Comparison of IOL position algorithms for prediction of postoperative IOL position (mLP) in terms of preoperatively available information on the subset of our data that included data on LT
| Formula origin | IOL model | Formula | Offset | SD |
|---|---|---|---|---|
| Olsen and Hoffmann20 | ZA9003 N=20 | ACD+0.45×LT | 0 | 0.18 |
| Preussner et al23 | CT+4.60×(AL/23.6)0.7 | 0.17 | 0.22 | |
| Olsen19 | 4.98+0.19×AL+0.49×ACD+0.28×LT−0.41×CR+0.028×Rx | 0 | 0.18 | |
| This paper | 3.93+0.40×ACD | 0 | 0.17 | |
| Olsen and Hoffmann20 | ZCB00 N=27 | ACD+0.47×LT | 0 | 0.31 |
| Preussner et al23 | CT+4.60×(AL/23.6)0.7 | −0.06 | 0.23 | |
| Olsen19 | 5.13+0.19×AL+0.49×ACD+0.28×LT−0.41×CR+0.028×Rx | 0 | 0.22 | |
| This paper | 3.55+0.46×ACD | 0 | 0.18 |
In the formulas constants meant to be adjustable were adjusted by MLR regression (there is no issue with collinearity) to give zero offset. In the formula by Preussner et al, the constants are not meant to be adjustable, leading to non-zero offset.
ACD, anterior chamber depth; CR, corneal radius; CT, corneal thickness; IOL, intraocular lens; LT, lens thickness; mLP, middle lens position; MLR, multiple linear regression.