通过探索 IOL 和角膜屈光度对屈光影响的影响，提供一种简化的方法来优化晶状体常数，使人工晶状体 (IOL) 计算公式的平均预测误差 (ME) 为零，而无需对公式本身进行编程有效镜头位置的变化。

优化公式的理论发展和对记录数据集的回顾性临床评估。

对 8878 名白内障患者的回顾性数据进行了检查，使用 4 个 IOL 模型和 6 个 IOL 度数计算公式进行术前和术后测量。使用示意性眼模型来研究有效晶状体位置（ELP）的微小变化对术后等效球面（SE）屈光的影响。研究了角膜曲率 (K) 和 IOL 度数 (P) 对 SE 的影响。使用理论厚晶状体模型设计了一个公式，将 IOL 度数计算公式的平均预测误差归零。这是通过增加预测的 ELP 来实现的，然后可以将其转换为 IOL 常数的增量。该方法使用不同的 IOL 模型和单常数优化 IOL 计算公式在记录的现实术后数据集上进行了测试。

对于 ELP 的微小变化，IOL 屈光力与术后屈光变化之间存在指数关系。将 ME 归零所需的 ELP 调整等于 ME 与以下表达式的平均值之间的比率：0。0006*（磷2+2K*磷）在所考虑的数据集上。使用该公式获得的值的准确性已在记录的术后数据集以及已发表和未发表的公式中得到证实。

所提出的方法允许没有特殊专业知识的外科医生优化 IOL 常数，从而使记录数据集上的 ME 无效，而无需编码不同的公式。个体眼睛的影响与植入的人工晶状体的平方度数成正比。

"点击查看英文标题和摘要"

PURPOSE

To provide a simplified method to optimize lens constants to zero the mean prediction error (ME) of an intraocular lens (IOL) calculation formula, without the need to program the formula itself, by exploring the influence of IOL and corneal power on the refractive impact of variations in effective lens position.

Theoretical development of an optimized formula and retrospective clinical evaluation on documented datasets.

Retrospective data from 8878 patients with cataracts with pre- and postoperative measurements available using 4 IOL models and 6 IOL power calculation formulas were examined. A schematic eye model was used to study the impact of small variations in effective lens position (ELP) on the postoperative spherical equivalent (SE) refraction. The impact of keratometry (K) and IOL power (P) on SE was investigated. A theoretical thick lens model was used to devise a formula to zero the average prediction error of an IOL power calculation formula. This was achieved by incrementing the predicted ELP, which could then be translated into an increment in the IOL constant. This method was tested on documented real-life postoperative datasets, using different IOL models and single-constant optimized IOL calculation formulas.

For small variations in ELP, there was an exponential relationship between IOL power and the resultant postoperative refractive variation. The ELP adjustment necessary to zero the ME equated to a ratio between the ME and the mean of the following expression: 0.0006*(P2+2K*P) on the considered datasets. The accuracy of the values obtained using this formula was confirmed on documented postoperative datasets, and on published and nonpublished formulas.

The proposed method allows surgeons without special expertise to optimize an IOL constant to nullify the ME on a documented dataset without coding the different formulas. The influence of individual eyes is proportional to the squared power of the implanted IOL.