[HPV primary test in the cervical cancer screening: reproducibility assessment and investigation on cytological outcome of Hybrid Capture 2 borderline samples].

Objectives: to evaluate the reproducibility of the High-Risk Hybrid Capture 2 (HC2-HR, Qiagen) test within the frame of cervical cancer screening with HPV, for samples with values very close to the cut-off (borderline), that is 0.80≤RLU/CO≤0.99 (RLU/CO: relative light units/cut-off) and to assess any possible presence of high grade lesions.

Design: all borderline samples collected from January 2014 to August 2015 were repeated with HR-HC2 test within two weeks from the first test. For HPV-positive samples, cytology and colposcopy results (if present) were analysed.

Setting And Participants: national and international directions recommend the use of HPV testing as a primary test for cervical cancer screening for women aged 30/34-64 years. In Tuscany Region (Central Italy) the primary screening with HPV test was introduced in 2013 and the HPV test currently used, HR-HC2, has a positive cut-off≥1 RLU/CO.

Main Outcome Measures: proportion of HPV-positive tests at repetition, then stratified by type of material (ThinPrep, STM) and for the following subclasses of RLU/CO: 0.80-0.89 and 0.90-0.99.

Results: 253 samples were borderline at first test with HR-HC2 (about 0.4% of all samples tested). About 83% (209/253) of the samples was confirmed to be negative at the HPV test repetition: 88% for RLU/CO=0.80-0.89 and 76% for RLU/CO=0.90-0.99 (p-value=0.014). Median RLU/CO value of 44 HPV-positive samples at repetition is 1.4, with a maximum RLU/CO value of 16.7. In addition, 90% of HPV-positive samples at repetition has a RLU/CO values below or equal to 3.6. Even the used sampling system affects reproducibility: 26.4% (33/125) of the samples resulted positive at the repetition were in ThinPrep vs. 8.6% (11/128) of those in STM (p-value<0.001). The cytology result of HPV positive samples at repetition showed 84% (37/44) of normal samples and 7 low-grade lesions. No high-grade lesion was found in people who performed a colposcopy.

Conclusion: as expected, the reproducibility decreases for values closest to the cut-off (0.90-0.99), confirming the intrinsic limit to all analytical methods near to the cut-off. Moreover, even if the difference in reproducibility between the two systems of sampling is statistically significant, in samples that resulted positive at repetition high-grade lesions were not found at the time of HPV screening test.