Pixium Vision — Prima 2 18-month data shows strong promise

Pixium previously released positive 12-month data on the five PRIMA-FS subjects implanted with the 378-pixel Prima chip in July 2019. Since then, the company has worked to transition these subjects to the second-generation Prima 2 system consisting of second-generation transparent augmented reality (AR) glasses and a second-generation pocket computer employing improved analytics. These transparent AR glasses allow for the combination of both prosthetic vision and natural residual (peripheral) vision, and the improved pocket computer and its software provides enhanced algorithms, designed to incorporate more advanced image processing and artificial intelligence (AI) functionality. Altogether, these external hardware improvements aim to provide better functional visual experience with the 378-pixel Prima chip implant. As a reminder, the paused US feasibility study and the planned upcoming PRIMAvera pivotal study will use the Prima 2 system including the 378-pixel chip and the second-generation AR glasses and pocket computer.

The 18-month data released on 31 March shows several promising aspects. On the safety side, there were no indications of any ocular health or tolerability issues. More importantly, the use of the new Prima 2 visual system components has led to some measurable improvements in VA, in part due to some of the features on the external device components, which include improved magnification capabilities. The company reported 18-month data on four of the five EU patients implanted in PRIMA-FS (one of the five patients implanted has passed away due to health reasons completely unrelated to Prima implantation or usage).

Pixium reported that of the four remaining subjects, with the Prima 2 system activated, among them they reported between three and seven lines of improvement on the VA scale using the Landolt C optotype (the type of figures or symbols used to measure VA), compared to baseline. Effective device-assisted prosthetic VA for the four subjects was between LogMAR 0.5 (approximately 20/60, or c 33% of normal VA expected in healthy subjects) and LogMAR 0.69 (approximately 20/100, or c 20% of normal VA). Each 0.1 increment on the LogMAR scale represents the next lower VA line of the VA chart (ie the higher the LogMAR value, the lower the effective VA). Altogether, these measures are markedly superior to the baseline results, even given that they were assisted to a degree by the device’s magnification features.

Baseline VA was measured on the implanted eye shortly after surgery but without activation of the Prima system’s glasses or pocket computer (and hence, the Prima system was inactive); baseline VA is expected to be very comparable to pre-implantation VA. We note that the study’s inclusion criteria required entry VA in the implanted eye to be no better than LogMAR 1.3 (20/400 on Snellen scale, or 5% of normal VA). Baseline VA among the four subjects was between LogMAR 1.3 and LogMAR 1.4 (approximately 20/500, or c 4% of normal VA).

In the 12-month data published in a recent American Academy of Ophthalmology journal article,1 it was reported that the three subjects with optimally-positioned Prima implant chips had Prima-activated prosthetic VA between 20/460 (4.3% of normal VA) and 20/550 (3.6% of normal VA). Note these measures were taken with the initial-generation glasses and pocket computer and the magnification and AI capabilities associated with Prima 2 were not available. In addition, the first-generation glasses were opaque, so patients were unable to combine their natural remaining residual vision with the prosthetic vision provided by the Prima chip. Hence, the 12-month post-implantation VA results using the initial Prima glasses and pocket computer are not directly comparable to the 18-month results using the Prima 2 system.

Even if part of the improvement in VA versus baseline is due to the Prima 2 external devices’ magnification capabilities, these results nonetheless represent significant improvements in the ability of the patients to resolve visual details. Further, we are reassured that there is no degradation in Prima prosthetic visual performance between months 12 and 18 (and management indicates that this remains the case in some early 24-month data), as there had been some speculation that advanced Dry-AMD (which attacks and damages photoreceptor cells in the retina) could eventually provoke atrophic damage to the retinal ganglion cells (that the Prima system relies upon for providing the patient’s prosthetic vision).

We note that assessments of the device’s ability to provide improvements in functional tasks (ie whether the device assists patients in accomplishing day-to-day activities such as locating items, identifying details or maintaining independence) will be very useful for both regulators and potential insurers, and we believe the company hopes to provide some functional use or quality-of-life (QoL) measures in the 24-month data, but this will depend on the evolution of the current COVID-19 pandemic situation (as in-person follow-up visits are generally not feasible during the current pandemic).

While this is dependent on variables and factors such as sample size, in general, we anticipate that to demonstrate statistical significance in the pivotal trial and obtain regulatory approval the Prima 2 device should show at least three lines of VA improvement versus baseline, in addition to showing safety. We are encouraged that in PRIMA-FS, where patients were required to have a very advanced degree of geographic atrophy to be admissible,2 the 18-month data has shown three to seven lines of improvement vs baseline. We believe that this level of VA amelioration can potentially allow the company to propose somewhat less restrictive inclusion criteria in terms of Dry-AMD disease severity at baseline for the upcoming PRIMAvera pivotal study, while maintaining a sufficient buffer to allow it to demonstrate statistical significance. Relaxing the Dry-AMD baseline severity criteria should facilitate patient recruitment and may lead to earlier study data. We estimate that the PRIMAvera study will involve approximately 50 patients to satisfy European regulators (or up to 70 patients if the design is also accepted by the FDA as a registration-enabling pivotal trial for the US market). At this time, we are not revising our timing assumptions stated in our 26 March 2020 update note (as the COVID-19 pandemic and the timing of improvements or stabilization may affect filing and pivotal study start timelines). We continue to expect that Pixium will file the pivotal study with at least the European regulator in mid-2020 or Q320.

We are not revising our financial forecasts and they were discussed in our 26 March 2020 update note. However, the level of VA improvement shown in the 18-month data and the continued safety of the Prima implant provides us with an increased level of confidence that the product’s prosthetic visual benefits are meaningful and that it may eventually obtain regulatory approval. As the sample size is small (four remaining patients in PRIMA-FS, and two subjects have been implanted to date in the US feasibility study), and there is limited objective data on functional or QoL improvements from the Prima 2 system (which may influence approval likelihood and certainly reimbursement prospects), there remains ongoing development risk. Nonetheless, we believe that there is justification in increasing our probability of success to 20%, from 15% previously.

Our valuation for Pixium Vision is based on a rNPV approach, employing a 12.5% cost of capital and is based on the Prima opportunity in dry-AMD. We now apply a probability of success estimate for Prima-AMD in our model of 20% (from 15% previously). We now obtain a pipeline rNPV (enterprise value, excluding net cash) of €98.0m versus €69.5m previously.

Pixium had €1.0m net cash (€6.8m gross cash offset by €2.6m in refundable advances and €3.2m in long-term debt) on 31 December 2019 and, given our projection of a €0.47m monthly cash burn rate in H120 and the estimated conversion of €0.3m of part of its European Select Growth Opportunities Fund (ESGO) debt tranches to equity in Q120, we estimate Q120 net debt of €0.1m. After including €0.1m in estimated Q120 net debt, we obtain an equity valuation of €97.9m, or €3.85 per share (versus €2.78 previously).

As discussed in our 26 March 2020 note, we believe the measures previously announced by Pixium to control its cash burn rate should enable it to weather the COVID-19 storm at least into Q420 and retain a positive gross cash balance, even in the absence of an equity offering (which would be challenging to pursue or heavily dilutive given current market conditions).

We continue to assume that Pixium will need to raise €50m (including the remaining or unused €7.5m in tranches from ESGO funding facility totalling €10m) to bring Prima to launch (now in H223). If the company secures an additional tranche from ESGO (two tranches of €1.25m have been issued to date), we expect it will have sufficient funds into 2021. Our model continues to assume that the company will raise €24.7m in net illustrative debt in 2020 and an additional €25m in illustrative debt in 2021.