Risk evaluation for the development of cervical intraepithelial neoplasia: development and validation of risk-scoring schemes.

Cervical cancer screening guidelines do not comprehensively define what constitutes high risk. This study developed and validated simple risk-scoring schemes to improve Papanicolaou smear screening for women at high risk. Four cumulative risk score (CRS) schemes were derived respectively for the development of cervical intraepithelial neoplasia grade 1 (CIN1) and grade 2 or worse (CIN2+) using community-based case-control data (n = 1523). By calculating the area under the receiver operating characteristic (AU-ROC) curve, these schemes were validated in a Papanicolaou smear follow-up cohort (n = 967) and a hospital-based cytology screening population (n = 217). A high DNA load of high-risk human papillomavirus (HR-HPV) was the main predictor for CIN1 and CIN2+, although age, married status combined with the number of sexual partners, active and passive smoking and age at sexual debut also affected associated lesions. In the training set, only the HPV-testing-contained CIN2+ CRS scheme presented an excellent discrimination for identifying CIN2+ (AU-ROC = 0.866). Using a CRS cutoff value of 4 to identify CIN2+, the sensitivity and specificity of predicting CIN2+ for the 3- and 5-year follow-ups were 100% and 90.8%, and 83.3% and 90.4%, respectively, in the validation cohort. In the hospital-based validation population, the CRS scheme showed comparable discrimination for CIN2+ detection (sensitivity 88.2% and specificity 84.6%). Women with CRS ≥ 4 had a 5.4% and 9.1% of 3- and 5-year cumulative incidence, respectively, and a 40.5-fold hazard ratio of developing CIN2+. In conclusion, combined with HR-HPV testing and verified risk factors, a simple CRS scheme could effectively improve the implementation of CIN2+ screening.