Eyeworld

EW FEATURE 62 IOL calculations February 2015 by Michelle Dalton EyeWorld Contributing Writer Physicians offer their thoughts on potential formulas S ince the introduction of the IOL, researchers and clinicians alike have been trying to determine the best formula to use to im- plant the best possible lens for a par- ticular patient, based on numerous characteristics that include refractive error, axial length, ocular comorbidi- ties, and previous surgery. Improved conditions for accurate IOL power calculation include the advent of small incision surgery, corneal scan- ning devices, IOL manufacturing improvements, laser biometry, and enhanced capsulorhexis techniques. Determining which of the many potential formulas is best depends on all the above, experts say. "All formulas are still far from perfect, as the rate of eyes within 0.5 D of the refractive target is only about 75%," said Giacomo Savini, MD, G.B. Bietti Eye Foundation, Rome, Italy. With one-quarter of patients left with a suboptimal result, "I am investigating the role of corne- al asphericity and lens thickness even in thin lens third-generation formulas," he said, adding that he presented some data on the topic at the 2014 ASCRS•ASOA Symposium & Congress. 1 IOL power calculation formulas started as theoretical, with the Fyodorov, Colenbrander, and Binkhorst formulas in the late 60s and early-to-mid 70s, said Thomas Olsen, MD, University Eye Clinic, Aarhus, Denmark. The fourth- generation formulas used today include the Holladay 2, Preussner, and Olsen, he said. Dr. Olsen is the author of the Olsen formula and the originator of the PhacoOptics PC software (IOL Innovations, Aarhus, Denmark) using the newly described C-constant approach for the predic- tion of the IOL position. 'Normal' eyes Healthy eyes (those without any comorbidities or previous surgery) can have fairly accurate readings from the third-generation formulas, including Hoffer Q, Holladay 1, and SRK/T, Dr. Savini said. "I also rely on the Haigis formu- la, although I find it difficult to op- timize its constants by myself. I am interested in investigating whether the C-concept by Dr. Olsen will car- ry any advantage, but until evidence of any advantage from different studies is available, I will go on with the standard formulas," he said. A modern IOL power calcula- tion formula should fulfill a num- ber of requirements in order to be safe and effective, Dr. Olsen said, including that it should be accurate in "normal" eyes. "In fact, almost any formula would work within the 'normal' range by establishing average values for the outcome based on the clin- ical environment," he said. For ex- ample, the SRK I formula used "little optical modeling" and was effective in the normal range, but "not so in the extreme range." Noting that modern diagnostic instruments can provide signifi- cantly more data to help determine which lens would be best for an in- dividual patient also means the term "normal" may need to be explored further, Dr. Olsen said. "Do we limit it to normal axial length? Normal K-reading? Normal anterior segment? What about the posterior curvature of the cornea?" he said. "The challenge is to take advantage of the more detailed in- formation in our clinical practice." Dr. Savini agreed, noting that differing axial lengths require different formulas. "For very short eyes I use the Hoffer Q formula, which has been shown to be the best in these cases by many studies," Dr. Savini said. "For long eyes, I rely on the SRK/T for the same reason. In long eyes, I also follow the approach suggested a few years ago by Prof. Haigis, who dramatically changes the constants of the IOLs." He does not use the Holladay 2 formula as "it does not offer any ad- vantage and it requires many more data [and time] points compared IOL power calculations and biometry AT A GLANCE • The Fyodorov, Colenbrander, and Binkhorst formulas were among the first, created in the late 60s and early-to-mid 70s. • Differing axial lengths require different formulas. • Fourth-generation formulas used today include the Holladay 2, Preussner, and Olsen. Ray tracing software allows accurate no-history IOL power calculation even in cases like decentered radial keratotomy with residual corneal astigmatism. Source: Giacomo Savini, MD to the others," he said. "I am also awaiting the results of the Hoffer H5 formula, which considers gender and race, but this will still take some time." Non-virgin eyes Patients who present post-laser vision surgery, either LASIK or PRK, pose a different challenge, Dr. Savini said. Data he presented at the 2014 ASCRS•ASOA Symposium & Con- gress indicated the Savini method combined with double-K SRK/T, the Seitz/Speicher method combined with the same double-K SRK/T, the Masket method with SRK/T, and the Shammas no history method were the most accurate, he said. 1 "These methods allow me to have the same accuracy as with un- operated eyes, i.e., more than 70% of cases within 0.5 D of intended refraction," Dr. Savini said. Dr. Savini does not change his formula of choice when other mor- bidities such as corneal disorders, diabetes, or glaucoma are present. Ray tracing The advantage of ray tracing is that it only requires Snell's law and needs no other assumptions than the physical shape and refractive index of the material, Dr. Olsen said.

• Cover