye in the superficial capillary plexus (12.9 mm1 in normal left eye vs 12.5 mm1 in affected right eye, reduction 3.1%) and the deep capillary plexus (17.6 mm1 in normal left eye vs 17.8 mm1 in affected right eye, difference þ0.8%). The patient returned after 1 month and noted that the paracentral scotomas were present but improved. The OCT showed persistent hyperreflective lesions in the inner nuclear layer with corresponding thinning. Repeat OCT 4
angiography did not show any reduction of the vascularity of the deep capillary plexus based on quantitative capillary density comparisons between the affected and unaffected areas. CASE 3:
A 31-year-old man (Figure 2) presented with acute loss of vision in the superior visual field of his left eye after left-sided head trauma while intoxicated with alcohol. Past medical history was notable for a transposition of the great vessels, surgically repaired when he was an infant. BCVA was 20/20 in the right eye and 20/201 in the left eye. Ophthalmoscopic examination of the right eye was unremarkable except for 1 dot-blot hemorrhage inferiorly Ophthalmoscopic examination of the left eye revealed attenuation of the inferotemporal retinal arteriole consistent with a BRAO. Corresponding fluorescein angiography (Figure 2) confirmed a peripheral filling defect of the inferotemporal retinal arteriole. Spectraldomain OCT (Figure 2) revealed a band of hyperreflectivity at the level of the inner nuclear layer consistent with paracentral acute middle maculopathy. OCT angiography
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
FIGURE 1. Case 2. En face optical coherence tomography (OCT) angiography and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the right eye of a 72-year-old man. (First row, first from left) A 3 3 3 mm macular cube OCT with en face image of the deep capillary plexus in the right eye shows prominent hyperreflective lesions corresponding to hyperreflective, plaque-like bands at the level of the inner nuclear layer on B-scan spectral-domain optical coherence tomography, consistent with paracentral acute middle maculopathy (Second row, second from left). (Second row, first from left) Corresponding infrared image is shown. White arrows demonstrate corresponding lesions. (First row, second and third from left) A 3 3 3 mm macular cube OCT angiography with en face projection capturing the deep capillary plexus of the right and left eye, respectively, demonstrates patent perfusion superiorly in the right eye. (Third row, first from left). A 6 3 6 mm macular cube OCT angiography with en face projection of the deep capillary plexus clearly demonstrates perfusion within the brightly colored paracentral acute middle maculopathy lesions. Adjacent areas are perfused and may appear dark because of a contrast effect. (Third row, second and third from left) OCT angiography scans of the deep capillary plexus have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1); on quantitative analysis, no significant difference was noted in the capillary density of either the superficial or deep plexus in the normal vs the affected eyes. (Bottom row, first from left) A 3 3 3 mm macular cube OCT angiography with en face projection capturing the deep capillary plexus of the right eye 1 month after initial presentation demonstrates normal perfusion. (Bottom row, second from left) OCT of the right eye at 1 month follow-up demonstrating persistent multifocal hyperreflectivity of the inner nuclear layer with attenuation.
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
5
FIGURE 2. Case 3. Fluorescein angiography, en face optical coherence tomography (OCT) angiography, and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the setting of branch retinal artery occlusion in the left eye of a 31-year-old man. (First row, first from left) Fluorescein angiography shows an infratemporal peripheral filling defect, as demonstrated by small white arrows. (First row, second from left) A 3 3 3 mm macular cube OCT with en face image of the deep capillary plexus in the affected left eye shows a prominent area of hyperreflectivity inferiorly corresponding to the paracentral acute middle maculopathy lesion. (Second row) Spectral-domain OCT (B-scan) of the inferior macula shows an acute focal hyperreflective, plaque-like band at the level of the inner nuclear layer consistent with paracentral acute middle maculopathy. (Bottom row, first and second from left) A 3 3 3 mm macular cube OCT angiography with en face projection capturing the deep capillary plexus of the right and left eye demonstrates patent perfusion of the deep capillary plexus in each eye. (Bottom row, third and fourth from left) For quantitative analysis, OCT angiography scans of the deep capillary plexus have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1). Qualitative and quantitative analysis was not significantly different.
(Third row, second from left) revealed patent flow and perfusion in the deep capillary plexus of the left eye, corresponding to the paracentral acute middle maculopathy lesion on spectral-domain OCT. Vessel density was calculated for the superficial capillary plexus and deep 6
capillary plexus and results for each eye were compared. There was minimal difference in the superficial capillary plexus (14.9 mm1 in the normal right eye vs 14.4 mm1 in the affected left eye, reduction 3.4%) and in the deep capillary plexus (19.3 mm1 in the
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
FIGURE 3. Case 6. En face optical coherence tomography (OCT) angiography and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the setting of central retinal artery occlusion in the left eye of a 82-year-old man. (First row, first from left) Spectral-domain OCT shows diffuse hyperreflectivity at the level of the inner nuclear level, consistent with paracentral acute middle maculopathy. (Second row, first from left) Follow-up OCT 3 months later shows diffuse thinning of the
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
7
normal right eye vs 18.7 mm1 in the affected left eye, reduction 3.1%). CASE 6:
An 82-year-old man (Figure 3) presented with sudden loss of vision in the left eye. Past medical history was notable for hypertension, hyperlipidemia, and type 2 diabetes. He reported a temporal headache on the left side but denied any other symptoms of giant cell arteritis. BCVA was 20/30 in the right eye and count fingers at 4 feet in the left eye. Ophthalmoscopic examination of the right eye was within normal limits except for a flat benign choroidal nevus. Ophthalmoscopic examination of the left eye revealed attenuated retinal vessels in the absence of embolic material, associated with cotton-wool spots and a cherry-red spot in the macula. Fluorescein angiography revealed delayed retinal vascular flow. Spectral-domain OCT (Figure 3, Top left) showed diffuse hyperreflective band-like lesions at the level of the inner nuclear layer in the left eye, consistent with an acute CRAO associated with paracentral acute middle maculopathy. Erythrocyte sedimentation rate was within normal limits (26 mm/hr), but C-reactive protein was elevated to 3.2 mg/L, raising suspicion for a diagnosis of giant cell arteritis. The patient was admitted, given a 5-day course of intravenous solumedrol, and discharged home on a prednisone taper. A temporal artery biopsy on the left side showed no evidence of giant cell arteritis on pathology review. Upon follow-up 2 months later, the patient’s visual acuity had improved to 20/30 in the left eye. Spectral-domain OCT (Figure 3, Second row, first from left) showed diffuse thinning of the inner nuclear layer, corresponding to the prior hyperreflective paracentral acute middle maculopathy lesions. OCT angiography (Figure 3, Fourth row, second from left) obtained at this time demonstrated projection artifact of the superficial capillary plexus and only modest perfusion detectible in the deep capillary plexus of the left eye. Vessel density was calculated for the superficial capillary plexus and deep capillary plexus of each eye. In the affected left eye, there was attenuation
of both the superficial capillary plexus (12.2 mm1 in the normal right eye vs 9.1 mm1 in the affected left eye, reduction 29.5%) and, more significantly, the deep capillary plexus (15.8 mm1 in the normal right eye vs 9.2 mm1 in the affected left eye, reduction 52.7%). The amount of projection artifact was quantified by recalculating the retinal capillary density and subtracting the density of the superficial plexus from the deep plexus using imaging software. The resulting deep capillary plexus was further reduced by 17.2%. CASE 8:
A 9-year-old boy (Figure 4) presented with a 6month history of a red spot in the central field of the right eye. By history the scotoma developed immediately after the right eye was traumatized by the suction vacuum drain of a swimming pool. Past medical history was notable for alpha-thalassemia trait. BCVA was 20/20-1 in the right eye and 20/20 in the left eye. Humphrey visual field 10-2 testing revealed a paracentral scotoma in the right eye. Ophthalmoscopic retinal examination of each eye was unremarkable. Spectral-domain OCT (Figure 4, Top right) and en face (Figure 4, Top left) analysis revealed severe thinning of the inner nuclear layer and outer plexiform layer in a patchy parafoveal distribution, consistent with old paracentral acute middle maculopathy lesions in the right eye. With en face OCT angiography analysis, the superficial capillary plexus (Figure 4, Second row, first from left) appeared to be within normal limits; but the deep capillary plexus (Figure 4, Third row, first from left) demonstrated patchy areas of perifoveal capillary attenuation and pruning, consistent with deep capillary ischemia. Vessel density was calculated for the superficial capillary plexus and deep capillary plexus of each eye. In the affected right eye, there was minimal reduction in the superficial capillary plexus (14.9 mm1 in the normal left eye vs 14.5 mm1 in the affected right eye, reduction 2.7%) and significant attenuation of the deep capillary plexus (19.3 mm1 in the normal left eye vs 16.7 mm1 in the affected right eye, reduction 14.4%).
inner nuclear layer. (First row, second from left and Second row, second from left) A 3 3 3 mm macular cube OCT with en face image of the deep capillary plexus in the right and left eye shows normal perfusion of the deep capillary plexus in the right eye and a profound loss of the deep capillary plexus in the left eye. A 3 3 3 mm macular cube OCT angiography with en face projection capturing the superficial capillary plexus (Third row, first and second from left) and deep capillary plexus (Fourth row, first and second from left) of the right and left eye, respectively, demonstrates projection artifact from the superficial capillary plexus in the left eye, as evidenced by large-caliber vessels not characteristic of the deep capillary plexus that correspond to identical vessels in the superficial capillary plexus, as well as marked reduction in perfusion of the deep capillary plexus (Fourth row, second from left). For quantitative analysis, OCT angiography scans of the superficial capillary plexus (Third row, third and fourth from left) and deep capillary plexus (Fourth row, third and fourth from left) have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1). Note significant projection artifact in the deep capillary plexus from the superficial capillary plexus in the affected left eye (Fourth row, fourth from left). (Bottom row, first from left and Bottom row, second from left) Projection artifact from the superficial capillary plexus was removed using Fiji and GIMP imaging software, and the resulting binarized and skeletonized images of the deep capillary plexus of the normal right eye and affected left eye are pictured. Note that the normal right eye is essentially unchanged after subtraction (Bottom left), while the left eye appears significantly more attenuated (Bottom right).
8
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
FIGURE 4. Case 8. En face optical coherence tomography (OCT) angiography and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the setting of a history of a suction injury to the right eye 6 months prior from the vacuum suction of a pool drain in a 9-year-old boy. (First row, first and second from left) A 3 3 3 mm macular cube OCT with en face imaging of the deep capillary plexus in the right and left eye, respectively, shows dark patchy areas of absent vascular markings in a perifoveal distribution in the right eye and a normal deep retinal plexus in the left eye. (First row, third from left) Spectral-domain OCT B-scan image of the right eye shows severe thinning of the inner nuclear layer and outer plexiform layer in a multifocal distribution, consistent with a history of paracentral acute middle maculopathy. A 3 3 3 mm macular cube OCT angiography with en face projection capturing the superficial capillary plexus (Second row, first and second from left) and deep capillary plexus (Third row, first and second from left) of the right and left eye, respectively, demonstrates patchy perifoveal loss of flow within the deep capillary plexus of the right eye (Third row, first from left). For quantitative analysis, OCT angiography scans of the superficial capillary plexus (Second row, third and fourth from left) and deep capillary plexus (Third row, third and fourth from left) have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1) and shows significant attenuation of the capillary density of the deep plexus in the affected right eye (Third row, third from left).
DISCUSSION IN THIS STUDY, WE PERFORMED EN FACE OCT ANGIOGRAPHY
analysis of the superficial and deep retinal capillary plexus in normal eyes and in eyes with paracentral acute middle maculopathy. OCT angiography is an effective tool in the qualitative evaluation of paracentral acute middle maculopathy lesions.12,13 As described above, we demonstrate that focal acute paracentral acute middle maculopathy lesions show relative preservation of perfusion of the deep retinal capillary plexus, while old VOL. -, NO. -
lesions and eyes with diffuse paracentral acute middle maculopathy (acute and old) associated with CRAO exhibit marked reduction in perfusion. The association of paracentral acute middle maculopathy with deep capillary ischemia is not surprising, since the lesion always resides in the inner nuclear layer flanked by the intermediate and deep retinal capillary plexus. The subsequent development of permanent inner nuclear layer thinning is consistent with an ischemic infarct. Recently, there have been reports associating paracentral acute middle maculopathy with various retinal vascular diseases,
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
9
including diabetic retinopathy,14,15 retinal artery occlusion,14 central retinal vein occlusion,16 sickle cell retinopathy,15 and Purtscher retinopathy,14 further supporting an ischemic pathogenesis. Our report describes a cohort of patients with paracentral acute middle maculopathy similarly affected by retinal vascular disease in all but 1 case. We demonstrate 4 cases of paracentral acute middle maculopathy secondary to CRAO, 2 cases of paracentral acute middle maculopathy secondary to BRAO, 1 case of bilateral paracentral acute middle maculopathy in a patient with sickle cell retinopathy, and 1 case of isolated paracentral acute middle maculopathy occurring secondary to a suction injury. Patients in 4 cases (1–4) underwent OCT angiography in the acute stage; the remaining 4 patients (Cases 5–8) underwent OCT angiography at the old stage with atrophy of the inner nuclear layer. In our study, the average vessel density of the superficial capillary plexus in affected eyes was similar, although slightly attenuated, compared to that of the unaffected fellow eyes (reduction 6.0%, P ¼ .33), and the difference did not reach statistical significance. However, the deep capillary plexus was significantly attenuated compared with the normal fellow eye (19.4% reduction, P ¼ .04). Focal acute lesions demonstrated preservation of perfusion of the deep capillary plexus (18.4 mm1) as compared with the normal fellow eyes (18.2 mm1); the significant reduction in the density of the deep capillary retinal plexus is representative of old paracentral acute middle maculopathy lesions and lesions associated with CRAO. Followup analysis of Case 2 demonstrated preserved vascular density of the deep capillary plexus. OCT at this time showed some persistent hyperreflectivity of the inner nuclear layer with associated thinning. These findings suggest that the loss of capillary density seen in old focal cases may occur at some indeterminate point after injury. Cases of diffuse paracentral acute middle maculopathy associated with CRAO showed a marked loss of the deep capillary plexus in patients with both acute (Cases 1 and 4) and old (Cases 5 and 6) disease, with varying degrees of projection artifact from the superficial capillary plexus. We believe this may be secondary to profound capillary nonperfusion caused by large vessel occlusion. Our original calculation of the deep capillary plexus density in these cases likely highly overestimates the true degree of perfusion. To quantify the degree of projection artifact for these 4 cases, we used imaging software to subtract the vessel density of the superficial capillary plexus from the deep capillary plexus, both in eyes with paracentral acute middle maculopathy and in fellow control eyes. The percent reduction in vessel density of the deep capillary plexus was greater in the affected eyes vs normal eyes in Cases 1, 4, and 5, suggesting greater projection artifact from the superficial plexus in the eyes with paracentral acute middle maculopathy vs the normal fellow eyes. The patient in Case 6 had a markedly attenuated superficial capillary plexus; thus, for this case the percent reduction was not greater in the affected eye than in the control eye after 10
correction of the projection artifact. This methodology for correcting projection artifact has not previously been published, and is therefore not yet validated. Because this method is still preliminary, we report both sets of vessel density calculations. Of note, the BCVA in Case 1 was remarkably good, at 20/50. This is likely the case secondary to immediate reperfusion of the central retinal artery. It is also possible, although there are no data available to validate this claim, that CRAO occurring in conjunction with paracentral acute middle maculopathy may have a better prognosis, as the superficial capillary plexus is spared, with selective infarction of the deep capillary plexus. The preservation of perfusion in the acute stages of ischemic injury, and the subsequent damage caused by reperfusion, have been well described. Ischemiareperfusion injury is a condition in which blood flow returns to a tissue after a period of ischemia. Reperfusion limits immediate retinal damage but is accompanied by mechanisms such as generation of excessive reactive oxygen species,17 nitric oxide neurotoxicity,18,19 and inflammation.20 A recent study demonstrated that ischemia-reperfusion in rat retinas caused capillary dropout after 8–14 days of reperfusion and concluded these vascular changes occurred after neuronal loss.21 Another recent study22 demonstrated that the number of functional vessels decreased after ischemia and reperfusion, and further showed that loss of perfusion occurred predominantly in the deep capillary plexus rather than in capillaries located at the retinal surface. We surmise that the pathogenesis of paracentral acute middle maculopathy may be related to ischemia and reperfusion injury, accounting for the immediate restoration in microcirculation seen in some of the acute and focal paracentral acute middle maculopathy lesions and the subsequent legacy of atrophy of the inner nuclear layer associated with loss of the deep capillary plexus. In the case of diffuse acute paracentral acute middle maculopathy occurring in the setting of CRAO (ie, large vessel obstruction), no capillary reperfusion may occur, causing severe acute impairment of flow within the deep retinal capillary plexus that was observed in this study. In all of our cases, ischemia of the intermediate capillary plexus was likely a significant contributing factor in the development of paracentral acute middle maculopathy23; but it is impossible to separate the intermediate from the deep capillary plexus at this time using presently available OCT angiography technology. As this imaging field advances, with improved segmentation technology, we may be able to determine whether the intermediate or the deep retinal capillary plexus is more important in the etiology of paracentral acute middle maculopathy. Interestingly, mild attenuation of the superficial capillary plexus was identified in certain old cases of paracentral acute middle maculopathy in this study (Cases 5, 6, and 7), indicating that while ischemia is predominantly present at the level of the intermediate and deep capillary plexus,
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
the trilaminar arrangement of the retinal capillary plexuses is not completely independent. The advantages of our method included rapid and noninvasive acquisition of OCT angiography images, depth-resolved microvascular detection, and semiautomated analysis with quantitated calculation technique. In particular, our method of skeletonization using Fiji, with subsequent calculation of vessel density using GIMP, takes only a few minutes and yields identical results each time it is repeated for a specific image, and is thus clinically applicable. But our study also had several limitations. Technical limitations included segmentation errors in scans and low quality, caused by movement artifact or retinal pathology, requiring manual segmentation. Three out of 8 cases required manual segmentation. Case 1 was resegmented owing to errors in automatic segmentation, possibly due to motion artifact during initial acquisition. Motion artifact can be addressed by including faster eye tracking to improve the image quality, especially in subjects with poor fixation and reduced visual acuity. Cases 5 and 6 were resegmented owing to severe thinning of the inner nuclear layer, leading to inadvertent incorporation of the superficial vasculature. As mentioned in the cases with profound attenuation of the deep capillary plexus, projection artifact from the superficial capillary plexus confounded imaging and quantitated analysis. Our method of subtraction of the projected superficial capillary plexus from the deep capillary plexus likely removed a proportion of the true deep plexus as well, possibly leading to an underestimation of the true plexus. Further, OCT angiography could only be used to map vessels with flow; other vessels may have been present in areas of capillary attenuation but their flow was too slow to be visualized. We did not correct for ocular magnification in the calculation of vessel density because axial length measurements were not obtained for all patients. Our method of calculating vessels by automated binarization may cause vessels to be incorrectly removed or added based on image quality and artifact. Finally, we were only able to obtain follow-up analysis on Case 2. Our reliability measurements are therefore limited, and longer follow-up with multiple highquality scans would be necessary to help predict changes in vessel density over the evolution of a paracentral acute middle maculopathy lesion from acute to old. Parafoveal vessel density in healthy eyes has previously been reported. Using an adaptive optics scanning laser ophthalmoscope, and outlining the centerlines of the ves-
sels to obtain their measurements, Tam and associates10 found an average vessel density of 31.6 mm1. Popovic and associates24 found a mean capillary density of 38.0 mm1 using a high-resolution, wide-field, dualconjugate adaptive optics instrument and skeletonizing the vessel signature to a 1-pixel-wide tracing. There are no reports of normal vessel density analysis using splitspectrum amplitude decorrelation angiography. Further work is needed to define normal vessel density of the superficial and deep retinal capillary plexus using this system to better reference our values to the normal population. Although we have provided broad clinical and anatomic evidence to strongly support deep (and intermediate) retinal capillary ischemia as the cause of paracentral acute middle maculopathy in this study and many prior studies,1,2,12–14,23 we cannot definitely exclude a primary neuropathic mechanism with secondary vascular loss. Showing causality would require animal models of retinal capillary occlusion and subsequent demonstration of the characteristic findings of paracentral acute middle maculopathy with multimodal imaging. Since focal lesions show relative preservation of perfusion, OCT angiography may not enhance the diagnosis of acute lesions, although longitudinal follow-up with quantitative OCT angiography may provide a valuable tool to assess long-term changes and the time course of capillary bed remodeling after vascular injury. Case 2 illustrates that deep capillary bed loss in the old phase may not be observable after only 1 month post acute injury. Our study included a relatively small number of patients with paracentral acute middle maculopathy with retrospective analysis. Larger prospective investigations may serve to more reliably confirm our conclusions. In summary, the superficial and deep retinal capillary networks can be visualized noninvasively and with high resolution using OCT angiography, in both normal eyes and eyes affected by paracentral acute middle maculopathy. An objective vessel density can be calculated to quantify the degree of ischemia. While focal acute paracentral acute middle maculopathy lesions may maintain perfusion, old paracentral acute middle maculopathy lesions lead to a profound loss of the deep (and intermediate) capillary plexus. Paracentral acute middle maculopathy lesions occurring in the setting of severe ischemia, as found in CRAO, lead to a profound loss of the deep capillary plexus in both acute and old cases. Ischemia-reperfusion injury may be responsible for the pathogenesis of some paracentral acute middle maculopathy lesions.
FUNDING/SUPPORT: NOVARTIS, MISSION VIEJO, CA (A.G.); ALLERGAN, IRVINE, CA (A.G.); ALCON, FORT WORTH, TX (AG); Harold and Pauline Price Foundation, Research to Prevent Blindness, New York, NY (M.G.B.); Optovue, Fremont, CA (D.S.), Genentech (D.S.), Regeneron (D.S.). Financial disclosures: Alain Gaudric: Allergan, Irvine, CA; Novartis, Mission Viejo, CA; Thrombogenics, Leuven, Belgium; Michael B. Gorin: Harold and Pauline Price Foundation, Research to Prevent Blindness, New York, NY; SriniVas Sadda: Optos, Dunfermline, United Kingdom; Carl Zeiss Meditec, Jena, Germany; David Sarraf: Optovue, Fremont, CA; Genentech, South San Francisco, CA; Regeneron, Tarrytown, NY research grants. The following authors have no financial disclosures: Julia Nemiroff, Laura Kuehlewein, Ehsan Rahimy, Irena Tsui, and Rishi Doshi. All authors attest that they meet the current ICMJE criteria for authorship. The authors acknowledge Fei Yu, Stein Eye Institute, University of California Los Angeles, Los Angeles, California, for statistics assistance.
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
11
REFERENCES 1. Sarraf D, Rahimy E, Fawzi AA, et al. Paracentral acute middle maculopathy: a new variant of acute macular neuroretinopathy associated with retinal capillary ischemia. JAMA Ophthalmol 2013;131(10):1275–1287. 2. Tsui I, Sarraf D. Paracentral acute middle maculopathy and acute macular neuroretinopathy. Ophthalmic Surg Lasers Imaging Retina 2013;44(6 Suppl):S33–S35. 3. Rahimy E, Sarraf D. Paracentral acute middle maculopathy spectral-domain optical coherence tomography feature of deep capillary ischemia. Curr Opin Ophthalmol 2014;25(3): 207–212. 4. Weinhaus RS, Burke JM, Delori FC, Snodderly DM. Comparison of fluorescein angiography with microvascular anatomy of macaque retinas. Exp Eye Res 1995;61(1):1–16. 5. de Carlo TE, Bonini Filho MA, Chin AT, et al. Spectraldomain optical coherence tomography angiography of choroidal neovascularization. Ophthalmology 2015;122(6): 1228–1238. 6. Kim DY, Finger J, Zawadzki RJ, et al. Optical imaging of the chorioretinal vasculature in the living human eye. Proc Natl Acad Sci U S A 2013;110(35):14354–14359. 7. Choi W, Mohler KJ, Potsaid B, et al. Choriocapillaris and choroidal microvasculature imaging with ultrahigh speed OCT angiography. PLoS One 2013;8(12):e81499. 8. Nagiel A, Sadda SR, Sarraf D. A promising future for optical coherence tomography angiography. JAMA Ophthalmol 2015; 133(6):629–630. 9. Schindelin J, Arganda-Carreras I, Frise E, et al. Fiji: an opensource platform for biological-image analysis. Nat Methods 2012;9(7):676–682. 10. Tam J, Martin JA, Roorda A. Noninvasive visualization and analysis of parafoveal capillaries in humans. Invest Ophthalmol Vis Sci 2010;51(3):1691–1698. 11. Zheng D, LaMantia A-S, Purves D. Specialized vascularization of the primate visual cortex. J Neurosci 1991;11(8): 2622–2629. 12. Christenbury JG, Klufas MA, Sauer TC, Sarraf D. OCT angiography of paracentral acute middle maculopathy associated with central retinal artery occlusion and deep capillary ischemia. Ophthalmic Surg Lasers Imaging Retina 2015;46(5): 579–581.
12
13. Philipakis E, Dupas B, Bonnin P, Hage R, Gaudric A, Tadayoni R. Optical coherence tomography angiography shows deep capillary plexus hypoperfusion in incomplete central retinal artery occlusion. Retin Cases Brief Rep 2015;9(4): 333–338. 14. Yu S, Pang CE, Gong Y, et al. The spectrum of superficial and deep capillary ischemia in retinal artery occlusion. Am J Ophthalmol 2015;159(1):53–63. 15. Chen X, Rahimy E, Sergott RC, et al. Spectrum of retinal vascular diseases associated with paracentral acute middle maculopathy. Am J Ophthalmol 2015;160(1):26–34. 16. Rahimy E, Sarraf D, Dollin ML, Pitcher JD, Ho AC. Paracentral acute middle maculopathy in nonischemic central retinal vein occlusion. Am J Ophthalmol 2014;158(2): 372–380. 17. Szabo ME, Droy-Lefaix MT, Doly M, Carre´ C, Braquet P. Ischemia and reperfusion-induced histologic changes in the rat retina. Demonstration of a free radical-mediated mechanism. Invest Ophthalmol Vis Sci 1991;32(5):1471–1478. 18. Hangai M, Yoshimura N, Hiroi K, Mandai M, Honda Y. Inducible nitric oxide synthase in retinal ischemiareperfusion injury. Exp Eye Res 1996;63(5):501–509. 19. Neufeld AH, Kawai S, Das S, et al. Loss of retinal ganglion cells following retinal ischemia: the role of inducible nitric oxide synthase. Exp Eye Res 2002;75(5):521–528. 20. Tsujikawa A, Ogura Y, Hiroshiba N, et al. Retinal ischemiareperfusion injury attenuated by blocking of adhesion molecules of vascular endothelium. Invest Ophthalmol Vis Sci 1999;40(6):1183–1190. 21. Zheng L, Gong B, Hatala DA, Kern TS. Retinal ischemia and reperfusion causes capillary degeneration: similarities to diabetes. Invest Ophthalmol Vis Sci 2007;48(1):361–367. 22. Nakahara T, Hoshino M, Hoshino S, Mori A, Sakamoto K, Ishii K. Structural and functional changes in retinal vasculature induced by retinal ischemia-reperfusion in rats. Exp Eye Res 2015;135:134–145. 23. Rahimy E, Kuehleweinm L, Sadda S, et al. Paracentral acute middle maculopathy: what we knew then and what we know now. Retina 2015;35(10):1921–1930. 24. Popovic Z, Knutsson P, Thaung J, Owner-Petersen M, Sjo¨strand J. Noninvasive imaging of human foveal capillary network using dual-conjugate adaptive optics. Invest Ophthalmol Vis Sci 2011;52(5):2649–2655.
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
Biosketch Julia Nemiroff is a third year ophthalmology resident at the Stein Eye Institute, University of California Los Angeles. She received her undergraduate degree in biology from New York Institute of Technology. She completed medical school at New York University School of Medicine, and internship at Winthrop University Hospital. She serves on the board of directors of Unite For Sight.
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
12.e1
To assess microvascular blood flow of the deep retinal capillary plexus in eyes with paracentral acute middle maculopathy using optical coherence tomography (OCT) angiography. DESIGN: Retrospective, multicenter observational case series. METHODS: Clinical and multimodal imaging findings from 8 patients with paracentral acute middle maculopathy were reviewed and analyzed. OCT angiography scans were analyzed and processed, and vessel density was calculated. RESULTS: Eight patients (7 male, 1 female, aged 9–82 years) were included. OCT angiography was obtained at either the acute (4 cases) or old stage (4 cases). Scans of the deep capillary plexus showed preservation of perfusion in acute lesions and capillary attenuation in old cases. Cases of central retinal artery occlusion showed marked loss of the deep capillary plexus. The mean vessel density of the superficial capillary plexus in normal fellow eyes was 12.8 ± 1.8 mmL1 vs 12.1 ± 1.9 mmL1 in eyes with paracentral acute middle maculopathy (reduction L6.0%, P [ .08). The mean vessel density of the deep capillary plexus in normal fellow eyes was 17.5 ± 1.4 mmL1 vs 14.7 ± 3.5 mmL1 in eyes with paracentral acute middle maculopathy (reduction L19.4%, P [ .04). This significant difference was representative of the eyes with old lesions. CONCLUSION: Paracentral acute middle maculopathy lesions correspond to preservation of perfusion in focal acute lesions and to pruning of the plexus in old cases. Cases of central retinal artery occlusion demonstrate marked hypoperfusion of the deep capillary plexus. Our
Accepted for publication Oct 29, 2015. From the Stein Eye Institute, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, California (J.N., I.T., M.B.G., D.S.); Institute for Ophthalmic Research, Center for Ophthalmology, Eberhard Karls University Tu¨bingen, Tu¨bingen, Germany (L.K., S.S.); Retina Service, Wills Eye Hospital, Philadelphia, Pennsylvania (E.R.); Kaiser Permanante, Tustin, California (R.D.); Ophtalmologie, Hopital Lariboisiere, AP-HP, Universite´ Paris 7 – Sorbonne Paris Cite´, Paris, France (A.G.); and Greater Los Angeles Veterans Affairs Healthcare Center, Los Angeles, California (D.S.). Inquiries to David Sarraf, Retinal Disorders and Ophthalmic Genetics Division, Stein Eye Institute, David Geffen School of Medicine at UCLA, 100 Stein Plaza, Los Angeles, California 90095; e-mail: dsarraf@ ucla.edu 0002-9394/$36.00 http://dx.doi.org/10.1016/j.ajo.2015.10.026
Ó
2015 BY
study further supports an ischemic pathogenesis of this retinal vasculopathy. (Am J Ophthalmol 2015;-: -–-. Ó 2015 by Elsevier Inc. All rights reserved.)
P
ARACENTRAL ACUTE MIDDLE MACULOPATHY IS A
recently described entity in patients presenting with an acute-onset paracentral scotoma. Spectraldomain optical coherence tomography (OCT) reveals hyperreflective band-like lesions at the level of the inner nuclear layer. Although these lesions resolve, patients are left with atrophy of the inner nuclear layer, resulting in a permanent paracentral visual field defect. Paracentral acute middle maculopathy can be idiopathic, or it can be secondary to local retinal vascular or systemic disease. Deep retinal capillary ischemia has been proposed as a causative factor in the development of these lesions, as the intermediate and deep retinal capillary plexuses flank the inner and outer boundaries of the inner nuclear layer, respectively.1–3 Fluorescein angiography, the reference standard for visualizing the retinal vasculature, has limited depth resolution.4 With OCT angiography, it is possible to obtain high-resolution, depth-resolved en face images of the retinal microvasculature by calculating motion contrast in OCT B-scans acquired repeatedly at the same location. OCT angiograms are co-registered with OCT B-scans from the same location, allowing simultaneous visualization of structure and blood flow.5–8 We used OCT angiography to study the retinal capillary microvasculature at different levels of the retina in 8 cases of paracentral acute middle maculopathy occurring in various clinical scenarios.
METHODS APPROVAL FOR THIS STUDY WAS OBTAINED FROM THE UNI-
versity of California Los Angeles (UCLA) Institutional Review Board committee. Research adhered to the tenets of the Declaration of Helsinki and was conducted in accord with regulations set forth by the Health Insurance Portability and Accountability Act. We retrospectively identified 8 patients from 3 tertiary referral centers. All were evaluated between August 2014
ELSEVIER INC. ALL
RIGHTS RESERVED.
1
TABLE 1. Clinical Characteristics and Associated Pathology in Patients Presenting With Paracentral Acute Middle Maculopathy Case
Sex
Age (y)
Affected Eye
1 2 3
Male Male Male
66 72 31
OD OD OS
4 5
Male Female
63 57
6 7 8
Male Male Male
82 16 9
BCVA
Ophthalmologic Diagnosis
Systemic Disease
20/50 20/20-1 20/20-1
CRAO BRAO BRAO
OS OS
20/30 20/300
CRAO CRAO
OS OU OD
CF at 4 ft 20/20 20/20-1
HTN; type II DM; hypothyroidism HTN; atrial fibrillation; mitral valve repair; pacemaker Transposition of the great vessels (status post repair); head trauma 2 days prior Left middle cerebral artery infarction; HTN Internal carotid artery dissection; cerebral vascular occlusion coinciding with ocular event; meningioma; hyperlipidemia HTN; type II DM Sickle cell disease (hemoglobin type SC) Alpha-thalassemia trait
CRAO, NPDR SCR Trauma
BCVA ¼ best-corrected visual acuity; BRAO ¼ branch retinal artery occlusion; CF ¼ count fingers; CRAO ¼ central retinal artery occlusion; DM ¼ diabetes mellitus; HTN ¼ hypertension; NPDR ¼ nonproliferative diabetic retinopathy; SCR ¼ sickle cell retinopathy.
and July 2015 and diagnosed with paracentral acute middle maculopathy. Written informed consent was obtained for each patient. We reviewed the clinical and multimodal imaging data for each patient and obtained OCT angiography analysis. The diagnostic criteria of paracentral acute middle maculopathy included a history of acute-onset paracentral scotoma with or without decline in visual acuity and a nonprogressive course. All patients demonstrated characteristic abnormalities with spectral-domain OCT, including either the acute finding of hyperreflective, plaque-like bands within the inner nuclear layer or old lesions demonstrating thinning of the inner nuclear layer. High-resolution digital color imaging, red-free photography, spectral-domain OCT, and OCT angiography were performed at the time of presentation for each patient. Fluorescein angiography was obtained in 7 of the 8 cases. The AngioVue OCT angiography system (Optovue, Inc, Fremont, California, USA) operates at 70, 000 A-scans per second to acquire OCT angiography volumes consisting of 304 3 304 A-scans in approximately 2.6 seconds. It uses a split-spectrum amplitude decorrelation angiography software algorithm and orthogonal registration and merging of 2 consecutive scan volumes to obtain 3 3 3 mm and 6 3 6 mm OCT angiography volumes of both eyes of each patient. OCT angiograms were co-registered with OCT B-scans, to allow visualization of both retinal vasculature and structure, and were performed in both the normal and affected eyes of each patient. The OCT angiography software was used to segment the superficial capillary plexus and deep capillary plexus in 3 3 3 mm scans. Consistently in each patient, the superficial capillary plexus slab was taken from the internal limiting membrane (offset 3 mm) to the inner plexiform layer (offset 15 mm). The deep capillary plexus slab was taken from the inner plexiform layer (offset 15 mm) to the outer plexiform layer (offset 70 mm). In cases where thinning of the inner nuclear layer caused failure of automatic segmentation, a 2
thinner 30 mm band was manually adjusted to include the deep capillary plexus. OCT angiography analysis of the superficial and deep capillary plexus was performed independently by a trained reader to assess for capillary attenuation and pruning involving either plexus. Quantitative analyses were performed using the publicly available software Fiji ImageJ 2.0.0-rc-29/1.49q (http://fiji. sc/Fiji)9 and GNU Image Manipulation Program GIMP 2.8.14 (http://gimp.org). Fiji was used to binarize and skeletonize the en face image of the superficial and deep retinal capillary plexus, showing the blood vessels as a 1-pixelwide line. GIMP was used to count the number of black pixels and total pixels. Vessel density was then calculated as [(pixels of vessels) (3/304)]/(area in mm2) in mm1.10,11 In cases of partial or complete projection artifact from the superficial capillary plexus, Fiji was used to binarize the superficial and deep capillary plexus images. GIMP was then used to subtract the black pixels of the superficial capillary plexus from the deep capillary plexus image. The resulting deep capillary plexus was skeletonized using Fiji software, and pixels were calculated using GIMP. Vessel density calculation was then repeated in the manner described above.
RESULTS NINE EYES WITH PARACENTRAL ACUTE MIDDLE MACULOP-
athy (8 patients: 7 male, 1 female) were identified and enrolled in this study. Clinical characteristics are summarized in Table 1. Patient ages ranged from 9 to 82 years (mean 49.5 years, median 60). Isolated band-like hyperreflective lesions in the middle retinal layers consistent with acute paracentral middle maculopathy lesions were observed with spectral-domain OCT imaging in 5 patients at baseline. Three patients presented with patchy
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
thinning of the inner nuclear layer on spectral-domain OCT, consistent with old paracentral acute middle maculopathy. Retinal vascular etiologies leading to paracentral acute middle maculopathy included central retinal artery occlusion (CRAO, n ¼ 4), branch retinal artery occlusion (BRAO, n ¼ 2), suction eye injury (n ¼ 1), and sickle cell retinopathy (n ¼ 1). OCT angiography was obtained at the acute stage of the lesion in 4 cases and at the old stage in 4 cases. The superficial capillary plexus demonstrated minimally attenuated perfusion in affected eyes. Focal acute cases (n ¼ 2, Cases 2 and 3) demonstrated patent perfusion of the deep capillary plexus within the lesions, whereas focal old cases (n ¼ 2, Cases 7 and 8) showed pruning of the deep capillary plexus within the lesions. One month follow-up analysis was obtained for Case 2 and demonstrated normal vascular density of the deep capillary plexus. Four cases of diffuse paracentral acute middle maculopathy associated with CRAO (Cases 1 and 4 with acute lesions, and Cases 5 and 6 with old lesions) demonstrated profound to complete absence of the deep capillary plexus with significant projection artifact from the superficial capillary plexus. Quantitative analysis was carried out in all cases. Because the patient in Case 7 was affected bilaterally, and thus had no control fellow eye, Case 7 was excluded from statistical analysis. Table 2 includes the vessel density measurements for each eye (n ¼ 9), and Table 3 shows a summary of the statistical analysis that was carried out comparing the patients with unilateral disease (n ¼ 7). The mean vessel density of the superficial capillary plexus in normal fellow eyes was 12.8 6 1.8 mm1 vs a mean of 12.1 6 1.9 mm1 in eyes with paracentral acute middle maculopathy (reduction 6.0%, P ¼ .33). The mean vessel density of the deep capillary plexus in normal fellow eyes was 17.5 6 1.4 mm1 vs a mean of 14.7 6 3.5 mm1 in eyes with paracentral acute middle maculopathy (reduction 19.4%, P ¼ .04). Cases 1 through 4 were scanned at the acute stage of the paracentral acute middle maculopathy lesion (n ¼ 4), and Cases 5 through 8 were scanned at the old stage of the lesion (n ¼ 4). Focal acute lesions (n ¼ 2, Cases 2 and 3) demonstrated preservation of perfusion of the deep capillary plexus (18.4 mm1) compared with the normal fellow eyes (18.2 mm1), whereas eyes with focal old lesions (n ¼ 1, Case 8) demonstrated attenuation of the deep capillary plexus (16.7 mm1) compared with the normal fellow eye (19.3 mm1) (Table 3). Localized vascular density analysis was also performed and compared between normal eyes and eyes with focal acute lesions. In Case 2, 33% of the 3 3 3 mm cube was involved, and the deep capillary density measured 6.3 mm1 in the affected eye vs 6.5 mm1 for the corresponding area in the unaffected eye. In Case 3, 25.6% of the 3 3 3 mm cube was involved, and capillary density in the affected area measured 4.8 mm1 vs 5.2 mm1 in the corresponding area of the unaffected, contralateral eye. This analysis confirmed the preservation of perfusion VOL. -, NO. -
TABLE 2. Retinal Capillary Density as Measured on Optical Coherence Tomography Angiography images in Paracentral Acute Middle Maculopathy Vessel Density (mm1) Superficial Capillary Plexus
Case
Deep Capillary Plexus
Affected Eye
Acute vs Chronic PAMM
Normal Eye
PAMM Eye
Normal Eye
PAMM Eye
OD OD OS OS OS OS OD OS OD
Acute Acute Acute Acute Chronic Chronic Chronic Chronic Chronic
12.2 12.9 14.9 9.7 13.1 12.2 N/A N/A 14.9
11.8 12.5 14.4 11.7 10.9 9.1 15.0 13.1 14.5
17.1 17.6 19.3 16.4 16.8 15.8 N/A N/A 19.3
15.7 17.8 18.7 13.9 11.0 9.2 21.3 19.4 16.7
1 2 3 4 5 6 7 8
N/A ¼ not applicable; PAMM ¼ paracentral acute middle maculopathy.
within the deep retinal capillary plexus for acute focal lesions. The 4 cases of CRAO with both acute (n ¼ 2) and old (n ¼ 2) paracentral acute middle maculopathy lesions showed partial or complete loss of the deep capillary plexus with overlying projection artifact from the superficial capillary plexus (Cases 1, 4, 5, and 6). The degree of projection artifact was obtained by recalculating the capillary density and subtracting the density of the superficial capillary plexus from that of the deep capillary plexus. A summary of calculated vessel density of the resulting deep capillary plexus is shown in Table 4. The mean vessel density of the deep plexus after subtraction in normal fellow eyes was 11.4 mm1 vs 7.5 mm1 in eyes with paracentral acute middle maculopathy and CRAO (reduction 41.2%) despite a relatively similar superficial vessel density in both groups. Representative cases (2, 3, 6, and 8) are described in detail below with corresponding figures. CASE 2:
A 72-year-old man (Figure 1) presented with a 3day history of seeing a ‘‘red line’’ across his right eye. Past medical history was notable for hypertension, atrial fibrillation, mitral valve repair, and pacemaker placement. Bestcorrected visual acuity (BCVA) was 20/20-1 in the right eye and 20/20-2 in the left eye. Fundus examination showed a patch of retinal whitening along the superior arcade and a normal retinal examination of the left eye. Spectral-domain OCT (Figure 1, Bottom row) showed numerous dense hyperreflective foci of the inner nuclear layer in band-like patches throughout the superior macula, consistent with paracentral acute middle maculopathy. An occult BRAO was suspected.
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
3
TABLE 3. Mean Parafoveal Vessel Density of the Superficial Capillary Plexus and Deep Capillary Plexus in Eyes Affected by Paracentral Acute Middle Maculopathy Compared With the Normal Fellow Eye and Separated Into Acute Versus Chronic Lesions
Variable 1
Superficial capillary plexus normal eye (mm ) Superficial capillary plexus PAMM eye (mm1) Superficial capillary plexus percent reduction (%) Deep capillary plexus normal eye (mm1) Deep capillary plexus PAMM eye (mm1) Deep capillary plexus percent reduction (%)
Mean 6 SD Vessel Density
Mean 6 SD Vessel Density in Acute Lesions
Mean 6 SD Vessel Density in Chronic Lesions
N¼7
N¼4
N¼3
12.8 6 1.8 12.1 6 1.9 6.0 6 14.9 (P ¼ .33a) 17.5 6 1.4 14.7 6 3.5 19.4 6 20.1 (P ¼ .04a)
12.4 6 2.1 12.6 6 1.2 þ2.1 6 10.9 17.6 6 1.2 16.5 6 2.1 6.8 6 7.4
13.4 6 1.4 11.5 6 2.7 16.9 6 13.4 17.3 6 1.8 12.3 6 3.9 36.3 6 19.7
PAMM ¼ paracentral acute middle maculopathy. P values from paired t test.
a
TABLE 4. Vascular Density of the Deep Capillary Plexus After Subtraction of Projection Artifact From the Superficial Capillary Plexus Measured on Optical Coherence Tomography Angiography in Eyes Affected by Paracentral Acute Middle Maculopathy and Central Retinal Artery Occlusion and in Normal Fellow Eyes Vessel Density of the Deep Capillary Plexus Normal Eye
PAMM Eye
Case
PAMM Eye
Acute vs Chronic PAMM
Original (mm-1)
After Subtractiona (mm-1)
PR (%)
Original (mm-1)
After Subtractiona (mm-1)
PR (%)
1 4 5 6
OD OS OS OS
Acute Acute Chronic Chronic
17.1 16.4 16.7 15.8
11.8 11.8 9.2 12.8
36.6 32.2 58.0 20.5
15.7 13.9 11.0 9.2
10.0 8.0 4.4 7.7
44.5 54.4 86.1 17.2
PAMM ¼ paracentral acute middle maculopathy; PR ¼ percent reduction (as calculated between original vessel density and vessel density after subtraction). a Subtraction of the superficial capillary plexus projection artifact from the deep capillary plexus.
On en face OCT angiography, the superficial capillary plexus of the affected right eye appeared to be within normal limits. The deep capillary plexus (Figure 1, First row, second from left) demonstrated patent structure with careful analysis of the en face images, and patent flow and perfusion with careful study of the OCT angiograms, corresponding to the hyperreflective paracentral acute middle maculopathy lesions with adjacent areas of hyporeflectivity, likely due to a contrast effect of the adjacent bright hyperreflective plaques. Vessel density was calculated for the superficial and deep retinal capillary plexus. There was minimal difference in vessel density between each eye in the superficial capillary plexus (12.9 mm1 in normal left eye vs 12.5 mm1 in affected right eye, reduction 3.1%) and the deep capillary plexus (17.6 mm1 in normal left eye vs 17.8 mm1 in affected right eye, difference þ0.8%). The patient returned after 1 month and noted that the paracentral scotomas were present but improved. The OCT showed persistent hyperreflective lesions in the inner nuclear layer with corresponding thinning. Repeat OCT 4
angiography did not show any reduction of the vascularity of the deep capillary plexus based on quantitative capillary density comparisons between the affected and unaffected areas. CASE 3:
A 31-year-old man (Figure 2) presented with acute loss of vision in the superior visual field of his left eye after left-sided head trauma while intoxicated with alcohol. Past medical history was notable for a transposition of the great vessels, surgically repaired when he was an infant. BCVA was 20/20 in the right eye and 20/201 in the left eye. Ophthalmoscopic examination of the right eye was unremarkable except for 1 dot-blot hemorrhage inferiorly Ophthalmoscopic examination of the left eye revealed attenuation of the inferotemporal retinal arteriole consistent with a BRAO. Corresponding fluorescein angiography (Figure 2) confirmed a peripheral filling defect of the inferotemporal retinal arteriole. Spectraldomain OCT (Figure 2) revealed a band of hyperreflectivity at the level of the inner nuclear layer consistent with paracentral acute middle maculopathy. OCT angiography
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
FIGURE 1. Case 2. En face optical coherence tomography (OCT) angiography and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the right eye of a 72-year-old man. (First row, first from left) A 3 3 3 mm macular cube OCT with en face image of the deep capillary plexus in the right eye shows prominent hyperreflective lesions corresponding to hyperreflective, plaque-like bands at the level of the inner nuclear layer on B-scan spectral-domain optical coherence tomography, consistent with paracentral acute middle maculopathy (Second row, second from left). (Second row, first from left) Corresponding infrared image is shown. White arrows demonstrate corresponding lesions. (First row, second and third from left) A 3 3 3 mm macular cube OCT angiography with en face projection capturing the deep capillary plexus of the right and left eye, respectively, demonstrates patent perfusion superiorly in the right eye. (Third row, first from left). A 6 3 6 mm macular cube OCT angiography with en face projection of the deep capillary plexus clearly demonstrates perfusion within the brightly colored paracentral acute middle maculopathy lesions. Adjacent areas are perfused and may appear dark because of a contrast effect. (Third row, second and third from left) OCT angiography scans of the deep capillary plexus have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1); on quantitative analysis, no significant difference was noted in the capillary density of either the superficial or deep plexus in the normal vs the affected eyes. (Bottom row, first from left) A 3 3 3 mm macular cube OCT angiography with en face projection capturing the deep capillary plexus of the right eye 1 month after initial presentation demonstrates normal perfusion. (Bottom row, second from left) OCT of the right eye at 1 month follow-up demonstrating persistent multifocal hyperreflectivity of the inner nuclear layer with attenuation.
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
5
FIGURE 2. Case 3. Fluorescein angiography, en face optical coherence tomography (OCT) angiography, and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the setting of branch retinal artery occlusion in the left eye of a 31-year-old man. (First row, first from left) Fluorescein angiography shows an infratemporal peripheral filling defect, as demonstrated by small white arrows. (First row, second from left) A 3 3 3 mm macular cube OCT with en face image of the deep capillary plexus in the affected left eye shows a prominent area of hyperreflectivity inferiorly corresponding to the paracentral acute middle maculopathy lesion. (Second row) Spectral-domain OCT (B-scan) of the inferior macula shows an acute focal hyperreflective, plaque-like band at the level of the inner nuclear layer consistent with paracentral acute middle maculopathy. (Bottom row, first and second from left) A 3 3 3 mm macular cube OCT angiography with en face projection capturing the deep capillary plexus of the right and left eye demonstrates patent perfusion of the deep capillary plexus in each eye. (Bottom row, third and fourth from left) For quantitative analysis, OCT angiography scans of the deep capillary plexus have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1). Qualitative and quantitative analysis was not significantly different.
(Third row, second from left) revealed patent flow and perfusion in the deep capillary plexus of the left eye, corresponding to the paracentral acute middle maculopathy lesion on spectral-domain OCT. Vessel density was calculated for the superficial capillary plexus and deep 6
capillary plexus and results for each eye were compared. There was minimal difference in the superficial capillary plexus (14.9 mm1 in the normal right eye vs 14.4 mm1 in the affected left eye, reduction 3.4%) and in the deep capillary plexus (19.3 mm1 in the
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
FIGURE 3. Case 6. En face optical coherence tomography (OCT) angiography and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the setting of central retinal artery occlusion in the left eye of a 82-year-old man. (First row, first from left) Spectral-domain OCT shows diffuse hyperreflectivity at the level of the inner nuclear level, consistent with paracentral acute middle maculopathy. (Second row, first from left) Follow-up OCT 3 months later shows diffuse thinning of the
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
7
normal right eye vs 18.7 mm1 in the affected left eye, reduction 3.1%). CASE 6:
An 82-year-old man (Figure 3) presented with sudden loss of vision in the left eye. Past medical history was notable for hypertension, hyperlipidemia, and type 2 diabetes. He reported a temporal headache on the left side but denied any other symptoms of giant cell arteritis. BCVA was 20/30 in the right eye and count fingers at 4 feet in the left eye. Ophthalmoscopic examination of the right eye was within normal limits except for a flat benign choroidal nevus. Ophthalmoscopic examination of the left eye revealed attenuated retinal vessels in the absence of embolic material, associated with cotton-wool spots and a cherry-red spot in the macula. Fluorescein angiography revealed delayed retinal vascular flow. Spectral-domain OCT (Figure 3, Top left) showed diffuse hyperreflective band-like lesions at the level of the inner nuclear layer in the left eye, consistent with an acute CRAO associated with paracentral acute middle maculopathy. Erythrocyte sedimentation rate was within normal limits (26 mm/hr), but C-reactive protein was elevated to 3.2 mg/L, raising suspicion for a diagnosis of giant cell arteritis. The patient was admitted, given a 5-day course of intravenous solumedrol, and discharged home on a prednisone taper. A temporal artery biopsy on the left side showed no evidence of giant cell arteritis on pathology review. Upon follow-up 2 months later, the patient’s visual acuity had improved to 20/30 in the left eye. Spectral-domain OCT (Figure 3, Second row, first from left) showed diffuse thinning of the inner nuclear layer, corresponding to the prior hyperreflective paracentral acute middle maculopathy lesions. OCT angiography (Figure 3, Fourth row, second from left) obtained at this time demonstrated projection artifact of the superficial capillary plexus and only modest perfusion detectible in the deep capillary plexus of the left eye. Vessel density was calculated for the superficial capillary plexus and deep capillary plexus of each eye. In the affected left eye, there was attenuation
of both the superficial capillary plexus (12.2 mm1 in the normal right eye vs 9.1 mm1 in the affected left eye, reduction 29.5%) and, more significantly, the deep capillary plexus (15.8 mm1 in the normal right eye vs 9.2 mm1 in the affected left eye, reduction 52.7%). The amount of projection artifact was quantified by recalculating the retinal capillary density and subtracting the density of the superficial plexus from the deep plexus using imaging software. The resulting deep capillary plexus was further reduced by 17.2%. CASE 8:
A 9-year-old boy (Figure 4) presented with a 6month history of a red spot in the central field of the right eye. By history the scotoma developed immediately after the right eye was traumatized by the suction vacuum drain of a swimming pool. Past medical history was notable for alpha-thalassemia trait. BCVA was 20/20-1 in the right eye and 20/20 in the left eye. Humphrey visual field 10-2 testing revealed a paracentral scotoma in the right eye. Ophthalmoscopic retinal examination of each eye was unremarkable. Spectral-domain OCT (Figure 4, Top right) and en face (Figure 4, Top left) analysis revealed severe thinning of the inner nuclear layer and outer plexiform layer in a patchy parafoveal distribution, consistent with old paracentral acute middle maculopathy lesions in the right eye. With en face OCT angiography analysis, the superficial capillary plexus (Figure 4, Second row, first from left) appeared to be within normal limits; but the deep capillary plexus (Figure 4, Third row, first from left) demonstrated patchy areas of perifoveal capillary attenuation and pruning, consistent with deep capillary ischemia. Vessel density was calculated for the superficial capillary plexus and deep capillary plexus of each eye. In the affected right eye, there was minimal reduction in the superficial capillary plexus (14.9 mm1 in the normal left eye vs 14.5 mm1 in the affected right eye, reduction 2.7%) and significant attenuation of the deep capillary plexus (19.3 mm1 in the normal left eye vs 16.7 mm1 in the affected right eye, reduction 14.4%).
inner nuclear layer. (First row, second from left and Second row, second from left) A 3 3 3 mm macular cube OCT with en face image of the deep capillary plexus in the right and left eye shows normal perfusion of the deep capillary plexus in the right eye and a profound loss of the deep capillary plexus in the left eye. A 3 3 3 mm macular cube OCT angiography with en face projection capturing the superficial capillary plexus (Third row, first and second from left) and deep capillary plexus (Fourth row, first and second from left) of the right and left eye, respectively, demonstrates projection artifact from the superficial capillary plexus in the left eye, as evidenced by large-caliber vessels not characteristic of the deep capillary plexus that correspond to identical vessels in the superficial capillary plexus, as well as marked reduction in perfusion of the deep capillary plexus (Fourth row, second from left). For quantitative analysis, OCT angiography scans of the superficial capillary plexus (Third row, third and fourth from left) and deep capillary plexus (Fourth row, third and fourth from left) have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1). Note significant projection artifact in the deep capillary plexus from the superficial capillary plexus in the affected left eye (Fourth row, fourth from left). (Bottom row, first from left and Bottom row, second from left) Projection artifact from the superficial capillary plexus was removed using Fiji and GIMP imaging software, and the resulting binarized and skeletonized images of the deep capillary plexus of the normal right eye and affected left eye are pictured. Note that the normal right eye is essentially unchanged after subtraction (Bottom left), while the left eye appears significantly more attenuated (Bottom right).
8
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
FIGURE 4. Case 8. En face optical coherence tomography (OCT) angiography and spectral-domain OCT and quantitated capillary density maps of paracentral acute middle maculopathy in the setting of a history of a suction injury to the right eye 6 months prior from the vacuum suction of a pool drain in a 9-year-old boy. (First row, first and second from left) A 3 3 3 mm macular cube OCT with en face imaging of the deep capillary plexus in the right and left eye, respectively, shows dark patchy areas of absent vascular markings in a perifoveal distribution in the right eye and a normal deep retinal plexus in the left eye. (First row, third from left) Spectral-domain OCT B-scan image of the right eye shows severe thinning of the inner nuclear layer and outer plexiform layer in a multifocal distribution, consistent with a history of paracentral acute middle maculopathy. A 3 3 3 mm macular cube OCT angiography with en face projection capturing the superficial capillary plexus (Second row, first and second from left) and deep capillary plexus (Third row, first and second from left) of the right and left eye, respectively, demonstrates patchy perifoveal loss of flow within the deep capillary plexus of the right eye (Third row, first from left). For quantitative analysis, OCT angiography scans of the superficial capillary plexus (Second row, third and fourth from left) and deep capillary plexus (Third row, third and fourth from left) have been transformed into binary, skeletonized vessel maps using Fiji software to assess the vessel density as total vessel length per area (mmL1) and shows significant attenuation of the capillary density of the deep plexus in the affected right eye (Third row, third from left).
DISCUSSION IN THIS STUDY, WE PERFORMED EN FACE OCT ANGIOGRAPHY
analysis of the superficial and deep retinal capillary plexus in normal eyes and in eyes with paracentral acute middle maculopathy. OCT angiography is an effective tool in the qualitative evaluation of paracentral acute middle maculopathy lesions.12,13 As described above, we demonstrate that focal acute paracentral acute middle maculopathy lesions show relative preservation of perfusion of the deep retinal capillary plexus, while old VOL. -, NO. -
lesions and eyes with diffuse paracentral acute middle maculopathy (acute and old) associated with CRAO exhibit marked reduction in perfusion. The association of paracentral acute middle maculopathy with deep capillary ischemia is not surprising, since the lesion always resides in the inner nuclear layer flanked by the intermediate and deep retinal capillary plexus. The subsequent development of permanent inner nuclear layer thinning is consistent with an ischemic infarct. Recently, there have been reports associating paracentral acute middle maculopathy with various retinal vascular diseases,
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
9
including diabetic retinopathy,14,15 retinal artery occlusion,14 central retinal vein occlusion,16 sickle cell retinopathy,15 and Purtscher retinopathy,14 further supporting an ischemic pathogenesis. Our report describes a cohort of patients with paracentral acute middle maculopathy similarly affected by retinal vascular disease in all but 1 case. We demonstrate 4 cases of paracentral acute middle maculopathy secondary to CRAO, 2 cases of paracentral acute middle maculopathy secondary to BRAO, 1 case of bilateral paracentral acute middle maculopathy in a patient with sickle cell retinopathy, and 1 case of isolated paracentral acute middle maculopathy occurring secondary to a suction injury. Patients in 4 cases (1–4) underwent OCT angiography in the acute stage; the remaining 4 patients (Cases 5–8) underwent OCT angiography at the old stage with atrophy of the inner nuclear layer. In our study, the average vessel density of the superficial capillary plexus in affected eyes was similar, although slightly attenuated, compared to that of the unaffected fellow eyes (reduction 6.0%, P ¼ .33), and the difference did not reach statistical significance. However, the deep capillary plexus was significantly attenuated compared with the normal fellow eye (19.4% reduction, P ¼ .04). Focal acute lesions demonstrated preservation of perfusion of the deep capillary plexus (18.4 mm1) as compared with the normal fellow eyes (18.2 mm1); the significant reduction in the density of the deep capillary retinal plexus is representative of old paracentral acute middle maculopathy lesions and lesions associated with CRAO. Followup analysis of Case 2 demonstrated preserved vascular density of the deep capillary plexus. OCT at this time showed some persistent hyperreflectivity of the inner nuclear layer with associated thinning. These findings suggest that the loss of capillary density seen in old focal cases may occur at some indeterminate point after injury. Cases of diffuse paracentral acute middle maculopathy associated with CRAO showed a marked loss of the deep capillary plexus in patients with both acute (Cases 1 and 4) and old (Cases 5 and 6) disease, with varying degrees of projection artifact from the superficial capillary plexus. We believe this may be secondary to profound capillary nonperfusion caused by large vessel occlusion. Our original calculation of the deep capillary plexus density in these cases likely highly overestimates the true degree of perfusion. To quantify the degree of projection artifact for these 4 cases, we used imaging software to subtract the vessel density of the superficial capillary plexus from the deep capillary plexus, both in eyes with paracentral acute middle maculopathy and in fellow control eyes. The percent reduction in vessel density of the deep capillary plexus was greater in the affected eyes vs normal eyes in Cases 1, 4, and 5, suggesting greater projection artifact from the superficial plexus in the eyes with paracentral acute middle maculopathy vs the normal fellow eyes. The patient in Case 6 had a markedly attenuated superficial capillary plexus; thus, for this case the percent reduction was not greater in the affected eye than in the control eye after 10
correction of the projection artifact. This methodology for correcting projection artifact has not previously been published, and is therefore not yet validated. Because this method is still preliminary, we report both sets of vessel density calculations. Of note, the BCVA in Case 1 was remarkably good, at 20/50. This is likely the case secondary to immediate reperfusion of the central retinal artery. It is also possible, although there are no data available to validate this claim, that CRAO occurring in conjunction with paracentral acute middle maculopathy may have a better prognosis, as the superficial capillary plexus is spared, with selective infarction of the deep capillary plexus. The preservation of perfusion in the acute stages of ischemic injury, and the subsequent damage caused by reperfusion, have been well described. Ischemiareperfusion injury is a condition in which blood flow returns to a tissue after a period of ischemia. Reperfusion limits immediate retinal damage but is accompanied by mechanisms such as generation of excessive reactive oxygen species,17 nitric oxide neurotoxicity,18,19 and inflammation.20 A recent study demonstrated that ischemia-reperfusion in rat retinas caused capillary dropout after 8–14 days of reperfusion and concluded these vascular changes occurred after neuronal loss.21 Another recent study22 demonstrated that the number of functional vessels decreased after ischemia and reperfusion, and further showed that loss of perfusion occurred predominantly in the deep capillary plexus rather than in capillaries located at the retinal surface. We surmise that the pathogenesis of paracentral acute middle maculopathy may be related to ischemia and reperfusion injury, accounting for the immediate restoration in microcirculation seen in some of the acute and focal paracentral acute middle maculopathy lesions and the subsequent legacy of atrophy of the inner nuclear layer associated with loss of the deep capillary plexus. In the case of diffuse acute paracentral acute middle maculopathy occurring in the setting of CRAO (ie, large vessel obstruction), no capillary reperfusion may occur, causing severe acute impairment of flow within the deep retinal capillary plexus that was observed in this study. In all of our cases, ischemia of the intermediate capillary plexus was likely a significant contributing factor in the development of paracentral acute middle maculopathy23; but it is impossible to separate the intermediate from the deep capillary plexus at this time using presently available OCT angiography technology. As this imaging field advances, with improved segmentation technology, we may be able to determine whether the intermediate or the deep retinal capillary plexus is more important in the etiology of paracentral acute middle maculopathy. Interestingly, mild attenuation of the superficial capillary plexus was identified in certain old cases of paracentral acute middle maculopathy in this study (Cases 5, 6, and 7), indicating that while ischemia is predominantly present at the level of the intermediate and deep capillary plexus,
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
the trilaminar arrangement of the retinal capillary plexuses is not completely independent. The advantages of our method included rapid and noninvasive acquisition of OCT angiography images, depth-resolved microvascular detection, and semiautomated analysis with quantitated calculation technique. In particular, our method of skeletonization using Fiji, with subsequent calculation of vessel density using GIMP, takes only a few minutes and yields identical results each time it is repeated for a specific image, and is thus clinically applicable. But our study also had several limitations. Technical limitations included segmentation errors in scans and low quality, caused by movement artifact or retinal pathology, requiring manual segmentation. Three out of 8 cases required manual segmentation. Case 1 was resegmented owing to errors in automatic segmentation, possibly due to motion artifact during initial acquisition. Motion artifact can be addressed by including faster eye tracking to improve the image quality, especially in subjects with poor fixation and reduced visual acuity. Cases 5 and 6 were resegmented owing to severe thinning of the inner nuclear layer, leading to inadvertent incorporation of the superficial vasculature. As mentioned in the cases with profound attenuation of the deep capillary plexus, projection artifact from the superficial capillary plexus confounded imaging and quantitated analysis. Our method of subtraction of the projected superficial capillary plexus from the deep capillary plexus likely removed a proportion of the true deep plexus as well, possibly leading to an underestimation of the true plexus. Further, OCT angiography could only be used to map vessels with flow; other vessels may have been present in areas of capillary attenuation but their flow was too slow to be visualized. We did not correct for ocular magnification in the calculation of vessel density because axial length measurements were not obtained for all patients. Our method of calculating vessels by automated binarization may cause vessels to be incorrectly removed or added based on image quality and artifact. Finally, we were only able to obtain follow-up analysis on Case 2. Our reliability measurements are therefore limited, and longer follow-up with multiple highquality scans would be necessary to help predict changes in vessel density over the evolution of a paracentral acute middle maculopathy lesion from acute to old. Parafoveal vessel density in healthy eyes has previously been reported. Using an adaptive optics scanning laser ophthalmoscope, and outlining the centerlines of the ves-
sels to obtain their measurements, Tam and associates10 found an average vessel density of 31.6 mm1. Popovic and associates24 found a mean capillary density of 38.0 mm1 using a high-resolution, wide-field, dualconjugate adaptive optics instrument and skeletonizing the vessel signature to a 1-pixel-wide tracing. There are no reports of normal vessel density analysis using splitspectrum amplitude decorrelation angiography. Further work is needed to define normal vessel density of the superficial and deep retinal capillary plexus using this system to better reference our values to the normal population. Although we have provided broad clinical and anatomic evidence to strongly support deep (and intermediate) retinal capillary ischemia as the cause of paracentral acute middle maculopathy in this study and many prior studies,1,2,12–14,23 we cannot definitely exclude a primary neuropathic mechanism with secondary vascular loss. Showing causality would require animal models of retinal capillary occlusion and subsequent demonstration of the characteristic findings of paracentral acute middle maculopathy with multimodal imaging. Since focal lesions show relative preservation of perfusion, OCT angiography may not enhance the diagnosis of acute lesions, although longitudinal follow-up with quantitative OCT angiography may provide a valuable tool to assess long-term changes and the time course of capillary bed remodeling after vascular injury. Case 2 illustrates that deep capillary bed loss in the old phase may not be observable after only 1 month post acute injury. Our study included a relatively small number of patients with paracentral acute middle maculopathy with retrospective analysis. Larger prospective investigations may serve to more reliably confirm our conclusions. In summary, the superficial and deep retinal capillary networks can be visualized noninvasively and with high resolution using OCT angiography, in both normal eyes and eyes affected by paracentral acute middle maculopathy. An objective vessel density can be calculated to quantify the degree of ischemia. While focal acute paracentral acute middle maculopathy lesions may maintain perfusion, old paracentral acute middle maculopathy lesions lead to a profound loss of the deep (and intermediate) capillary plexus. Paracentral acute middle maculopathy lesions occurring in the setting of severe ischemia, as found in CRAO, lead to a profound loss of the deep capillary plexus in both acute and old cases. Ischemia-reperfusion injury may be responsible for the pathogenesis of some paracentral acute middle maculopathy lesions.
FUNDING/SUPPORT: NOVARTIS, MISSION VIEJO, CA (A.G.); ALLERGAN, IRVINE, CA (A.G.); ALCON, FORT WORTH, TX (AG); Harold and Pauline Price Foundation, Research to Prevent Blindness, New York, NY (M.G.B.); Optovue, Fremont, CA (D.S.), Genentech (D.S.), Regeneron (D.S.). Financial disclosures: Alain Gaudric: Allergan, Irvine, CA; Novartis, Mission Viejo, CA; Thrombogenics, Leuven, Belgium; Michael B. Gorin: Harold and Pauline Price Foundation, Research to Prevent Blindness, New York, NY; SriniVas Sadda: Optos, Dunfermline, United Kingdom; Carl Zeiss Meditec, Jena, Germany; David Sarraf: Optovue, Fremont, CA; Genentech, South San Francisco, CA; Regeneron, Tarrytown, NY research grants. The following authors have no financial disclosures: Julia Nemiroff, Laura Kuehlewein, Ehsan Rahimy, Irena Tsui, and Rishi Doshi. All authors attest that they meet the current ICMJE criteria for authorship. The authors acknowledge Fei Yu, Stein Eye Institute, University of California Los Angeles, Los Angeles, California, for statistics assistance.
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
11
REFERENCES 1. Sarraf D, Rahimy E, Fawzi AA, et al. Paracentral acute middle maculopathy: a new variant of acute macular neuroretinopathy associated with retinal capillary ischemia. JAMA Ophthalmol 2013;131(10):1275–1287. 2. Tsui I, Sarraf D. Paracentral acute middle maculopathy and acute macular neuroretinopathy. Ophthalmic Surg Lasers Imaging Retina 2013;44(6 Suppl):S33–S35. 3. Rahimy E, Sarraf D. Paracentral acute middle maculopathy spectral-domain optical coherence tomography feature of deep capillary ischemia. Curr Opin Ophthalmol 2014;25(3): 207–212. 4. Weinhaus RS, Burke JM, Delori FC, Snodderly DM. Comparison of fluorescein angiography with microvascular anatomy of macaque retinas. Exp Eye Res 1995;61(1):1–16. 5. de Carlo TE, Bonini Filho MA, Chin AT, et al. Spectraldomain optical coherence tomography angiography of choroidal neovascularization. Ophthalmology 2015;122(6): 1228–1238. 6. Kim DY, Finger J, Zawadzki RJ, et al. Optical imaging of the chorioretinal vasculature in the living human eye. Proc Natl Acad Sci U S A 2013;110(35):14354–14359. 7. Choi W, Mohler KJ, Potsaid B, et al. Choriocapillaris and choroidal microvasculature imaging with ultrahigh speed OCT angiography. PLoS One 2013;8(12):e81499. 8. Nagiel A, Sadda SR, Sarraf D. A promising future for optical coherence tomography angiography. JAMA Ophthalmol 2015; 133(6):629–630. 9. Schindelin J, Arganda-Carreras I, Frise E, et al. Fiji: an opensource platform for biological-image analysis. Nat Methods 2012;9(7):676–682. 10. Tam J, Martin JA, Roorda A. Noninvasive visualization and analysis of parafoveal capillaries in humans. Invest Ophthalmol Vis Sci 2010;51(3):1691–1698. 11. Zheng D, LaMantia A-S, Purves D. Specialized vascularization of the primate visual cortex. J Neurosci 1991;11(8): 2622–2629. 12. Christenbury JG, Klufas MA, Sauer TC, Sarraf D. OCT angiography of paracentral acute middle maculopathy associated with central retinal artery occlusion and deep capillary ischemia. Ophthalmic Surg Lasers Imaging Retina 2015;46(5): 579–581.
12
13. Philipakis E, Dupas B, Bonnin P, Hage R, Gaudric A, Tadayoni R. Optical coherence tomography angiography shows deep capillary plexus hypoperfusion in incomplete central retinal artery occlusion. Retin Cases Brief Rep 2015;9(4): 333–338. 14. Yu S, Pang CE, Gong Y, et al. The spectrum of superficial and deep capillary ischemia in retinal artery occlusion. Am J Ophthalmol 2015;159(1):53–63. 15. Chen X, Rahimy E, Sergott RC, et al. Spectrum of retinal vascular diseases associated with paracentral acute middle maculopathy. Am J Ophthalmol 2015;160(1):26–34. 16. Rahimy E, Sarraf D, Dollin ML, Pitcher JD, Ho AC. Paracentral acute middle maculopathy in nonischemic central retinal vein occlusion. Am J Ophthalmol 2014;158(2): 372–380. 17. Szabo ME, Droy-Lefaix MT, Doly M, Carre´ C, Braquet P. Ischemia and reperfusion-induced histologic changes in the rat retina. Demonstration of a free radical-mediated mechanism. Invest Ophthalmol Vis Sci 1991;32(5):1471–1478. 18. Hangai M, Yoshimura N, Hiroi K, Mandai M, Honda Y. Inducible nitric oxide synthase in retinal ischemiareperfusion injury. Exp Eye Res 1996;63(5):501–509. 19. Neufeld AH, Kawai S, Das S, et al. Loss of retinal ganglion cells following retinal ischemia: the role of inducible nitric oxide synthase. Exp Eye Res 2002;75(5):521–528. 20. Tsujikawa A, Ogura Y, Hiroshiba N, et al. Retinal ischemiareperfusion injury attenuated by blocking of adhesion molecules of vascular endothelium. Invest Ophthalmol Vis Sci 1999;40(6):1183–1190. 21. Zheng L, Gong B, Hatala DA, Kern TS. Retinal ischemia and reperfusion causes capillary degeneration: similarities to diabetes. Invest Ophthalmol Vis Sci 2007;48(1):361–367. 22. Nakahara T, Hoshino M, Hoshino S, Mori A, Sakamoto K, Ishii K. Structural and functional changes in retinal vasculature induced by retinal ischemia-reperfusion in rats. Exp Eye Res 2015;135:134–145. 23. Rahimy E, Kuehleweinm L, Sadda S, et al. Paracentral acute middle maculopathy: what we knew then and what we know now. Retina 2015;35(10):1921–1930. 24. Popovic Z, Knutsson P, Thaung J, Owner-Petersen M, Sjo¨strand J. Noninvasive imaging of human foveal capillary network using dual-conjugate adaptive optics. Invest Ophthalmol Vis Sci 2011;52(5):2649–2655.
AMERICAN JOURNAL OF OPHTHALMOLOGY
--- 2015
Biosketch Julia Nemiroff is a third year ophthalmology resident at the Stein Eye Institute, University of California Los Angeles. She received her undergraduate degree in biology from New York Institute of Technology. She completed medical school at New York University School of Medicine, and internship at Winthrop University Hospital. She serves on the board of directors of Unite For Sight.
VOL. -, NO. -
OCT ANGIOGRAPHY OF DEEP RETINAL CAPILLARY ISCHEMIA
12.e1
