n/a
n/a
HPV positive women had cytological triage in 3 randomised controlled trials (RCTs) that had follow up for ≥2 rounds of screening and were included in a pooled analysis of invasive cancer incidence. In all these RCTs, HPV-positive women with abnormal cytology (although at different cut-off, ranging from ASC-US+ to HSIL+) were immediately referred to colposcopy while the remaining were invited for HPV re-testing (interval 6-18 months) and referred if still positive (frequently also if cytology become abnormal). This approach can be considered as validated.
In all such studies cytology was interpreted blindly to HPV result. If HPV is used as a stand-alone primary test then the cytology reader will be aware of HPV positivity. In a study nested in the NTCC RCT cytology was interpreted a-posteriori with knowledge of HPV positivity. The relative sensitivity for CIN2+ of having HPV testing and informed cytology ≥ASC-US vs. stand alone cytology ≥ASCUS was 1.58 (95% CI 1.22-2.01) while the relative immediate referral was 0.95 (0.86-1.04). The absolute cross-sectional sensitivity of informed cytology (85.6%; 76.6-92.1) among HPV women was indeed similar to that of p16 over-expression observed in another study nested in NTCC while specificity was better (66% vs. 57%). Data from the Finnish RCT also suggest greater sensitivity of informed vs. blind cytology.
In a first survey of routine activity in Italy (immediate referral of HPV+ women if their informed cytology was ASC-US+ and repeat HPV test after 1 year in the remaining with referral if still positive), the proportion of HPV-positive women judged to have cytology varied strongly between the 10 study centres (range 20.0%-56.9%) so as the proportion of CIN2+ detected immediately or including 1-year HPV re-testing (range 49.0%-94.3%). This suggests strong variability in the criteria of interpretation and in sensitivity of triage cytology. However, immediate referral (because cytology was ≥ASC-US) had limited effect on the overall referral (including HPV repeat) and sensitivity of cytology had no effect on overall sensitivity. This shows that such a protocol is robust to variability in cytology interpretation.
n/a