Establishment and validation of a risk prediction model for high-grade cervical lesions

Abstract

ObjectiveTo establish and validate a risk prediction model for cervical high-grade squamous intraepithelial lesions (HSIL). MethodsThis retrospective study included patients who underwent cervical biopsies at the Cervical Disease Centre of Maternal and Child Hospital of Hubei Province between January 2021 and December 2021. ResultsA total of 1630 patients were divided into the HSIL + cervical lesion group (n = 186) and the ≤ LSIL cervical lesions group (n = 1444). LSIL, ASC-H, HSIL and SCC, high-risk HPV, HPV16, HPV18/45, multiple HPV strains, acetowhite epithelium, atypical vessels, and mosaicity were independently associated with HSIL + lesions. These factors were used to establish a risk prediction model with a demonstrated area under the curve (AUC) of 0.851 and a C-index of 0.829. Calibration curve analysis showed that the model performed well, with a mean absolute error (MAE) of 0.005. The decision curve showed that the model created by combining the risk factors was more specific and sensitive than each predictive variable. ConclusionThe model for predicting HSIL demonstrated promising predictive capability and might help identify patients requiring biopsy and treatment.