Comparison of different commercial methods for HPV detection in follow-up cytology after ASCUS/LSIL, prediction of CIN2-3 in follow up biopsies and spontaneous regression of CIN2-3.

OBJECTIVE: Different Human Papilloma Virus (HPV) tests are currently used. An integrated comparison of the Amplicor, Cobas4800, PreTect HPV-Proofer and APTIMA HPV tests has not been done. METHODS: We compared the high-risk HPV detection power of these HPV tests in 528 consecutive population-based follow-up Liquid-Based Cytology samples (LBC) after ASCUS/LSIL index cytology. Their sensitivity and specificity to detect HPV in LBC, their predictive values of histopathologic CIN2-3 in follow-up punch biopsies and CIN2-3 regression in the subsequent cones was assessed. The HPV subtypes detected by the Linear Array genotyping-test (LA), PreTect HPV-Proofer and Cobas4800 were also compared. The follow-up histopathology was consensus expert-reviewed and Ki67/p16-supported. The predictive values of the HPV results in LBC by the different tests for presence of CIN2-3 in follow-up biopsies, and regression in subsequent cones, was assessed. RESULTS: Amplicor, Cobas4800 and APTIMA show good agreement for HPV-positivity/negativity. PreTect HPV-Proofer has many discrepancies versus any of the other methods. The sensitivities for Amplicor, Cobas4800 and APTIMA to detect CIN2-3 were very high (96-100%), but rather low for PreTect HPV-Proofer (53%). Specificity in case of CIN1 or less in follow-up biopsies of Amplicor and Cobas4800 is lower than APTIMA and highest for PreTect HPV-Proofer. HPV subtyping by LA agreed in 90% with Cobas4800 but 70% with PreTect HPV-Proofer. CONCLUSIONS: The Amplicor, Cobas4800 and APTIMA give comparable results but PreTect HPV-Proofer differs from the other tests, with low sensitivity but higher specificity. None of the methods predicted regression of CIN2-3.