Development of a nomogram for predicting postoperative recurrence of cervical intraepithelial neoplasia using immunohistochemical and clinical parameters.

BACKGROUND

We aimed to develop a nomogram to predict abnormal follow-up results of co-testing for cytology and human papillomavirus (HPV) in cervical intraepithelial neoplasia (CIN) patients after conization.

RESEARCH DESIGN AND

METHODS

Two hundred sixty-three patients initially diagnosed as CIN2+ were recruited. Data on immunohistochemical (IHC) staining scores, along with demographic and clinical information were collected. Using least absolute shrinkage and selection operator (LASSO) regression analysis, variables were identified for inclusion. A predict model and nomogram were developed through multi-factor logistic regression. The goodness-of-fit test was applied across different cohorts to construct the calibration curve of the model, and the predictive effect was evaluated by the receiver operating characteristic curve. Decision curve analysis was performed to determine the net benefit.

RESULTS

Five predictor variables, including protein expression score, vaginal infection, HPV coinfection, and cone height were screened and plotted as a nomogram. The calibration curves showed a good fit. The area under the curve of the model was 0.835 for the training cohort and 0.728 for the internal test cohort. The decision curve analysis indicated that the nomogram provides significant net advantages for clinical use.

CONCLUSION

A practical nomogram predict model was developed to predict abnormal follow-up outcomes in CINs after conization.