Advertisement

Not all HPV nucleic acid tests are equal: only those calibrated to detect high grade lesions matter for cervical screening

Open ArchivePublished:November 03, 2017DOI:https://doi.org/10.1016/j.cmi.2017.10.023
      Dear Sir,
      We were interested to read the recent article by de Thurah et al. [
      • de Thurah L.
      • Bonde J.
      • Lam J.U.H.
      • Rebolj M.
      Concordant testing results between various human papillomavirus assays in primary cervical cancer screening: systematic review.
      ] in your journal examining discordance in human papillomavirus (HPV) results across a range of different assays when compared with the Hybrid Capture 2 (HC2) test. HC2 is the original benchmark technology for HPV-based cervical screening upon which improved assay systems have now been developed. However, HC2 has not satisfied the technical selection criteria for primary population-based screening assays in either the Netherlands or Australia, which are the first two countries to have explicitly considered and published assay requirements for population-based HPV screening programmes. In fact only three out of the nine assays in this systematic review would satisfy the criteria of both countries as being suitable for primary population-based cervical screening.
      The value of continuing to benchmark the screening results of an HPV assay to HC2, in terms of concordance of specimens without high-grade cervical lesions, remains a basic science question rather than a question of clinical utility. The value of an HPV assay in a population-based screening programme rests with its ability to detect high-grade cervical lesions, such as cervical intraepithelial neoplasia grade 2 or above (≥CIN2), rather than with the detection of HPV infection per se. Assays specifically calibrated to detect oncogenic HPV at thresholds associated with the presence of such lesions are the only assays that should be considered for use in population-based screening, as detection of HPV at lower thresholds is not an aim of screening and will lead to over-investigation.
      Although de Thurah et al. have undertaken a very well-planned systematic review of concordance between HC2 and nine other assays for detection of any HPV, they have deliberately excluded much of the data relating to assay concordance in women with HPV-positive, histologically confirmed ≥CIN2 lesions. The authors' rationale for this was that in many of the studies ≥CIN2 lesions were vastly over-represented when compared with the rates of these lesions in the general population. Together with the selection of any HPV infection, rather than HPV infection associated with confirmed ≥CIN2 lesions, this deliberate exclusion unfortunately calls into question the relevance of the systematic review findings for application to primary cervical screening as implied in the article's title.
      It is an HPV assay's ability to accurately predict the presence of high-grade cervical disease that matters for its potential utility in a screening programme. This can be appreciated by considering the findings of Rebolj et al. [
      • Rebolj M.
      • Bonde J.
      • Preisler S.
      • Ejegod D.
      • Rygaard C.
      • Lynge E.
      Differential detection of human papillomavirus genotypes and cervical intraepithelial neoplasia by four commercial assays.
      ]. They compared four different HPV assays (HC2, CLART, cobas and APTIMA) and found that of 651 screened women who had a positive HPV result, only 29% produced a positive result in all four assays. Even when HPV-positive women with ≤CIN1 (similar to atypical squamous cells of undetermined significance or below) lesions were examined for concordance of HPV results, only 34% were HPV-positive for all four assays. However, when the authors examined HPV-positive women with a histologically confirmed ≥CIN2 lesion, over 84% of these women had positive HPV results in all four assays, with an additional 12% positive for three assays, and the remaining 4% positive for two assays (Fig. 1).
      Figure thumbnail gr1
      Fig. 1Distribution of histologically confirmed cervical intraepithelial neopolasia (CIN) lesions in human papillomavirus-positive women with screening samples at ages 30–65 years (adapted from ref.
      [
      • Rebolj M.
      • Bonde J.
      • Preisler S.
      • Ejegod D.
      • Rygaard C.
      • Lynge E.
      Differential detection of human papillomavirus genotypes and cervical intraepithelial neoplasia by four commercial assays.
      ]
      ).
      Because a high degree of concordance for suitably clinically calibrated assays is necessary for a test to be appropriate for screening programme use, there are now established criteria to determine such concordance, the Meijer Criteria, which explicitly set out the required level of concordance for such assays for ≥CIN2 compared with the established reference standard HC2-based [
      • Meijer C.J.
      • Berkhof J.
      • Castle P.E.
      • Hesselink A.T.
      • Franco E.L.
      • Ronco G.
      • et al.
      Guidelines for human papillomavirus DNA test requirements for primary cervical cancer screening in women 30 years and older.
      ]. Two of the most established HPV (PCR-based) assays, the Roche cobas 4800 and the Abbott RealTime, have been demonstrated to have very low discordance, 1.6% [
      • Heideman D.A.
      • Hesselink A.T.
      • Berkhof J.
      • van Kemenade F.
      • Melchers W.J.
      • Daalmeijer N.F.
      • et al.
      Clinical validation of the cobas 4800 HPV test for cervical screening purposes.
      ] and 1.1% [
      • Carozzi F.M.
      • Burroni E.
      • Bisanzi S.
      • Puliti D.
      • Confortini M.
      • Giorgi Rossi P.
      • et al.
      Comparison of clinical performance of Abbott realtime high risk HPV test with that of hybrid capture 2 assay in a screening setting.
      ], respectively, compared with the HC2 result when samples have histologically confirmed ≥CIN2 lesions.
      Clinicians should be reassured that the emerging HPV-based national screening programmes have established clear criteria for HPV-based tests to be suitable for use in screening. The review of de Thurah et al. usefully reminds us why such criteria are critical and that all and any HPV detection is not clinically useful. The use of HPV assays without appropriate sensitivity and specificity for ≥CIN2 could result in lesions being missed or alternatively in over-investigation. These are significant harms, so careful screening test selection and monitoring of performance and quality are required to ensure that the real-world implementation of HPV-based screening achieves the exciting promise of improved cervical cancer prevention.

      Transparency declaration

      DH, JMLB, and MS are employees of VCS Ltd, which has received support from Roche Molecular Systems for HPV testing as part of the Compass Trial, a large randomized trial of primary HPV screening.

      References

        • de Thurah L.
        • Bonde J.
        • Lam J.U.H.
        • Rebolj M.
        Concordant testing results between various human papillomavirus assays in primary cervical cancer screening: systematic review.
        Clin Microbiol Infect. 2018; 24: 29-36
        • Rebolj M.
        • Bonde J.
        • Preisler S.
        • Ejegod D.
        • Rygaard C.
        • Lynge E.
        Differential detection of human papillomavirus genotypes and cervical intraepithelial neoplasia by four commercial assays.
        J Clin Microbiol. 2016; 54: 2669-2675
        • Meijer C.J.
        • Berkhof J.
        • Castle P.E.
        • Hesselink A.T.
        • Franco E.L.
        • Ronco G.
        • et al.
        Guidelines for human papillomavirus DNA test requirements for primary cervical cancer screening in women 30 years and older.
        Int J Cancer. 2009; 124: 516-520
        • Heideman D.A.
        • Hesselink A.T.
        • Berkhof J.
        • van Kemenade F.
        • Melchers W.J.
        • Daalmeijer N.F.
        • et al.
        Clinical validation of the cobas 4800 HPV test for cervical screening purposes.
        J Clin Microbiol. 2011; 49: 3983-3985
        • Carozzi F.M.
        • Burroni E.
        • Bisanzi S.
        • Puliti D.
        • Confortini M.
        • Giorgi Rossi P.
        • et al.
        Comparison of clinical performance of Abbott realtime high risk HPV test with that of hybrid capture 2 assay in a screening setting.
        J Clin Microbiol. 2011; 49: 1446-1451

      Linked Article