Construction and Validation of a Model for Predicting Cervical Intraepithelial Neoplasia Grade II+: A Cross-Sectional...

Area of Science:

• Gynecology
• Oncology
• Medical Informatics

Background:

• Cervical cancer is a leading global malignancy in women.
• Early screening is crucial, but effective prediction models for cervical lesions are lacking.
• Accurate prediction of cervical intraepithelial neoplasia grade II or higher (CINII+) is vital for timely intervention.

Purpose of the Study:

• To develop a predictive model for CINII+ using machine learning.
• To compare the efficacy of models integrating ThinPrep cytologic test (TCT) + human papillomavirus (HPV) testing with clinical data versus TCT + traditional clinical data.
• To evaluate the predictive performance of different machine learning models for CINII+ detection.

Main Methods:

• Collected clinical data from women undergoing cervical cancer screening (2020-2024).
• Applied ten machine learning algorithms to build two models: Model 1 (TCT+HPV+clinical data) and Model 2 (TCT+traditional clinical data).
• Evaluated model performance using AUC, calibration curves, and decision curve analysis.

Main Results:

• HPV positivity, TCT indicating HSIL, colposcopy results, and early age of pregnancy were identified as predictors of CINII+.
• Model 1 (TCT+HPV+clinical data) showed significantly higher predictive efficacy than Model 2 (TCT+clinical data).
• The difference in AUC between the models was statistically significant (P=0.006 in training, P=0.035 in testing).

Conclusions:

• The TCT+HPV-integrated model demonstrated superior performance in predicting CINII+ compared to the TCT-only model.
• Incorporating HPV testing into routine screening enhances early diagnostic accuracy for cervical lesions.
• This study supports the integration of HPV testing for improved cervical cancer screening outcomes.