EyeWorld Asia-Pacific December 2018 Issue

64 EWAP DEVICES December 2018 by Michelle Stephenson EyeWorld Contributing Writer New methods for IOL calculations S everal new IOL calcu- lation formulas are gaining popularity. According to Doug- las Koch, MD , Houston, formu- las can be divided into four cat- egories: vergence formulas, which can be subclassified according to the number of variables they take into account; ray-tracing formu- las; artificial intelligence formu- las; and regression formulas. “In addition to superb new vergence formulas, particularly the Barrett Universal II, two ap- proaches that are getting a lot of attention now are the ray-tracing formulas and the artificial intel- ligence approaches,” Dr. Koch said. “Ultimately, I think the ray- tracing formulas will be what we end up using, but they have yet to demonstrate superior outcomes, primarily because they don’t have a better way of calculating effec- tive lens position than any of the other approaches.” Ray-tracing formulas are based on pure optics, and they take into account such factors as irregular astigmatism of the cor- nea, cornea asphericity, and IOL parameters, such as anterior and posterior curvature and spherical aberration. “Theoretically, that would give them an advantage and would give them a universal appeal because they could be used in normal eyes but also in eyes with complex corneas that make vergence approaches theoretically more difficult,” he said. Artificial intelligence ap- proaches, like Radial Basis Func- tion (RBF) from Warren Hill, MD , use big data. “This is an ingenious method of generating internal data relationships that look at different combinations of data and, by doing that, can slot where any given patient’s eye might fall. Using just three key data points—corneal curvature, anterior chamber depth, and axial length— one can determine for that set of measurements the optimal IOL power,” Dr. Koch explained. According to Elizabeth Yeu, MD , Norfolk, Virginia, the Barrett Universal formula has improved her outcomes, particularly for eyes shorter than 22 mm and eyes longer than 25.5 mm. “In very long eyes, the Barrett Universal performs on its own extremely well. I think that formula has become better over time. While it’s not artificial intelligence, it is refined based on results, so there is some level of artificial tweaking that has occurred since its intro- duction years ago,” she said. She noted that the Hill-RBF and the Barrett Universal are her two go-to formulas. “We’ve had a little bit of a sneak peek because Dr. Hill incorporates our patient outcomes information as part of his data systems used to refine the RBF formula. We had a first user’s look at it to see if it was further refined in its results as compared to the original RBF, and it has proven to lead to better outcomes,” she said. IOL calculation preferences For standard cases, Dr. Koch uses the Barrett Universal II, the Hill- RBF, and the Holladay 1. Dr. Yeu agreed. “For average axial length, I use Holladay 1 because I still love it. I also use Hill-RBF and Barrett Universal,” she said. For high myopes, Dr. Koch has been using the Holladay 1 with the Wang-Koch modifica- tion. “However, I think that the Barrett and the Hill- RBF are also doing a good job now with the long eyes. For the short eyes, we’re still stymied. We published a paper that showed roughly 70–75% accuracy within 0.5 D for short eyes, and none of the for- mulas stood out. 1 In fact, the old Holladay 1 did as well as, if not better than, any of the other for- mulas. So the short eyes remain a challenging niche that we have not sorted out yet,” he said. Dr. Yeu said the Hill-RBF and the Barrett Universal are compara- ble for very high myopes with an axial length of more than 27 mm. “For eyes shorter than 22 mm, the Barrett Universal, in my hands, has been the most accurate, with the Hill-RBF as a backup to make sure that it doesn’t deviate too