Comparison of two sample collection devices for anal cytology in HIV- positive men who have sex with men: Cytology-brush and Dacron-swab.

2021 
Objective HIV-positive men who have sex with men (MSM) are a vulnerable group for anal cancer (AC), a cancer with a well-described precursor lesion, which can be detected early in screening programs using anal liquid-based cytology (aLBC). We aim to compare two aLBC sample collection devices: cytology brush (CB) and Dacron swab (DS). Methods Retrospective analysis of two consecutive study periods, the first using CB and the second DS. Participants underwent an aLBC, a human papillomavirus (HPV) DNA test and a high-resolution anoscopy (HRA), and a biopsy was performed for suspicious lesions. The sensitivity and specificity of aLBC, area under the receiver operating characteristic (ROC) curve (AUC), and concordance between cytology and HRA were assessed using Cohen's kappa coefficient. Results A total of 239 participants were enrolled (CB group, 120; DS group, 119). aLBC was benign in 46% of samples, and high-grade squamous intraepithelial lesion (HSIL) was detected in 11.7%. Prevalence of biopsy-proven HSIL was 15.3%. No differences in cytological and histological results were observed between the groups. aLBC-HRA concordance was weak for benign results (CB group, k = 0.309; DS group, k = 0.350) as well as for HSIL (k = 0.321 and 0.387, respectively). Sensitivity and specificity were 100% and 51.4%, respectively, in the CB group and 88% and 54.3% in the DS group (AUC = 0.711 and 0.759, respectively, P-value = .514). Representation of the transformation zone (TZ) was adequate in 83.3% of samples in the CB group and 50.4% in the DS group (P-value Conclusion Our data suggest that both devices had similar accuracy to detect anal HSIL, although samples collected with CB are more likely to have an adequate TZ representation, the presence of which could be an indicator of sample quality.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    29
    References
    0
    Citations
    NaN
    KQI
    []