For nearly a century, the slit lamp has remained the cornerstone of the ophthalmic examination and has remained largely unchanged despite rapid advances in diagnostic imaging and digital technologies. In this Q&A, John Berestka, MD—cofounder and chief medical officer of Lightfield Medical—discusses the clinical motivation, engineering challenges, and physician perspective behind developing an automated slit lamp designed to modernize anterior-segment imaging. He also shares his vision for how standardized imaging and artificial intelligence could expand access to eye care, improve clinical efficiency, and shape the next generation of ophthalmic practice.
Ophthalmology Management: As cofounder and CMO at Lightfield Medical, what initially drew you to the challenge of automating the slit lamp, and what unmet clinical needs were you hoping to address?
Chip Berestka, MD: The slit lamp is the heart and soul of ophthalmology. It is used in almost every eye exam. The slit-lamp principle predates electricity in the clinic. The design used in today’s slit lamp dates to Goldmann’s 1933 version, and nothing has really changed since then. It takes 1 or 2 years of residency training to really become proficient at the slit lamp. Mastery of the fine muscle movements that control the optics leads to an ophthalmologist’s love affair with the slit lamp. But this love affair has its costs. Even with attention to ergonomics, using the slit lamp continually stresses the lumbar and cervical spine. Nine percent of ophthalmologists have to cut their surgical careers short because of spinal disability.1,2 In my own case, slit lamp use caused a severe C-7 radiculopathy. I also developed bilateral sciatica. At times, I had radicular pain going down 3 limbs, and I knew ophthalmology needed a better option.
I watched my retina colleagues with jealousy: Retina specialists could use optical coherence tomography (OCT) and Optos images to do 90% of their exams. And as I thought about it, every aspect of the eye exam has been automated and digitized except the slit lamp. As a result, we can delegate refractions, topographies, visual fields, biometries, OCTs, ultrasounds, and retinal photographies to our technicians. But the slit lamp is the analogue bottleneck of the clinic. The patient and the doctor need to be sitting across from each other in the same place at the same time.
OM: Developing an automated slit lamp requires translating the expertise of experienced eyecare professionals into technology. What were some of the biggest clinical and technical hurdles your team encountered during development, and how did you overcome them?
CB: I filed my first patent in 2014 but found that camera systems that could quickly obtain images at multiple focal planes were incredibly expensive and computationally constrained. I travelled to Kiel, Germany, and Tenerife, Spain, to explore the top offerings in light-field industrial cameras but was frustrated by the quality and resolution of the images. Ophthalmologists need both high resolution and a high dynamic range.
I was paired with Jan Bonel through The Winning Pitch Challenge, and he had worked with Pablo Artal, PhD, and Harilaos Ginis, PhD. They initially thought that automated slit-lamp photography would be much easier than the high-end aberrometry and scatter analysis optics they had helped pioneer.
It took them over 18 months to get our first really useful images. In the process, though, they had completely redesigned the slit-lamp from first principles. Our optics are nonparfocal, and our design is the first breakthrough in slit-lamp design since Goldmann. We have been granted 6 US patents and 15 more in the pipeline.
OM: How did your experience as a physician influence the design decisions behind the automated slit lamp, particularly with regard to image quality, workflow integration, and usability in real-world practice settings?
CB: Image quality is the sine qua non for clinical use. We ran a 2-year clinical trial in my office to continuously refine our parameters and techniques to get images that would be useful to ophthalmologists.
Along the way, we realized that our technology could capture images that simply are not possible with conventional slit-lamp design. Our automated slit lamps use digital light engines, liquid lenses, and high-speed cameras that can create and capture structured light. As a result, we can capture compound slits that are a combination of a thin slit and a thick slit. Compound slits cut the acquisition, transmission, storage, and review time of a slit-lamp exam in half.
We can also reimagine the way we capture retroillumination images. As a cornea specialist, I frequently rely on retro imaging to catch subtle cases of Fuchs’ dystrophy, anterior basement membrane dystrophy, and PSC cataracts that would otherwise go undetected. Our team of engineers pioneered a better way of capturing retroillumination images that uses light that straddles the visible and infrared spectrum. We get beautiful 360° intrapupillary retroillumination images without the familiar white-bar-of-light artifact that makes retro images very difficult in undilated pupils. Because we use near-infrared illumination, we get partial dilation in undilated pupils—a nonmydriatic slit-lamp camera, in a sense.
This nonmydriatic retroillumination technology will improve patient care beyond traditional ophthalmology. The majority of US eye care is delivered at optical chains where patients are often not dilated. As a result, corneal dystrophies and posterior cataracts often go undiagnosed. Optical chains with our technology would likely catch every case of Fuchs’ dystrophy.
OM: Artificial intelligence (AI)-enabled ophthalmic imaging tools are increasingly being evaluated for their potential to improve access to care. How do you see the automated slit lamp impacting screening, teleophthalmology, and care delivery in underserved or remote communities?
CB: One of my motivations in developing this technology was to help bring eye care to underserved parts of America and the world. The challenges are both distance and scale. Starlink has largely solved the problem of distance, and AI will help us solve the challenge of scale. Federal studies predict a 30% ophthalmologist shortage by 2035, with a 71% predicted shortage in the rural United States.3
AI in the interpretation of retinal images is already quite advanced. There are numerous retinal AI companies, and Project Orbis is already using the technology. There is no AI yet for slit-lamp images. The problem is a lack of a homogeneous image set. Bolted-on cameras for slit lamps have existed for years, but good focus is often lacking and the slit lighting, slit sizes, slit angles, and background illumination are always varied.
Captura will provide the first homogeneous library of images suitable for AI training. We are working with a Spanish company to develop AI identification of a number of conditions, with Fuchs’ dystrophy as our first target. We have been approached by a number of other AI companies.
OM: Looking ahead, what role do you believe automated imaging platforms and AI will play in the future of ophthalmology, and what lessons from the development of Lightfield Medical’s automated slit lamp could help guide the next generation of ophthalmic technologies?
CB: I see the combination of automated imaging platforms and AI improving a doctor’s productivity and daily experience. AI will initially be able to distinguish normal from abnormal. Then AI will be more specific and tell the doctor: “Hey, look over here.” Ultimately, it will suggest a diagnosis for the doctor. AI won’t replace the doctor, but it will give the doctor superpowers. I think AI will augment doctors, not eliminate them.
We’ve learned from our optical physicists that the best breakthroughs come from stepping back and questioning everything, using Elon Musk’s mantra to “design from first principles.”
OM: Is there anything else you would like to share with the Ophthalmology Management audience?
CB: The road to becoming a physician inventor is long and arduous. Before embarking, talk to as many people as you can. Proceeding is probably not worth it if the market is too small or if the improvement is only on the margins. My first 2 startup ideas were failures; luckily I was able to pivot back to my day job without too much pain.
Once you are sure you have something substantial and revolutionary, get a patent, and then explore The Winning Pitch Challenge. The connections and advice I received there changed my life. OM
References
1. Sivak-Callcott JA, Diaz SR, Ducatman AM, Rosen CL, Nimbarte AD, Sedgeman JA. A survey study of occupational pain and injury in ophthalmic plastic surgeons. Ophthalmic Plast Reconstr Surg. 2011;27(1):28-32. doi:10.1097/IOP.0b013e3181e99cc8
2. Tan NE, Wortz BT, Rosenberg ED, Radcliffe NM, Gupta PK. Digital survey assessment of factors associated with musculoskeletal complaints among US ophthalmologists. Clin Ophthalmol. 2021;15:4865-4874. Published 2021 Dec 30. doi:10.2147/OPTH.S341516
3. Berkowitz ST, Finn AP, Parikh R, Kuriyan AE, Patel S. Ophthalmology workforce projections in the United States, 2020 to 2035. Ophthalmology. 2024;131(2):133-139. doi:10.1016/j.ophtha.2023.09.018
