The Argus® II Retinal Prosthesis System

https://doi.org/10.1016/j.preteyeres.2015.09.003Get rights and content

Abstract

The Argus® II Retinal Prosthesis System (Second Sight Medical Products) is the first prosthetic vision device to obtain regulatory approval in both Europe and the USA. As such it has entered the commercial market as a treatment for patients with profound vision loss from end-stage outer retinal disease, predominantly retinitis pigmentosa. To date, over 100 devices have been implanted worldwide, representing the largest group of patients currently treated with visual prostheses.

The system works by direct stimulation of the relatively preserved inner retina via epiretinal microelectrodes, thereby replacing the function of the degenerated photoreceptors. Visual information from a glasses-mounted video camera is converted to a pixelated image by an external processor, before being transmitted to the microelectrode array at the macula. Elicited retinal responses are then relayed via the normal optic nerve to the cortex for interpretation.

We reviewed the animal and human studies that led to the development of the Argus® II device. A sufficiently robust safety profile was demonstrated in the phase I/II clinical trial of 30 patients. Improvement of function in terms of orientation and mobility, target localisation, shape and object recognition, and reading of letters and short unrehearsed words have also been shown. There remains a wide variability in the functional outcomes amongst the patients and the factors contributing to these performance differences are still unclear. Future developments in terms of both software and hardware aimed at improving visual function have been proposed. Further experience in clinical outcomes is being acquired due to increasing implantation.

Introduction

The Argus® II Retinal Prosthesis System (Second Sight Medical Products Inc., Sylmar, California, USA) is a commercially available device that aims to restore a basic level of vision to patients with profound vision loss from outer retinal dystrophies. The device elicits visual perceptions by means of electrical stimulation of the residual diseased retina. It was the first device to go into widespread clinical use with regulatory permission in multiple countries.

The discovery by Foerster (1929) that it is possible to elicit transient and reproducible visual percepts (known as phosphenes) upon direct electrical stimulation of the visual pathway took place almost a century ago. Since then extensive research efforts have been focused on understanding and controlling such phosphene responses. The ability to reliably elicit and modify phosphenes in a controllable way by manipulating the stimulating parameters, such that they reflect the surrounding visual scenes, has been the common goal of all electrical stimulation-based visual prostheses.

The Argus® II has become the most widely used and most successful retinal prosthesis currently available in terms of regulatory approval. Since obtaining the CE mark in 2011 and FDA approval as a humanitarian device in 2013, commercial implantation has begun in many countries worldwide. Use of the device has been predominantly for patients with profound vision loss from retinitis pigmentosa and to a lesser extent, choroideremia as well as for a planned cohort with extensive geographic atrophy from age-related macular degeneration (AMD) (ClinicalTrials.gov Identifier: NCT02227498). To date, over 100 devices have been implanted and the number is likely to increase. With its integration into clinical practice, it seems timely to review the pioneering work leading up to the regulatory approval of this product, as well as the subsequent clinical outcomes with the use of this device. In particular, we will evaluate the practical implications of using this device for the patients in real-life settings based on published literature.

The issues of electrical stimulation safety, choice of stimulating wave forms, biocompatibility and hermeticity in the development of the device will not be discussed in this review as these topics have already been covered comprehensively elsewhere (Humayun, 2001, Margalit et al., 2002). Stronks and Dagnelie (2014) gave a didactic account of how different stimulating parameters such as amplitude and frequency affects the phosphene brightness and size, while Ahuja et al. (2013) discussed the factors affecting electrode thresholds in detail. These topics will therefore also not be discussed in any detail.

In 1952, Hodgkin and Huxley first described the electrical nature of signal propagation in all nervous systems by the means of action potentials (Hodgkin and Huxley, 1952). During electrical stimulation of any neural tissue with an external electrode, the injection of electrical charges creates a localised depolarisation and subsequent initiation of action potentials. As such, electrical stimulation at any point along the visual pathway could elicit visual phosphenes. Commencing with cortical stimulation by Brindley et al., in 1968 (Brindley and Lewin, 1968a) and later by others (Dobelle, 2000, Normann et al., 2009), electrical stimulation has also been described at the levels of lateral geniculate nucleus (Panetsos et al., 2009, Panetsos et al., 2011), the optic nerve (Sakaguchi et al., 2009, Sakaguchi et al., 2012; Wang et al., 2011) and the retina.

Of all of the anatomical sites listed above, retinal stimulation (e.g. the epiretinal implant Argus® II and the subretinal implant alpha-IMS) has been the most successful. There are many reasons for this and they can be best summarised as: a) greater accessibility at lower surgical risk than the intracranial visual pathways; b) straightforward monitoring of the device by direct visualisation; and c) potentially predictable and reproducible retinotopy by applying stimulation at a pre-processing site.

With the advent of modern vitreoretinal surgical techniques, access to the retina and the subsequent implantation of stimulating electrodes are comparatively easier than other sites of implantation. This is exemplified by the widespread implantation of the Argus® II System in many countries by many different surgeons over a relatively short period, at a level of surgical morbidity acceptable to regulators (Humayun et al., 2012; Rizzo et al., 2014). Despite the relative accessibility and safety discussed here, implantation still requires advanced vitreoretinal surgical skills. Complications and problems were also easily identified during the phase I/II clinical trials due to the ability to directly visualise the device (Humayun et al., 2012).

The other advantage of a retinal prosthesis is the theoretically predictable retinotopy by stimulating the visual system at a site before significant processing of the signal has occurred. Brindley and Lewin (1968b) have demonstrated that although stimulation of cortical electrodes gave rise to phosphenes in locations in agreement with the classic Holmes' retinotopic map of the visual cortex (Holmes, 1945), many of the phosphenes were complex and non-discrete in nature. This was predominantly thought to be due to the fact that the retina, as well as the rest of the pre-cortical visual pathway, carried out significant processing of the signal. This was borne out in the discovery of organisational processing in the retina, as demonstrated by the dichotomous centre/surround responses of the retinal ganglion cell (RGC) receptive fields to light stimuli. Furthermore, as there are around 120 million photoreceptors while only 1.5 million ganglion cells, many photoreceptors converge onto a single bipolar cell especially at the periphery, with further convergence taking place from the bipolar cells to RGCs (Kolb, 2003). Conversely within the macular region of the retina, the photoreceptor:bipolar cell:RGC ratio approaches 1:1:1, with minimal convergence. It is thus envisaged that focal electrical stimulating patterns with a multi-electrode array in the macular region would more likely manifest retinotopic correlations along the visual pathway. There is, however, a particular limitation to this rationale, due to the arcuate displacement of axons and the piling up of ganglion cell bodies when approaching the fovea.

The retina as a viable site for electrical stimulation to generate phosphene perception was first demonstrated by contact lens electrode stimulation in Retinitis Pigmentosa (RP) patients (Potts and Inoue, 1969, Potts and Inoue, 1970, Potts et al., 1968). Given the advantages of retinal stimulation, there are two main physical approaches to access the retina: the epiretinal approach – whereby the multi-electrode array is placed on the retinal surface in direct contact with the nerve fibre layer; and the subretinal approach – whereby the array is placed underneath the retina and is in closest contact with the bipolar cells. Both approaches have achieved reliable phosphene activation and have shown comparable functional improvements in human clinical trials. The Argus® II Retinal Prosthesis System adopts the epiretinal approach, as do the Epi-Ret3 (Menzel-Severing et al., 2012) and the Intelligent IRIS Implant (Velikay-Parel et al., 2009, Velikay-Parel et al., 2013), which were developed by two German consortium groups. The subretinal approach is adopted by the alpha-IMS device (Stingl et al., 2013).

All current retinal prostheses (including the Argus® II Retinal Prosthesis) work by electrically eliciting patterned focal responses in the residual inner retina. Ideal candidates for treatment would therefore have conditions where the outer retina (i.e. photoreceptors and/or retinal pigment epithelium) has been destroyed by any mechanism, while the inner retina (e.g. bipolar cells, RGCs, horizontal cells and amacrine cells) remains relatively intact. The largest single condition that manifests this combination of outer retinal loss with relative inner retina preservation is RP (Milam et al., 1998).

RP denotes a group of hereditary outer retinal dystrophies, affecting around 1 in 4000 live births and more than a million people worldwide (Hartong et al., 2006). Affected individuals suffer from progressive visual loss which can be profound (0.5% with no light perception and 25% with ≤20/200 vision in both eyes) (Grover et al., 1999). Post mortem histological studies of eyes of patients with moderate to severe RP have shown that even though all cellular layers of the retina underwent degeneration and cell loss with disease progression, the bipolar cell layer and the RGC layer remained relatively unaffected, with 78% and 30% preservation respectively, even in cases of severe RP (Santos, 1997, Stone, 1992). Treatment options for RP, other than for the associated cataract, epiretinal membrane and macular oedema, are limited (Guadagni et al., 2015). As such, they represent an ideal group of patients who may benefit from retinal prosthesis treatment.

Owing to the exploratory nature of the study, the recruitment criteria for entry into the Argus® II phase I/II clinical trial was RP patients with logMAR 2.9 (bare light perception) vision or worse (Humayun et al., 2012). If visual loss in the eyes was asymmetrical, the worse seeing eye was chosen as the study eye to minimise potential harm to the patient.

Although called retinal prostheses and popularly known as the ‘bionic eye’, devices such as the Argus® II at best predominantly attempt to replace photoreceptor function. As such, they depend on some native residual function of the inner retina and optic nerve. The success of a retinal prosthesis, therefore, depends on how well it is able to replace the functions of the degenerated or absent photoreceptors, namely: a) the efficient capture of the visual image; b) the transduction of the captured image into meaningful neural electric signals; and c) the subsequent activation of the residual inner retina (bipolar cells and RGCs), from where visual information can be relayed by the optic nerve to the visual cortex.

To achieve these goals, the Argus® II Retinal Prosthesis System employs 3 external components and 3 internal components. The 3 external components are: (see Fig. 1)

  • 1.

    A glasses mounted video-camera – for real-time image capture.

  • 2.

    A portable computer (the Visual Processing Unit, VPU) – for processing of the captured scenes and translation into electrical stimulating parameters conveying spatial–temporal information.

  • 3.

    An external coil (built into the side arm of the glasses) – for wireless transmission of the processed data from the VPU and electrical power to the internal components using radiofrequency (RF) telemetry.

The 3 internal components entail: (see Fig. 2)

  • 1.

    An internal coil – as a wireless receiver of RF telemetry, converting radio waves back to electrical signals to recover both data and electrical power.

  • 2.

    An inbuilt Application-Specific-Integrated-Circuit (ASIC) – for generating appropriate electrical pulses in accordance with the stimulating parameter data recovered from the internal coil, which are then relayed to the multi-electrode array.

  • 3.

    A 60-channel microelectrode (6 × 10) epiretinal array – consisting of 60 platinum electrodes (diameter = 200 μm) spaced 575 μm (centre-to-centre) apart, embedded in a thin film of polyimide. Each microelectrode is connected to the ASIC in a parallel circuit via a metallised polymer connecting cable, such that each electrode can be activated independently according to the stimulating parameters. The array comes into direct contact with the retinal surface, allowing injection of electrical charges locally to stimulate the underlying retinal tissues (see Fig. 3).

Surgical implantation of the Argus® II Retinal Prosthesis involves the standard vitreoretinal surgery techniques of pars plana vitrectomy (Machemer et al., 1971, Machemer et al., 1972) and scleral buckling procedures (Friedman, 1958, Schepens, 1957). If the patient is phakic, lensectomy is usually performed from the outset, as subsequent cataract formation would render clinical monitoring difficult.

A standard 3-port pars plana vitrectomy is first performed, with removal of the posterior hyaloid face to prevent future development of an epiretinal membrane. Any pre-existing epiretinal membrane is removed at the time of surgery in order to optimise electrical contact between the microelectrodes and the retinal surface. A 360° conjunctival peritomy is performed to allow isolation of all 4 recti muscles in preparation for placing the encircling band carrying the extraocular portion of the device.

The internal coil and ASIC are sealed in protective hermetic cases, which have a concave under surface, conforming to the curvature of the globe.

These are placed flush on bare sclera surface and sutured onto the sclera, usually in the supero-temporal quadrant of the globe, at a pre-determined distance from the limbus (approximately 5 mm) depending on the axial length of the globe (see Fig. 4). A 5 mm pars plana sclerotomy in the supero-temporal quadrant allows introduction of the microelectrode array into the vitreous cavity (see Fig. 5) – this is the only intraocular portion of the device. With appropriate scleral placement of the extraocular part of the device, the microelectrode array would rest naturally on the retinal surface at the posterior pole with minimal tension. Gentle manipulation of the array position is possible to optimise placement in the macular region. Once the array position is satisfactory, a spring-tensioned, titanium retinal tack is inserted at the heel of the array, to ensure close apposition of the array and the retinal surface (see Fig. 6). The sclerotomy is then sutured close around the traversing cable connecting the array to the ASIC to avoid scleral leakage and hypotony.

The internal coil and ASIC cases on the sclera are further stabilised by an encircling band, which passes under each of the 4 recti muscles around the globe before being gently tightened and held with a Watzke's sleeve. Finally, an allograft (e.g. Tutoplast®) or autologous fascia-lata patch is sutured over the hermetic cases before the conjunctival closure over the device. Surgical time generally falls between 1.5 and 4 h (Rizzo et al., 2014).

Research into the possibility of retinal prosthetic vision began in the early 1990s with Mark Humayun, Robert Greenberg and Eugene de Juan at Johns Hopkins University. They first demonstrated that focal electrical stimulation with a platinum electrode could elicit localised retinal responses in isolated animal retinas (Humayun et al., 1994) (discussed in detail in section 2.1.1). The group subsequently moved to the University of Southern California (USC), and started a collaboration with the Second Sight company that would eventually lead to the development of the Argus I and Argus® II retinal prostheses.

The prototype retinal prosthesis, the Argus I, began its phase I clinical trial involving 6 patients in 2002 (Humayun et al., 2003, Humayun et al., 2005, Yanai et al., 2003). The main differences between the first generation device and the Argus® II device are:

  • 1.

    The stimulating array of Argus I consisted of 16 microelectrodes (4 × 4 configuration), of either 260 μm or 520 μm in diameter, or both sizes alternating in a checkerboard pattern, with a centre-to-centre inter-electrode separation of 800 μm (see Fig. 7 inset) (Horsager et al., 2009).

  • 2.

    In the Argus I system, the hermetic casing containing the internal coil and ASIC was placed subcutaneously in the temporal bone recess, with the connecting cable leaving the periorbital space via a lateral canthotomy and tunnelled along the temporal bone subcutaneously to reach the temporal recess (see Fig. 7). This approach was similar to that of the cochlear implant and the alpha-IMS subretinal implant (Zrenner et al., 2011), and required dissection of the temporal region with the assistance of maxillofacial/otolaryngology expertise and extended surgical time. As such, this approach has been revised in the subsequent Argus® II design to simplify the implantation.

  • 3.

    In the Argus I system the external coil is situated over the temporal bone, held magnetically to the subcutaneous internal coil.

Initial results from this clinical trial with a follow-up period of up to 33 months supported safety and long term functioning of the device. A wide range of electrode thresholds were observed both within and across the subjects, but many electrodes were able to elicit visual percepts within the safety charge density limit (Brummer and Turner, 1977, Brummer et al., 1983).

Variability in the performances across the subjects was also noted, but was generally encouraging with the subjects being able to enumerate and localise high contrast objects with greater accuracy than by chance. Two subjects were also able to orientate shape (in the form of letter ‘L’) and identify 3 common objects (i.e. plate, cup and knife) with greater accuracy than by chance. Furthermore, using high contrast square wave gratings, one subject was able to differentiate the orientation of the gratings in 4 directions (vertical, horizontal, diagonal to right, diagonal to left) significantly better than by chance. The best level of resolution achievable was equivalent to logMAR 2.21 vision, in keeping with the theoretical resolution possible with the 4 × 4 array (Caspi et al., 2009). The ability to carry out these tasks supported the notion that the subjects are capable of interpreting patterned electrical stimulation. Based on these results, the next generation retinal prosthesis – the Argus® II – with 60 microelectrodes, was developed.

Section snippets

Proof of concept

The development of the Argus® II Retinal Prosthesis, as with all retinal prostheses, was based on the following assumptions:

  • 1.

    The inner retina (RGCs with/without bipolar cells) of end-stage RP patients remains functionally intact for transmission of information along the visual pathway to the visual cortex.

  • 2.

    The residual inner retina can be activated focally with localised electrical stimulation to elicit discrete phosphenes without causing damage or toxicity to the retina.

  • 3.

    Retinotopy is relatively

Clinical outcomes

Despite a plethora of theoretical predictions and extensive animal testing, confirmation of successful stimulation of the inner retina with matching perceptions by the patients has only become available since the commencement of long-term human implantation trials. The Argus I trial began in 2002 and the international multi-centre phase II clinical trial of the Argus® II began in 2007 (ClinicalTrials.gov Identifier: NCT00407602).

Future developments

Future developments in the Argus® II Retinal Prosthesis System have been described and envisaged at both the software level and the hardware level of the device.

Conclusion

The Argus® II Retinal Prosthesis System has played an important role in establishing retinal prostheses as a viable and potentially beneficial treatment option in blinding outer retinal conditions. The ability of this device to provide stable, chronic retinal stimulation in a relatively safe manner over many years has been recognised and has led to regulatory approval across many countries. However, despite the increasing volume of published outcomes from clinical trials using the Argus® II

References (127)

  • R.E. Marc et al.

    Neural remodeling in retinal degeneration

    Prog. Retin. Eye Res.

    (2003)
  • E. Margalit et al.

    Retinal prosthesis for the blind

    Surv. Ophthalmol.

    (2002)
  • S. Rizzo et al.

    The Argus II retinal prosthesis: 12-month outcomes from a single-study center

    Am. J. Ophthalmol.

    (2014)
  • M. Abramian et al.

    Activation of ganglion cell axons following epiretinal electrical stimulation with hexagonal electrodes

    Conf. Proc. IEEE Eng. Med. Biol. Soc.

    (2010)
  • M.M. Abramian et al.

    Activation of retinal ganglion cells following epiretinal electrical stimulation with hexagonally arranged bipolar electrodes

    J. Neural Eng.

    (2011)
  • M.M. Addi et al.

    Charge recovery during concurrent stimulation for a vision prosthesis

    Conf. Proc. IEEE Eng. Med. Biol. Soc.

    (2008)
  • A.K. Ahuja et al.

    Blind subjects implanted with the Argus II retinal prosthesis are able to improve performance in a spatial-motor task

    Br. J. Ophthalmol.

    (2011)
  • M. Arsiero et al.

    Subjects blinded by outer retinal dystrophies are able to recognize shapes using the Argus II Retinal Prosthesis System

  • X. Beebe et al.

    Charge injection limits of activated iridium oxide electrodes with 0.2 ms pulses in bicarbonate buffered saline

    IEEE Trans. Biomed. Eng.

    (1988)
  • M.R. Behrend et al.

    Resolution of the epiretinal prosthesis is not limited by electrode size

    IEEE Trans. Neural Syst. Rehabil. Eng.

    (2011)
  • G.S. Brindley

    The number of information channels needed for efficient reading

    J. Physiol.

    (1964)
  • G.S. Brindley et al.

    The visual sensations produced by electrical stimulation of the medial occipital cortex

    J. Physiol.

    (1968)
  • G.S. Brindley et al.

    The sensations produced by electrical stimulation of the visual cortex

    J. Physiol.

    (1968)
  • S.B. Brummer et al.

    Electrochemical considerations for safe electrical stimulation of the nervous system with platinum electrodes

    IEEE Trans. Biomed. Eng.

    (1977)
  • S.B.S. Brummer et al.

    Criteria for selecting electrodes for electrical stimulation: theoretical and practical considerations

    Ann. N. Y. Acad. Sci.

    (1983)
  • A. Caspi et al.

    Feasibility study of a retinal prosthesis: spatial vision with a 16-electrode implant

    Arch. Ophthalmol.

    (2009)
  • K. Cha et al.

    Simulation of a phosphene-based visual field: visual acuity in a pixelized vision system

    Ann. Biomed. Eng.

    (1992)
  • S.-J. Chen et al.

    Neural responses elicited by electrical stimulation of the retina

    Trans. Am. Ophthalmol. Soc.

    (2006)
  • L. daCruz et al.

    The Argus II epiretinal prosthesis system allows letter and word reading and long-term function in patients with profound vision loss

    Br. J. Ophthalmol.

    (2013)
  • G.G. Dagnelie et al.

    Paragraph text reading using a pixelized prosthetic vision simulator: parameter dependence and task learning in free-viewing conditions

    Invest. Ophthalmol. Vis. Sci.

    (2006)
  • G.G. Dagnelie et al.

    Real and virtual mobility performance in simulated prosthetic vision

    J. Neural Eng.

    (2007)
  • G.G. Dagnelie et al.

    (16:30) Electrically elicited ERGs and VEPs in Argus II recipients: a window into signal processing in the degenerated RP retina

    Doc. Ophthalmol.

    (2014)
  • W.H. Dobelle

    Artificial vision for the blind by connecting a television camera to the visual cortex

    ASAIO J.

    (2000)
  • J.D. Dorn et al.

    The Argus™ II retinal prosthesis provides complex form vision for a subject blinded by retinitis pigmentosa

  • J.D. Dorn et al.

    The detection of motion by blind subjects with the epiretinal 60-electrode (Argus II) retinal prosthesis blind subjects and motion detection

    JAMA Ophthalmol.

    (2013)
  • O. Foerster

    Beiträge zur Pathophysiologie der Sehbahn und der Sehsphäre

    J. Psychol. Neurol.

    (1929)
  • M.W. Friedman

    The scleral buckling procedure

    Trans. Pac. Coast Otoophthalmol. Soc. Annu. Meet.

    (1958)
  • R.J. Greenberg et al.

    A computational model of electrical stimulation of the retinal ganglion cell

    IEEE Trans. Biomed. Eng.

    (1999)
  • S.S. Grover et al.

    Visual acuity impairment in patients with retinitis pigmentosa at age 45 years or older

    Ophthalmology

    (1999)
  • T. Guo et al.

    Selective activation of ON and OFF retinal ganglion cells to high-frequency electrical stimulation: a computational modeling study

  • D. Güven et al.

    Long-term stimulation by active epiretinal implants in normal and RCD1 dogs

    J. Neural Eng.

    (2005)
  • M. Hauer et al.

    An intraocular camera for retinal prostheses: restoring sight to the blind

  • Hauer, M.C., 2009. Intraocular camera for retinal prostheses: refractive and diffractive lens...
  • J.S. Hayes et al.

    Visually guided performance of simple tasks using simulated prosthetic vision

    Artif. Organs

    (2003)
  • A.L. Hodgkin et al.

    A quantitative description of membrane current and its application to conduction and excitation in nerve

    J. Physiol.

    (1952)
  • G. Holmes

    Ferrier lecture: the organization of the visual cortex in man on JSTOR

  • A. Horsager et al.

    Temporal interactions during paired-electrode stimulation in two retinal prosthesis subjects

    Invest. Ophthalmol. Vis. Sci.

    (2011)
  • A. Horsager et al.

    Spatiotemporal interactions in retinal prosthesis subjects

    Invest. Ophthalmol. Vis. Sci.

    (2010)
  • A. Horsager et al.

    Predicting visual sensitivity in retinal prosthesis patients

    Invest. Ophthalmol. Vis. Sci.

    (2009)
  • M. Humayun et al.

    Bipolar surface electrical stimulation of the vertebrate retina

    Arch. Ophthalmol.

    (1994)
  • Cited by (0)

    1

    Percentage of work contributed by each author in the production of the manuscript is as follows: Yvonne Hsu-Lin Luo: 50% (wrote the initial draft); Lyndon da Cruz: 50% (revised the manuscript).

    View full text