7. Adaptive Optics Retinal Imaging of Retinal Diseases

In addition to the above basic applications, ophthalmic AO systems are translating into clinical applications that are rapidly expanding. It is expected that AO technology will soon translate into clinical applications. The most promising application of AO retinal imaging is the detection of early signs of retinal diseases. AO retinal imaging indeed provided new information about the pathological changes of the retinal microstructure in several diseases, including inherited or acquired neuro-retinal degenerations [110–112] and vaso-occlusive diseases [113, 114]. Because the rate of disease progression is typically slow for most retinal degenerations, it is estimated that patients show a clinical loss of visual function years after the onset of the disease. If AO imaging can provide high-resolution measurements of the cone photoreceptor structure, micro-vasculature and nerve fiber bundles in patients, then it may open a new frontier for the early diagnosis of retinal diseases and for monitoring the efficacy of therapies at a cellular level. Researchers [7, 85, 110] have demonstrated that AO devices allow imaging of exactly the same retinal area over days or months in order to follow disease progression with microscopic accuracy.

Across the developed world, the major causes of vision loss are attributed to diabetic retinopathy (DR), age-related macular degeneration (AMD) and glaucoma [5]. Diagnosis usually occurs once the damage has already happened to some extent, because of the relatively low resolution of current retinal imaging techniques when used to detect abnormalities of the retinal microstructures and the relatively poor sensitivity of functional testing. Considering that structural damage of these microstructures precedes their functional impairment, the detection of pathological variations of photoreceptors, capillaries and nerve fiber bundles at the pre-clinical stage of the disease can be beneficial for early treatment and avoidance of serious visual loss. Furthermore, a better understanding of the microscopic alterations in retinal tissue may provide further insight into the mechanisms of disease progression and be helpful to identify new approaches for therapeutic intervention.

Diabetic retinopathy (DR) is a frequently occurring complication of diabetes mellitus, that is a metabolic disease in which the patient has a high serum glucose level. According to the World Health Organization, diabetes mellitus is responsible for about 12% of new cases of blindness between the ages of 45 and 74 years in the developed world. With the incidence of diabetes throughout the world projected to rise from 150 million to approximately 300 million by the year 2025, DR represents a major threat to the global population and will likely present ever-increasing burdens on the health care delivery system [115, 116]. DR can be classified as non proliferative (NPDR) or proliferative (PDR). NPDR is further graded as mild, moderate and severe according to the Early Treatment Diabetic Retinopathy Study (ETDRS) severity scale [117].

The earliest clinical pathological changes of DR occur in the microvascular structures [118, 119]. According to the most commonly accepted patho-physiological model (i.e., the microvascular theory) [120], DR consists of a microangiopathy that induces pathological changes of the vascular structures and the blood rheological properties as a consequence of chronic hyperglycaemia. It has been also postulated that DR is a multifactorial disease involving the retinal neuronal cells (neurodegenerative theory) [121–124]. The neurodegenerative changes are apoptosis of several populations of retinal cells, including photoreceptors, bipolar and ganglion cells. The functional and structural impairment of neural tissue has been theorized to participate in the generation of the earliest morphological alterations of the vascular structures [121–125]. The question of whether the photoreceptor loss could be initially determined by a neuronal or vascular breakdown during diabetes mellitus remains unsolved. Work is needed to understand whether the neurodegenerative changes of retinal cells may precede the vascular damages or the two processes are simultaneous. The exact nature of their interdependence is complex and still not known.

Noninvasive detection of the pathological signs of DR is usually performed by dilated fundus examination, colour fundus photography and more sophisticated retinal imaging techniques, such as Spectral Domain Optical Coherence Tomography (SD-OCT). Fluorescein angiography (FA) can be useful to assess the integrity of the blood retinal barrier as the amount of fluorescein leakage is related to the dysfunction of the retinal vascular endothelium. Even if FA represents an important diagnostic tool and improves the accuracy of laser treatment of DR complications (e.g., diabetic macular oedema and new vitreo-retinal vessels) [126, 127], on the other hand, it requires injection of a fluorescent dye agent that can lead to unintended systemic complications. AO retinal imaging could be useful to detect pathological changes early in the course of DR, such as microaneurysms, micro-haemorrhages and loss of photoreceptors [113, 114, 128–130] ( Figure 5). The detection of the early retinal changes related to DR might be important in the management of the patient requiring a better glyco-metabolic control. Authors [128, 129] described image processing and analysis algorithms to extract the capillary vessel information and provided excellent visualization of the parafoveal capillary network in healthy eyes using AOSLO devices. In a recent study [114], the capillary network was evaluated in patients with type 2 diabetes. Researchers found a higher capillary dropout and a higher tortuosity of the arterovenous channels in patients with diabetes and no diabetic retinopathy than in healthy controls. Tam et al.[113] analyzed the retinal microvasculature in a patient with type 1 diabetes and severe NPDR over a 16 months follow-up period. Longitudinal assessment of the capillaries showed microaneurysm formation and disappearance as well as the formation of tiny capillary bends similar in appearance to intraretinal microvascular abnormalities. In vivo imaging of the capillary network has been also shown using an AO-SD-OCT [131–133]. The 3D information provided by OCT represents a major advantage compared to en face imaging techniques. The AO-SD-OCT can provide a theoretical spot volume of 3 μm³, capable of reconstructing the entire retinal capillary network of the inner retina. On the other hand, OCT cannot distinguish the lumen from the vascular walls of larger vessels [134]. In a preliminary investigation on large retinal arterioles in patients with type 1 diabetes and NPDR, we acquired images of the vessel walls using an AO flood-illumination retinal camera ( Figure 6 and Movie 1 (movie showing a retinal artery, as seen in Figure 6). The blood vessel shows periodic twitching, probably corresponding to the cardiac cycle of the patient, as previously shown by Zhong et al.[135], showing the capability of AO ophthalmoscopy to evaluate them for monitoring both arterial lumen and walls in vaso-occlusive retinal diseases. Both axial and en face AO imaging techniques could therefore be complementary to noninvasively analyze the retinal vessels at high resolution. The detection of pre-clinical abnormalities of retinal microcirculation in patients with diabetes could represent a valuable advantage of AO retinal imaging in comparison with current noninvasive imaging modalities.

Figure 5

Figure 6

Non-necrotic photoreceptor loss, in addition to microangiopathy, has been considered to be responsible for the vision loss associated to DR [121–124]; on the other hand, a loss of cone photoreceptors has never been clinically demonstrated. An objective of our current work is to evaluate the photoreceptor mosaic geometry and reflectance: the procedure has the potential to provide additional and valuable information about the cellular changes of retinal pathologies. In a recent study [136] we found that, in a population of eleven patients with type 1 diabetes, the parafoveal cone photoreceptors showed a higher variation in intensity than in healthy controls at the same retinal eccentricity ( Figure 7). This phenomenon was particularly evident near the areas of intraretinal focal oedema ( Figure 8). The regional differences in image intensity with areas of cones brighter than adjacent areas were also previously seen in other retinal diseases, including cone-rod dystrophy [110]. The significance of this phenomenon is not yet clear. In general, there are various hypotheses on the intensity variation of cones in AO imaging, as discussed in the previous section [97–104]. Fluctuations in reflectivity may be caused by molecular changes within the cones that are due to phototransduction or to changes in outer segment length, related to disk shedding. It has been also theorized that areas of cones darker than adjacent cones may reflect a disruption of the waveguide properties of the cones themselves [105, 109] or to light interactions between the end of the OS tip and the pigments in macrophages or retinal pigment epithelial (RPE) cells [102, 103]. Technical factors, however, should be considered to contribute to the spatial variation of reflectance between adjacent areas of cones, as previously discussed [100, 102, 104].

Figure 7

Figure 8

Age Related Macular Degeneration (AMD) is the leading cause of blindness in the elderly across the developed world. It represents a deterioration of the macular area and usually affects older adults. A recent study found that the late stage AMD (the most disabling form of the pathology) is present in approximately 5% of the over 65 's and 12% of the over 80 's [137]. AMD is a multifactorial disease, involving ocular, systemic and genetic risk factors. The ocular risk factors include darker iris pigmentation and hyperopic refraction, while systemic risk factors include cigarette smoking, obesity, sunlight exposure and cardiovascular diseases [138, 139]. Genes influence several biological pathways related to AMD, including the immune processes, mechanisms involving collagen and glycosaminoglycans synthesis and angiogenesis. All these factors have been associated with the onset, progression and bilateral involvement of early, intermediate, and advanced states of AMD [140–146]. Genetic susceptibility can be influenced by the environmental factors: taken together, both factors are highly predictive of the onset, progression and response to treatments [147]. Several patho-biological pathways have been implicated in the pathogenesis of AMD: these include senescence, shown by a lipofuscin accumulation in the RPE cells, choroidal ischemia and oxidative damage [148, 149].

There are two clinical types of AMD, the “dry” and “wet” form. In the early stages of AMD, which is asymptomatic, insoluble extracellular aggregates called drusen accumulate in the retina. The late stage of dry AMD, which is also known as geographic atrophy (GA), is characterized by scattered or confluent areas of degeneration of RPE cells and the overlying retinal photoreceptors, which rely on the RPE for trophic support. The other late stage form of AMD, the wet form (10–15%), is typified by choroidal neo-vascularization (CNV), where newly immature blood vessels grow toward the outer retina from the underlying choroid leaking fluid below or within the retina [150, 151].

Retinal imaging for the management of AMD includes fluorescein angiography. In the wet AMD form, leakage of dye (hyperfluorescence) is noted and classified by location (subfoveal, juxtafoveal, or extrafoveal) and by type (classic, occult, or mixed). Indocyanine green angiography (ICG) uses an intravenous dye with different characteristics from fluorescein (e.g., less melanin absorbance). It improves identification and characterisation of neovascular variants of AMD (e.g., the polypoidal choroidal vasculopathy). SD-OCT enables high-resolution in vivo cross-sectional or volumetric tomographic visualisation of the retinal micro-architecture. It allows visualisation of the cross-sectional outline of the neovascular choroidal complex, but its internal structure cannot be well resolved and the neovascular components cannot be distinguished from the fibrous components, haemorrhages or dense exudates within the lesion. With the advent of the anti-VEGF therapy, SD-OCT imaging is widely used for the early diagnosis of CNV and for the treatment and re-treatment management.

In the early, asymptomatic, AMD stages (i.e., presence of drusen), the ability to predict the rate of progression is currently limited. By monitoring drusen over time, en face AO imaging ( Figure 9), also combined with AO-SD-OCT imaging, can theoretically detect both their progression, in terms of size, and analyze their direct effect on the overlying photoreceptor mosaic [152–154]. Godara et al.[152] showed, for a 45-year old female, that both the cone density and cone arrangement of the mosaic overlying the drusen were within normal limits. AO imaging revealed a regular photoreceptor mosaic with areas of hyper-reflectivity coinciding with the location of the drusen. The increased reflectivity associated with the drusen could be attributed to increased scatter from the RPE (due to decreased melanin or accumulation of some other waste material), to loss of outer segment pigment or loss of the photoreceptor outer segment. Boretsky et al.[154] identified several small drusen deposits that were not observed with standard wide field imaging techniques in early AMD. They also investigated large coalescent drusen and areas of geographic atrophy in advanced stage dry AMD, showing significant decrease in visible photoreceptor density. A sensitive, non-invasive, imaging tool could help to better recognize the earliest retinal changes and to identify patients who could progress rapidly and may benefit from a more intensive observation and management. Furthermore in the near future, an early diagnosis of the macular disease could have an important role in the evaluation of the effectiveness of new prevention strategies.

Figure 9

Glaucoma is the leading cause of irreversible, preventable blindness worldwide [155]. It has been estimated that over 11 million glaucoma sufferers worldwide are bilaterally blind from the disease [156]. Primary open angle glaucoma (POAG) is a chronic disease characterized by progressive loss of retinal ganglion cells, usually associated with ocular hypertension, that leads to structural damage of the inner retinal layers, as shown by progressive regional or diffuse thinning of the retinal nerve fiber layer (RNFL) [157]. Axonal tissue loss in the RNFL has been reported to be one of the earliest detectable glaucomatous changes, preceding morphologic changes of the optic nerve head (ONH), followed by functional loss, as shown by progressive visual field (VF) defects. The temporal sequence of glaucomatous structural/functional damage suggests that looking for structural changes at the ONH/RNFL level should theoretically allow an earlier diagnosis than detection of functional defects [158, 159].

Many imaging modalities have been used to analyze RNFL loss in glaucomatous eyes. Colour and red-free fundus photography [160] represent the standard approaches, but the changes of RNFL are not detectable until there is more than 50% nerve fiber loss. Scanning laser polarimetry, scanning laser ophthalmoscopy and SD-OCT are imaging modalities that allow a quantitative analysis of the ONH and RNFL. Commercially available OCT devices, however, cannot provide sufficiently clear images of individual nerve fiber bundles to identify the specific structural abnormality that underlies the pathogenesis of glaucoma [161, 162]. Adding AO to imaging systems such as flood-illuminated ophthalmoscopes, SLO equipment or OCT has recently allowed researchers to identify individual nerve fiber bundles [163–165], providing high-resolution images of both the RNFL and the ONH ( Figure 10). Takayama et al.[163], using an AOSLO, measured the individual nerve fiber bundles width in a population of twenty healthy adults. In all the eyes, the AOSLO images showed hyperreflective bundles, representing the nerve fiber bundles, in the RNFL. Dark lines among the hyperreflective bundles were considered to represent Müller cell septa. The width of nerve fiber bundles, at distances from the edge of the optic disc ranging between 1.00 and 6.00 mm, was 22 ± 6 μm. There were no significant differences among the bundle widths at these distances along the same meridian. The hyperreflective bundles on the temporal and nasal sides of the optic disc were, however, narrower (on average 20 μm width) than those above and below the optic disc (on average 30 μm width). In the central retina, the hyperreflective bundles nasal to the fovea were narrower than those above or below the fovea.

Figure 10

Kocaoglu et al.[165] used an AO-OCT to obtain images of the RFNL in four healthy subjects. The imaging sessions were confined to three locations: retinal eccentricities of 6 degree superior and inferior to the fovea, and 3 degree nasal to the fovea. The authors showed that the nerve fiber bundles reflect noticeably more light than the surrounding tissue, a factor of approximately two times more. As they approach the fovea, the nerve fiber bundles become thin (both in width and depth) and separate. Bundles at 3 degrees demonstrate a larger aspect ratio (width to thickness) than those at 6 degrees with average width and thickness ranging from 30–50 μm and 10–15 μm, respectively.

Currently, there are no studies on glaucomatous eyes or eyes with ocular hypertension. AO imaging could indeed be useful to evaluate the morphological characteristics of the RNFL in patients with ocular hypertension in order to identify possible risk factors implicated in the ONH damage progression. AO could also help to recognize early glaucomatous damages and to identify patients who could progress rapidly and also may benefit from more intensive observation and management. Moreover, AO imaging could have an important role in the evaluation of neuroprotection strategies.