Abstract / Summary
Effective triage of high-risk human papillomavirus (hrHPV)-positive women is critical for cervical cancer prevention. This study evaluated and compared PAX1 methylation ( PAX1 m ) and the p16/Ki67 immunohistochemistry dual-stain score (IDS) for risk stratification of high-grade cervical lesions. This retrospective study included 183 women with HPV infections. PAX1 m levels (ΔCp) were quantified, and a three-tier grading system (Grade 1: ΔCp>12.48; Grade 2:8.52~12.48; Grade 3: ΔCp≤8.52) was established based on optimal cut-offs for detecting cervical intraepithelial neoplasia (CIN) grade 2 or worse (CIN2 +) and CIN3+. IDS (0–6) was categorized as 0–2, 3–4, and 5–6. Correlations of PAX1 m Grade and IDS Subgroup with pathological outcomes (≤ CIN1, CIN2, CIN3 +) were analyzed using Kendall's tau-b statistics. Multivariate logistic regression analysis was performed to identify the independent predictors. Both PAX1 m grade and IDS subgroups increased significantly with lesion severity ( P < 0.001). PAX1 m Grade showed a stronger correlation (Kendall's tau-b = 0.454 vs. 0.418) with pathological outcomes than the IDS Subgroup. Multivariable analysis confirmed PAX1 m Grade 3 as a strong independent predictor for CIN3+ (adjusted odds ratio [OR] = 25.3, 95% confidence interval [CI]: 6.9–91.8), demonstrating a larger effect size than IDS Subgroups 3–4 (OR = 8.7) or HPV16/18 positivity (OR = 18.2). The combination of PAX1 m grade and IDS subgroups stratified the population into distinct risk phenotypes. Among women with PAX1 m Grade 3 and IDS 3–6, 55.1% had CIN3+ and 40.8% had CIN2. Among those with PAX1 m Grade 1 and an IDS of 0–4, 67.4% had ≤ CIN1. The intermediate PAX1 m Grade 2 group exhibited heterogeneous distribution across the IDS Subgroups, which the combination helped resolve. In this cohort, the three-tier PAX1 m grading system showed a potential advantage over the IDS in terms of lesion severity. The combination of PAX1 m Grade and IDS Subgroup allowed further stratification of risk, particularly within the intermediate-grade categories. These preliminary proof-of-concept findings support further evaluation of this combined biomarker strategy in prospective externally validated cohorts.