Practical model for residual/recurrent cervical intraepithelial lesions in patients with negative margins after cold-knife conization

This study aimed to identify reliable risk factors for residual/recurrent cervical intraepithelial lesions in patients with negative margins after cold-knife conization. 2352 women with HSIL (high-grade squamous intraepithelial lesions) were treated with cold-knife conization with a negative margin from January 2014 to December 2020 and followed up until December 2021. Univariate analysis was used to select discriminative clinical features. Multivariate logistic regression was used to built four predictive models based on the different combination of follow-up data (Model A: preoperative factors; Model B: first follow-up data; Model C: second follow-up data; Model D: both follow-up data). The accuracy, sensitivity, specificity, false-positive rate (FPR), false-negative rate (FNR), and area under the receiver operating characteristic curve (AUC) were evaluated on the validation cohort. The predictive power of risk factors was further validated using six machine learning algorithms. Model D demonstrated the highest AUC 0.91 (95% CI 0.87 to 0.96) in the validation cohort, while Model A-C achieved AUCs of 0.69 (95% CI 0.59 to 0.78), 0.88 (95% CI 0.80 to 0.95) and 0.89 (95% CI 0.81 to 0.97) respectively. Six machine learning methods achieved consistent results. Kaplan-Meier (KM) survival curves demonstrated that our models can effectively stratify patients with all models (p < 0.05 for all models). The identification model can serve as a complementary screening procedure for early detection or prediction of recurrence after cold-knife conization in HSIL patients.