ICRH Monograph Marc Arbyn Evidence regarding Human Papillomavirus Testing in Secondary Prevention of Cervical Cancer


November 13th Mr. Marc Arbyn defended his thesis: entitled: “Evidence regarding Human Papillomavirus Testing in Secondary Prevention of Cervical Cancer” in Ghent

Venue: auditorium C - UZ Campus Gent – De Pintelaan 185 – 9000 GENT, Belgium

Promotors/supervisors: Prof. dr. Marleen Temmerman, Prof. Dr. John-Paul Bogers  & Prof. dr. Steven Weyers.


Executive summary
The large majority of cervical cancers are caused by a persistent infection of the cervical
epithelium with human papillomavirus (HPV) types, which are transmitted mainly
through sexual contact. The carcinogenic potential has been demonstrated for a dozen of
HPV types, which therefor are termed as high-risk (hr).
The PhD thesis assesses the evidence for four clinical applications of hrHPV testing in
the framework of the secondary prevention of cervical cancer.
The four clinical applications of hrHPV testing which we evaluated subsequently as
considered historically in the scientific literature are: (1) triage of women with minor
abnormal cervical cytology, (2) follow-up of women treated for cervical pre-cancer, (3)
primary cervical cancer screening, and (4) HPV testing on self-samples collected by the
woman her-self.
We conducted systematic reviews and meta-analyses to assess a series a clinical
questions, which in essence compared the performance of a conventional and a new
method of intervention or testing. Two main designs were applied: (1) evaluation of the
absolute or relative diagnostic test accuracy (DTA) and (2) evaluation of the efficacy or
effectiveness of an intervention. DTA reviews were performed for all the four clinical
applications, where we tried to derive the average sensitivity and specificity of hrHPV
tests, as well as the relative sensitivity and specificity of hrHPV assays versus other
assays to detect cervical pre-cancer. Cervical pre-cancer is defined as histologically
confirmed presence of cervical intraepithelial neoplasia of grade II or III or
adenocarcinoma in situ of the cervix uteri. Questions related to the impact of hrHPVbased
versus cytology-based screening for cervical cancer were addressed by pooling
randomised trials and screening cohort studies.
Established procedures, recommended by the Cochrane Collaboration and healthtechnology
assessment guidelines, were followed to retrieve the key data from selected
relevant references from the scientific literature and to synthesize findings from many
studies into overall measures. The quality of selected studies and variability of results
were examined and accounted for as far as possible.
hrHPV testing is more accurate (more sensitive and similarly specific) than repeat
cytology to find underlying cervical pre-cancer in women with a cytological finding of
atypical squamous cells of undetermined significance (ASC-US). Triage of women with
low-grade squamous intraepithelial lesions (LSIL) with a hrHPV DNA test is sensitive
but poorly specific. The utility of hrHPV testing in women with LSIL is limited, since the
posttest risk of pre-cancer in case of LSIL is hardly higher than the pretest risk. Triage of
ASC-US or LSIL with a RNA test targeting all high-risk HPV types is as sensitive but
more specific than the Hybrid Capture II assay (the most used hrHPV DNA test). Cytoimmuno-
chemistry for p16 or p16/ki67 is substantially more specific but slightly less
sensitive to find underlying pre-cancer in women with ASC-US or LSIL. Isolated
HPV16/18 genotyping is insufficiently sensitive in triage of women with minor abnormal
cytology but distinction between HPV16/18 positive women, other hrHPV positive
womenstratifies risk allowing for specific follow-up recommendations for hrHPV
positive women.
Pretest-Posttest-Probability plots are useful to demonstrate the clinical utility of available
triage options.
Women treated for cervical pre-cancer are at significantly increased risk of recurrent
cervical (pre-) cancer compared to the general population and therefore require careful
follow-up. hrHPV testing after approximately halve a year post-treatment is more
sensitive and almost as specific as repeat cytology to predict residual of recurrent CIN2+.
Adding cytology (co-testing) increases the sensitivity but decreases the specificity of the
prediction of treatment failure. The risk of CIN3 in case of a double negative co-test 6
and 18 months after treatment is as low as for women with a negative cytological
screening test.
Primary cervical cancer screening with hrHPV tests is more sensitive but less specific for
detecting underlying cervical pre-cancer than cytology. Moreover, randomised trials have
consistently shown that the subsequent risk of CIN3 and of invasive cervical cancer is
lower five years after a negative hrHPV DNA test compared to three years after a
negative cytology result. Prevalence of transient hrHPV infection is high in women
younger than 30 years and no increased efficacy of hrHPV screening in this age group is
demonstrated. Improved protection against cervical cancer by hrHPV DNA screening
using HC2 or GP5+/6+ PCR-EIA has been directly observed in randomised trials. It has
been accepted that other hrHPV DNA tests, which demonstrate good reproducibility and
non-inferior cross-sectional sensitivity and specificity for CIN2+ compared to HC2 or
GP5+/6+ PCR-EIA, can also be applied in primary cervical cancer screening. The
following hrHPV DNA assays fulfil these equivalency criteria; cobas-4800, Abbott
RealTime High Risk HPV, BD Onclarity, PapilloCheck, and the HPV-Risk assay. The
APTIMA HPV assay is highly reproducible, as sensitive and more specific than the
standard comparator assays, but –since it targets mRNA of hrHPV - requires additional
demonstration of longitudinal safety over 5-year or longer intervals. hrHPV positive
women require triage before referral for colposcopy. Reflex cytology followed, six to
twelve months later by a second cytology test or hrHPV retest when reflex-cytology is
negative, completed with HPV16/18 genotyping or not, show good sensitivity with
moderate to low specificity for detecting cervical pre-cancer among hrHPV positive
women. P16/Ki67 chemistry and certain methylation markers may provide promising
alternative triage options.
hrHPV DNA testing on self-samples is as accurate as on clinician-taken samples under
the condition that a validated PCR-based assay is used.
Discussion and conclusions
hrHPV testing can be recommended for triage of women with ASC-US and as test of cure
after treatment for cervical pre-cancer.
Considering available evidence, European guidelines propose cervical cancer screening
in the Member States with a clinically validated hrHPV DNA assay, followed by an
appropriate triage algorithm to manage hrHPV positive women. Validated PCR platforms
detecting hrHPV DNA can be used also on self-samples, which may generate
opportunities to reach women who do not participate in regular screening.
The existence of hrHPV negative cancers that may be missed by HPV-only screening and
the question whether co-testing with hrHPV and cytology may avoid these cases is the
object of intensive debate. Careful cost-effectiveness analyses together with new metaanalytical
tools for pooling of cumulative incidence data according to the baseline co-test
status, may open pathways to address these questions. Pooling of cumulative incidence
curves may also provide a key for defining the longitudinal criteria for new screening
tests that target other molecules than hrHPV DNA.

Authors & affiliation: 
Marc Arbyn Supervisors: Prof. dr. Marleen Temmerman, Prof. dr. Steven Weyers, Prof. dr. John Paul Bogers
Published In: 
ICRH Monographs
Publication date: 
Monday, November 13, 2017