In a relatively new field of research, the study population should be kept homogeneous and structured to achieve interpretable results. Patients with rod-cone or cone-rod degeneration, who represent the main patient group for subretinal implants, should not have additional ophthalmologic diseases. As the subretinal electronic receiver array requires a focused image, clear optic media are necessary. Also, inner retinal disease or disease of the optic nerve or central visual pathway would be a contraindication for the subretinal visual implant's successful functioning.
Because there was no pre-existing experience with a subretinal implant in humans, the quality of the chip-mediated vision to be expected prior to the pilot trial was unknown. The computations based on the spatial alignment of the micro-photodiodes and animal experiments estimated a resolution of approximately 1.0° to 0.25° visual angle or decimal acuity of 0.06 to 0.08;[4, 18-20] however, it was impossible to predict whether the chip-mediated vision would allow reading, independent orientation in space or just light perception in blind patients. Thus, low pre-implantation baseline for remaining visual function in these patients was set to allow proper assessment of chip-mediated vision. This means blindness (at least monocular), in the sense of inability to localise light and objects and lack of independent visual mobility in space. All our study subjects were completely blind (no light perception) or had only an ability to distinguish light and darkness, being unable to correctly localise a light source.
Although it is also a disease affecting the outer retina, patients with macular degeneration were not included. Despite being heavily impaired, they typically do not become blind in the sense mentioned above but suffer from a central scotoma, not necessarily losing their mobility.
Blindness as an inclusion criterion was necessary also from an ethical point of view. In a first pilot trial, the potential risk of harming patients had to be as low as possible. Also for safety reasons only one eye—the one with the least function—received the implant.
Clear optic media are important not only for a sharp image falling onto the chip, but also for the best possible vision for the surgeon implanting a new device. To avoid post-operative or age-related lens opacification, an obligatory cataract operation was included prior to the study procedures. Vitrectomy with silicone oil tamponade is a part of the implantation surgery, thus homogeneity of the vitreous can be assumed as well.
In the absence of photoreceptor input, it is still an unmet challenge to examine the function of the inner retina and of the rest of the visual pathway that at present cannot be solved with certainty. Nevertheless, there are some ways for estimating whether there is sufficient function of these systems.
The first requirement is that the subjects have a fully developed and functioning central visual pathway. This is the case if a normally functioning visual system with the ability to learn reading had been established in childhood. Congenital or early-childhood blindness are contraindications for electronic visual implants. Because the visual pathway keeps developing post-natally up to school age or early teens, our inclusion criteria allowed only subjects who were able to use some useful vision at least up to the twelfth year of life and had learned to read and to move independently.
Second, the bipolar and ganglion neurons must still be able to transmit signals to the brain. How can this be tested if there are no functioning photoreceptors to pass on the signal? After the loss of photoreceptors in retinitis pigmentosa, a slow progressive degeneration of bipolar and ganglion cells can follow, although a surprisingly large population of ganglion cells and bipolar cells survives many years of complete blindness, possibly by lateral signal processing in the inner retina. The central question to be answered is whether or not residual bipolar and ganglion cells are capable of properly forwarding electrical signals. Retinal excitability can be measured by means of electrically evoked phosphenes by a corneal electrode[23, 24] and such a measurement can provide valid information about the functionality of the inner retina. Recent results point to the possibility of using additional information about this functionality by pupillography, in cases where the electrically evoked phosphenes threshold cannot be determined.
Additionally, the retinal vasculature must still allow sufficient perfusion for the inner retinal layers. We observed this in two patients of the pilot study who had considerable retinal ischemia in the region of the active microchip, when no implant-mediated visual perception was possible. Thus, a fluorescein angiographic evaluation of the retinal vasculature was made obligatory prior to inclusion in the study. Last but not least, the retinal thickness measurements by optical coherence tomography (OCT) should show neither oedema nor an extremely atrophic retina (which would suggest a profound bipolar or ganglion cell depletion). A recent subject whose retina exhibited almost no layering when imaged using spectral domain OCT had only a minimal implant-mediated perception without shape recognition or measurable acuity. Although based only on this one case, we assume that a visible layering of the inner retina in a high-resolution OCT is a good marker for its functionality. A correlation between electrical excitability of the inner retina and its thickness has been described by other research groups.
Attention should also be paid to pigment clumps on the posterior pole of the eye prior to study inclusion. It is well known that the bone spicule pigment contains migrated retinal epithelium cells, with intra-retinal deposits of pigment.[21, 27] Therefore, large pigment clumps may reduce the light falling onto the chip surface, because they would be localised intra-retinally on top of the implant. Moreover, due to different adhesion properties, the clumps may present a surgical obstacle for a smooth implantation of the device in the subretinal space. Therefore, we avoided including individuals with heavily clumped pigmentation in the target area of the subretinal chip.
Our team developed and implemented a new planning procedure to include all the previously mentioned features of the retina to define the most appropriate position of the implant.
In summary, we define our current most important eye-related inclusion criteria as follows:
- hereditary retinal degeneration of the outer retinal layers with the retinal vessels still perfused and pigment of mild to moderate density
- functionality of the inner retina as measured by transcorneal electrical excitability and layering visible in a spectral domain OCT
- no other ophthalmologic disease with relevant effects upon visual function and
- blindness (at least monocular) or visual functions not appropriate for navigation/orientation but a reading visual acuity in earlier life and appropriate visual functions up to 12 years of age.
In a study that involves a surgical procedure, the general health of the patient deserves special attention. The complete surgical implantation includes extraorbital cable placement and intraorbital microdevice placement, resulting in a surgery of about six to eight hours. The obligatory explantation with the device used in the pilot study because of the transdermal cable-bound power supply required another two to four hours of surgical procedure under general anaesthesia. Systemic diseases or conditions that might imply considerable risks with regard to the surgical interventions and long anaesthesia, as well as pregnancy, nursing or age below 18 and above 78 are exclusion criteria for study participation.
Last but not least, we regard the psychological precondition of the patient as well as his/her attitude to the study participation as very important. A blind or nearly blind person can find him/herself in critical psychological situations. In hereditary retinal disorders that lead to blindness mostly in the patient's middle age, the patient has experienced his/her visual impairment at young adult age with the knowledge of progression to blindness without an available treatment. Naturally, various psychological reactions can be evoked by the condition. Analogous to the grief cycle model, visual loss brings stages of denial, anger, bargaining, depression and acceptance as well as significant anxiety. An electronic visual implant can restore only parts of the earlier visual functions and bring a blind patient to the very low or low vision range. To avoid false or unrealistic expectations concerning the benefit of study participation, disappointments or biased results, we only include blind individuals who have accepted their blindness and who have no neurological or psychiatric problems. This is ensured by a psychiatric examination as a part of the screening procedures.
The study was designed as a prospective pilot clinical study based on randomised intra-individual implants in a double-masked fashion. For better practicability, a single-masked design is used currently in the clinical trial. In our trials, each patient received an implant, while sham surgery was not performed.
Intervention takes place as an intra-individual function of activity or inactivity of the retinal implant. Implant power ‘ON’ or ‘OFF’ modality is unknown to the patient and each test is always performed under these two conditions via two consecutive tests in a randomised order.
It should be pointed out that the patient is often aware of the condition (ON or OFF) despite the masked design, as he/she is able to judge whether he/she has a perception or not. Nevertheless, we strongly recommend sticking to this study design because additional information about the state of the implant power could eventually bias the patient's performance.
In the initial pilot study with a wire-bound device, a study time of four weeks had been set, followed by explantation of the subretinal device. This time limitation was due to the transdermal cable (Figure 2A) and to safety aspects of the implant, as the reaction of the human tissue to the new technology was not known. As no side effects or serious adverse events and no infections resulted from the implant, after the eighth subject of the pilot trial, the study period was prolonged in an amendment to four months duration.
In the short time of four weeks, as much information as possible had to be obtained. For this reason, a high frequency of visits was established, consisting of daily clinical and functional measurements for many hours. This is feasible only with a low number of study patients and the study physicians cannot plan more than two (optimally only one) patients to be implanted and tested in parallel.
During these weeks or later months, the study team has to consider also a variety of logistic aspects. As the study population is strictly defined, patients had to be recruited from large distances. Logistics, accommodation and personal assistance for blind subjects must be available in an unfamiliar city for the study period.
Refraction correction after implantation
As mentioned above, the implantations were combined with a preceding cataract surgery and intra-operative silicone oil tamponade. The silicon oil changes the eye's refraction and pseudophakia precludes accommodation. Depending on the distance to the visual stimulus, appropriate refractive correction is necessary to guarantee a focused image on the microphotodiodes. This must be determined prior to the functional tests; however, an autorefractometer can fail in the case of highly reflective subretinal implants and also a retinoscopy over a subretinal chip can lead to irregular reflexes. Thus, a very careful retinoscopy is recommended or, if available, refraction with a manual refractometer, where the examiner can directly check the reflection image quality on the microchip.
As soon as chip-mediated vision allows the patient to see grids or optotypes, an additional subjective refraction can be performed.