A Competing-Risks Approach to the Progression, Regression and Persistence of High-Grade Cervical Dysplasia in Patients over 30 Years Old-A Prospective Study.

Background/Objectives: In Romania, where cervical cancer incidence remains among the highest in the European Union, a risk-based management strategy could support more precise allocation of limited resources. The aim of this study was to test the prognostic utility of immediate pre-treatment and post-treatment risk predictions, derived from the American Society of Colposcopy and Cervical Pathology (ASCCP) risk-based management guidelines for the prediction of progression, regression or persistence of high-grade cervical dysplasia. Methods: In this prospective cohort study, we included 223 patients aged over 30 years who underwent self-referred or targeted screening with or without histologically confirmed cervical intraepithelial neoplasia (CIN) of any grade. We employed Fine and Gray's subdistribution hazard model that evaluated the cumulative incidence function for each specific outcome, treating other outcomes as competing events. These outcomes were further stratified depending on the type of high-grade dysplasia. Results: The immediate post-treatment risk was significantly associated with subsequent progression of cervical dysplasia. For a cut-off of 60%, the immediate post-treatment risk was able to significantly predict the progression of both CIN2+ and CIN3+. On the other hand, the immediate pre-treatment risk > 60% was significantly associated with progression of CIN3+, but not of CIN2+. Also, the immediate pre-treatment risk was significantly associated with regression, but this observation did not persist at the >60% threshold. Both pre- and post-treatment risk > 60% were strongly associated with persistence across histologic subgroups. Conclusions: The ASCCP-derived immediate risk estimates, especially post-treatment risk > 60%, proved effective in predicting progression and persistence of high-grade cervical dysplasia.