Journal of Medical Virology 83:471–482 (2011)

Analytical and Clinical Performances of a

Restriction Fragment Mass Polymorphism Assay for
Detection and Genotyping of a Wide Spectrum of
Human Papillomaviruses
Hyo-Pyo Lee,1 Soo-Ok Kim,2 Tae Sook Hwang,3 Jae-Man Bae,1 Soo Nyung Kim,1
Jae Won Kim,4 Sun Young Hwang,2 Han Sung Lee,2 Soo-Kyung Shin,2 Woojae Cho,2 and
Sun Pyo Hong 2*
1
Department of Obstetrics and Gynecology, KonKuk University School of Medicine, Seoul, Korea
2
R&D Center, GeneMatrix Inc, Yongin, Kyeonggi-do, Korea
3
Department of Pathology, KonKuk University School of Medicine, Seoul, Korea
4
Department of Obstetrics and Gynecology, Seoul National University School of Medicine, Seoul, Korea

Matrix-assisted laser desorption ionization time- KEY WORDS: HPV; genotype; MALDI-TOF;
of-flight (MALDI-TOF) mass spectrometry-based mass spectrometry; cervical
restriction fragment mass polymorphism (RFMP) cancer
assay was adapted to human papillomavirus
(HPV) genotyping. The analytical sensitivity and
the clinical utility were evaluated by testing
INTRODUCTION
defined HPV genome equivalents and a total of
426 specimens composed of normal cytology, Persistent human papillomavirus (HPV) infection is
atypical squamous cells of undetermined signi- the primary cause of cervical cancer, and there is an
ficance, low grade squamous intraepithelial increasing need for more accurate, broad-spectrum and
lesion, high grade squamous intraepithelial high-throughput methods for HPV genotyping to use as
lesion and invasive squamous cell carcinoma. a screening tool for early detection and intervention
The RFMP assay was able to detect 38.4–114.6 [Laconi et al., 2001; Terry et al., 2001; Munoz et al.,
genomic equivalents of a wide variety of HPV 2003]. More than 100 different HPV genotypes have
types. The RFMP assay detected 34 different been identified, of which approximately 45 can infect
HPV genotypes in cervical samples of which 8% the mucosa of the anogenital tract [Steben and Duarte-
were found to be multiple-type infections. The Franco, 2007]. Based on their oncogenicity, the anogen-
high-risk HPV positivity rate according to the ital HPV types have been subdivided into low-risk HPV,
histological diagnosis was 7.9% (8/101), 31.7% which are found mainly in genital warts, high-risk HPV,
(38/120), 50% (55/110), 86% (37/43), 96.2% (50/ which are associated frequently with invasive cervical
52) in normal, atypical squamous cells of unde- cancer, and probable high-risk HPV types [Walboomers
termined significance, low grade squamous et al., 1999]. Several studies have demonstrated the
intraepithelial lesion, high grade squamous potential benefit of incorporating HPV genotyping into
intraepithelial lesion and squamous cell carci-
noma subgroups, respectively. Diagnostic sensi-
Abbreviations: HPV, human papillomavirus; MALDI-TOF,
tivities/specificities for the cervical lesions of matrix-assisted laser desorption/ionization time of flight; RFMP,
squamous cell carcinoma and high grade squa- restriction fragment mass polymorphism.
mous intraepithelial lesion or worse histology Hyo-Pyo Lee and Soo-Ok Kim contributed equally to the study.
were found to be 96.2%/92.1% and 91.6%/ Grant sponsor: National Research Foundation of Korea (NRF)
92.1%, respectively. The sensitivity, accuracy, funded by the Ministry of Education, Science and Technology
wide range of genotype identification and high- (2010K001128-00210).
throughput capacity with cost-effectiveness of *Correspondence to: Sun Pyo Hong, PhD, Gene Matrix Inc,
Yongin, Korea. E-mail: [email protected]
the test consumables make the RFMP assay
Accepted 21 October 2010
suitable for mass screening and monitoring of
DOI 10.1002/jmv.21992
HPV-associated cervical cancer. J. Med. Virol. Published online in Wiley Online Library
83:471–482, 2011. ß 2011 Wiley-Liss, Inc. (wileyonlinelibrary.com).

ß 2011 WILEY-LISS, INC.

472 Lee et al.

screening programs to identify women who are at risk of lesions, and 52 of squamous cell carcinoma. The control
developing invasive cervical cancer and for managing samples (normal, n ¼ 101) were chosen randomly from
patients with equivocal cytological abnormalities age-matched women who had no past or current evi-
[Clavel et al., 2004; Bulkmans et al., 2005]. With the dence of cervical neoplastic lesions and had no clinical
development of type-specific HPV vaccines, determining symptoms of any sexually transmitted infections.
the HPV genotypes present or associated with exposure
may be important for assessing the utility of particular Standards and Plasmid Controls
vaccines [Laconi et al., 2001; Terry et al., 2001; Munoz
Plasmids with type-specific inserts of the high-risk
et al., 2003].
HPV type 16 (HPV-16), -18, -31, -33, -35, -39, -45, -51,
The restriction fragment mass polymorphism (RFMP)
-52, -56, -58, -68, -82, probable high-risk HPV-26, -53,
assay is based on nucleic acid target amplification and
-66, low-risk HPV-6, -11, -34, -40, -43, -44, -54, -55, -61,
mass detection using matrix-assisted laser desorption
-72, -81, -83, non-established-risk HPV-62, -69, -74, -84
ionization time-of-flight (MALDI-TOF) mass spectrom-
were cloned into the pCR-Script Amp cloning vector
etry of oligonucleotides excised by type IIS restriction
(Stratagene, La Jolla, CA) and used for assessing the
enzyme digestion [Kim et al., 2005b]. The use of a type
lower limit of detection of the RFMP assay. The DNA
IIS restriction enzyme makes the assay independent of
concentration of all plasmid stock solutions was deter-
the occurrence of restriction sites, and universal to
mined using the PicoGreen double-stranded DNA quan-
different genotypes because the cleavage sites of these
titation reagent kit (Molecular Probes, Eugene, OR).
enzymes are distal to the recognition sites incorporated
Defined dilution series of 10–500 plasmids/reaction
into the amplification primers. The RFMP strategy
were made in phosphate-buffered saline (pH 8.0) with
has proved to be an accurate and reliable assay for
a background of human DNA (10 ng/ul of human DNA;
genotyping hepatitis C virus, identifying antiviral drug
Sigma–Aldrich, Stockholm, Sweden). WHO inter-
resistant hepatitis B virus as well as human genomic
national standards for HPV DNA (types 16 and 18) were
variations [Kim et al., 2005a; Ahn et al., 2006; Lee et al.,
purchased from National Institute for Biological
2006; Paik et al., 2006; Yeon et al., 2006]. In particular,
Standards and Control (NIBSC, Hertfordshire, UK)
RFMP was shown to be more sensitive and to have
for calibrating the concentration of the type-specific
superior quantitation of mixtures containing multiple
plasmid controls. Ten thousand copies of plasmid
viral genotypes without the need for population-based
DNA containing the entire HPV 16 genome (ATCC
cloning and subsequent sequencing [Lee et al., 2006;
45113D) and distilled water with 10 ng/ul of human
Hong et al., 2008]. The RFMP protocol has been
DNA extracted from human peripheral blood leukocytes
described previously for high resolution HPV genotyp-
were included in all reactions from the extraction step
ing and exploits differences in the molecular masses of
as positive and negative controls, respectively. To avoid
oligonucleotides derived from 22-bp genotype-specific
carry-over contamination, the procedures recom-
motifs in the HPV L1 gene [Hong et al., 2008]. In this
mended by Kwok and Higuchi [1989] were followed
study, 426 clinical specimens with various cervical cytol-
strictly.
ogies were analyzed by the RFMP assay and the assay
performance was compared with results from direct or
RFMP Assay for HPV Genotyping
clonal sequencing assays. In addition, the sensitivity of
the assay was determined using dilutions of the WHO DNA was extracted from the entire cervical sample by
standards for HPV DNA. the QIAamp DNA Mini Kit (Qiagen, Chatworth, CA)
according to the manufacturer’s instructions. HPV DNA
MATERIALS AND METHODS was amplified with PGMY09/11 primer pools, compris-
ing of two non-degenerate pools of L1 consensus primers
Specimens
[Gravitt et al., 2000]. The second round primer pairs to
Cervical cell specimens referred for cervical cancer generate the RFMP product comprised of a sense primer
screening were collected for this study from August 2008 PV-rfmpF specific to bases 6,584 to 6,603 (50 -GCM-
to March 2010 at the participating organizations. The CAGGGHCAYAAGGATGAATGG-30 ) and an antisense
experimental protocol conforms to the ethical guidelines primer PV-rfmpR specific to bases 6,657 to 6,626
of the 1975 Declaration of Helsinki as reflected by prior (50 -GTACTDCKDGTRGTATCHACMACGGATGTAAC-
approval by the organizations’ Institutional Review AAA-30 ). Underlined is a five-nucleotide sequence
Boards. All participants were informed and written (ggatg) embedded in the primers to introduce a
consent of patients was obtained. The case group FokI site (a neoschizomer of BtsCI) in amplicon.
(aged 24–65) was allocated using liquid-based cytology Nucleotide sequence positions were numbered accord-
(SurePathTM, TriPath Imaging Inc., Burlington, NC) ing to HPV genotype 16 (GenBank accession number;
based on the new Bethesda system [Solomon et al., K02718). Restriction enzyme digestion of PCR products
2002] and subsequent histological evaluation of punch was performed by mixing the PCR mixtures with 10 ml
biopsy samples obtained at colposcopy as follows; 120 of buffer containing 50 mM potassium acetate, 20 mM
specimens of atypical squamous cells of indeterminate Tris-acetate, 10 mM magnesium acetate, 1 mM dithio-
significance, 110 of low-grade squamous intraepithelial threitol, and 1 unit of FokI and BtsCI as described
lesions, 43 of high-grade squamous intraepithelial previously [Hong et al., 2008]. The reaction mixtures
J. Med. Virol. DOI 10.1002/jmv
Clinical Evaluation of RFMP HPV Genotyping Assay 473

were incubated at 378C for 1 hr. The resulting digest obtained by reference to the LANL HPV database
was purified by vacuum filtration through 96- or sequences for 45 anogenital HPV genotypes without
384-well Oasis1 mElution Plates (Waters, Tokyo, duplication. The 45 anogenital HPV genotypes were
Japan). The desalted reaction mixtures were resuspend- assigned to 14 high-risk, three probable high-risk,
ed with matrix solution containing 15 mg/ml 3-hydrox- 15 low-risk, and 13 non-established-risk HPV types
ypicolinic acid, 0.023 M ammonium citrate, and 12% according to their reported carcinogenicity (Table I).
acetonitrile; they were then spotted in 3-ml volumes International standard HPV DNA of 5
106 genomic
on a polished MTP AnchorChipTM plate (Bruker equivalents was diluted serially in phosphate-buffered
Daltonics, Bremen, Germany). Mass spectra were saline (pH 8.0) with a carrier of human peripheral blood
acquired with the aid of installed software (flexcontrol DNA. When the diluted sample series was tested by the
3.0) on a Microflex linear MALDI-TOF mass spectrom- RFMP assay, the analytical sensitivity was as low as
etry (Bruker Daltonics). The RFMP assays were per- 90 genomic equivalents per reaction. Analysis of the
formed independently three times by different dilution series of various HPV genotype-containing
investigators. plasmids calibrated against the WHO HPV inter-
national standard revealed detection limits per reaction
Sequencing Analysis ranging from 38.4 to 114.6 copies by the RFMP assay
(Table II). When mixed genotype pools consisting of two
PCR products amplified with PGMY09/11 were puri-
or three HPV types were tested, the RFMP assay had
fied with QIAquick PCR purification Kit (Qiagen)
detection limits of 107.4–121 copies of each HPV type per
according to the manufacturer’s instructions or cloned
reaction (Table II).
into the pCR-Script Amp cloning vector (Stratagene).
Purified PCR products or cloned DNAs were sequenced
by the BigDye terminator v3.1 Cycle Sequencing kit and Diagnostic Performance of the RFMP
ABI PRISM 310 genetic Analyzer (Applied Biosystems, Assays on Clinical Samples
Foster City, CA) using 5 pmols of either PGMY09/11
Representative results for RFMP assays of high-risk
primers or vector-specific primers in both directions,
HPV types and probable high-risk HPV types are shown
respectively. Approximately 10 clones per sample were
in Figure 1. The genotype distribution of 275 HPV cer-
selected randomly for clonal sequencing analysis. HPV
vical samples as determined by RFMP assay is shown in
genotype was determined by comparing the sequencing
Figure 2. All the samples were genotyped successfully by
results with HPV reference sequences from Los Alamos
RFMP, and the results were consistent over three inde-
National Laboratories (LANL) HPV Database (http://
pendent runs conducted by different investigators.
hpv-web.lanl.gov).
Overall, the HPV positivity rate for all specimens was
64.5% (275/426) and the high-risk HPV positivity rate
Statistics
was 44.1% (188/426). The prevalence of HPV genotypes
Data were analyzed by Cohen’s Kappa coefficient test was found to be as follows; HPV 16 (17.4%), 52 (15.1%),
for concordance for qualitative items and Cochran 53 (6.4%), 62 (5.0%), 58 (4.7%), 81 (3.7%), 56 (3.3%), 66
Mantel Haenszel chi-square test for the association (3%), 18 (3%), 44 (2.7%), 31 (2.3%), 43 (2%), 51 (2%), 61
between two qualitative variables using the statistical (2%), and 72 (2%) constituting 74.6% of HPV positive
package SAS (version 8; SAS Institute Inc., Cary, NC). samples, as shown in Figure 2. The HPV positivity rates
P-values of <0.05 were considered as statistically increased with the progressive severity level of the cer-
significant. vical lesion. Likewise, the high-risk HPV positivity rate
increased with worsening histological diagnosis; 7.9%
RESULTS (8/101) in controls, 31.7% (38/120) in atypical squamous
cells of undetermined significance, 50% (55/110) of low
RFMP HPV Genotyping Assay
grade squamous intraepithelial lesion, 86% (37/43)
From the alignment of the reference sequences of the of high grade squamous intraepithelial lesion and
L1 major capsid gene of all available HPV genotypes 96.2% (50/52) of squamous cell carcinoma subgroups
from the LANL HPV database (http://hpv-web.lanl.gov), (Table III). Mixed genotype infection could be detected
sequences corresponding to nucleotides 6,604–6,625 by recognition of peak combinations matched to pre-
(numbered according to the HPV 16; GenBank accession dicted mass patterns of more than two genotypes, as
number K02718), surrounded by conserved regions shown in Figure 3. Of interest, 8% (24/275) of the
were chosen as genotype-specific motifs. Primers flank- samples were determined to be infected with mixed
ing the variable regions, PV-rfmpF/PV-rfmpR, were genotypes. As with single-genotype infections, the pre-
designed and modified to introduce upon amplification, dominate genotypes in mixed-genotype infections were
a type IIS restriction endonuclease recognition sequence 52 and 16, which were encountered in three and six
50 of the variable motif. Cleavage of an 80 bp amplified cases, respectively. Approximately 25% (6/24) of the
segment with FokI and BtsCI leads to excision of two sets mixed-genotype infections found in RFMP analysis
of 7mer/12mer/13mer fragments reflecting the pattern were determined to be single genotype (three cases) or
of sequence variations in the 22-bp-genotype-specific to be ambiguous (three cases) by direct sequencing.
motif (Table I). The combined mass patterns were Direct sequencing was not able to identify specific
J. Med. Virol. DOI 10.1002/jmv
474 Lee et al.
TABLE I. Expected Mass Patterns of HPV Genotypes in RFMP Assays

H, high risk; PH, probable high risk; L, low risk; NE, risk not yet established.

J. Med. Virol. DOI 10.1002/jmv

 TABLE II. Limit of Detections of RFMP Assay for Various HPV Genotypes

HPV Copies/ Tested/ Calculated HPV Copies/ Tested/ Calculated HPV Copies/ Tested/ Calculated
type PCR detected sensitivity type PCR detected sensitivity type PCR detected sensitivity
16 500 5/5 47.6 5/5 500 5/5 92.2 11 500 5/5 38.9
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 5/5 50 4/5 50 5/5
25 3/5 25 3/5 25 4/5
10 1/5 10 2/5 10 3/5
52 500 5/5 107.4 66 500 5/5 90.3 84 500 5/5 107.4
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 4/5 50 4/5 50 4/5
25 2/5 25 2/5 25 2/5
10 2/5 10 1/5 10 2/5
58 500 5/5 99.9 26 500 5/5 92.2 5/5 500 5/5 76.0
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 4/5 50 4/5 50 4/5
25 1/5 25 3/5 25 4/5
10 1/5 10 2/5 10 2/5
56 500 5/5 57.8 62 500 5/5 92.2 82 500 5/5 47.6
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 5/5 50 4/5 50 5/5
25 3/5 25 3/5 25 3/5
10 3/5 10 2/5 10 1/5
18 500 5/5 76.0 81 500 5/5 105.2 74 500 5/5 61.7
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 4/5 50 4/5 50 5/5
25 4/5 25 3/5 25 3/5
10 2/5 10 3/5 10 4/5
31 500 5/5 67.8 44 500 5/5 76.0 83 500 5/5 114.6
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 4/5 50 4/5 50 3/5
25 4/5 25 4/5 25 3/5
10 1/5 10 2/5 10 1/5
51 500 5/5 57.8 43 500 5/5 103.1 34 500 5/5 105.2
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 5/5 50 3/5 50 4/5
25 3/5 25 4/5 25 3/5
10 3/5 10 1/5 10 3/5
33 500 5/5 52.9 61 500 5/5 52.9 69 500 5/5 92.2
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 5/5 50 5/5 50 4/5
25 3/5 25 3/5 25 3/5
10 2/5 10 2/5 10 2/5
45 500 5/5 52.9 72 500 5/5 38.4 16 þ 18 500 5/5 107.4
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 5/5 50 5/5 50 4/5
25 3/5 25 4/5 25 2/5
10 2/5 10 4/5 10 2/5
35 500 5/5 59.2 6 500 5/5 92.2 43 þ 45 500 5/5 114.6
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 4/5 50 4/5 50 3/5
25 5/5 25 3/5 25 3/5
10 2/5 10 2/5 10 1/5
68 500 5/5 79.5 5/5 500 5/5 57.8 11 þ 16 þ 18 500 5/5 121.0
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 4/5 50 5/5 50 3/5
25 3/5 25 3/5 25 4/5
10 1/5 10 3/5 10 2/5
39 500 5/5 67.8 40 500 5/5 90.6 16 þ 18 þ 43 500 5/5 121.0
250 5/5 250 5/5 250 5/5
100 5/5 100 5/5 100 5/5
50 4/5 50 4/5 50 3/5
25 4/5 25 4/5 25 4/5
10 1/5 10 4/5 10 2/5
Defined dilutions of HPV plasmids of the 32 HPV types in human genomic DNA backgrounds were made from 500 copies to 10 copies per plasmid/PCR
and analytical sensitivities were calculated by Probit analysis at a 95% detection level. Calculated sensitivity was expressed as copies/reaction.

J. Med. Virol. DOI 10.1002/jmv

476 Lee et al.

Fig. 1. Representative RFMP HPV genotyping results. RFMP results for high-risk group; HPV types 16,
18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 68, 82 and probable high-risk group; HPV types 66, 53. X and Y axes
represent relative peak intensity (r.i.) and mass-to-charge ratio (m/z), respectively. Oligonucleotide stand-
ards of 6mer (50 -ACGTAC-30 ; 1,762.2 Da) and 16mer (50 -ACGTACGTACGTACGT-30 ; 4,881.2 Da) are used
for mass calibration standards.

J. Med. Virol. DOI 10.1002/jmv

Clinical Evaluation of RFMP HPV Genotyping Assay 477

Fig. 2. Overall HPV distribution determined by the RFMP assay in 275 HPV positive specimens. A total
of 300 genotypes were counted from 275 HPV positive samples, of which 24 cases showed mixed infection;
two genotypes were found in 23 samples and three types in one sample. Frequency fractions of HPV
genotypes found as single or mixed HPV types are indicated by gray and black bars, respectively. NE, risk
not yet established.

genotypes in the multiple genotype infections, or ident- genotype 16 was confirmed by clonal sequencing (seven
ified only the major HPV type, while RFMP determined and three clones for HPV 31 and 16, respectively), con-
HPV type 16 mixed with 31 or HPV type 33 with 11 in sistent with the RFMP genotype result. As shown in
case 1 and 2, respectively. This was further corroborated Figure 3B, the direct sequencing detected only HPV
by clonal sequencing (Fig. 3). genotype 33 but visual inspection of the electrophero-
The accuracy of HPV genotyping using the RFMP gram showed several possible base polymorphisms at
assay was estimated by sequencing analysis of the L1 nucleotide positions at 6,610 (C/T), 6,616 (C/A), 6,619 (C/
gene. Overall, the RFMP assay showed concordance T), 6,622 (C/G), 6,623 (G/T), and 6,625 (A/G) though it
with clonal sequencing except in six cases, which were was not clear whether these were true nucleotide mixes
found to be mixed-genotype infections in RFMP analysis or background contribution. In this case, the RFMP
but single genotype or ambiguous by direct sequencing assay identified the genotype 33 mixed with genotype
(Kappa ¼ 0.969, 95% CI: 0.941–0.969). As shown in 11, which were confirmed by clonal sequencing; nine
Figure 3A, the direct sequencing showed base polymor- clones of HPV type 33 (6,604-TATTTGTTGGGGC-
phisms at nucleotide positions 6,604 (C/T), 6,616 (C/T), AATCAGGTA-6,625) and one clone of HPV type 11
6,619 (C/T), 6,621 (A/T), 6,622 (G/A), and 6,623 (C/T), (6,604-TATTTGCTGGGGAAACCACTTG-6,625.
which can be interpreted as HPV type 16 (6,604- Peak patterns in the RFMP assay that were incon-
CATTTGTTGGGGTAACCAACTA-6,625), HPV type 31 sistent with the molecular mass patterns predicted from
(6,604-TATTTGTTGGGGCAATCAGTTA-6,625), HPV the up-to-date databases may reflect potential genotype
type 31 variant (6,604-TATTTGTTGGGGCAATCTGT- variants. To investigate this further, the samples with
TA-6,625), or HPV type 60 (6,604-TATTTGTTGGGG- aberrant peaks were sequenced. As shown in Figure 3,
TAACCAATTA-6,625) from alignment with the refer- the sample showing the mass deviation from the con-
ence sequences from the LANL HPV database (http:// ventional HPV 31 type, could be explained by 9 Da mass
hpv-web.lanl.gov). A mixture of genotype 31 variant and increase of the internal downward 12mer fragment

TABLE III. Incidence of HPV Genotypes Based on Histopathological Diagnosis

Histology

RFMP results (n ¼ 426) Normal (n ¼ 101) ASCUS (n ¼ 120) LSIL (n ¼ 110) HSIL (n ¼ 43) SCC (n ¼ 52)
High risk 8 (7.9%) 38 (31.7%) 55 (50.0%) 37 (86.0%) 50 (96.2%)
Probable high risk 2 (2.0%) 6 (5.0%) 14 (12.7%) 3 (7.0%) 1 (1.9%)
Low risk 6 (5.9%) 16 (13.3%) 11 (10.0%) 0 (0.0%) 0 (0.0%)
NE 5 (5.0%) 8 (6.7%) 14 (12.7%) 0 (0.0%) 1 (1.9%)
HPV positive 21 (20.8%) 68 (56.7%) 94 (85.4%) 40 (93.0%) 52 (100%)
HPV negative 80 (79.2%) 52 (43.3%) 16 (14.6%) 3 (7.0%) 0 (0.0%)
ASCUS, atypical squamous cells of undetermined significance; LSIL, low grade squamous intraepithelial lesion; HSIL, high grade squamous
intraepithelial lesion; SCC, squamous cell carcinoma; NE, risk not yet established.

J. Med. Virol. DOI 10.1002/jmv

478 Lee et al.

Fig. 3. Mixed genotype infections detected by the RFMP assay and confirmation with clonal sequencing.
Top, RFMP analysis; Middle, direct sequencing; Bottom, clonal sequencing. Sequences from 6,509 to 6,624
in HPV L1 gene were depicted under the electropherogram. Figures in parentheses were detected clone
numbers in clonal analysis. X and Y axes represent relative peak intensity (r.i.) and mass to charge ratio (m/z),
respectively. HPV genotypes 31 variant (31v) mixed with 16 and genotypes 33 with 11 were presented in A and
B, respectively, where peaks unique to HPV genotypes 31 and 33 were underlined for easy recognition.

(3,661–3,770) and 9 Da decrease of upward 13mer frag- matograms in six cases. When a total of 426 test results
ment. An A to T conversion at position 6,621 could lead to were considered, overall concordance (Kappa value)
the observed mass deviation and this was confirmed between the RFMP and sequencing assays was 0.969
subsequently by clonal sequencing (Fig. 3). Thus, the (95% CI: 0.941–0.969). In an attempt to associate the
RFMP assay has practical usefulness for detection of high-risk HPV positivity with the histological subgroups
mixed genotype HPV infections as well as screening of cervical lesions, table analysis was carried out on
of genotype variants that have base variations in the squamous cell carcinoma, high grade squamous intra-
genotype determinant motif, with reduced likelihood epithelial lesion and worse (termed ‘‘HSIL’’), low grade
of false negative or positive results that can occur squamous intraepithelial lesion or worse (termed
with existing methods using hybridization between ‘‘LSIL’’), atypical squamous cells of undetermined sig-
the target sequences and fixed probe sequences. nificance or worse (termed ‘‘ASCUS’’) histology sub-
Of 275 HPV-positive samples identified by RFMP, the groups versus normal histology (Table IV). Odds ratios
sequencing results could not be deduced from chro- for the development of cervical lesions of ASCUS,

TABLE IV. Relationship Between Cervical Histology and the High-Risk HPV Identified by RFMP

Lesions vs. normal Sensitivity (%) Specificity (%) PPV (%) NPV (%) Odds ratio
(n ¼ 101) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
SCC (n ¼ 52) 96.2 (89.2–98.9) 92.1 (88.5–93.5) 86.2 (80.0–88.7) 97.9 (94.1–99.4) 290.63 (64.0–1,271.7)
HGSIL  (n ¼ 95) 91.6 (86.7–94.8) 92.1 (87.5–95.1) 91.6 (86.7–94.8) 92.1 (87.5–86.1) 126.42 (45.9–347.7)
LGSIL  (n ¼ 205) 69.3 (66.3–71.1) 92.1 (86.1–95.8) 94.7 (90.6–97.2) 59.6 (55.7–62.0) 26.20 (12.2–56.3)
ASCUS  (n ¼ 325) 55.4 (53.4–56.6) 92.1 (85.7–95.9) 95.7 (92.3–97.8) 39.1 (36.4–40.7) 14.43 (6.9–30.2)
PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval. ‘‘’’ represents a histology group that is corresponding to or
worse than.

J. Med. Virol. DOI 10.1002/jmv

Clinical Evaluation of RFMP HPV Genotyping Assay 479

LSIL, HSIL and squamous cell carcinoma were 14.43 adversely patient management decisions. The HC II
(95% CI: 6.6–30.2), 26.2 (95% CI: 12.2–56.3), 126.4 (95% assay does not permit the identification of type-specific
CI: 45.9–347.7), and 290.6 (95% CI: 64.0–1,271.7), genotypes, but only defines samples as positive for high-
respectively. The correlation of the high-risk HPV detec- risk or low-risk HPV types and cannot provide infor-
tion with histology results is detailed in Table IV with a mation about true persistence of the virus. Since the
result of ASCUS, LSIL, HSIL, or squamous cell determination of high-risk HPV persistence is the most
carcinoma defined as ‘‘disease positive’’ for calculating important predictor of the development of cervical car-
clinical sensitivity and specificity. Diagnostic sensi- cinoma, it is of value to identify exactly the specific HPV
tivities of the high-risk HPV positivity determined by genotype [Perrons et al., 2005].
RFMP for ASCUS, LSIL, HSIL and squamous cell The RFMP assay exploits differences in the molecular
carcinoma were 55.4, 69.3, 91.6, and 96.2%, respectively masses of oligonucleotides comprising the HPV geno-
while the diagnostic specificities were all 92.1%. The type-specific variations in L1 gene. The assay is based on
highest positive predictive value of 95.7% was observed amplification and mass detection of oligonucleotides
for the development of cervical lesions of ASCUS in the excised from type IIS enzyme digestion using MALDI-
presence of the high-risk HPV genotype. The highest TOF mass spectrometry. Mass spectrometry assesses
negative predictive value of 97.9% was observed for the directly the mass of the PCR product, whereas other
development of squamous cell carcinoma in the presence technologies only measure indirectly PCR products,
of the high-risk HPV genotype. either through hybridization or by sequencing reac-
tions, which use PCR products as templates.
Procedures have been developed in which PCR products
DISCUSSION
are used as templates to which oligonucleotide primers
Infection with high-risk HPV is the most important are hybridized, base-extended and then analyzed by
risk factor for cervical cancer [Syrjanen et al., 2004]. mass spectrometry [Sauer et al., 2006]. These assays
Using modern HPV detection methods, 95–100% of squ- can be useful, but they fail to employ one of the key
amous cell carcinoma and 75–95% of high-grade cervical advantages of mass spectrometry that the analysis of
intraepithelial neoplasia lesions have detectable HPV PCR products can be direct. In RFMP assay, mass of
DNA [Munoz et al., 2003]. Research into HPV-induced PCR products is determined directly, rather than their
cervical cancer has demonstrated that genotypes 16 and identity being interpreted on the basis of fluorescent or
18 cause approximately 70% of cases worldwide [Sideri radioactive reporter tags. Both DNA strands can be
et al., 2009]. Studies of the oncogenic potential of these analyzed in parallel, providing a level of internal con-
HPV types have demonstrated clearly that high-risk firmation not achievable by other methods. The use of a
HPVs are a necessary cofactor for the development of type IIS restriction enzyme makes the assay independ-
cervical cancer [Kailash et al., 2006; Castellsagué, ent of restriction sites within the HPV genome and
2008]. Further information, such as the viral type(s) suitable for many different viral genotypes because
involved in the infection, their persistence, viral load these enzymes cleave DNA at a fixed distance from
and E6/E7 oncogene expression, have been proposed to the recognition sites incorporated into the amplification
increase the positive predictive value of a virological primers.
parameter in the detection of developing lesions Diagnostic sensitivities and specificities of the high-
[Bosch et al., 1995; Chang et al., 1997; Kaufman risk HPV positivity determined by RFMP for cervical
et al., 1997]. lesions of cervical cancer itself and HSIL were 96.2%/
Presently, HPV genotyping is performed largely by 92.1% and 91.6%/92.1%, respectively. Moreover, when
hybridization-based assays [van Ham et al., 2005; Jiang HPV types 26, 53, and 66, designated as probable high
et al., 2006]. HPV cannot be cultured reliably in a labora- risk, were included as the high-risk HPV positivity, the
tory setting; therefore, diagnosis HPV relies on molecu- diagnostic sensitivity was increased to 95.8% 96.2%
lar methods that detect HPV DNA or RNA in cervical/ for cervical cancer and HSIL, respectively. Pooled
vaginal samples [Koliopoulos et al., 2009]. HPV DNA analysis of 11 case–control studies from nine countries
testing using the Digene Hybrid Capture assay (HC II) involving 1,918 women with histological confirmed squ-
was more sensitive than conventional Pap testing for amous cell carcinoma and 1,928 control women provided
detection of high-grade lesions and cancers, but less robust estimates of the risk of the cancer linked 30 HPV
specific [Mayrand et al., 2007]. Recent reports have types [Munoz et al., 2003]. In that study, the estimate of
shown significant analytical inaccuracy and reproduci- the pooled odds ratio for squamous cell carcinoma associ-
bility problems of the HC II near to the cut-off, ated with HPV positivity was 158.2 (95% CI: 113.4–
due mainly to the cross-reactivity of its high-risk 220.6). Compared to those estimates, odds ratio for
probe cocktail [Terry et al., 2001; Poljak et al., 2002; the cervical cancer with high-risk positivity estimated
Federschneider et al., 2004; Seme et al., 2006]. Another by the RFMP assay was 290.63 (95% CI: 64.0–1,217.7).
study showed that there were at least 22 different low- When HPV genotypes 26, 53, and 66 were supplemented
risk HPV genotypes that cross-hybridize with the cur- to high-risk group, significantly tighter association of
rent HC II high-risk probes used [Matsukura and the cervical cancer was observed (odds ratio ¼ 464.1;
Sugase, 2001]. This cross-reaction by the high-risk 95% CI: 71.0–2,886.2). Since no HPV negative case was
HC II probes can lead to false positive results and affect found in cervical cancer group, exact odds for the cancer
J. Med. Virol. DOI 10.1002/jmv
480 Lee et al.

with HPV positivity assessed by the RFMP assay could and specific target amplification could be validated sim-
not be determined but was estimated to be more than ultaneously with mass analysis, providing a level of
194.3 (95% CI: 31.6–1,164.1) assuming that one nega- internal confirmation not achievable by other methods.
tive test had been detected in the cancer cases. The As a result, the specificity of the analysis is increased,
possible explanation for the greater correlation of cancer which minimizes false-positive signals. In contrast,
development with detection and genotyping by the methods that rely on fluorescence estimation do not
RFMP assay, compared to the PCR followed by oligohy- achieve this level of specificity as total fluorescent back-
bridization method, could be ascribed to higher speci- ground is detected rather than limiting background to a
ficity and the capacity to identify a broader span of specific molecular mass. The uniform lower limit of
genotypes, leading to fewer positive results in controls detection among wide range of HPV types tested could
and higher positivity in patients. Thus, the RFMP be ascribed to the nested-PCR with primers common
method has both greater clinical sensitivity and speci- to each HPV type and the nature of the RFMP assay
ficity for detecting and genotyping HPV in various levels assigning genotypes from the specific mass. Therefore,
of cervical lesions and comparable or superior analytical the sensitive and specific RFMP assay should have
sensitivity for the number of HPV copies that could be significant applicability to the facile detection of HPV-
detected, compared with most other reported methods infection.
[Kocjan et al., 2008; Poljak et al., 2009]. By combining the merits of unique assay chemistry
The high-risk HPV types detected most frequently and the nature of MALDI-TOF mass spectrometry, the
were HPV 16 (17.4%), 52 (15.1%), 58 (4.7%), 56 (3.3%), RFMP assay is able to screen for HPV genotypes in a
18 (3.0%), and 31 (2.3%). It is noteworthy that HPV 52 robust high-throughput manner capable of analyzing
was the most common and detected more frequently 384 samples in 3 hr automatically with the aid of the
than HPV 16 in 224 cervical cancer free samples Bruker software (flexcontrol 3.0), which is almost
(17.4% vs. 14.3%). Similar findings have been reported 10 times faster than existing hybridization methods.
among cervical cancer-free women in Korea and other In terms of cost-effectiveness, the direct cost per test
Asian countries such as China and Taiwan [de Sanjosé (reagents and labor) of the RFMP assay was estimated to
et al., 2007], suggesting a geographical difference of be <$10 including viral DNA extraction, PCR, restric-
HPV genotype epidemiology should be taken into con- tion digestion, desalting, and matrix for mass analysis,
sideration in HPV diagnosis and future HPV vaccine which is much cheaper than the sequencing or hybrid-
development. ization assays that are approximately $50–100 per test.
In the first WHO international collaborative study for These costs do not include the capital equipment, which
the assessment of the performance of various HPV are slightly greater for the RFMP method. However,
detection assays, a panel of HPV genotypes was tested with the advent of many diagnostic assays operating
in a number of laboratories, of which approximately half on a mass spectrometer platform such as clinical
of participating laboratories failed to detect high con- genotyping and microorganism identification and the
centrations of HPV 31 and, to a lesser extent, to detect gradual spread of compact mass spectrometer into
HPV 35, 52, and 6 [Quint et al., 2006]. In addition, the laboratories as medical devices (i.e., Bruker Microflex),
WHO international collaborative study performed in the burden of the amortization cost should be sub-
2007 emphasized that technical improvement of the stantially reduced [Nagy et al., 2009].
current assays was deemed imperative to enable sensi- In conclusion, the RFMP assay for HPV detection
tive detection of HPV39 and 68 [Pagliusi et al., 2006]. and genotyping enables efficient detection and
The RFMP assay showed the existence of 34 kinds of identification of 45 anogenital HPV types including
HPV genotypes in a total of 275 HPV positive cervical 14 high-risk, 3 probable high-risk, 15 low-risk, and
samples (100% concordance in triplicate runs performed 13 non-established-risk HPV types. The RFMP assay
independently) and detected reproducibly about 50–100 has the analytical sensitive to detect 38.4–114.6
genomic equivalents of a wide variety of HPV types genomic equivalents of a wide variety of HPV types as
calibrated against WHO international standard HPV well as double or triple HPV type mixed infections. The
DNA (Fig. 2 and Tables I and II) that include HPV genotype results of RFMP assay were in full agreement
39, 68, and mixed genotypes composed of two or three with direct or clonal sequencing analyses except the few
HPV types. The analytical sensitivity of the current samples with a low viral load or mixed infections sus-
standard of care, the Digene HCII, extends only down pected to be beyond the limit of analysis of the compari-
to about 5,000 copies [Snijders et al., 2003]. Thus, the son method. The high-risk positivity determined by
RFMP assay is two log10-fold more sensitive than the RFMP method provided excellent measures for diagnos-
current test that is, standard of care. The specificity of tic sensitivity, specificity, and predictive values for
the RFMP assay devolves likely from the requirement detecting highs levels of cervical lesions such as high
that backgrounds have the identical molecular weight grade squamous intraepithelial lesion or squamous cell
as the target sequence. This stringent requirement carcinoma, though it should be validated further by
ensues from the direct analysis of the restriction subsequent controlled-studies with additional cases.
enzyme-digested nested PCR product by MALDI-TOF The sensitivity, genotype accuracy, broad-spectrum of
mass spectrometry. Further, both strands of genotype- genotype coverage, and capability for high-throughput
specific DNA fragments were analyzed in parallel, analysis should make the RFMP assay suitable for
J. Med. Virol. DOI 10.1002/jmv
Clinical Evaluation of RFMP HPV Genotyping Assay 481

robust virus screening in laboratories for early Kocjan BJ, Seme K, Poljak M. 2008. Detection and differentiation of
diagnosis and intervention of HPV-associated cervical human papillomavirus genotypes HPV-6 and HPV-11 by FRET-
based real-time PCR. J Virol Meth 153:245–249.
malignancies. Koliopoulos G, Valasoulis G, Zilakou E. 2009. An update review on HPV
testing methods for cervical neoplasia. Expert Opin Med Diagn
ACKNOWLEDGMENTS 3:123–131.
We gratefully acknowledge the valuable assistance Kwok S, Higuchi R. 1989. Avoiding false positives with PCR. Nature
339:237–238.
of Dr. William Folk at the University of Missouri, MI, Laconi S, Greco M, Pellegrini-Bettoli P, Rais M, Laconi E, Pani P. 2001.
USA and Dr. Scott Bowden at Victorian Infectious One-step detection and genotyping of human papillomavirus in
Diseases Reference Laboratory, Australia for reviewing cervical samples by reverse hybridization. Diagn Mol Pathol 10:
200–206.
the manuscript.
Lee CH, Kim S-O, Byun KS, Moon MS, Kim E-O, Yeon JE, Yoo W, Hong
REFERENCES SP. 2006. Predominance of hepatitis B virus YMDD mutants is
prognostic of viral DNA breakthrough. Gastroenterology 130:
Ahn SH, Kim do Y, Chang HY, Hong SP, Shin JS, Kim YS, Kim H, Kim 1144–1152.
JK, Paik YH, Lee KS, Chon CY, Moon YM, Han KH. 2006. Matsukura T, Sugase M. 2001. Relationships between 80 human pap-
Association of genetic variations in CCR5 and its ligand, illomavirus genotypes and different grades of cervical intraepithe-
RANTES with clearance of hepatitis B virus in Korea. J Med lial neoplasia: Association and causality. Virology 283:139–147.
Virol 78:1564–1571. Mayrand M-H, Duarte-Franco E, Rodrigues I, Walter SD, Hanley J,
Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J, Ferenczy A, Ratnam S, Coutlee F, Franco EL. the Canadian
Schiffman MH, Moreno V, Kurman R, Shah KV. 1995. Prevalence of Cervical Cancer Screening Trial Study Group. 2007. Human pap-
human papillomavirus in cervical cancer: A worldwide perspective. illomavirus DNA versus papanicolaou screening tests for cervical
International biological study on cervical cancer (IBSCC) Study cancer. N Engl J Med 357:1579–1588.
Group. J Natl Cancer Inst 87:796–802. Munoz N, Bosch FX, de Sanjose S, Herrero R, Castellsague X, Shah KV,
Bulkmans NWJ, Rozendaal L, Voorhorst FJ, Snijders PJF, Meijer Snijders PJF, Meijer CJLM. the International Agency for Research
CJLM. 2005. Long-term protective effect of high-risk human pap- on Cancer Multicenter Cervical Cancer Study Group. 2003.
illomavirus testing in population-based cervical screening. Br J Epidemiologic classification of human papillomavirus types associ-
Cancer 92:1800–1802. ated with cervical cancer. N Engl J Med 348:518–527.
Castellsagué X. 2008. Natural history and epidemiology of HPV infec- Nagy E, Maier T, Urban E, Terhes G, Kostrzewa M. 2009. Species
tion and cervical cancer. Gynecol Oncol 110:S4–S7. identification of clinical isolates of bacteroides by matrix-assisted
Chang DY, Chen RJ, Lee SC, Huang SC. 1997. Prevalence of single and laser-desorption/ionization time-of-flight mass spectrometry. Clin
multiple infection with human papillomaviruses in various grades Microbiol Infect 15:796–802.
of cervical neoplasia. J Med Microbiol 46:54–60. Pagliusi SR, Dillner J, Pawlita M, Quint WGV, Wheeler CM, Ferguson
Clavel C, Cucherousset J, Lorenzato M, Caudroy S, Nou JM, M. 2006. Chapter 23: International Standard reagents for harmo-
Nazeyrollas P, Polette M, Bory JP, Gabriel R, Quereux C, nization of HPV serology and DNA assays – An update. Vaccine
Birembaut P. 2004. Negative human papillomavirus testing in 24:S193–S200.
normal smears selects a population at low risk for developing Paik YH, Han KH, Hong SP, Lee HW, Lee KS, Kim SO, Shin JE, Ahn
high-grade cervical lesions. Br J Cancer 90:1803–1808. SH, Chon CY, Moon YM. 2006. The clinical impact of early detection
de Sanjosé S, Diaz M, Castellsagué X, Clifford G, Bruni L, Muñoz N, of the YMDD mutant on the outcomes of long-term lamivudine
Bosch FX. 2007. Worldwide prevalence and genotype distribution of therapy in patients with chronic hepatitis B. Antivir Ther 11:
cervical human papillomavirus DNA in women with normal 447–455.
cytology: A meta-analysis. Lancet Infec Dis 7:453–459. Perrons C, Jelley R, Kleter B, Quint W, Brink N. 2005. Detection of
Federschneider JM, Yuan L, Brodsky J, Breslin G, Betensky RA, Crum persistent high risk human papillomavirus infections with hybrid
CP. 2004. The borderline or weakly positive Hybrid Capture II HPV capture II and SPF10/LiPA. J Clin Virol 32:278–285.
test: A statistical and comparative (PCR) analysis. Am J Obstet Poljak M, Kocjan BJ, Kovanda A, Lunar MM, Lepej SZ, Planinic A,
Gynecol 191:757–761. Seme K, Vince A. 2009. Human papillomavirus genotype specificity
Gravitt PE, Peyton CL, Alessi TQ, Wheeler CM, Coutlee F, Hildesheim of hybrid capture 2 low-risk probe cocktail. J Clin Microbiol
A, Schiffman MH, Scott DR, Apple RJ. 2000. Improved amplifica- 47:2611–2615.
tion of genital human papillomaviruses. J Clin Microbiol 38:357– Poljak M, Marin IJ, Seme K, Vince A. 2002. Hybrid capture II HPV test
361. detects at least 15 human papillomavirus genotypes not included in
Hong SP, Shin S-K, Lee EH, Kim EO, Ji SI, Chung HJ, Park SN, Yoo W, its current high-risk probe cocktail. J Clin Virol 25:89–97.
Folk WR, Kim S-O. 2008. High-resolution human papillomavirus Quint WGV, Pagliusi SR, Lelie N, de Villiers E-M, Wheeler CM. the
genotyping by MALDI-TOF mass spectrometry. Nat Protoc 3:1476– World Health Organization Human Papillomavirus DNA
1484. International Collaborative Study Group. 2006. Results of the first
Jiang H-L, Zhu H-H, Zhou L-F, Chen F, Chen Z. 2006. Genotyping of World Health Organization International Collaborative Study of
human papillomavirus in cervical lesions by L1 consensus PCR and detection of human papillomavirus DNA. J Clin Microbiol 44:571–
the Luminex xMAP system. J Med Microbiol 55:715–720. 579.
Kailash U, Soundararajan CC, Lakshmy R, Arora R, Vivekanandhan S, Sauer S, Reinhardt R, Lehrach H, Gut IG. 2006. Single-nucleotide
Das BC. 2006. Telomerase activity as an adjunct to high-risk human polymorphisms: Analysis by mass spectrometry. Nat Protoc 1:
papillomavirus types 16 and 18 and cytology screening in cervical 1761–1771.
cancer. Br J Cancer 95:1250–1257. Seme K, Fujs K, Kocjan BJ, Poljak M. 2006. Resolving repeatedly
Kaufman RH, Adam E, Icenogle J, Lawson H, Lee N, Reeves KO, Irwin borderline results of hybrid capture 2 HPV DNA test using polymer-
J, Simon T, Press M, Uhler R, Entman C, Reeves WC. 1997. ase chain reaction and genotyping. J Virol Methods 134:252–256.
Relevance of human papillomavirus screening in management of Sideri M, Cristoforoni P, Casadio C, Boveri S, Igidbashian S, Schmitt
cervical intraepithelial neoplasia. Am J Obstet Gynecol 176:87–92. M, Gheit T, Tommasino M. 2009. Distribution of human papillo-
Kim HS, Han KH, Ahn SH, Kim EO, Chang HY, Moon MS, Chung HJ, mavirus genotypes in invasive cervical cancer in Italy: A represen-
Yoo W, Kim SO, Hong SP. 2005a. Evaluation of methods for tative, single institution case series. Vaccine 27:A30–A33.
monitoring drug resistance in chronic hepatitis B patients during Snijders PJ, Brule AJvd, Meijer CJ. 2003. The clinical relevance of
lamivudine therapy based on mass spectrometry and reverse human papillomavirus testing: Relationship between analytical
hybridization. Antivir Ther 10:441–449. and clinical sensitivity. J Pathol 201:1–6.
Kim YJ, Kim S-O, Chung HJ, Jee MS, Kim BG, Kim KM, Yoon J-H, Lee Solomon D, Davey D, Kurman R, Moriarty A, O’Connor D, Prey M,
H-S, Kim CY, Kim S, Yoo W, Hong SP. 2005b. Population genotyp- Raab S, Sherman M, Wilbur D, Wright T , Young N. Members ftFG,
ing of hepatitis C virus by matrix-assisted laser desorption/ioniz- the Bethesda Workshop. 2002. The 2001 Bethesda System:
ation time-of-flight mass spectrometry analysis of short DNA Terminology for reporting results of cervical cytology. JAMA 287:
fragments. Clin Chem 51:1123–1131. 2114–2119.

J. Med. Virol. DOI 10.1002/jmv

482 Lee et al.

Steben M, Duarte-Franco E. 2007. Human papillomavirus infection: van Ham MAPC, Bakkers JMJE, Harbers GK, Quint WGV, Massuger
Epidemiology and pathophysiology. Gynecol Oncol 107:S2–S5. LFAG, Melchers WJG. 2005. Comparison of two commercial assays
Syrjanen S, Shabalova I, Petrovichev N, Kozachenko V, Zakharova T, for detection of human papillomavirus (HPV) in cervical scrape
Pajanidi J, Podistov J, Chemeris G, Sozaeva L, Lipova E, Tsidaeva I, specimens: Validation of the Roche AMPLICOR HPV Test as a
Ivanchenko O, Pshepurko A, Zakharenko S, Nerovjna R, Kljukina means to screen for HPV genotypes associated with a higher risk
L, Erokhina O, Branovskaja M, Nikitina M, Grunberga V, of cervical disorders. J Clin Microbiol 43:2662–2667.
Grunberg A, Juschenko A, Tosi P, Cintorino M, Santopietro R, Walboomers JMM, Jacobs MV, Manos MM, Bosch FX, Kummer JA,
Syrjanen K. 2004. Acquisition of high-risk human papillomavirus Shah KV, Snijders PJF, Peto J, Meijer CJLM, Muñoz N. 1999.
infections and Pap smear abnormalities among women in the New Human papillomavirus is a necessary cause of invasive cervical
Independent States of the Former Soviet Union. J Clin Microbiol cancer worldwide. J Pathol 189:12–19.
42:505–511. Yeon JE, Yoo W, Hong SP, Chang YJ, Yu SK, Kim JH, Seo YS, Chung
Terry G, Ho L, Londesborough P, Cuzick J, Mielzynska-Lohnas I, HJ, Moon MS, Kim SO, Byun KS, Lee CH. 2006. Resistance to
Lorincz A. 2001. Detection of high-risk HPV types by the hybrid adefovir dipivoxil in lamivudine resistant chronic hepatitis B
capture 2 test. J Med Virol 65:155–162. patients treated with adefovir dipivoxil. Gut 55:1488–1495.

J. Med. Virol. DOI 10.1002/jmv