Applicability Of Iol Calculation Formulas Designed For Lasik To Smile Surgery: A Comparative Study

Being the predominant refractive surgeries in China, a considerable cohort of individuals has experienced Femtosecond Laser-assisted In Situ Keratomileusis (FS-LASIK) and Small Incision Lenticule Extraction (SMILE) procedures.
Despite their prevalence, there exists a notable gap in understanding optimal intraocular lens (IOL) power calculation formulas for patients with a prior history of FS-LASIK or SMILE surgeries.

Tianjin Key Laboratory of Ophthalmology and Visual Science, Tianjin Eye Institute, Tianjin Eye Hospital, Tianjin, China

A total of 196 eyes from 100 postoperative refractive surgery patients were included in the study, with 50 individuals (96 eyes) having undergone FS-LASIK and 50 individuals (100 eyes) having undergone SMILE.
By using a theoretical surgical model, we conducted extensive preoperative and postoperative measurements, including demographic data, corneal biometric parameters, and axial length.
Various formulas, including the Barrett Universal II (BUII) formula, as a baseline, were employed to calculate IOL power for the patients.

The Barrett True K (BTK) formula, demonstrated an mean absolute error (AE) within 0.5 D for both FS-LASIK and SMILE groups (0.29±0.36 D and 0.36±0.54 D, respectively).
Notably, the FS-LASIK group showed 89% of results differing by less than 0.25 D compared to preoperative BUII results. The Barrett True K No History (BTKNH) formula, which also incorporates measured posterior corneal curvature, performed similarly to BTK in both groups.
Additionally, the Masket formula, relying on refractive changes based on empirical experience, displayed promising potential for IOL calculations in SMILE patients compared with BTK (p=0.329).

Our study suggests that, in individuals who have undergone SMILE surgery, employing EVO-TK, BTK-K, and Masket-K formulas may lead to more precise calculation outcomes. Presently, AI-based formulas do not exhibit notable advantages over traditional formulas.
Nevertheless, integrating historical data into these formulas has the potential to improve their overall performance.