The ESCRS Symposium focused on innovations in the digital cataract and refractive surgery space
Earlier this year, the European Society of Cataract and Refractive Surgeons (ESCRS) hosted its annual meeting in Vienna, Austria. Leaders from that congress made their way to the United States for the American Academy of Ophthalmology (AAO) meeting in San Francisco, California.
During the refractive surgery subspecialty day on Friday, 3 November, participants were able to attend the ESCRS symposium online or in-person. Oliver Findl, MD, president of the ESCRS, and Burkhard Dick, MD, secretary of the ESCRS, moderated the event; they also served as panelists alongside Thomas Kohnen MD, PhD, FEBO andRudy Nuijts, MD. Their discussion addressed innovations that are rapidly changing the surgical landscape for experts in the cataract and refractive space, from artificial intelligence (AI) to machine learning and beyond.
“We have to make sure to responsibly apply these tools – and at the same time to ensure that (unlike in the Terminator series) the machines do not take over,” said Dr Dick. “In medicine, humans will always care for humans.” He said that, in the cataract and refractive surgeon's operating room, these new technologies seem to offer numerous advantages, but their potential should be discussed with a vital caveat: “as it appears now.” To illustrate this, he pointed to AI and the technologies that can be used in conjunction with it.
“As ESCRS President Oliver Findl has rightfully stated, the only thing spreading faster than artificial intelligence application development is the hype surrounding it,” Dr Dick stated. AI presents challenges for professionals in many sectors, so it’s important to approach its application with case-specific problems in mind. While ophthalmologists may not encounter copyright obstacles when using AI, as some in the editorial space have, they may instead face inconsistent outcomes among patient groups, as in recent research about machine learning and glaucoma.
However, Dr Dick continued, AI applications do have the potential “to make our interventions even safer, even more effective than they already are by acting as an additional eye, and an additional memory, for the surgeon.” Biomorphometric detection can serve as one example. “Before starting surgery, the best available information about the individual patient's risk for intra- and perioperative complications is necessary and thus the best possible data flow is desirable,” he said. Physicians can look to surgery-specific models that can provide perioperative risk stratification, for complications in specific procedures as well as risk factors for individual patients. “Recent studies have shown that AI models for perioperative risk stratification have excellent performance in evaluating the risk of postoperative complications,” he noted.
Biomorphometric detection can play a decisive role for some patients, Dr Dick said. Toric intraocular lens (IOL) alignment is one procedure where an AI-enhanced practice can provide significant benefit. “The exact positioning of the toric IOL in relation to the target axis in a patient with visually significant preoperative astigmatism is crucial. With a deviation by 30 degrees, the astigmatism-correcting effect of the lens is gone,” Dr Dick said. Cyclorotation, which occurs when a patient moves from an upright position to a supine position, can be a major factor in this deviation. “An image-guided system like the Callisto eye Z Align (Carl Zeiss Meditec) superimposes the target axis into the surgical microscope while scleral and limbal blood vessels are matched with a preoperatively captured image. Another system, IntelliAxis-L (Lensar, Inc.) creates capsulotomy markings with a femtosecond laser and compensates for cyclorotation by matching a preopratively captured image of the iris with the intraoperative image through iris registration,” he said. These technologies can significantly reduce misalignment – in the case of the femtosecond laser-assisted capsular marking, Dr Dick reported, the mean misalignment was just 1.71 degrees.
Next-generation innovations in the cataract and refractive space go beyond the existing applications of AI. Commenting on his presentation, which focused on machine learning, Dr Nuijts showcased how new technologies can better leverage existing collections of data amongst physicians and researchers. “Machine learning can help estimate the probability of posterior capsule rupture more reliably and objectively,” Dr Nuijts stated. One of the landmark ESCRS projects, the European Registry of Quality Outcomes for Cataract and Refractive Surgery (EUREQUO), serves to further innovation in this space. According to Dr Nuijts, “Probabilistic classifiers constructed on data of the EUREQUO appear to estimate the probability of posterior capsule rupture more reliably than existing scoring systems in the literature.”
These machine learning-enabled methods have distinct advantages over traditional tools, which create room for marked improvements in patient satisfaction. According to Dr Nuijts, machine learning can also help physicians stay ahead of possible surgical complications or adverse effects when there are multiple ocular comorbidities at play. “It is important to consider interactions between risk factors when estimating the probability of posterior capsule rupture,” he said.