Recent Advances in Clinical Applications of Imaging in Retinal Diseases

Abstract

Many diseases that cause visual impairment, as well as systemic conditions, manifest in the posterior segment of the eye. With the advent of high-speed, high-resolution, reliable, and noninvasive imaging techniques, ophthalmologists are becoming more dependent on ocular imaging for disease diagnosis, classification, and management in clinical practice. There are rapid advances on the indications of multimodal retinal imaging techniques, including the application of ultra-widefield fundus angiography, fundus autofluorescence, optical coherence tomography, as well as optical coherence tomography angiography. This review summarizes and highlights the clinical applications, latest indications, and interpretations of multimodal imaging in age-related macular degeneration, polypoidal choroidal vasculopathy, diabetic macular edema, central serous chorioretinopathy, diabetic retinopathy, retinal vein occlusion, and uveitis.

INTRODUCTION

Ophthalmic imaging is becoming essential in guiding the diagnosis, classification, and management of a vast array of retinal diseases. Recently, there are several important updates on the clinical applications and interpretations of retinal imaging in age-related macular degeneration (AMD), polypoidal choroidal vasculopathy (PCV), diabetic macular edema (DME), central serous chorioretinopathy (CSC), diabetic retinopathy (DR), retinal vein occlusion (RVO), and uveitis. Advancements in imaging technology enable high-speed and high-resolution image acquisition by reduction of signal noise as well as motion artifacts. The enhanced quality of noninvasive optical coherence tomography (OCT) and OCT angiography (OCTA) images leads to improved reliability and standards for diagnosing and classifying chorioretinal diseases, which gradually reduces the dependency for invasive techniques. Nevertheless, dye-based angiography still plays a crucial role in examining the retinal and choroidal circulation, and fundus autofluorescence (FAF) allows the functional evaluation of retinal pigment epithelium (RPE). Ultra-widefield imaging technology provides the opportunity for documenting the full extent of tissue damages caused by the underlying disease, which could be important for treatment planning and response monitoring. This article firstly introduces several advances in retinal imaging techniques that are rapidly gaining popularity among ophthalmologists to utilize in the clinic, including ultra-widefield dye-based angiography, FAF, swept source OCT and OCTA. This is followed by elaboration of the up-to-date indications of retinal imaging that led to recent paradigm changes in diagnosis, classification, and management of common and important conditions including PCV, DR, DME, RVO, uveitis, and CSC. The objective of this review article is to appeal to ophthalmologists on the novel interpretations of retinal imaging in these conditions, which often requires simultaneous scrutinization of multimodal imaging findings.

INTRODUCTION TO RECENT ADVANCES IN VARIOUS IMAGING MODALITIES

Ultra-widefield Imaging

Ultra-widefield imaging makes use of the scanning laser ophthalmoscopy technology that captures around 200 degrees of the retina with or without mydriasis. Image distortion may occur in up to 5% to 10% of images because of 2-dimensional representation of a 3-dimensional surface, which may need to be rectified when documenting the measurement of lesion size and extent, although this has been addressed in some devices.^1,2 Eyelash artifact may also appear during image acquisition. Nevertheless, ultra-widefield retinal imaging technology provides unprecedented opportunity to visualize the peripheral retina. There are several examples in clinical practice: ultra-widefield fundus angiography (FA) had been used to assess peripheral retinal vasculature such as nonperfusion areas and retinal neovascularization in DR, RVO, and retinal vasculitis. Ultra-widefield imaging has been helpful in pediatric patients who may be uncooperative during imaging acquisition, for conditions such as retinopathy of prematurity and Coats disease.³ Ultra-widefield indocyanine green angiography (ICGA) has improved the understanding of peripheral exudative hemorrhagic chorioretinopathy, a disease predominantly found anterior to the globe equator, as study has found that polypoidal lesions were found in up to 70% of these cases, making peripheral exudative hemorrhagic chorioretinopathy being considered a variant of PCV.⁴

Fundus Autofluorescence

In AMD, FAF can identify areas of RPE atrophy and RPE rip. It demonstrates subretinal drusenoid deposits, which have high association with late AMD and retinal angiomatous proliferation. FAF also provides prognostic value, as banded and diffuse patterns of geographic atrophy seen on FAF in AMD are associated with higher rates of disease progression. It is an important, noninvasive ancillary imaging modality that helps confirm the diagnosis of CSC and its subtypes by revealing the pattern of hyper/hypoautofluorescence such as gravitational RPE defects. Furthermore, it is useful for observing disease progression, such as distinguishing active versus inactive uveal effusion syndrome and to monitor response to treatment, for example, changes in autofluorescence lesions after chemotherapy in primary vitreoretinal lymphoma.^5,6 FAF is able to assess for macular displacement following retinal detachment (RD) repair surgery. The presence of hyperautofluorescent lines running parallel to retinal blood vessels, referred to as RPE ghost vessels or retinal vessel printings, are of functional significance due to their association with postoperative symptoms of distortion.⁷ One limitation of FAF is that quantitative comparison of autofluorescence between different raw FAF images may have low reliability due to the high noise-to-signal ratio. Furthermore, capturing FAF images prior to performing FA is preferred to avoid relatively increased autofluorescence secondary to excitation of the photoreceptors.⁸

In combination with ultra-widefield technology, ultra-widefield FAF can be used to evaluate AMD, inherited retinal dystrophy, and RD. In AMD, peripheral FAF abnormalities can be classified into granular, mottled, or nummular, which correspond to peripheral drusen, cobblestone degeneration, and RPE depigmentation on fundus photo, respectively. A prospective study has shown that peripheral FAF abnormalities are common in older patients and neovascular AMD compared to dry AMD, this finding may help refine the phenotyping of AMD in the future.⁹ In inherited retinal dystrophy, ultra-widefield FAF may predict visual function as peripheral FAF abnormalities correlate with the size of visual field defect in perimetry and amplitude of rod and cone functions in full-field electroretinography results.¹⁰ After RD repair surgery, FAF can evaluate early postoperative RPE function over previous detached retina.^11,12 Postoperatively, hypoautofluorescence was found over area of detached retina and hyperfluorescent lines may correlate with demarcation lines or residual subretinal fluid. Moreover, persistent postoperative granular autofluorescence in the macula-off RD was associated with worse vision.¹¹

Swept Source Optical Coherence Tomography

Swept source OCT (SSOCT) is a recent advancement in OCT technology, it uses a tunable laser for frequency-swept light source and photodetector as opposed to broadband light source and spectrometer camera used in spectral domain OCT (SDOCT). SSOCT enables faster image acquisition and more resistance to motion artifact due to higher scan speed. In addition, the longer wavelength (1040 nm) used in SSOCT, compared to SDOCT (840 nm), means that deeper penetration and high depth of resolution can be achieved. As a result, SSOCT can provide high-resolution images of the vitreous, retina, and choroid simultaneously without sensitivity roll-off.¹³ One limitation of SSOCT, however, is a slight decrease in lateral resolution compared with SDOCT due to the longer wavelength used.¹³ The widespread use of SSOCT is currently limited by its higher cost compared to SDOCT. SSOCTA may visualize the choroidal vasculature and type 1 choroidal neovascularization (CNV) better than SDOCTA.¹⁴ In addition, en face SSOCT images enable better assessment of the changes in choroidal vasculature, providing new insights for pachychoroid spectrum diseases such as CSC¹⁵ and PCV.¹⁶

En face Optical Coherence Tomography

En face OCT produces layer by layer coronal images of the retina as opposed to conventional cross-sectional OCT images. En face OCT images are digitally reconstructed using high-density volumetric OCT scanning and can be used to assess a variety of retinal diseases affecting both inner and outer retinal layers.^17–21 The clinical application of en face OCT includes the classification of lamellar hole and macular pseudohole,¹⁸ the measurement of geographic atrophy extent in AMD,²¹ assessment of perivenular ischemia in RVO,¹⁹ and diagnosis of PCV.²² Furthermore, en face OCT has improved our understanding of the pathogenesis of vitreomacular diseases and the effect of internal limiting membrane peeling.^17,23 In a retrospective study, the adhesive zone (A-zone) on en face OCT was found to be associated with metamorphopsia score in eyes with epiretinal membrane and this parameter may be used to predict functional outcome in epiretinal membrane.²⁴

Optical Coherence Tomography Angiography

Optical coherence tomography angiography is a noninvasive imaging tool that detect flow signals to provide images of different retinal and choroidal vascular layers simultaneously. Without the injection of dye, OCTA utilizes intrinsic motion contrast to map vascular plexuses. In addition, OCTA software allows automatic quantitative measurement of vascular density at different retinal levels, including the superficial capillary plexus, intermediate capillary plexus, and deep capillary plexus with high repeatability and reproducibility.²⁵ Recently, the use of widefield SSOCTA technology has enabled the acquisition of wider OCTA images up to 12×12 mm (as compared to 6×6 mm in conventional OCTA). The use of widefield OCTA (which consists of five 12×12 mm images centered on macula, superonasal, inferonasal, superotemporal, and inferotemporal) can be used to detect nonperfusion area and neovascularization in DR patients noninvasively.²⁶

CLINICAL APPLICATIONS OF IMAGING IN RETINAL DISEASES

Non-ICGA Diagnosis of PCV

PCV is a subtype of neovascular AMD (nAMD), characterized by type 1 neovascularization associated with aneurysmal dilation (polypoidal lesions) and branching vascular network.²⁷ Conventionally, ICGA is the gold standard to diagnose PCV and guide photodynamic therapy (PDT). Based on patients undergoing ICGA, the prevalence of PCV among nAMD is up to 50%.^28,29 However, ICGA is an invasive investigation that requires intravenous dye injection. Recently, there is a shift to using SDOCT, a quick and noninvasive investigation, in the diagnosis and monitoring of nAMD and PCV. Several studies have shown that non–ICGA-based features, such as fundus photo appearance and OCT features, have a high sensitivity and specificity to differentiate PCV from nAMD.^30–32 In 2021, the Asia-Pacific Ocular Imaging Society (APOIS) proposed a non–ICGA-based diagnostic criteria for PCV.²² The new diagnostic criteria consist of 9 features, with 6 based on cross-sectional OCT, 2 on fundus photography, and 1 on en face OCT (Table 1). There are 3 major criteria [sub-RPE ring lesion, sharp-peaked pigment epithelial detachment (PED), and complex RPE elevation on en face OCT] and 4 minor criteria (orange nodule, thick choroid with dilated Haller layer, complex PED, and double layer sign). These new diagnostic criteria achieved high sensitivity, specificity, and positive predictive value when compared to ICGA. Specifically, the presence of all 3 major criteria achieved an area under receiver operating characteristic curve of 0.90.³²

TABLE 1 - Diagnostic Features of PCV on Various Non-ICGA Imaging Modalities

SDOCT	Sharply peaked PED
	PED notch
	Sub-RPE ring-like lesion
	Double layer sign
	Fluid compartment
En Face OCT	Pachyvessels
	RPE elevations
Fundus photography	Subretinal orange nodules
	Extensive subretinal hemorrhage ≥4DD

DD indicates disc diameter; ICGA, indocyanine green angiography; PCV, polypoidal choroidal vasculopathy; PED, pigment epithelial detachment; RPE, retinal pigment epithelium; SDOCT, spectral domain optical coherence tomography.

As anti–vascular endothelial growth factor (anti-VEGF) monotherapy can be the first-line treatment for both nAMD and PCV, the differentiation in treatment-naive eyes may not be crucial in the clinical setting.³³ However, in eyes with poor response to anti-VEGF monotherapy, rescue PDT may be required in PCV cases.^34–36 Hence, the APOIS workgroup followed up with a set of non-ICGA treatment criteria to diagnose PCV in eyes with suboptimal response to anti-VEGF monotherapy and the use of OCT alone to guide PDT treatment. The combination of sharp-peaked PED and sub-RPE ring like lesion on OCT and orange nodule on fundus photo was found to have an area under receiver operating characteristic curve of 0.85 and specific (0.85) in identifying PCV.³² In addition, OCT-guided PDT planning, which was determined by the extent of PED, double layer sign, and sharp-peaked PED, was found to cover 100% of polypoidal lesions and 90% of branching neovascular network consistently.³⁷

Optical coherence tomography angiography can be used in conjunction with OCT to improve the diagnostic accuracy of PCV. On OCTA, polypoidal lesions may appear as localized sub-RPE hyperflow signal on cross-sectional images and a focal hyperflow sign in the outer retina slab on en face images.³⁸ SSOCTA may have advantages over SDOCTA as it has deeper penetrance and produce images with higher resolution.³⁹ On SSOCTA, polypoidal lesions appear as glomeruli-like tangled vascular structure next to branching neovascular network. This finding suggests that polypoidal lesions are neovascularization rather than an aneurysmal structure.⁴⁰ In fact, the APOIS workgroup recommended the term polypoidal lesions to describe these lesions based on OCT and histological analysis.²²Figure 1 demonstrates a case of PCV with features on OCT and OCTA images, in correlation with ICGA findings.

FIGURE 1: A 65-year-old woman diagnosed with polypoidal choroidal vasculopathy in her left eye. Fundus photo (A) showed macular hemorrhage with an orange nodule. Spectral domain optical coherence tomography (C) found a sharp pigment epithelial detachment with adjacent double layer sign. Optical coherence tomography angiography (D) showed compact, tangled vessels (circled in red) that corresponded to the polypoidal lesions on indocyanine green angiography (B). The branching neovascular network was also revealed on optical coherence tomography angiography.

It is important to note that the set of non-ICGA diagnostic criteria is limited by its low negative predictive value of 0.68²² and the set of non-ICGA treatment criteria is limited by its low sensitivity of 0.65,³⁷ meaning that these criteria cannot be used to exclude PCV in treatment-naive eyes and may miss some PCV in anti-VEGF nonresponders. In addition, combined SDOCT and OCTA may further improve the diagnostic accuracy,³⁸ but OCTA is not widely available, and accurate segmentation of the lesion may be challenging.³⁹ In the future, ICGA may remain as the most validated gold standard in the diagnosis of PCV, but non-ICGA diagnostic features may offer a quick and noninvasive way to distinguish PCV from nAMD and guide PDT treatment in real-life clinical practice.^22,37

Hotspot in ICGA

ICGA is the gold standard diagnostic tool for PCV. However, hotspots on ICGA can still represent lesions other than PCV (Table 2).⁴¹

TABLE 2 - Differential Diagnoses of ICGA Hotspots

PCV

Retinal angiomatous proliferation

Choroidal neovascularization

Central serous chorioretinopathy

Retinal microaneurysms or macroaneurysms

Punched out RPE defect/choroidal vascular knuckle

Late RPE staining with ICG

ICGA indicates indocyanine green angiography; PCV, polypoidal choroidal vasculopathy; RPE, retinal pigment epithelium.

Figure 2 shows a case of hotspots on ICGA due to microaneurysms rather than polypoidal lesions.

FIGURE 2: Left-eye fundus photo (A) of a 72-year-old woman showed exudative maculopathy. Spectral domain optical coherence tomography (B) found intraretinal fluid, irregular retinal pigment epithelial layer, and hyperreflective dots. Indocyanine green angiography (D) found multiple, tiny hotspots (red arrows) that were nonpulsatile and not surrounded by dark halo. Optical coherence tomography angiography (C) did not show any signal that denoted vascular network in the outer retina. Fluorescein angiography (E–F) showed mild late leakage and multiple staining which were also present in the periphery (red arrows on bottom right). With multimodal imaging, the diagnosis of diabetic macular edema and drusenoid pigment epithelium due to concurrent dry age-related macular degeneration was established.

These conditions can be mistaken for PCV and thus the EVEREST criteria are one of the widely-adopted diagnostic criteria in clinical practice.⁴¹ Characteristics, such as presence of hypofluorescent halo, branching vascular network, pulsatile polyp, and fundus photography showing orange subretinal nodule and associated submacular hemorrhage, help distinguish PCV from other differential diagnoses on ICGA.

Central Serous Chorioretinopathy

CSC is characterized by serous RD caused by leakage of fluid through the RPE. The term “central” refers to the form of the disease causing visual symptoms due to the presence of serous detachments in the macula.⁴² CSC usually affects working-age patients with typical onset between 39 and 51 years.⁴³ While acute onset CSC may spontaneously resolve with gradual visual improvement, nevertheless, chronic CSC can be a sight-threatening disease leading to legal blindness. In a study that retrospectively observed CSC patients for a 10-year period, Mrejen et al⁴⁴ reported 79.7% of patients maintained driving-standard vision with best-corrected visual acuity (VA) of 20/40 or better in at least 1 eye, and 12.8% were legally blind with best-corrected VA of 20/200 or worse in both eyes. VA at the 10-year follow-up visit was significantly worse than the first visit, and poorer vision was associated with the presence of cystoid macular degeneration, CNV, outer retinal disruption on OCT, and FAF changes.⁴⁴ Therefore, several variants of CSC manifestations had poor prognosis. Nonetheless, there is high discordance even among experienced retina specialists in the classification of CSC. Singh et al⁴⁵ conducted a survey and found 36 different terms to classify CSC with very low agreement among the graders. Hence, retinal imaging has an important role in providing a more objective and standardized classification for CSC that will ultimately lead to precise prognostication and refined selection criteria for clinical trials in the future.

In the absence of a gold standard diagnostic and classification method for CSC, a standardized classification system was recently proposed by the consensus of experts from the Central Serous Chorioretinopathy International Group consisting of several prespecified criteria according to the findings of serous RD and RPE alterations detected through multimodal imaging.⁴⁶

Multimodal imaging includes FA, ICGA, FAF, structural OCT, and OCTA. FA identifies focal leaks through RPE defects, serous PED, and other RPE alterations. ICGA reveals delays in choroidal filling, dilated choroidal vessels, and areas of mid-phase choroidal hyperfluorescence.⁴⁷ These choroidal abnormalities may precede serous RD and often are present in fellow eyes unaffected by CSC. FAF helps identify RPE alteration and extent of the disease.⁴² OCT identifies serous RD, subretinal fibrin, outer retinal abnormalities, serous RPE detachments, increased choroidal thickness, and dilated choroidal vessels.⁴⁷ OCTA is a novel tool that enhanced the sensitivity and specificity in detecting CNV when combined with OCT. CNV had been considered an uncommon occurrence in CSC, with an estimated incidence of 2% to 9% of CSC eyes based on FA investigation.^48,49 Conventionally, detection of CNV by FA is difficult because the typical appearance of stippled, granular hyperfluorescence and occult leakage of type 1 CNV in FA can be masqueraded by multifocal leakages and stains from RPE and window defects.⁵⁰ In a study of 305 CSC eyes, Ng et al⁵¹ found 20.7% had CNV, detected with multimodal imaging including OCTA. Other smaller studies using OCTA detected 20% to 58% CNV among CSC patients.^52–56 These results suggest that OCTA can be a more effective tool than dye-based angiography for detecting CNV associated with CSC (Fig. 3). Table 3 outlines the major and minor diagnostic criteria, and Table 4 lists the classification algorithm. The broad range of CSC findings identifiable with novel multimodal imaging tools refined CSC classification with respect to disease severity and duration resulting from enhanced interobserver agreement and validity in several observational studies.^57–59

FIGURE 3: A 55-year-old man with complex central serous chorioretinopathy based on diffuse retinal pigment epithelium changes on ultra-widefield fundus photo (A–B) and gravitational hyperautofluorescence tract. Multifocal leakages were found on fluorescein angiography and indocyanine green angiography (C–D). Optical coherence tomography (E–F) revealed serous retinal detachment, and intraretinal cyst overlying a shallow irregular pigment epithelial detachment was correlated with optical coherence tomography angiography (G) signal of a secondary choroidal neovascularization.

TABLE 3 - Proposed Major and Minor Diagnostic Criteria by Central Serous Chorioretinopathy International Group⁴⁵

Major Criteria (must have BOTH)	Minor Criteria (must have at least 1)
(1) Presence or evidence of prior serous retinal detachment documented on OCT involving the posterior pole unrelated to another disease process	(1) SFCT of 400 μm or more
(2) At least 1 area of RPE alteration on FAF, SD OCT, or infrared imaging	(2) Mid-phase hyperfluorescent placoid areas on ICGA
	(3) One or more focal leaks on FA

FA indicates fluorescein angiography; FAF, fundus autofluorescence; ICGA, indocyanine green angiography; OCT, optical coherence tomography; RPE, retinal pigment epithelium; SDOCT, spectral domain OCT; SFCT, subfoveal choroidal thickness.

TABLE 4 - Multimodal Imaging Classification by Central Serous Chorioretinopathy International Group⁴⁵

Simple (total area of RPE alteration ≤2DA)	Complex (total area of RPE alteration >2 DA or multifocal)	Atypical
Primary (first-known episode of SRF) or Recurrent [presence of SRF with history or signs of resolved episode(s)] or Resolved (absence of SRF)		Bullous variant or RPE tear or Association with other retinal diseases
±Persistent (SRF >6 mo)
±Outer retinal atrophy (ONL thinning and/or ELM disruption and/or EZ attenuation
	±intraretinal fluid
±CNV

+ if fovea is involved (STF, out retinal atrophy, PED).

CNV indicates choroidal neovascularization; DA, disc area; ELM, external limiting membrane; EZ, ellipsoid zone; ONL, outer nuclear layer; PED, pigment epithelial detachment; RPE, retinal pigment epithelium; SRF, subretinal fluid.

DR and DME

The screening of DR in the primary care setting is conventionally done by mydriatic standard 7-field Early Treatment Diabetic Retinopathy Study (ETDRS) photographs. But nonmydriatic ultra-widefield imaging may be an alternative DR screening tool in the future as it has been shown to be comparable to dilated fundal examination in the grading of DR.⁶⁰ In addition, ultra-widefield image and FA can detect peripheral predominant lesions in DR, which is associated with greater risk of DR progression.^61,62

DME is the most common sight-threatening complication of DR, and ~50% of DME eyes showed suboptimal visual gain despite anti-VEGF injection.⁶³ Retinal imaging can be used to determine the extent and prognosis of DME and help retinal physicians in patients counseling and making individualized treatment decision. OCT provides quantitative and qualitative biomarkers that are associated with visual and anatomical outcome of DME.⁶⁴ Quantitative parameters include the central subfield thickness (CST), subfoveal choroidal thickness (SFCT), and choroidal vascularity index (CVI). Qualitative biomarkers include morphology of DME, disorganization of inner retinal layers (DRIL), hyperreflective retinal foci (HRF), and disruption of the ellipsoid zone (EZ) and external limiting membrane (ELM).

The CST has been used to quantify the degree of DME and monitor treatment response in clinical trials.^63,65,66 However, there are conflicting reports on whether baseline CST is predictive of visual prognosis in DME. In addition, change in CST has been shown to be only weakly associated with change in VA in DME eyes treated with anti-VEGF injections.^64,67 In summary, CST should not substitute VA as the outcome measurement in DME.

Based on OCT, the morphology of DME can be classified into diffuse retinal thickening, cystoid macular edema (ME), and serous RD. There are conflicting reports regarding the prognostic significance of morphology of DME.^68–70 On the other hand, in a recent retrospective study, Szeto et al⁶⁴ reported the presence of cystoid ME and serous RD was associated with greater reduction in CST after anti-VEGF treatment but found no association with posttreatment vision. Future prospective studies are required to address the significance of DME morphology in the management of DME.

Enhanced-depth imaging OCT has enabled the visualization of the choroidal vessels. As the choroidal vasculature delivers oxygen and nutrients to the outer retinal layers, it was hypothesized that a decrease in choroidal thickness and vascularity may be associated with worse outcome DME. The SFCT is a surrogate marker of choroidal vascularity, and CVI is a quantitative parameter to assess the choroidal vascularity. It has been shown that higher CVI is associated with thicker SFCT and is a more robust marker of choroidal vascularity.⁷¹ Currently there are conflicting reports on the SFCT value in DME. SFCT was reported to be reduced significantly in DME in some studies^72–74 but such observation was not reproduced in other studies.^74,75 On the other hand, it was shown that CVI was reduced in the presence of DR regardless of DME status.⁷⁶ The significance of SFCT and CVI as biomarkers for DME remains controversial and requires further studies to clarify.

Qualitative OCT biomarkers have been shown to be associated with prognosis of DME. DRIL, defined as indistinguishable boundaries between ganglion cell—inner plexiform layer complex and the inner nuclear—outer plexiform complex, has been shown to be a robust biomarker for poorer vision and treatment responsiveness to anti-VEGF therapy.^64,77,78 The visibility of the ELM and EZ on OCT may represent the integrity of the Muller cells and photoreceptors, respectively. It is, therefore, not surprising that disruption to ELM and EZ was shown to be associated with worse baseline and posttreatment vision in DME.^64,79–81 HRF, which may represent extravasated lipoproteins or activated microglial cells, through compromised ELM and EZ, may migrate into the choroid as hyperreflective choroidal foci.⁸² The clinical significance of HRF and hyperreflective choroidal foci is still under debate, as some studies showed that they were associated with worse vision and vision gain after treatment,^64,82–85 while some studies showed that eyes with higher number of baseline HRF were associated with greater reduction in CST after treatment.^86,87 The extent and severity of these biomarkers can be reduced with anti-VEGF or intravitreal steroid treatment, and reversal of these biomarkers are associated with greater improvement in vision.^64,79,88,89

Diabetic macular ischemia (DMI), traditionally assessed by FA, can be defined as enlargement of the foveal avascular zone (FAZ).⁹⁰ DMI may be observed in up to 20% of eyes without clinical signs of DR, making it a potential biomarker for early DR.^91,92 OCTA can assess the FAZ noninvasively and allow concurrent assessment of DMI in different retinal layers, including the superficial capillary plexus, intermediate capillary plexus, and deep capillary plexus, separately.^93,94 In addition, quantitative metrics can be derived from OCTA, such as FAZ area and circularity, vessel density, and vessel tortuosity.⁹⁴ Quantitative OCTA metrics can predict DR progression and DME development, in addition to established risk factors.^95,96 DMI at the level of deep capillary plexus has been reported to be independently associated with worse vision^92,97,98 and poorer response to anti-VEGF treatment in DME eyes.⁹⁹

Retinal Vein Occlusion

RVO, including both branch retinal vein occlusion and central retinal vein occlusion, is the second commonest retinal vascular disorder. ME and neovascularization are potential vision-threatening complications of RVO.¹⁰⁰ The use of ultra-widefield dye-based angiography has enabled detection of untreated nonperfusion area in peripheral retina in RVO eyes with prior laser photocoagulation. This is of clinical importance as untreated peripheral nonperfusion may drive neovascularization and ME, and prompt treatment with laser photocoagulation may stabilize vision.¹⁰¹

Although multiple studies have validated the use of OCT biomarkers as predictors of DME treatment outcome, there is currently conflicting evidence on whether OCT biomarkers can be used as prognostic factors of RVO-associated ME.^102,103 In a retrospective study, Babiuch and colleagues reported that the presence of baseline DRIL was associated with less vision gain after anti-VEGF therapy. On the other hand, in a post hoc analysis of a prospective randomized controlled trial involving 202 eyes treated with intravitreal ranibizumab, disruption of ELM was found to be independently associated with baseline vision only but not with subsequent vision.¹⁰² The use of OCT biomarkers in the prediction of treatment response in RVO-associated ME hence warrants further investigation.

Automated OCTA metrics, including global vascular density and foveal avascular area, were correlated with peripheral retinal nonperfusion detected on FA.¹⁰⁴ Hence, OCTA may help identify high-risk RVO patients to undergo FA, as untreated peripheral nonperfusion in RVO may lead to recurrent ME and retinal neovascularization.^101,104 In addition, FAZ area and vascular densities in superficial capillary plexus and deep capillary plexus are surrogate markers of macular ischemia and have been reported to be negatively correlated with VA in RVO eyes.^104–106

Uveitis

Ultra-widefield FA enables the assessment of the peripheral extent of retinal vascular changes caused by uveitis that is not revealed by conventional FA. Leakages along the peripheral retinal vessels and staining of perivascular sheath are associated with retinal vasculitis.¹⁰⁷ Sometimes, peripheral retinal arteriolar or vein occlusions and retinal neovascularization can occur.¹⁰⁷ Evaluation of the full extent of capillary nonperfusion or ischemia is important when planning to perform sectoral panretinal photocoagulation, with the aim to reduce the risk of vitreous hemorrhage leading to vision loss in patients with retinal vasculitis.¹⁰⁸ Ultra-widefield FA is also indicated for thorough investigation of multifocal choroiditis and chorioretinal lesions that may arise from infectious etiology, including tuberculosis, syphilis, and toxoplasmosis.¹⁰⁷ Chorioretinal lesions due to lymphoma may mimic uveitis as masquerade syndrome. In addition, FA can be used to assess optic disc edema, ME, as well as ischemia.

Vasculitis of large choroidal vessels underlies Vogt-Koyanagi-Harada disease (VKH), Behcet disease, birdshot chorioretinopathy, toxoplasma chorioretinitis, sarcoidosis-related uveitis, multifocal choroiditis, placoid chorioretinitis, and serpiginous chorioretinopathy.¹⁰⁷ In these conditions, ICGA is able to reveal hypofluorescent spots due to blockage from choroidal infiltrates or flow voids from inflammation of the choroidal vasculature.¹⁰⁸ Impaired perfusion of the choriocapillaris can also be seen by OCTA as flow deficit. Choriocapillaris flow deficit can be quantified as flow deficit density (the percentage of image area of flow void to the total image area), total flow deficit area, mean flow deficit size, and the total number of flow deficits (above a predefined physiological flow void area).¹⁰⁹ The evaluation of choriocapillaris flow impairment is a potentially useful objective marker for noninvasive and close monitoring of treatment response and drug dosage titration (Fig. 4). However, quantitative analysis of the choriocapillaris using the current imaging technologies is challenging, mainly due to the limited resolution and reliability of OCTA scans in visualizing the extremely compact microvasculature within an ultrathin layer that anteriorly lies very close to the light scattering RPE/Bruch membrane complex and posteriorly without distinct boundary. OCTA detects blood flow as to generate a signal, thus, the dark region represents areas with blood flow below the OCTA device’s decorrelation threshold, which can be caused by true loss of vessels, reduced blood flow or artifact.²⁴ It is important to be aware of the caveats in the interpretation of images, such as signal attenuation by speckle noise, pathologic lesions, eye blinking, undulations of the RPE causing segmentation error, projection artifact, and unreliable thresholding strategy used in flow deficit binarization.⁵⁸ There is no standard algorithm for OCTA imaging technique for studying the choriocapillaris. Therefore, assessment of interobserver and intraobserver repeatability, comparison between different scan areas, with different flow segmentation strategies and with control subjects, are important for the validation of quantitative metrics of choriocapillaris flow deficit. Furthermore, OCTA is able to detect secondary CNV associated with uveitis that is often difficult to distinguish from exudations due to inflammatory changes on fundoscopic exam and FA.

FIGURE 4: Left eye of a 26-year-old man diagnosed with acute posterior multifocal placoid pigment epitheliopathy (APMPPE) before corticosteroid (week 0) with weekly monitoring for treatment response and corticosteroid dosage adjustment using ultra-widefield fundus photo (top row) and optical coherence tomography angiography images of the choriocapillaris flow deficit (dark areas). The gradual resolution of placoid lesions on the fundus corresponded with reduction of flow deficit areas of the choriocapillaris layer.

Summary

Technological advancements allow high-resolution, depth-resolved, and noninvasive OCT images of chorioretinal tissue as well as microangiography. Furthermore, ultra-widefield dye-based angiography and FAF enable unprecedented panoramic assessment of chorioretinal lesions, vascular changes, ischemia, and RPE defect. The application of advanced retinal imaging techniques has had a major impact on clinical management with paradigm changes in the diagnosis, subtypes classification, and prognostication of retinal diseases. In some clinical settings with limited accessibility to ICGA, the use of SDOCT is an important asset in diagnosing PCV. When IGCA is available, there are some caveats when interpreting the presence of hot spots due to a number of differential diagnoses. In CSC, there is increasing consensus on standardized diagnostic and classification criteria by simultaneous interpretation of multimodal imaging. Structural classification of DME subtypes on OCT has implications on anti-VEGF treatment responses and prognosis, and OCTA metrics are potentially useful biomarkers to predict the onset and progression of DME and DR with improving validity from clinical studies. The use of ultra-widefield FA enables detection of untreated nonperfused areas in peripheral retina in eyes affected by RVO and retinal vasculitis, to allowed targeted treatment with laser photocoagulation. Ultra-widefield dye-based angiography assessment of peripheral chorioretinal lesions aids the diagnosis of infectious, inflammatory, and neoplastic etiologies. In addition to retinal imaging, mapping and spatial characterization of visual function by microperimetry also provides insightful information in retinal diseases.¹¹⁰ Microperimetry is useful for monitoring the progression of early/intermediate AMD and to identify prognostic factors. It is especially important for monitoring geographic atrophy and DME that spares the central retina in the initial stages. Fundus tracking and precise point-wise follow-up examinations enhance the retest reliability of patients with unstable fixation when performing microperimetry.¹¹⁰ Future development in artificial intelligence and consensus meetings will further enhance the clinical application of these imaging modalities for precise diagnosis and personalized management of retinal diseases.