Estimation of intraocular lens position from full crystalline lens geometry: towards a new generation of intraocular lens power calculation formulas

Abstract

In a cataract surgery, the opacified crystalline lens is replaced by an artificial intraocular lens (IOL). To optimize the visual quality after surgery, the intraocular lens to be implanted must be selected preoperatively for every individual patient. Different generations of formulas have been proposed for selecting the intraocular lens dioptric power as a function of its estimated postoperative position. However, very few formulas include crystalline lens information, in most cases only one-dimensional. The present study proposes a new formula to preoperatively estimate the postoperative IOL position (ELP) based on information of the 3-dimensional full shape of the crystalline lens, obtained from quantitative eye anterior segment optical coherence tomography imaging. Real patients were measured before and after cataract surgery (IOL implantation). The IOL position and the postoperative refraction estimation errors were calculated by subtracting the preoperative estimations from the actual values measured after surgery. The proposed ELP formula produced lower estimation errors for both parameters -ELP and refraction- than the predictions obtained with standard state-of-the-art methods, and opens new avenues to the development of new generation IOL power calculation formulas that improve refractive and visual outcomes.

Figure 1

Raw optical coherence tomography (OCT) images for subject S#1 OD, including the definition of some biometric parameters. (a) Preoperative measurement. (b) Postoperative measurement. ACD_pre = anterior chamber depth, EPP = equatorial plane position, CT = corneal thickness.

Figure 2

3-D models for S#1 (OD). (a) Preoperative measurement, including the part of the crystalline lens visible through the pupil (green) and its full shape estimation. (b) Postoperative measurement, including the intraocular lens (IOL) in purple. (c) Both models superimposed. IOL = intraocular lens.

Figure 3

Graphical comparison of 3-D models for a thin lens (S#1 OD) and a thick lens (S#3 OD) using the real geometries obtained from the analysis. Cornea was used for registration.

Figure 4

Linear regression between EPP/(LT·ACD_pre) and ELP/ACD_pre. r = 0.98, p = 3.6·10⁻⁹; $\frac{\mathbf{LP}}{\mathbf{AC}{\mathbf{D}}_{\mathbf{pre}}}=\mathbf{0}\mathbf{.250}\mathbf{(}\mathbf{\pm }\mathbf{0}\mathbf{.07}\mathbf{)}$ + $\mathbf{8}\mathbf{.497}\mathbf{(}\mathbf{\pm }\mathbf{0}\mathbf{.43}\mathbf{)}\cdot \frac{\mathbf{E}\mathbf{P}\mathbf{P}}{\mathbf{(}\mathbf{L}\mathbf{T}\mathbf{\cdot }\mathbf{A}\mathbf{C}{\mathbf{D}}_{\mathbf{p}\mathbf{r}\mathbf{e}}\mathbf{)}}\mathbf{.}$ EPP = equatorial plane position, LT = lens thickness, ACD_pre = anterior chamber depth preoperative, ELP = estimated lens position, LP = lens position.

Figure 5

Mean of the absolute value of the estimation error (MAE, color bars) and of the refractive error (MARE, black dotted line) across the 12 eyes, for the compared estimation methods.