Diogo Lopes MD
The new extended beta version of the Hill Radial basis function (Hill-RBF) IOL calculation method appears to predict postoperative visual acuity with a degree of accuracy that is at least equal to third- and fifth-generation IOL calculation formulas in patents with short eyes, according to a study presented by Diogo Lopes MD, Hospital Garcia de Orta Almada, Portugal.
“This is the first study comparing the new extended beta version of Hill-RBF calculator with third- and fifth-generation formulas in short axial length eyes. The Hill-RBF and Holladay I formulas had the highest percentage of eyes within 0.50D of the target and the lowest median absolute errors,” Dr Lopes told the 22nd ESCRS Winter Meeting in Belgrade, Serbia.
The study involved 37 eyes of 31 cataract patients with axial lengths no greater than 22.0mm. Dr Lopes and his associates compared achieved outcomes with those predicted by four IOL power calculation formulas – Hill-RBF, Hoffer Q, Barrett Universal II and Holladay 1. They obtained biometric data with the IOLMaster 500 in all patients.
The IOL models implanted in the study included the Alcon SA60AT in four eyes, the AMO Sensar AAB00 in five eyes, the AMO Tecnis PCB00 in 12 eyes and the Bausch + Lomb Akreos MI60 in 16 eyes. All patients had undergone uneventful cataract surgery within the previous two years. The researchers calculated refractive prediction with each of the four formulas in all eyes and compared the predicted with the actual refractive outcome to give the refractive prediction error, Dr Lopes explained.
They found that although the Hill-RBF and Holladay formula presented a lower median absolute error (0.34D and 0.33D respectively) than Hoffer Q and Barrett (0.44D and 0.41D respectively), there was no statistically significant differences in the median absolute error between the four formulas (P>0.05).
However, he pointed out that the Hill-RBF and Holladay I formulas were able to achieve the highest percentage of eyes within 0.50D of the target, 70% and 68% respectively. Prediction error ranged from -0.46D to 1.39D with the Hill- RBF method, from -0.84D to 1.33D with the Holladay I formula, from -0.99D to 1.31D with the Barrett calculator and -1.15D to 1.00D in Hoffer Q.
“Although we have not used optimised lens constants and different IOL types, the Hill-RBF method performed at least at the same level as the other three formulas,” Dr Lopes added.
Diogo Lopes: firstname.lastname@example.org