Pixium Vision — Encouraging signs in Prima EU feasibility study

Pixium has been reporting encouraging signs of visual response and safety in the EU feasibility study for its Prima implant in recent ophthalmology and retina conferences in the US and Europe. The firm is gearing to report full interim data, including quantifiable objective end points, towards year-end 2018 or January 2019. The data, if positive, can be used to support a regulatory filing to commence an EU pivotal study in mid-2019 (the first implantations could occur in Q319), which, in our view, can support a potential EU market approval and launch in H222.

Pixium’s Prima is a miniaturised photovoltaic wireless sub-retinal implant that is implanted underneath the retina in a surgical procedure that may take less than 90 minutes under local anaesthesia. The current Prima iteration under human clinical development is a 2mm x 2mm wireless chip consisting of 378 electrodes (pixels) in total, with each pixel being roughly 100 microns (0.1mm) in length and width. Each photovoltaic pixel is independently controlled and self-powered by near-infrared light projected from glasses worn by the patient (the glasses consist of a camera and digital mirror projector, which emit a near infrared light pattern through the patient’s eyes, designed to be processed by the Prima pixels).

Located underneath the retina, the pixels embedded on the device aim to stimulate the patient’s bipolar cells, which are located mid-stream in physiological visual signal processing. In normal visual function, photoreceptor cells (located on the outer portion of the retina, or closer to the choroid) send information to bipolar cells (located within the retina), which then relay information into retinal ganglion cells (RGCs, which are on the inner portion of the retina), and onto the brain through the optic nerve. The Pixium system is designed to restore the function of individuals whose retinal photoreceptors have been damaged by retinal disease such as severe geographic atrophy associated with Dry-ARMD. The Prima system is powered by pulsed near infrared light projected through a miniaturised projector integrated in a pair of augmented reality-like glasses (incorporating a mini-camera) worn by the patient.

In late 2017, Pixium started the five-patient, single-site,1 36-month European feasibility study for its Prima device in patients with advanced Dry-ARMD. In January 2018, it announced the first human Prima activation as part of this feasibility study following a device implantation approximately one month previously (as per study protocol). In July 2018, Pixium announced that it had completed the fifth and final implantation, and on 30 August 2018 it confirmed that all five implantations in the EU study resulted in successful consecutive activations and light perception, including the perception of white-yellow patterns with adjustable brightness, in areas where no central vision remained prior to implantation. Following activation, all patients proceeded to the visual re-education stage of the study, implemented as per study protocol, which is intended to assist patients in interpreting the new light perception patterns emitted by Prima.

Pixium presented preliminary impressions on the Prima implantations, including initial experiences of the involved surgical technique and observations on patient responses, at recent conferences including EURETINA and the American Academy of Ophthalmology (AAO) annual meetings.

In September 2018, Pixium’s medical advisory board and study investigators announced the positive review of the first clinical results of the study, which includes an assessment of the safety data and patients’ performance during the first months of rehabilitation. In October 2018, Pixium indicated that the implant remains stable under the retina and the implantation procedure did not affect the residual peripheral vision. Study investigators specified that while the study was still ongoing, early results indicate that photovoltaic restoration of central vision with Prima appears to be demonstrated with overall perception and resolution being close to the expected performance with the current pixel size of the Prima device. During the training and rehabilitation follow-up visits to date (as of October 2018), a majority of the (five) patients were able to correctly identify different visual patterns, including letters and numbers and other symbols. Study investigators reported during the AAO 2018 meeting that some patients have reported VA measures of up to 20/460 (4.4% of the 20/20 optimal VA observed in healthy individuals), which appears to be better than management anticipated.

All together, we view these successful activations as encouraging and an early suggestion of proof-of-concept that the device can interface with retinal cells to restore some visual perception in an area where vision had been lost due to prolonged degenerative disease.

Full interim (six-month, post-implantation) data from the EU feasibility study, which should include safety and some functional vision measures, are expected near the end of Q418 or in early January 2019. The primary end point is the elicitation of visual perception through the Prima implant as measured through micro-perimetry using the Octopus Visual Field device.

If the trends revealed to date remain positive, the study data can be used to enable the design of the protocol for a larger, multi-centre, CE Mark-enabling European pivotal study, which can proceed in parallel to the 36-month EU feasibility study (once the six-month feasibility study’s interim analysis is completed). Initial implantations as part of the EU pivotal study could start in Q319.

A single-centre, five-patient US feasibility trial (PRIMA FS-US), conducted at the University of Pittsburgh Medical Center, is actively recruiting and screening potentially eligible patients. We believe the first implantations are expected to occur in early 2019. This is a few months later than originally anticipated, as US regulatory requirements limiting the amount of information the firm can disclose to prospective patients on the implant’s ability to provide possible visual benefit have made recruitment more challenging. Once interim results from the European feasibility study are released in early 2019, we believe the public data (likely suggestive of visual perception provided by the implant) can be used to encourage the recruitment of patients for the US feasibility study. The study’s primary end point will be elicitation of visual perception of the Prima device, while secondary end points will include VA, measured by methods such as Early Treatment Diabetic Retinopathy Study and Freiburg Visual Acuity & Contrast Test scales. Pixium believes that 12-month safety and performance data on all five patients will likely be sufficient for US regulators to allow a larger US (pilot) study to be started. We anticipate that study data from the US feasibility study would be available in H120 and that recruitment for the US pilot study can begin in H220 (vs our prior estimate of H120).

The current Prima iteration in clinical trials (378 electrode) uses electrodes (or ‘pixels’) that are individually approximately 100 microns (0.1mm) in length, but the company and its researchers have been researching higher-density chips that use smaller individual electrodes and that can hold higher electrode/pixel densities, and that can theoretically provide higher visual resolution when implanted in patients. A higher level of VA can potentially extend the market reach of Prima to patients with less severe forms of atrophic ARMD, as the current 378-electrode iteration is only appropriate for those who already have severe forms of geographic atrophy (and incoming VA of under 5%).

Using the current manufacturing process used for Prima, it may be possible to reduce the individual electrode size down to 50 or 75 microns, whereas using even smaller electrode sizes (such as 10 microns, which would result in up to 40,000 pixels for a 2mm x 2mm chip) would require a different manufacturing process.

However, using higher-density Prima chips may entail some added risk, as the activation energy thresholds required to the device to function (as emitted through the pulsed near-IR light projected by the specialised AR glasses worn by the patient) will increase, given the need to stimulate a significantly increased amount of electrodes in the implant. Furthermore, even if a Prima device can theoretically emit signals corresponding to a higher level of resolution, the ability of the patient to resolve such fine details will depend on many factors, including the precision in the communication between the Prima chip and the external projection transmitted by the glasses worn by the patient; and the efficacy and precision of communication and interfacing between retinal cells and the electrical signals emitted by the Prima chip. Hence, it is not assured that a higher-density Prima chip would necessarily provide improved vision to the patient.

Another consideration is that even if the current EU feasibility study is successful, EU regulators may require yet another feasibility trial if a higher-density Prima device iteration is to be tested, prior to permitting a registration-enabling EU pivotal study. Management is keen to bring a Prima device to commercialisation and starting yet another EU feasibility study could add another 12 months (or longer) to the commercialisation timelines and is not management’s preferred approach.

Therefore, assuming the six-month data from the current EU feasibility study is positive, Pixium may discuss with regulators the possibility of using a mildly more-dense chip (with electrode lengths and widths of around 75 microns instead of 100 microns)2 that uses the same design/manufacturing process as the current (378-pixel) iteration, for a pivotal study. However, if EU regulators require another feasibility trial for such a change, then management will continue to use the 378-electrode iteration for the pivotal study phase, as its priority is to bring a Prima device to market (and then refine the device after approval if needed).

The use of significantly smaller electrode sizes (eg below 50 microns) entailing a different manufacturing process, which would definitely require another feasibility study, is not likely to be investigated in humans until the initial Prima device reaches commercialisation.

We also highlight that Pixium is working to develop advancements in the external glasses worn by the patient. The firm anticipates that future iterations of the glasses will be integrated with improved analytics and image processing functionality that can potentially improve the artificial vision and visual perception experienced by the patient, even in patients who will have been implanted with the first-generation Prima chip.

Once six-month data from the EU feasibility study has been attained (YE18 or early 2019), Pixium anticipates working with regulators on the design of a EU pivotal study in H119, with the aim of starting recruitment in mid-2019, with initial implantations potentially occurring before YE19. Previously, we assumed the pivotal study could commence in H119, but we now believe H219 is a more realistic estimate for initial implantations. We continue to estimate it will require 12 months of follow-up safety and efficacy data for European regulators to provide CE Mark approval.

We estimate that the EU pivotal study may require 40-60 patients. However, the true size will not be known until the current EU feasibility study is completed, as the final recruitment size for the pivotal study will likely depend upon the safety and level of visual improvement shown within the EU feasibility study.

We reiterate that to obtain CE Mark approval product safety is generally the primary consideration for regulators (it does not usually require demonstration of long-term clinical efficacy). We now estimate that 12-month data from the EU pivotal study (which we estimate is the minimum required for approval) will be available in H221 (from H121 previously), leading to potential EU commercialisation (CE Mark approval) in H222 (from H122 previously).

The US regulatory pathway for medical devices is more comprehensive. Our expectation is that following the attainment of 12-month data from the current US feasibility study (likely in H120), a larger US pilot study would need to be undertaken on a larger number of subjects (we estimate approximately 30 patients in total) prior to the start of a US pivotal study. We now estimate that US recruitment for this pilot study would start in H220.

Under an ideal scenario, Pixium could potentially also include data from sites participating in the EU pivotal trial as part of the US pilot study, which would reduce the need for duplicate or overlapping studies on similar patient populations. We assume this will be the case, thus allowing for the completion of the US pilot study in H220. We assume a US registration-enabling pivotal study would then start in 2021, which we believe will likely require 60-80 subjects and 18-24 months of follow up. Hence, we continue assume the earliest possible date for US approval and launch will be 2024.

Altogether, we expect that CE Mark clearance (and EU launch) would still occur 18-24 months earlier than US pre-market approval (PMA) and launch.

ARMD is the leading cause of blindness in adults over the age of 55 in western countries and is characterised by damage to the macular3 region of the retina, leading to central vision loss. ARMD patients generally maintain their peripheral vision. While the exact pathophysiology is not fully understood, ARMD is believed to be caused by oxidative stress, mitochondrial dysfunction and/or inflammatory processes. There are two forms of ARMD: dry (non-exudative, accounting for 80-90% of cases) and wet (exudative). Prima is intended for instances of Dry-ARMD where there is significant central or macular retinal atrophy.

The prevalence of ARMD in adults above the age of 45 is estimated at 8.0% and late-stage ARMD (with best-corrected vision acuity of 20/200, or 10% or worse) affects about 0.4% of individuals in this age group4 This represents about 815,000 people in Europe and 517,000 in the US. We assume that 30% of this late-stage subgroup would have sufficiently poor central vision to warrant potential consideration for Prima, and that 30% of these would meet all remaining inclusion criteria (including having the dry form of the disease with significant central retinal atrophy) and/or be suitable as potential responders (ie this considers that many of the ARMD patients are in poor general health and/or have concomitant eye diseases, such as glaucoma or poor optical media transparency, which would render them ineligible for Prima). Thus, we view the current target ARMD treatment population for Prima as about 73,200 in Europe and 46,500 in the US. We have adjusted our estimates mildly as we now anticipate, as stated previously, the EU launch will occur in H222 (vs H122, previously). We forecast a US launch in 2024.

As stated earlier, we anticipate that the Prima iteration to be launched will be either the current (100-micron electrode version containing 378 electrodes) version or one with slightly smaller electrodes and higher density (eg a 75-micron electrode version containing c 600 electrodes). The firm’s activities on substantially smaller electrode sizes (eg around 10 microns) carrying tens of thousands of total electrodes are more likely to be explored for a potential follow-on product and are not included in our forecasts. In an ideal and optimal scenario, once the first Prima iteration reaches the market, a next-generation Prima carrying tens of thousands of electrodes could theoretically deliver VA levels in the 25-50% range (20/80 to 20/40), which could make it potentially useable in a substantially larger segment of the Dry-ARMD population (that we anticipate for the current Prima iteration).

On 25 October 2018, Pixium provided an update on the status of its balance sheet and yearly cash burn rate through to 30 September 2018. It reported Q318 gross cash and equivalents of €18.4m, and a 9M18 operating cash burn rate of €5.24m, down 38% y-o-y. Costs were lower than the prior year period given that for most of 2017, Pixium had also been advancing the earlier-generation Iris II implant, which was halted to prioritise its resources on Prima. The company also announced that it received €2.1m in Q318 as part of a research tax credit (vs €1.7m in Q317).

Pixium’s monthly average burn rate was c €0.6m in 9M18, and we expect a comparable rate through H119. Although we also expect that the burn rate will increase in H219 once the EU pivotal study for Prima commences, we believe that Pixium’s funds on hand should be sufficient for the company to maintain its operations and fund its Prima strategy into early 2020.

The firm reported €8.2m in total gross debt on 30 June 2018 (€1.6m in conditional advances and €6.6m in long-term debt) and €16.7m in gross cash. While it did not provide a formal Q318 balance sheet, we forecast Q418e net cash of €6.8m.

With the start the EU pivotal study pushed into H219 (vs our prior estimate of H119), we have lowered our 2019 R&D expense assumption to €10.5m (from €15.0m). We now forecast 2018 and 2019 operating cash burn rates (excluding net interest) of €7.3m and €12.5m respectively, versus our prior estimates of €7.9m and €17.1m respectively.

We anticipate Pixium will seek to raise funds, likely in Q219 or mid-2019, in order to expand its financial runway to fund the EU pivotal study. Our model continues to estimate that Pixium will raise €20m in 2019, €30m in 2020 and €25m in 2021. As per usual Edison policy, our model assumes these sources will be in debt. We forecast that all this funding should enable Pixium to complete the registration-enabling Prima clinical studies in the EU to reach commercialisation in Europe. In addition, positive cash flows resulting from EU sales should enable the completion of the US pivotal study. We assume that Pixium will only start to become cash flow positive on a sustainable basis once Prima is launched (in H222).

We continue to value Pixium using an rNPV approach, employing a 12.5% cost of capital. Our valuation is based solely on the Prima opportunity in Dry-ARMD. We continue to apply a probability of success estimate for Prima-ARMD in our model of 15%. We have also adjusted our forex assumptions for US sales, by using a spot rate of $1.13/€ vs $1.16/€ previously.

After rolling forward our estimates and pushing back the EU Prima launch date to H222 (from H122, previously), we now obtain a pipeline rNPV (enterprise value) of €88.7m, down from €90.6m previously. After including €6.8m in estimated net cash at 31 December 2018, we obtain an equity valuation of c €95.5m, or €4.42 per share (compared to €4.75 previously). The lower per-share valuation is also due to the increase in share count from our 9 August 2018 research note (21.6m as of 30 November 2018, compared to 20.9m in our prior report).