Abstract:AIM: To explore preoperative risk factors affecting residual astigmatism after Toric intraocular lens(IOL)implantation, and construct and verify a clinically applicable prediction model.
METHODS: This retrospective study included patients who underwent Toric IOL implantation at Puyang Oilfield General Hospital and the Second People's Hospital of Puyang City from October 2023 to August 2025. Patients were categorized into two groups according to residual astigmatism at 3 mo postoperatively: >0.25 D group and ≤0.25 D group. Preoperative clinical data and ocular parameters of the two groups were collected and compared. Multivariate Logistic regression analysis was used to screen risk factors, and a nomogram prediction model was established accordingly. Receiver operating characteristic(ROC)curve, calibration curve, and Hosmer-Lemeshow test were adopted to evaluate discrimination, calibration, and clinical applicability of the model. Bootstrap method was applied for internal validation.
RESULTS: Totally 103 patients(103 eyes)were included, with 72 patients(72 eyes)in the group with residual astigmatism ≤0.25 D, including 38 males and 34 females, with an average age of 67.35±6.12 y; There were 31 cases(31 eyes)in group with residual astigmatism >0.25 D, including 19 males and 12 females, with an average age of 69.87±6.45 y. The proportion of patients with a history of diabetes, as well as the values for axial length, horizontal corneal diameter, flat corneal curvature(K1), steep corneal curvature(K2), astigmatic axis and proportion of lens nuclear hardness grade III-IV were significantly higher in the residual astigmatism group >0.25 D than those in the residual astigmatism ≤0.25 D group(all P<0.05). Multivariate analysis identified increased horizontal corneal diameter, higher K1 value, higher K2 value, and higher lens nuclear hardness(grade III-IV)as independent risk factors for postoperative residual astigmatism >0.25 D(all P<0.05). The nomogram constructed from these factors demonstrated that the area under the ROC curve(AUC)was 0.794, with a sensitivity of 80.31% and a specificity of 75.20%. The concordance index(C-index)of internal validation was 0.865. Calibration curve and Hosmer-Lemeshow test(χ2=5.231, P=0.734)indicated good consistency between predicted and actual values of the model.
CONCLUSION: The nomogram model established based on horizontal corneal diameter, K1, K2 and lens nucleus hardness exhibits favorable predictive performance. It can effectively predict the risk of residual astigmatism prior to Toric IOL surgery, providing reference for individualized clinical surgical planning.
