Objectives: To study the impact of different cervical cancer screening strategies including HPV testing.
Methods: A randomised controlled trial with a conventional arm (conventional cytology) and an experimental arm following two phases (first HPV testing+conventional cytology, second HPV testing alone). In phase one, different protocols were applied to different age groups (25-34 and 35-60). Published data on test accuracy during the phase one of recruitment are summarised.
Results: 45,307 women were recruited in phase one (about 95,000 overall). In the age group 35-60, HPV testing (by Hybrid Capture 2) alone at 2 RLU cut-off increased sensitivity vs. conventional cytology (relative sensitivity 1.41; 95% CI: 0.98-1.02) with a small loss in Positive Predictive Value (PPV; relative PPV 0.75; 95% CI: 0.45-1.25). Adding liquid-based cytology as screening test and referring to colposcopy women positive to either only marginally increased sensitivity but strongly reduced PPV. In the age group 25-34, similar results (relative sensitivity vs. conventional cytology 1.58; 95% CI: 1.032.44; relative PPV 0.78; 95% CI: 0.72-1.16) were obtained, despite 14% of women were HPV positive, with a strategy based on HPV alone as screening test, triaging HPV positive women by cytology, directly referring those ASCUS+ to colposcopy and repeating both tests after 1 year in those with normal cytology.
Conclusions: HPV testing, if used as screening test, should be applied alone, with cytology triage essential in younger women but preferable at all ages. Follow-up data will allow analysis of the safety of prolonging screening intervals and the relative persistence of lesions detected with different methods.