Machine Learning–based Prediction of LASIK Console Inputs for Aspheric Planning (Q-factor, Defocus, Astigmatism): A Translational Methods Study

Aspheric planning in laser refractive surgery remains difficult: surgeons often rely on empirical nomo-grams or simple linear regression for defocus and astigmatism, while console Q-factor modulation yields a variably predictable effect on asphericity and an inconsistent cross-effect on defocus. This translational methods proof-of-concept frames planning as supervised prediction of console-programmable inputs (de-focus, astigmatism, Q-factor) and evaluates competing models; it is not a clinical effectiveness study. We analyzed an anonymized, retrospective, single-platform dataset of 2,448 complete-case treatments. Multi-output regressors (linear and nonlinear) were trained and compared using prespecified metrics (R2, MAE/MSE) and residual-distribution visualization/calibration. Actuator–response checks related programmed inputs to changes in defocus (Z20) and primary spherical aberration (Z40). External validation used a temporally later, device-shift cohort (n=147). Linear regression predicted defocus and astigmatism well (e.g., defocus R2=0.98) but degraded for Q-factor (R2=0.47), whereas nonlinear models improved Q-factor error and calibration. Actuator–response analyses showed strong coupling for defocus input (R2=0.97), moderate coupling of Q-factor to ΔZ40 (R2=0.51), and a weak Q→defocus cross-effect (R2=0.12). On external validation, the best model generalized: Defocus MAE 0.22 D (R2=0.98) and Q-factor MAE 0.21 (R2=0.81). Supervised nonlinear multi-output models achieve lower error and better calibration for Q-factor than linear baselines, supporting a metric-driven pathway toward more reliable control of low-order refractive targets and primary asphericity. Potential clinical implications include tissue sparing, improved contrast, and near-vision gains. Prospective, human-in-the-loop evaluation with safety and patient-reported endpoints is warranted.