Evaluating the Predictive Performance of miR-124-2 and FAM19A4 for Cervical Lesions in a Single Center from Romania: A Prospective Study

<b>Background/Objectives</b>: Molecular triage strategies for cervical lesions based on miR-124-2 and FAM19A4 are poorly studied in various populations. The aim of this prospective study was to evaluate the individual and combined predictive performance of these two markers for the prediction of various histological categories of cervical lesions. <b>Methods</b>: An FAM19A4 and miR124-2 methylation analysis was performed on 70 samples from patients negative for intraepithelial lesion or malignancy (NILM), cervical intraepithelial neoplasia (CIN)1-3 and squamous cervical carcinoma (SCC), along with human papillomavirus (HPV) genotyping and cytological and histopathological assessment. Descriptive statistics examined clinical associations, while sensitivity analysis evaluated the predictive performance of these markers individually and combined. <b>Results</b>: The sensitivity of miR-124-2 was 28.1%, while its specificity was 86.8% for SCC. The ROC values ranged between 0.25 and 0.62 for the evaluated histological categories, suggesting a poor to moderate predictive performance. FAM19A4 had a sensitivity of 36% for predicting CIN3 and SCC, as well as a high specificity for CIN3 and SCC (88.9%), with ROC values between 0.35 and 0.73 for the evaluated histological categories. The combined tests improved the PPV for higher-risk lesions (CIN3, SCC), but did not significantly improve the ROC values. FAM19A4 achieved the best performance for the prediction of CIN2+ (ROC: 0.64) and CIN3+ lesions (ROC: 0.73). <b>Conclusions</b>: We hypothesize that while not suitable as stand-alone diagnostic tools, such biomarkers may aid in stratifying patients and optimizing referral decisions, pending further validation in larger, population-based cohorts.