Downloaded from bjo.bmj.com on June 14, 2014 - Published by group.bmj.com
Review
Retinal implants: a systematic review Alice T Chuang,1,2 Curtis E Margo,3 Paul B Greenberg1,2 1
Section of Ophthalmology, VA Medical Center, Providence, Rhode Island, USA 2 Division of Ophthalmology, Alpert Medical School, Brown University, Providence, Rhode Island, USA 3 Departments of Ophthalmology and Pathology, Morsani College of Medicine, University of South Florida, Tampa, Florida, USA Correspondence to Dr Paul B Greenberg, Section of Ophthalmology, VA Medical Center, 830 Chalkstone Ave, Providence, RI 02908, USA; [email protected] Received 10 May 2013 Revised 22 November 2013 Accepted 16 December 2013 Published Online First 8 January 2014
ABSTRACT Retinal implants present an innovative way of restoring sight in degenerative retinal diseases. Previous reviews of research progress were written by groups developing their own devices. This systematic review objectively compares selected models by examining publications describing five representative retinal prostheses: Argus II, Boston Retinal Implant Project, Epi-Ret 3, Intelligent Medical Implants (IMI) and Alpha-IMS (Retina Implant AG). Publications were analysed using three criteria for interim success: clinical availability, vision restoration potential and long-term biocompatibility. Clinical availability: Argus II is the only device with FDA approval. Argus II and Alpha-IMS have both received the European CE Marking. All others are in clinical trials, except the Boston Retinal Implant, which is in animal studies. Vision restoration: resolution theoretically correlates with electrode number. Among devices with external cameras, the Boston Retinal Implant leads with 100 electrodes, followed by Argus II with 60 electrodes and visual acuity of 20/1262. Instead of an external camera, Alpha-IMS uses a photodiode system dependent on natural eye movements and can deliver visual acuity up to 20/546. Long-term compatibility: IMI offers iterative learning; EpiRet 3 is a fully intraocular device; Alpha-IMS uses intraocular photosensitive elements. Merging the results of these three criteria, Alpha-IMS is the most likely to achieve long-term success decades later, beyond current clinical availability.
INTRODUCTION
To cite: Chuang AT, Margo CE, Greenberg PB. Br J Ophthalmol 2014;98:852–856. 852
In degenerative retinopathy such as retinitis pigmentosa and age-related macular degeneration, blindness is caused by photoreceptor cell damage. Any neuronal remodelling that does occur has not yet been shown to definitely affect vision processing ability.1 2 On this basis, vision-restoring implants have been developed to interface with various parts of the visual pathway, particularly the retina.3–5 Current retinal implants permit rudimentary greyscale light perception, allowing patients to perform simple visual tasks and independently navigate.6 7 To date, several reviews of retinal implants have been published.1 3 5 However, they have been authored by groups developing their own devices. While these researchers may have an understanding of the advantages and limitations of their specific prostheses, emphasis is often placed on their respective designs. Furthermore, most retinal implant devices are still in a stage of development requiring significant investment both financially and intellectually, limiting data publication.3 These trends underscore the need for a systematic review that provides a neutral third-party analysis of current prostheses using a priori criteria of success. To this end, we evaluate herein the state of current
retinal prostheses with particular emphasis on their clinical availability,1 vision restoration8 and longterm biocompatibility9 in patients with vision loss due to degenerative retinopathy.
METHODS Literature review was performed using the PubMed e-database through April 2013. Initial search terms were ‘retinal implant’ and ‘retina prosthesis’. Additional publications were identified from references of the preliminary literature and subsequent updates. Devices selected for analysis do not constitute a comprehensive list of worldwide research but were chosen for significant progress in one or more of the following domains: clinical availability, vision restoration or potential for long-term biocompatibility. The first selection criterion was progress in clinical trial availability, arguably the most robust yardstick for measuring technical headway. Those already tested in humans are closer to making a significant impact in clinical medicine. While this criterion may not determine long-term success, it means technical refinements are at a more advanced stage, usually at the human interface level. As regards to the duration of vision loss, neural remodelling may increase after decades of disuse.1 5 Longer wait periods may therefore decrease suitability for implantation, as patients might require extensive training to successfully use gained vision. Another vital trait of implants was vision restoration. The primary focus was clinical improvement in patients’ vision, but these data were not available for all devices. In its absence, technical details were used to predict visual outcomes. Parameters include pixel number, pixel density, the field of vision7 10 11 and how closely the implant interfaces with remaining retinal pathways. Implant locations include subretinal and epiretinal. Subretinal devices are inserted between the pigment epithelium and outer retina, while epiretinal devices are designed to interact with ganglion cells of the inner retina. Additionally, image capture can rely either on external cameras or implanted photodiode arrays.3 External cameras relay visual information to electrodes, the number and density of which determine pixel resolution.12 Photodiode arrays are directly stimulated by ambient light transmitted to the retina.2 Lastly, long-term biocompatibility was used to predict future success beyond what has already received FDA approval and CE marking. Implantation and postoperative serious adverse effects (SAE) were reviewed, and potential complications were also anticipated based on principles in biomedical device design. Early models depended on wired data and power transfer, including both transcleral and transdermal cables.5 7 9 Fully intraocular implants are now possible, which lowers the
Chuang AT, et al. Br J Ophthalmol 2014;98:852–856. doi:10.1136/bjophthalmol-2013-303708
Downloaded from bjo.bmj.com on June 14, 2014 - Published by group.bmj.com
Review risk of infection and implant debris, as permanently disrupting barriers between the intraocular space and outside environment is associated with increased risk of infection.13 Iterative learning is also possible, with individual calibration for each patient.5 Options that improve patient satisfaction may determine the ultimate fate of marketed devices, especially decades later, when retinal implants become prevalent.
RESULTS Table 1 summarises characteristics of the five devices selected for significant development in clinical availability, vision restoration potential or long-term biocompatibility.
Argus II Clinical availability Of the analysed retinal implants, Argus II was the first approved device in clinical trials in both the US (FDA phase 4 postmarket surveillance) and Europe ( phase IV, European CE marking).1 14 The first-generation model, Argus I, consisted of a 16-electrode array and was implanted in six subjects. Argus II consists of a 60-electrode array and has been implanted in 30 subjects for up to 38.3 months, with 94.4% of electrodes retaining function throughout their study.7 Twenty-nine patients continued home use of the device.7 20
Vision restoration Testing included locating objects, identifying motion direction, reading letters and orientation and mobility tasks.7 21 Of 28 patients, 16–28 performed better with the implant on versus off with an estimated visual field of 20°. The highest achieved visual acuity ranged from logMAR 1.6 to 2.9, with letter reading measured at 20/1262.7 Furthermore, tracing paths on touchscreens with auditory feedback had higher accuracy but longer times
with the implant on; this demonstrates potential learning and reactivating of the visual pathway.14
Long-term biocompatibility The Argus II design incorporates transcleral cables, which may increase risk of infection.22 23 Image processing is as follows: data from the external camera are transferred to a processor and transmitter coil, then wirelessly to an electronics case. The electronics case is connected by a transcleral cable to the epiretinal implant, which is held in place with retinal tacks.7 While external electronics allow simpler technical updates and troubleshooting, long-term infectious complications may be more frequent with transcleral cables.1 In 30 subjects, there were 9 SAE, including conjunctival erosion, endophthalmitis and hypotony. All were successfully treated without permanent complications, and only one patient did not continue in the trial. SAE rate was decreased in later surgeries.7 20
Boston Retinal Implant Vision restoration While not yet in humans trials, the Boston Retinal Implant may have the highest pixel resolution among devices using external cameras. This 100-electrode array prosthesis has been implanted in two minipigs for 3 and 5 months, respectively.3 9 Subretinal insertion is surgically more difficult, but this allows smoother integration with existing neuronal pathways.5 Power and data delivery successfully elicited electric current response in the electrodes at the eye surface, with reference electrodes at the ear.3 9
Long-term biocompatibility In future research, glasses frames with an external camera will wirelessly provide image data and power to an electronics case. Transcleral cables connect this electronic case to the subretinal
Table 1 Summary of retinal implants Name Clinical trial availability
Argus II (second sight)1
7 14
FDA approved (phase 4). Europe CE mark. 30 humans total. 9 SAEs
Vision restoration and technical specifications Tests Visual testing: object/motion mobility 20/1262 Image capturing External camera Array location Epiretinal Device Retinal tacks stabilisation Electrode 60 electrodes, 200 mm specifications diameter Intraocular 3 mm diameter dimensions Extraocular Glasses frame and handheld dimensions video processor
Long-term biocompatibility
Cables may increase infection. Hypotony, conjunctival erosion, dehiscence, endophthalmitis, retinal detachments reported
Boston Retinal Implant project3
9
Epi-Ret 313
15 16
Intelligent medical implants (IMI)5 12 17 18
Alpha-IMS (Retina Implant AG)2
6 8 10 19
2 minipigs, 3–5 months Wore through conjunctiva. No SAEs reported
6 humans, 28 days. No SAEs reported
3 humans, 30 months. No SAEs reported
Europe CE mark. Current trial: 19 humans, 3–9 months. 1 SAE
Electric stimulation
Electric stimulation
Electric stimulation
External camera Subretinal Physiologic closure of subretinal space 100 electrodes, 400 mm diameter 5 mm diameter×10 mm Primary metal coil: 19 mm radius. Hermetic case: 11×11×2 mm Hermetic casing eroded conjunctiva
External camera Epiretinal Retinal tacks
External camera Epiretinal Handheld
25 electrodes, 100 mm diameter 40 mm×3 mm×10 mm
49 electrodes. 100–360 mm diameters 60 mm×1 mm×10 mm
Visual testing: object/motion localisation 20/546 Multiphotodiode array Subretinal Physiological closure of subretinal space 1500 photodiodes, 15×30 mm
N/A
Glasses frame
22 mm cable to power control unit. Handheld controller
Fully intraocular. Loose tacks, epiretinal gliosis, increased IOP reported
Iterative (
Review
Retinal implants: a systematic review Alice T Chuang,1,2 Curtis E Margo,3 Paul B Greenberg1,2 1
Section of Ophthalmology, VA Medical Center, Providence, Rhode Island, USA 2 Division of Ophthalmology, Alpert Medical School, Brown University, Providence, Rhode Island, USA 3 Departments of Ophthalmology and Pathology, Morsani College of Medicine, University of South Florida, Tampa, Florida, USA Correspondence to Dr Paul B Greenberg, Section of Ophthalmology, VA Medical Center, 830 Chalkstone Ave, Providence, RI 02908, USA; [email protected] Received 10 May 2013 Revised 22 November 2013 Accepted 16 December 2013 Published Online First 8 January 2014
ABSTRACT Retinal implants present an innovative way of restoring sight in degenerative retinal diseases. Previous reviews of research progress were written by groups developing their own devices. This systematic review objectively compares selected models by examining publications describing five representative retinal prostheses: Argus II, Boston Retinal Implant Project, Epi-Ret 3, Intelligent Medical Implants (IMI) and Alpha-IMS (Retina Implant AG). Publications were analysed using three criteria for interim success: clinical availability, vision restoration potential and long-term biocompatibility. Clinical availability: Argus II is the only device with FDA approval. Argus II and Alpha-IMS have both received the European CE Marking. All others are in clinical trials, except the Boston Retinal Implant, which is in animal studies. Vision restoration: resolution theoretically correlates with electrode number. Among devices with external cameras, the Boston Retinal Implant leads with 100 electrodes, followed by Argus II with 60 electrodes and visual acuity of 20/1262. Instead of an external camera, Alpha-IMS uses a photodiode system dependent on natural eye movements and can deliver visual acuity up to 20/546. Long-term compatibility: IMI offers iterative learning; EpiRet 3 is a fully intraocular device; Alpha-IMS uses intraocular photosensitive elements. Merging the results of these three criteria, Alpha-IMS is the most likely to achieve long-term success decades later, beyond current clinical availability.
INTRODUCTION
To cite: Chuang AT, Margo CE, Greenberg PB. Br J Ophthalmol 2014;98:852–856. 852
In degenerative retinopathy such as retinitis pigmentosa and age-related macular degeneration, blindness is caused by photoreceptor cell damage. Any neuronal remodelling that does occur has not yet been shown to definitely affect vision processing ability.1 2 On this basis, vision-restoring implants have been developed to interface with various parts of the visual pathway, particularly the retina.3–5 Current retinal implants permit rudimentary greyscale light perception, allowing patients to perform simple visual tasks and independently navigate.6 7 To date, several reviews of retinal implants have been published.1 3 5 However, they have been authored by groups developing their own devices. While these researchers may have an understanding of the advantages and limitations of their specific prostheses, emphasis is often placed on their respective designs. Furthermore, most retinal implant devices are still in a stage of development requiring significant investment both financially and intellectually, limiting data publication.3 These trends underscore the need for a systematic review that provides a neutral third-party analysis of current prostheses using a priori criteria of success. To this end, we evaluate herein the state of current
retinal prostheses with particular emphasis on their clinical availability,1 vision restoration8 and longterm biocompatibility9 in patients with vision loss due to degenerative retinopathy.
METHODS Literature review was performed using the PubMed e-database through April 2013. Initial search terms were ‘retinal implant’ and ‘retina prosthesis’. Additional publications were identified from references of the preliminary literature and subsequent updates. Devices selected for analysis do not constitute a comprehensive list of worldwide research but were chosen for significant progress in one or more of the following domains: clinical availability, vision restoration or potential for long-term biocompatibility. The first selection criterion was progress in clinical trial availability, arguably the most robust yardstick for measuring technical headway. Those already tested in humans are closer to making a significant impact in clinical medicine. While this criterion may not determine long-term success, it means technical refinements are at a more advanced stage, usually at the human interface level. As regards to the duration of vision loss, neural remodelling may increase after decades of disuse.1 5 Longer wait periods may therefore decrease suitability for implantation, as patients might require extensive training to successfully use gained vision. Another vital trait of implants was vision restoration. The primary focus was clinical improvement in patients’ vision, but these data were not available for all devices. In its absence, technical details were used to predict visual outcomes. Parameters include pixel number, pixel density, the field of vision7 10 11 and how closely the implant interfaces with remaining retinal pathways. Implant locations include subretinal and epiretinal. Subretinal devices are inserted between the pigment epithelium and outer retina, while epiretinal devices are designed to interact with ganglion cells of the inner retina. Additionally, image capture can rely either on external cameras or implanted photodiode arrays.3 External cameras relay visual information to electrodes, the number and density of which determine pixel resolution.12 Photodiode arrays are directly stimulated by ambient light transmitted to the retina.2 Lastly, long-term biocompatibility was used to predict future success beyond what has already received FDA approval and CE marking. Implantation and postoperative serious adverse effects (SAE) were reviewed, and potential complications were also anticipated based on principles in biomedical device design. Early models depended on wired data and power transfer, including both transcleral and transdermal cables.5 7 9 Fully intraocular implants are now possible, which lowers the
Chuang AT, et al. Br J Ophthalmol 2014;98:852–856. doi:10.1136/bjophthalmol-2013-303708
Downloaded from bjo.bmj.com on June 14, 2014 - Published by group.bmj.com
Review risk of infection and implant debris, as permanently disrupting barriers between the intraocular space and outside environment is associated with increased risk of infection.13 Iterative learning is also possible, with individual calibration for each patient.5 Options that improve patient satisfaction may determine the ultimate fate of marketed devices, especially decades later, when retinal implants become prevalent.
RESULTS Table 1 summarises characteristics of the five devices selected for significant development in clinical availability, vision restoration potential or long-term biocompatibility.
Argus II Clinical availability Of the analysed retinal implants, Argus II was the first approved device in clinical trials in both the US (FDA phase 4 postmarket surveillance) and Europe ( phase IV, European CE marking).1 14 The first-generation model, Argus I, consisted of a 16-electrode array and was implanted in six subjects. Argus II consists of a 60-electrode array and has been implanted in 30 subjects for up to 38.3 months, with 94.4% of electrodes retaining function throughout their study.7 Twenty-nine patients continued home use of the device.7 20
Vision restoration Testing included locating objects, identifying motion direction, reading letters and orientation and mobility tasks.7 21 Of 28 patients, 16–28 performed better with the implant on versus off with an estimated visual field of 20°. The highest achieved visual acuity ranged from logMAR 1.6 to 2.9, with letter reading measured at 20/1262.7 Furthermore, tracing paths on touchscreens with auditory feedback had higher accuracy but longer times
with the implant on; this demonstrates potential learning and reactivating of the visual pathway.14
Long-term biocompatibility The Argus II design incorporates transcleral cables, which may increase risk of infection.22 23 Image processing is as follows: data from the external camera are transferred to a processor and transmitter coil, then wirelessly to an electronics case. The electronics case is connected by a transcleral cable to the epiretinal implant, which is held in place with retinal tacks.7 While external electronics allow simpler technical updates and troubleshooting, long-term infectious complications may be more frequent with transcleral cables.1 In 30 subjects, there were 9 SAE, including conjunctival erosion, endophthalmitis and hypotony. All were successfully treated without permanent complications, and only one patient did not continue in the trial. SAE rate was decreased in later surgeries.7 20
Boston Retinal Implant Vision restoration While not yet in humans trials, the Boston Retinal Implant may have the highest pixel resolution among devices using external cameras. This 100-electrode array prosthesis has been implanted in two minipigs for 3 and 5 months, respectively.3 9 Subretinal insertion is surgically more difficult, but this allows smoother integration with existing neuronal pathways.5 Power and data delivery successfully elicited electric current response in the electrodes at the eye surface, with reference electrodes at the ear.3 9
Long-term biocompatibility In future research, glasses frames with an external camera will wirelessly provide image data and power to an electronics case. Transcleral cables connect this electronic case to the subretinal
Table 1 Summary of retinal implants Name Clinical trial availability
Argus II (second sight)1
7 14
FDA approved (phase 4). Europe CE mark. 30 humans total. 9 SAEs
Vision restoration and technical specifications Tests Visual testing: object/motion mobility 20/1262 Image capturing External camera Array location Epiretinal Device Retinal tacks stabilisation Electrode 60 electrodes, 200 mm specifications diameter Intraocular 3 mm diameter dimensions Extraocular Glasses frame and handheld dimensions video processor
Long-term biocompatibility
Cables may increase infection. Hypotony, conjunctival erosion, dehiscence, endophthalmitis, retinal detachments reported
Boston Retinal Implant project3
9
Epi-Ret 313
15 16
Intelligent medical implants (IMI)5 12 17 18
Alpha-IMS (Retina Implant AG)2
6 8 10 19
2 minipigs, 3–5 months Wore through conjunctiva. No SAEs reported
6 humans, 28 days. No SAEs reported
3 humans, 30 months. No SAEs reported
Europe CE mark. Current trial: 19 humans, 3–9 months. 1 SAE
Electric stimulation
Electric stimulation
Electric stimulation
External camera Subretinal Physiologic closure of subretinal space 100 electrodes, 400 mm diameter 5 mm diameter×10 mm Primary metal coil: 19 mm radius. Hermetic case: 11×11×2 mm Hermetic casing eroded conjunctiva
External camera Epiretinal Retinal tacks
External camera Epiretinal Handheld
25 electrodes, 100 mm diameter 40 mm×3 mm×10 mm
49 electrodes. 100–360 mm diameters 60 mm×1 mm×10 mm
Visual testing: object/motion localisation 20/546 Multiphotodiode array Subretinal Physiological closure of subretinal space 1500 photodiodes, 15×30 mm
N/A
Glasses frame
22 mm cable to power control unit. Handheld controller
Fully intraocular. Loose tacks, epiretinal gliosis, increased IOP reported
Iterative (
