Construction and evaluation of Nomogram prediction model for postoperative recurrence of pterygium

  Objective  To analyze the risk factors of postoperative recurrence of pterygium patients in Gannan regions inhabited by ethnic groups, and to construct and evaluate the Nomogram prediction model based on results.

  Methods  A retrospective study was performed for 296 cases (296 eyes) of patients with excision of pterygium in Gannan regions inhabited by ethnic groups in the way of random cluster sampling, and the patients were divided into recurrence group and non-recurrence group according to recurrence status at one year after operation. Univariate and multivariate Logistic regression models were used to analyze the risk factors of recurrence of pterygium. On this basis, a Nomogram prediction model was constructed, and receiver operating characteristic (ROC) curve and H-L fitting curve were drawn to evaluate the discrimination and effectiveness of the model.

  Results  A total of 296 patients were included in this study, of which 45 recurred one year after operation, and the recurrence rate was 15.20%. The results of multivariate Logistic regression model showed that the longer daytime outdoor activities, short distance between the medial rectus muscle stop and the corneal limbus, short tear film rupture time and xerophthalmia were the independent risk factors affecting postoperative recurrence of patients with pterygium in the Gannan regionsinhabited by ethnic groups (P < 0.05). The area under the curve of ROC curve was 0.901, and the sensitivity and specificity were 86.67% and 88.84%, respectively (95%CI, 0.826 to 0.924); the H-L fitting curve showed there was a good fitting result (χ²=6.421, P=0.431).

  Conclusion  Longer daytime outdoor activities, short distance between the medial rectus muscle stop and the corneal limbus, short tear film rupture time and xerophthalmia are the independent risk factors affecting postoperative recurrence of patients with pterygium in the Gannan regions inhabited by ethnic groups. The Nomogram prediction model constructed on these results has good differentiation and effectiveness, and it can be used as an effective tool for clinical evaluation of postoperative recurrence risk of pterygium patients in Gannan regions inhabited by ethnic groups.