Science Corporation puts its PRIMA retinal chip into a Stargardt disease patient for the first time

In his first rehabilitation session, the patient saw a letter. In his second, he read a word, according to Science Corporation ^[6]. The report of the announcement does not include an acuity measurement or the size of the trial.

Hodak's framing on stage at Disrupt was wide. "The retina is part of the central nervous system," he said ^[11]. His description of what the patient actually has is narrow. It is "a two millimeter by two millimeter implant," and "it's like looking through a straw," he said ^[12]. He added: "It is not a big field of view color vision," but "it's black and white, but it is sharp," and the patient can "see a couple letters at a time" ^[13]. The user today is a Stargardt patient who wants sharp central vision for a few letters at a time.

What the patient did is read a word. What the pitch says is that the approach applies to multiple diseases ^[10]. There is a reason to expect the carry-over. The implant works like artificial photoreceptors ^[8], and Stargardt destroys the light-sensing cells of the outer retina while potentially sparing the inner retina that passes signals to the brain ^[9]. On the company's description, the chip replaces the part that fails. Two sessions show an early response. The size of the effect and whether it lasts need more sessions and more patients.

The roadmap is stated with confidence. Hodak said that "over the next five years, I think we'll be able to get pretty close to native acuity" ^[15]. That is a five-year forecast from the person running the program. He also said a next-generation implant is being worked on and will be in humans in the next year ^[16].

For anyone deciding how much weight a "platform" claim like this deserves, two questions sort it. Does the new case fail in the same place the product repairs? And is the evidence a measurement or an anecdote? If the failure point matches and the numbers are in, plan around the extension. If the failure point matches and the evidence is an anecdote, run a pilot and set a date for the measurement. A different failure point makes it a new product, whatever the evidence says.

Stargardt lands in the pilot cell. By the company's description the failure sits in the outer retina, where the chip substitutes, and the evidence is a letter and a word. We would treat the extension as pilot-grade. The cost of that caution is timing: the next-generation implant may reach patients before measurements on the current one are reported, and those numbers would then describe a device already being replaced.