We welcome the comment of Benevolo et al concerning our meta-analysis of the accuracy of a 5-type human papillomavirus (HPV) messenger ribonucleic acid (mRNA) assay and the 13-type HPV DNA test (Hybrid Capture 2 [HC2]) to detect cervical precancerous lesions (grade 2 cervical intraepithelial neoplasia or worse [CIN2+] or CIN3+) in women with a cytology result of atypical squamous cells of undetermined significance (ASC-US) or low-grade squamous intraepithelial lesions (LSIL).[1] We did not include their study[2] because of incomplete verification with a reference test, and because the randomness of the verification of triage test positive and negative results was not assured.

However, we agree that the positive predictive values (PPVs), if they result from the verification of the large majority of positive cases, do not suffer from verification bias. Moreover, intrastudy PPV ratios depend mainly on test accuracy because the underlying prevalence of disease is constant within a given study.

In our meta-analysis,[1] the test positivity rates pooled from the 5 studies that were included that evaluated both tests were 37% and 72%, respectively, for the mRNA and HC2 test (ratio, 0.51) in ASC-US and 39% and 73%, respectively, (ratio, 0.47) in LSIL. RNA triage reduced the burden for colposcopy referral to approximately one-half compared with triage with HC2. The pooled PPVs of the mRNA testing for underlying high-grade CIN were similar in ASC-US and LSIL (43% and 41%, respectively). The PPVs of HC2 were substantially lower (34% and 29%, respectively, in ASC-US and LSIL). Moreover, the PPV (posttest probability) of HC2 in the triage of LSILs hardly differed from the pretest probability (25%), thereby underscoring the generally low efficiency of this test to triage women with this cytological abnormality, as also concluded by Benevolo et al and from previous reviews.[3, 4] Another clinically relevant parameter, derived from diagnostic test accuracy reviews, is the risk of precancerous lesions among negative triage test findings (the complement of the negative predictive value). In our meta-analysis, the pooled risk for CIN2+ was 3% and 5%, respectively, when negative on HC2 and 8% and 10%, respectively, when RNA negative in women with ASC-US or LSIL cytology. Only in women with ASC-US and a negative HC2 test might the risk of underlying precancer be considered low enough to refer them back to the screening program.[4] Women with a negative 5-type RNA test cannot be reassured that they are free of precancer and therefore should be kept under surveillance.

Contrary to what Benevolo et al state, the range in interstudy variation of the PPV was not lower than for specificity, reflecting variability in local cytological standards and the inherent underlying prevalence of the disease. In primary screening for a rare disease, test negativity (the complement of test positivity) (FN + TN)/(TN+FN+FP+TP) approximates specificity (TN/[FP+TN]) since the FN and TP are small.[5] However, in triage of cytological abnormalities, this approximation does not hold. (Abbreviations: FN, TN, TP, FP: number of false-negatives, true-negatives, true-positives, false-positives).


  1. Top of page

Drs. Arbyn and Verdoodt were supported by a grant from the Belgian Foundation Against Cancer (Brussels, Belgium), the 7th Framework Programme of DG Research of the European Commission (PREHDICT and CoHeahr Projects, coordinated by the Free University of Amsterdam, the Netherlands [grant numbers 242061 and 603019]), and the International Agency for Research on Cancer (Lyon, France). Dr. Cuschieri was supported by a grant from and received consulting fees or honoraria from Gen-Probe. She has received travel support to past meetings from Gen-Probe and NorChip.

  • Marc Arbyn, MD, DrTMH

  • Unit of Cancer Epidemiology Scientific Institute of Public Health Brussels, Belgium

  • Freija Verdoodt, PhD

  • Unit of Cancer Epidemiology Scientific Institute of Public Health Brussels, Belgium

  • Kate Cuschieri, PhD

  • Scottish Human Papilloma Virus Reference Laboratory Royal Infirmary of Edinburgh Edinburgh, United Kingdom


  1. Top of page
  • 1
    Verdoodt F, Szarewski A, Halfon P, Cuschieri K, Arbyn M. Triage of women with minor abnormal cervical cytology: meta-analysis of the accuracy of an assay targeting messenger ribonucleic acid of 5 high-risk human papillomavirus types [published online ahead of print July 23, 2013]. Cancer (Cancer Cytopathol). doi: 10.1002/cncy.21325.
  • 2
    Benevolo M, Vocaturo A, Caraceni D, et al. Sensitivity, specificity, and clinical value of human papillomavirus (HPV) E6/E7 mRNA assay as a triage test for cervical cytology and HPV DNA test. J Clin Microbiol. 2011;49:2643-2650.
  • 3
    Arbyn M, Ronco G, Anttila A, et al. Evidence regarding human papillomavirus testing in secondary prevention of cervical cancer. Vaccine. 2012;30(suppl 5):F88-F99.
  • 4
    Arbyn M, Roelens J, Martin-Hirsch P, Leeson S, Wentzensen N. Use of HC2 to triage women with borderline and mild dyskaryosis in the UK. Br J Cancer. 2011;105:877-880.
  • 5
    Arbyn M, Ronco G, Cuzick J, Wentzensen N, Castle PE. How to evaluate emerging technologies in cervical cancer screening? Int J Cancer. 2009;125:2489-2496.