From: Machine learning adaptation of intraocular lens power calculation for a patient group
Parameter | Training set mean ± SD (range) | Validation set mean ± SD (range) | P value* |
---|---|---|---|
N, eye/patient | 1211/769 | 400/400 | |
Age (years) | 72.8 ± 8.7 (24–92) | 73.1 ± 9.8 (31–93) | 0.65 |
Axial length (mm) | 23.92 ± 1.54 (20.8–29.8) | 23.92 ± 1.53 (21.1–32.0) | 0.78 |
Anterior chamber depth (mm) | 3.15 ± 0.40 (2.12–4.54) | 3.15 ± 0.40 (2.15–4.06) | 0.97 |
Mean keratometry (D) | 44.35 ± 1.54 (37.3–49.7) | 44.13 ± 1.63 (36.8–48.5) | 0.02 |
Lens thickness (mm) | 4.63 ± 0.44 (3.0–5.8) | 4.63 ± 0.45 (3.3–5.9) | 0.75 |
Central corneal thickness (μm) | 518.4 ± 32.2 (365–634) | 518.8 ± 33.6 (392–606) | 0.84 |
White-to-white width (mm) | 11.7 ± 0.4 (10.2–13.0) | 11.7 ± 0.4 (10.2–12.9) | 0.35 |
IOL power (D) | 20.0 ± 3.9 (6.0–30.0) | 20.2 ± 3.8 (6.0–28.0) | 0.49 |
Postoperative MRSE (D) | − 0.39 ± 0.81 (− 4.88–1.25) | − 0.37 ± 0.86 (− 5.25–1.63) | 0.72 |