Abstract
Objectives:
To compare diagnostic strategies for detecting HSIL or worse (HSIL+) in women with high-grade cytology and negative colposcopy, and to assess whether sequential HPV testing, liquid-based cytology (LBC), and endocervical sampling improve diagnostic safety.
Methods:
This is a multicenter randomized clinical trial conducted at 4 referral centers between 2021 and 2022. Women with ASC-H or worse cytology and negative colposcopy underwent HPV testing and LBC. Participants were randomized to endocervical brushing (EB) or endocervical curettage (ECC). The reference standard was histopathology from excisional procedures or structured follow-up. Diagnostic performance (sensitivity, specificity, predictive values, and accuracy) was calculated for individual tests and predefined combinations (AND ALL and MAJORITY rules). Absolute differences between strategies were analyzed. Invalid results were classified as failures (false negatives or false positives) according to the final diagnosis.
Results:
Among 257 women, 35 cases of HSIL+ were identified. HPV testing showed the highest sensitivity (94.3%) and negative predictive value (98.6%), whereas LBC demonstrated higher specificity (84.9%) and accuracy (84.1%). The combination of HPV and LBC (AND rule) improved specificity to 91.7% with a sensitivity of 77.1%. Adding EB (MAJORITY rule) increased sensitivity to 100% (absolute increase 22.9%) but reduced specificity to 82.3%. ECC showed lower sensitivity (61.1%) and higher inadequacy rates (21.3%).
Conclusions:
In women with high-grade cytology and negative colposcopy, HPV testing followed by LBC provides a balanced diagnostic strategy. The addition of endocervical brushing may increase sensitivity in selected high-risk cases and improve the detection of HSIL+ after negative colposcopy.