Nishida Shohei, Inomata Yasuya, Hirata Akira
Inomata Eye Clinic, Kumamoto City, Kumamoto, Japan.
Hayashi Eye Hospital, Fukuoka City, Fukuoka, Japan.
Clin Ophthalmol. 2024 Aug 12;18:2253-2259. doi: 10.2147/OPTH.S471393. eCollection 2024.
We examined differences in the accuracy of three intraocular lens (IOL) calculation formulas: the traditional Sanders-Retzlaff-Kraff/Theoretical (SRK/T) formula; the Barrett Universal II (BU II) formula, which is a new-generation IOL calculation formula; and the postoperative spherical equivalent prediction using artificial intelligence and linear algorithms developed by Debellemanière, Gatinel, and Saad formula (PEARL-DGS [PEARL]) formula, and evaluated factors that cause postoperative refractive error (PE).
The study included 205 patients (205 eyes) with a mean age of 75.2 ± 8.7 years who underwent cataract surgery at our institution from December 2018 to October 2023. The PE of the three IOL calculation formulas was calculated and compared. Multivariate logistic regression analysis was performed with a PE higher than ±0.50 D as the dependent variable, and age, sex, axial length (AL), mean keratometry (mean K), anterior chamber depth (ACD), lens thickness (LT), and white-to-white (WTW) as independent variables.
The mean PE (ME) ± standard deviation of the SRK/T, BU II, and PEARL formulas was 0.11 ± 0.52, 0.11 ± 0.50, and 0.21 ± 0.50 D, respectively. MEs of the three IOL calculation formulas were significantly different from 0 (p < 0.01). The median absolute error (MedAE) was not significantly different among the three IOL calculation formulas (p = 0.83). The percentage of PE within ±0.50 D was not significantly different among the three IOL calculation formulas (p = 0.13). Multivariate logistic regression analysis showed that the significantly associated factors with PE higher than ±0.50 D were AL, ACD, and LT for the SRK/T formula, sex and LT for the BU II formula, and LT for the PEARL formula (all p < 0.05).
In the BU II and PEARL formulas, AL was excluded as a factor affecting PE, indicating that LT was a risk factor.
我们研究了三种人工晶状体（IOL）计算公式的准确性差异：传统的桑德斯 - 雷茨拉夫 - 克拉夫/理论（SRK/T）公式；新一代IOL计算公式巴雷特通用II（BU II）公式；以及使用德贝勒马尼耶、加蒂内尔和萨阿德开发的人工智能和线性算法进行术后球镜等效预测（PEARL - DGS [PEARL]）公式，并评估了导致术后屈光不正（PE）的因素。
本研究纳入了205例患者（205只眼），平均年龄为75.2±8.7岁，于2018年12月至2023年10月在我院接受白内障手术。计算并比较了三种IOL计算公式的PE。以PE高于±0.50 D为因变量，年龄、性别、眼轴长度（AL）、平均角膜曲率（平均K）、前房深度（ACD）、晶状体厚度（LT）和白对白（WTW）为自变量进行多因素逻辑回归分析。
SRK/T、BU II和PEARL公式的平均PE（ME）±标准差分别为0.11±0.52、0.11±0.50和0.21±0.50 D。三种IOL计算公式的ME均与0有显著差异（p < 0.01）。三种IOL计算公式的中位数绝对误差（MedAE）无显著差异（p = 0.83）。三种IOL计算公式中PE在±0.50 D以内的百分比无显著差异（p = 0.13）。多因素逻辑回归分析显示，SRK/T公式中与PE高于±0.50 D显著相关因素为AL、ACD和LT，BU II公式为性别和LT，PEARL公式为LT（均p < 0.05）。
在BU II和PEARL公式中，AL被排除为影响PE的因素，表明LT是一个危险因素。
