Reprinted from: Corrected Proof, Available online
19 November 2012
Comparison of cervical cancer screening
strategies incorporating different combinations
of cytology, HPV testing, and genotyping for
HPV 16/18: results from the ATHENA HPV study
J. Thomas Cox, Phillip
E.
Castle, Catherine M. Behrens, Abha Sharma,
Thomas C. Wright Jr, Jack Cuzick, Athena HPV Study Group
ELSEVIER
ISSN 0002-9378
www.AJOG.org
-
-
- - ---=~
~E
·:
S
:
E
.
£
'
RCH
____
---~---
---
W'~v..A.J.OG
.
.
org
GENERAl
,
GYNECOLOGY
Comparison of ceiVical cancer screening strategies
incorporating different combinations of cytology,
HPV testing, and genotyping for HPV 6/18:
results from the ATHENA HPV study
J.
Thomas Cox, MD; Phillip E. Castle, PhD, MPH; Catherine M. Behrens, MD, PhD; Abha Sharma, PhD;Thomas C. Wright Jr, MD; Jack Cuzick, PhD; and the Athena HPV Study Group
OBJECTIVE: The objective of the study was to compare 9 cervical can-cer screening strategies to the current screening standard (cytology with human papillomavirus [HPV] triage of atypical squamous cells of undetermined significance) for the detection of high-grade cervical
disease.
STUDY DESIGN: Women (n
= 34
,254) aged 30 years or older from the Addressing the Need for Advanced HPV Diagnostics (ATHENA) study underwent screening with cytology and HPV testing with simultaneous HPV16/18 genotyping; those with atypical squamous cells of undeter-mined significance cytology or greater or HPV-positive status were re-ferred for colposcopy.RESULTS: In general, screening strategies that offered greater sensitiv
-ity also required more referral to colposcopy. HPV testing was more sensitive than cytology for detection of cervical intraepithelial neoplasia grade 2 or greater, but strategies that depended on cytology for triage of HPV-positive women decreased this sensitivity. Various strategies of
cotesting with cytology increased sensitivity but did so by increasing testing. Strategies that included integrated HPV16/18 testing provided more efficient referral to colposcopy.
CONCLUSION: Strategies that maximize detection of women at greatest risk of cervical intraepithelial neoplasia grade 3 or greater by immediate referral to colposcopy, with follow-up testing of women at intermediate risk, maximize the benefits of cervical cancer screening while decreas-ing the potential harm. Incorporatdecreas-ing screendecreas-ing with HPV and triage of HPV-positive women by a combination of genotyping for HPV16/18 and cytology provided a good balance between maximizing sensitivity (ben-efit) and specificity by limiting the number of colposcopies (potential harm).
Key words: Addressing the Need for Advanced HPV Diagnostics, atypical squamous cells of undetermined significance, cotesting, human papillomavirus, human papillomavirus 16/18, low-grade squamous intraepitheliallesion
Cite this article as: Cox JT, Castle PE, Behrens CM, et al. Comparison of cervical cancer screening strategies incorporating different combinations of cytology, HPV testing, and genotyping for HPV 16/18: results from the ATHENA HPV study. Am J Obstet Gynecol 2012;208:xx-xx.
I
n March 2012, new primary cervicalscreening guidelines were jointly is-sued by the US Preventative Services
Task Force (USPSTF) 1 and a consortium
of the American Cancer Society (ACS),
the American Society for Colposcopy and Cervical Pathology (ASCCP), and the American Society of Clinical
Pathol-ogists (ASCP).2 Based on the evidence
that human papillomavirus (HPV)
test-ing is more sensitive and therefore pro-vides better negative predictive values (NPV) than cytology, these new guide-lines recommend that women aged 30 years and older be screened every 3 years From Student Health, University of California, Santa Barbara (retired), Santa Barbara, CA (Dr Cox); Albert Einstein College of Medicine, Bronx, NY (Dr
Castle); Roche Molecular Systems, Pleasanton, CA (Drs Behrens and Sharrna): the Department of Pathology, Columbia University Medical Center,
New York, NY (Dr Wright): and the Center for Cancer Prevention, Wolfson Institute of Preventive Medicine, Queen Mary University of London,
London, UK (Dr Cuzick).
Received July 20, 2012: revised Oct. 12, 2012: accepted Nov. 14, 2012.
This study was supported by Roche Molecular Systems.
J.T.C. and P.E.C. are compensated for their services as members of a Data Monitoring Board for HPV vaccines for Merck. J.T.C. has received
honoraria from Roche for assistance in the development of an educational slide set and for speaking on the cobas H PV Test. He has also received honoraria for service on the scientific advisory board for Gen-Probe and advisory boards for Graceway and Bradley Pharmaceuticals. P.E.C. has received HPV tests and testing for research at a reduced or no cost from Roche and Qiagen. T.C.W. is compensated for his services as a reference pathologist, clinical adviser, and speaker for Roche, Gen-Probe, BD Diagnostics, and lkonisys. He has received honoraria for serving on a scientific advisory board for HPV vaccines for Merck. J.C. has served on advisory boards for Roche, Gen-Probe, Abbott, and Qiagen. C. M. B. and AS. are employees of Roche Molecular Systems.
Presented at EUROGIN 2012, Prague, Czech Republic, July 8-11, 2011 .
Reprints: J. Thomas Cox, MD, 3345 Numancia St., Santa Ynez, CA 93460. [email protected].
0002-9378/$36.00 • © 2012 Mosby, Inc. All rights reserved. • http:/idx.chi.org/1 0.1 016:j.ajog.2Q12.11 .020
RES
. using cervical cytology alone or every 5
years using a combination of cervical
cy-tology and high-risk HPV testing (re-ferred to as cotesting).
The better NPV of HPV testing per-mits a safe extension of the screening in-terval, thereby reducing harms caused by screening. The ACS/ASCCP/ASCP guidelines endorsed the cotesting option as the preferred approach for women
aged 30 years and older,2 whereas the
USPSTF endorsed it as acceptable, 1 and
the American College of Obstetricians and Gynecologists (ACOG) expressed
support of these recommendations. 3
The ACS/ASCCP/ASCP guidelines rec-ommend that cytology-negative/HPV-positive women undergo follow-up in 12 months with repeat cytology and HPV testing or, alternatively, cytology-negative/ HPV-positive women can be genotyped for HPV 16 and HPV 18? With the latter option, women who are found to have ei-ther HPV 16 or HPV 18 are referred for colposcopy, whereas those without these highest-risk HPV types are cotested again in 12 months.
The Addressing the Need for Ad-vanced HPV Diagnostics (ATHENA) HPV study is a prospective 3 year cervi-cal cancer screening trial designed to compare the performance of the newly introduced cobas HPV Test (Roche Mo-lecular Diagnostics, Pleasanton, CA) both alone and in combination with cer-vical cytology among women aged 21 years and older in the United States. Based on the cross-sectional data from the ATHENA trial, the US Food and Drug Administration recently approved for use in the United States the cobas HPV Test, which detects 11 pooled high-riskHPV genotypes (HPV 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) and 1 pos-sible high-risk type (HPV 66) and con-currently provides separate results for
HPV 16 and HPV 18.4
-7
The current manuscript further ana-lyzes the enrollment results of the ATHENA trial to investigate alternative screening strategies to those endorsed by the most recent US cervical cancer screening guidelines. Ten different cervi-cal cancer screening strategies, including several that use HPV testing alone as the initial screening method, were
investi-gated. The performance of each strategy
for detection of cervical intraepithelial neoplasia grade 2 ( CIN2) or more severe or CIN3 or more severe was explored, as was the potential harm estimated by the number of tests and the number of col-poscopies (a metric used by the recent guideline process) needed to detect each high-grade lesion at baseline. These strategies included results of testing with cytology and/or various combinations of HPV testing, including HPV genotyping for HPV 16 and HPV 18.
MATERIAJ.S AND MFJ'HOLS Study protocol
As previously described, the ATHENA HPV study enrolled more than 47,000 women aged 21 years and older who pre-sented for cervical cancer screening; all eligible participants had both Papanico-laou testing (by liquid-based cytology, ThinPrep; Hologic, Bedford, MA) and HPV testing (by Amplicor HPV test, Lin-ear Array high-risk HPV genotyping test, and the cobas HPV Test, all from Roche
Molecular Systems).6 The protocol was
approved by the institutional review boards at all study sites, and all women provided written informed consent be-fore undergoing any study procedures. The current analysis focuses only on the subset of women aged 30 years old and older to compare alternative manage-ment strategies with those endorsed in the current guidelines.
All women in this subset who had
ei-ther abnormal cytology (atypical squa-mous cells of undetermined significance [ASC-US] or greater) or who tested pos-itive for HPV (by Amplicor or Linear Ar-ray test) were scheduled for colposcopy. In addition, to adjust for ascertainment bias, a randomly selected subset of
women who tested negative for both
cy-tology and HPV had colposcopy. Colpo-scopic biopsies were performed accord-ing to a standardized protocol, and a random biopsy was required in all women with adequate colposcopy in whom no lesion was seen; patients and colposcopists were blinded to the cytol-ogy and HPV results.
An expert central pathology review
( CPR) panel of 3 pathologists read all hi-1.e2 American Journal of Obstetrics & Gynecology MONTH 2012
opsies masked to any clinical data. Women achieving the study endpoint of CIN2 or more severe by CPR exited the study; those who did not reach this end-point proceeded to the 3 year follow-up phase of the study, scheduled to con-clude December 2012. The current anal-ysis is restricted to disease detected at en-rollment; disease detected over the
subsequent 3 years of follow-up will be
analyzed separately.
Screening strategies
The 10 screening strategies were evalu-ated based on review of the published cervical cancer screening literature and appear to be the strategies most likely to be considered potentially attractive by the clinical and public health commu-nities. Strategies 1 and 2 are cytology
screening strategies (Figure l). Strategy
1 consists of screening with cytology with reflex HPV testing (pooled high-risk HPV test for 14 genotypes) of ASC-US and referral of all women with HPV-positive ASC-US or low-grade squamous intraepithelial lesion (LSIL) or greater to colposcopy. Because this is the strategy most widely used in the United States, it serves as the comparator for the other 9 strategies.
Strategy 2 consists of screening with cytology alone, with referral of all women with ASC-US or greater to col-poscopy. Strategies 3, 4, and 5 (Figure 2) incorporate cotesting with both cytology and HPV testing. They vary as to whether genotyping for HPV 16 and HPV 18 is used and by the cytological threshold for referral to colposcopy. Strategies 6 through 10 (Figo.<:::e 3) use HPV testing alone (pooled high-risk HPV test with, or without, genotyping for HPV 16/18) as the initial screening test and differ by which triage tests are used to evaluate HPV-positive women. In the strategies described, women who did not meet the criteria for either imme-diate colposcopy or return to routine screening would be deferred to a 1 year
follow-up per the current guidelines2 as
indicated in ~:igures 2 and 3. Statistical analysis
For each screening strategy, the number
or without integrated HPV 16/18 geno-type detection) required at baseline was calculated, as was the number of colpos-copies required to detect 1 case of CIN2 or more severe or CIN3 or more severe. From the total catchment of CIN2 or more severe and CIN3 or more severe detected, the number of cases not iden-tified at baseline, and an estimate of the number that could potentially be identi-fied by each strategy at 12 month fol-low-up were calculated. The crude sensi-tivity and specificity for detection of CIN2 or more severe or CIN3 or more severe and its sensitivity and specificity relative to strategy 1 were also deter-mined (Tc<.bles 2 and 3).
The cobas HPV Test results were cate-gorized as follows: HPV positive (posi-tive for any of 14 high-risk HPV types); HPV negative (negative for all 14 high-risk HPV types); HPV 16/18 positive (positive for HPV 16 and/or HPV 18, re-gardless of the presence or absence of 12 other HPV types); positive for 12 other HPV types (positive for 1 or more of the 12 other HPV types and negative for HPV 16 and HPV 18).
For calculations of sensitivity and specificity, only those cases in which col-poscopy was performed and a valid bi-opsy result was obtained were consid-ered. Crude estimates are given because the intent was to report on the utility of the strategies as would be observed in a clinical situation. Verification bias ad-justment would not change the relative sensitivities or specificities of the various strategies or represent what happens in clinical practice.
:RES\!ITS
A total of 34,254 women aged 30 years or older were eligible for this analysis; the mean age was 44.7 years, and the demo-graphics are shown in Table 1. Among the eligible women, 2872 (8.4%) tested positive with the cobas HPV Test, and 1966 (5.7%) had abnormal cytology; 280 women were diagnosed with CIN2 or more severe and 189 with CIN3 or more severe. The most sensitive screening strategy was screening with HPV alone (pooled 14 high-risk types) with referral of all HPV-positive women to
colpos-Cytology primary screening options
Strategy 1: Cytology with reflex HPV (ASC-US triage) Routine follow-up
Pap Test Colposcopy
I
Routine
I
~
follow-upHPV Test ~ '---'
I
Colposcopy Strategy 2: Cytology Alone~
I
Routine screeningI
PapTest~~
~Asc-us
I
ColposcopyI
Strategies 1 and 2 screened initially with cytology only. Strategy 1, a recommended screening option for women aged 21 years or older, triages ASC-US by reflex HPV testing, sending to colposcopy ASC-US/HPV positive and more severe than ASC-US (ie, ASC with possible HSIL [ASC-H], LSIL, atypical glandular cells [AGC], and HSIL). Women with normal cytology or ASC-US/HPV negativity continued with routine screening. Strategy 2 sent all women with any abnormal cytology (ie, ASC-US or more severe) to colposcopy, and all women with normal cytology continued with routine screening. This is not a recommended screening option but is included in this study for comparison with other options.
ASC-US, atypical squamous cells of undetermined significance; HPV, human papillomavirus; HSIL, high-grade squamous intraepithelial lesion; LS/L, low-grade squamous intraepitheliallesion.
Cox. Cervical cancer screening strategies: evaluation of results from the A THEN A HPV study. A.m I Obstet Gyneco/2012.
copy (strategy 6), detecting 242 of CIN2 or more severe lesions (86.4%) and 170 (89.9%) of the CIN3 or more severe le-sions (Tables 2 and 3). However, this strategy also had the highest false-posi-tive rate for CIN3 or more severe
(38.0%). In terms of the utilization of colposcopy resources, it was almost as inefficient as the strategy of screening with cytology alone and referring all woman with ASC-US or more severe to colposcopy (strategy 2) since it required 9. 7 colposcopic evaluations to find a sin-gle case of CIN2 or more severe and 13.8 to find a single case of CIN3 or more severe.
All strategies depending on cytology alone or on cytology as the sole reflex test had the lowest sensitivities for detection of CIN2 or more severe and CIN3 or more severe. These included cytology with HPV triage of ASC-US (strategy 1; 51.4% and 56.1%, respectively) and cy-tology alone (strategy 2; 53.2% and 57.7%, respectively). Strategy 3 was also
cytology based because it screened with cytology and HPV testing ( cotesting) but used only the results of cytology and re-flex HPV testing of ASC-US to deter-mine immediate referral to colposcopy. Therefore, this strategy had an identical performance to strategy 1 during the first round of screening, except that cotesting also identified cytology-negative/HPV-positive women in need of additional follow-up in 12 months2 from which 72 additional cases of CIN3 or more severe could potentially be detected (subse-quently described as "cases identified for 12 month flu" [Tables 2 and 3)). HPV alone with cytology triage of HPV-posi-tive women (strategy 7) had the lowest cross-sec!ional sensitivity because the triage test negates the increased sensitiv-ity of HPV testing. In contrast to the lower sensitivity of each of these strate-gies, all but strategy 2 were among the most specific, with false-positive rates for detection of CIN3 or more severe
be-MONTH 2012 American Journal of Obstetrics & Gynecology 1.e3
..
I
Cotesting primary screening options
Strategy 3:
Cotesting with reflex for ASC-USHPVand Pap Test (Cotesting) Routine screening Colposcopy Repeat cotest in 12mo.
Strategy 4:
Cotesting with genotyping and cytology triage: HPV 16/18&
ASC-US HPV+ thresholdHPVwith 16/18 genotyping and Pap Test (Cotesting)
Routine screening
Colposcopy
Repeat cotest in 12 mo.
Strategy 5:
Cotesting with genotyping and cytology triage:HPVwith 16/18 genotyping and Pap Test (Cotesting)
HPV 16/18
&
LSIL thresholdASC-US or NILM/HPV+ but not 16/18 Routine screening Colposcopy Repeat cotest in 12 mo.
Strategies 3-5 screened initially with both cytology and testing for high-risk HPV. Strategies 4 and 5
also utilized the information provided when the HPV test also included separate results for HPV 16 and HPV 18 genotyping or, if not, when reflex genotyping could be done. Strategy 3, 1 of 2 recommended cotesting options for women aged 30 years or older, referred to colposcopy women with ASC-US HPV positivity and LSIL or more severe, irrespective of HPV result, whereas women with cytology-nega-tivity/HPV positivity had repeat cotesting in 12 months and women who were cytology negative/HPV negative and ASC-US/HPV negative continued routine screening. Strategy 4, also a recommended cotesting strategy, referred to colposcopy women who were ASC-US/HPV positive and LSIL or more severe irrespective of HPV result, as well as all women who were cytology negative/HPV 16/HPV 18 positive. Women who were cytology negative/HPV positive but not positive for HPV 16 or HPV 18 had repeat cotesting in 12 months, and women who were cytology negative/HPV negative and ASC-US/ HPV negative continued with routine screening. Strategy 5 was similar to strategy 4 except that the threshold for referral to colposcopy was LSIL or more severe or cytology negative/HPV 16/HPV 18 positive or ASC-US/HPV 16/HPV 18 positive, whereas ASC-US or negative cytology/HPV positive but
1.e4 American Journal of Obstetrics & Gynecology MONTH 2012
tween 8.7 (strategy 7) and 12.4 (strate-gies 1 and 3).
All 3 cotesting options approximately doubled the number of initial screening tests compared with the remaining strategies. Of the cotesting options, cotesting with genotyping triage and an ASC-US threshold (strategy 4) de-tected the most CIN2 or more severe and CIN3 or more severe, with the highest sensitivity (67.5o/o and 76.2o/o, respectively). However, because this option required 32o/o more colposco-pies than cotesting alone (strategy 3), the number of colposcopic evaluations to detect 1 CIN2 or more severe (ap-proximately 6) and 1 CIN3 or more se-vere (approximately 8) was similar. Of the cotesting options, strategy 4 also had the highest relative sensitivity for CIN2 or more severe and CIN3 or more severe ( 1.31 and 1.36, respec-tively) but also had the highest false-positive rate for CIN3 or more severe (18.5o/o). Strategy 5 was similar to strategy 4 but triaged non-HPV 16/ HPV 18 pooled positive women to col-poscopy using an LSIL threshold rather than an ASC-US threshold; this strat-egy was 5.3o/o less sensitive for CIN3 or more severe than strategy 4 but re-duced the number of colposcopies by 14.3o/o (1030 vs 1202). The cotesting strategies required between 5 and 6 colposcopies to detect 1 CIN2 or more severe, approximately the same as cy-tology with HPV triage of ASC-US (strategy 1 ).
All of the screening strategies based on
us-ing HPV alone (pooled 14 high-risk types, with or without HPV 16/18 genotyping) as the initial test required similar numbers of tests compared with the cytology-based strat-egies (from 34,254 to 37,126). Of the 5 HPV strategies, HPV alone (strategy 6) had the highest relative sensitivity and lowest relative specifi.city for CIN2 or more severe ( 1.68 and 0.71, respectively), whereas HPV with
cytol-not HPV 16/HPV 18 positive had 12 month fol-low-up.
ASC-US, atypical squamous cells of undetermined significance;
HPV, human papillomavirus; LSIL, low-grade squamous intraepl-thelialleslon.
Cox. Cervical cancer screening strategies: evaluation of results from the A THENA HPV study. Am J Obstet Gynecol2012.
IGUR
HPV screening algorithms
Strategy 6: HPValone
Strategy 7: H PV with reflex to cytology
HPVTest
Strategy 8: HPV with genotyping
HPVTest and 16/18 Genotyping Routine screening Colposcopy Follow-up in 12 mo.
Strategy 9: HPV with genotyping and reflex cytology: ASC-US threshold
HPVTest and 16/18 Genotyping Routine screening Colposcopy Follow-up
s::
in12mo. Pap Test I .---, <:ASc-us Ll_c_o_lp_o_sc_o_P_Y_Strategy 10: HPV with genotyping and reflex cytology: LSIL threshold
HPVTest
and 16/18
Genotyping
Routine screening
Strategies 6-1 0 screened initially with HPV testing. Strategy 6 referred all HPV-positive women to colposcopy and all negative women to routine screening. Strategy 7 reflex tested all HPV-positive women with cytology, referring to colposcopy only those with ASC-US or more severe, whereas those with negative cytology had follow-up in 12 months. Strategy 8 screened initially with a panel of HPV plus genotyping for HPV 16 and HPV 18, referring to colposcopy all women testing positive for HPV 16 and/or HPV 18 and to 12 month follow-up women positive for other HPV genotypes but not positive for HPV 16/HPV 18. Strategy 9 screened initially with a panel of HPV plus genotyping for HPV 16 and HPV 18, referring all women who were HPV 16/HPV 18 positive to colposcopy and reflex testing by cytology those HPV positive but not positive for HPV 16/HPV 18.
ogy triage (strategy 7) and HPV with gena-typing triage (strategy 8) had the highest rel-ative specificity (each 1.04) but the lowest relative sensitivity (0.92 and 0.85, res-pectively).
Using genotyping to triage HPV-pos-itive women to colposcopy (strategy 8) slightly decreased sensitivity for CIN2 or more severe compared with using cytol-ogy at an ASC-US threshold (strategy 7; 43.6% vs 47.5%) but increased sensitiv-ity for CIN3 or more severe (53.4% vs 51.9%), indicating a trend toward in-creased predictive value ofHPV 16/HPV 18 genotyping as a triage test compared with cytology.
Of all the screening strategies, HPV with cytology triage (strategy 7) and HPV with genotyping triage (strategy 8) required both the least number of col-poscopies (596 and 580, respectively) and the least number of colposcopies to detect 1 CIN2 or more severe (approxi-mately 5). However, both strategies had relatively low baseline sensitivity.
Strategies 8, 9, and 10 used HPV alone as the initial screen and incorporated genotyping triage for HPV-positive women. Strategies 9 and 10 added cytol-ogy triage for HPV -positive women who were HPV 16/HPV 18 negative to in-crease baseline detection of CIN2 or more severe caused by the other 12 HPV types, thereby reducing the number of cases deferred for identification to 12 month follow-up. This produced a gain
Women with any abnormal cytology ASC-US or more severe were also referred to colposcopy, whereas women who were cytology negative/ HPV positive for non-HPV 16/HPV 18 had 12 month follow-up. Strategy 1 0 screened initially with a panel of HPV plus genotyping for HPV 16 and HPV 18, referring all women who were HPV 16/HPV 18 positive to colposcopy and reflex testing by cy-tology those HPV positive but not positive for HPV 16/HPV 18. Women with LSIL or more severe were also referred to colposcopy, whereas women who were cytology negative/HPV positive and ASC-US/ HPV positiVe for non-HPV 16/HPV 18 had 12 month follow-up.
ASC-US, atypical squamous cells of undetermined significance; HPV, human papillomavirus.
Cox. Cervical cancer screening strategies: eval-uation of results from the ATHENA HPV study. Am J Obstet Gynecol 2012.
General Gynecology
Demographic and clinical characteristics of ATHENA
population aged 30 years or older at baseline
Characteristics Age, y Evaluable subjects {n=
34,254) Mean ::!: {SD) 44.7::!: (10.1) ... 30-39, n (%) 12,248 (35.8) :::::40, n (%) 22,006 (64.2) Race, n (%) ... White 28,821 (84.1)Black or African American 4503 (13.1)
...
Asian 503 (1.5)
American Indian or Alaskan native Native Hawaiian or other Pacific Islander 184 (0.5) 78 (0.2) Any combination/missinga 165 (0.5) ... Ethnicity, n (%) Hispanic or Latino 6144 (17 .9) ... Postmenopausal, n (%) 12,743 (37.2) ... HPV vaccine, n (%) 50 (0.1) lmmunocompromised or 224 (0.7) immunosuppressed, n (%) Family history of cervical disease related to cervical cancer, n (%)
...
Yes 1920 (5.6)
No 31 ,988 (93.4)
...
Unknown 346 (1.0)
Pap cytology test in past 5 y, n (%) 31 ,089 (90.8)
ATHENA, Addressing the Need for Advanced HPV Diagnostics; Pap, Papanicolaou.
" "Any combination/missing" refers to participants who selected more than 1 race or for whom the information was missing.
Cox. Cervical cancer screening strategies: evaluation of results from the ATHEN.l HPV study. Am] Obstet Gynecol
2012.
in sensitivity of approximately 14-20%
when compared with that achieved by strategy 8. Of these 2 strategies, strategy 9 was more sensitive than strategy 10 for CIN2 or more severe (63.6% vs 57.9%) and CIN3 or more severe (72.0% vs 66.7%) but slightly less specific (85.2% vs 88.0% and 85.7% vs 88.5%, respec-tively, for CIN2 or more severe and CIN3 or more severe) because of the lower threshold for referral to colposcopy of ASC-US, as opposed to LSIL.
Figure 'l demonstrates in graphical
form the trade-offs between sensitivity for CIN3 or more severe and the number of colposcopies that each strategy would
produce. Presenting the data in this manner delineates 3 groups. One in-cludes a single strategy, HPV alone, which was significantly more sensitive than all the other strategies but is also the least efficient, as measured by the num-ber of colposcopies. In contrast, the 5 strategies grouped with the lowest sensi-tivity for CIN3 or more severe (strategies
1, 2, 3, 7, and 8) required the lowest
number of colposcopies overall, with the
exception of cytology alone (strategy 2). The 4 strategies in the middle group (strategies 4, 5, 9, and 10) demonstrated a narrow range of sensitivity
(66.7-76.2%) but varied by 33o/o in the number
1.e6 American Journal of Obstetrics & Gynecology MONTH 2012
of colposcopies (810-1202). Within this group, cotesting with genotyping and cy-tology triage at the ASC-US HPV-posi-tive threshold (strategy 4) was the most sensitive but also required the most col-poscopies. HPV with genotyping and cy-tology triage at the LSIL threshold ( strat-egy 10) was the least sensitive and required the few est colposcopies. Cot est-ing with genotypest-ing and cytology triage with LSIL (strategy 5) and HPV with genotyping and reflex cytology triage with ASC-US (strategy 9) were very sim-ilar in both sensitivity and number of colposcopies, but strategy 5 required nearly twice the number of initial
screen-ing tests (Tables 1 and 3).
COyf},l tNT
The ideal cervical cancer screening strat-egy would provide maximum sensitivity to minimize missing disease as well as maximum specificity to minimize false positive results and overreferral. Unfor-tunately, cervical cancer screening strat-egies that maximize both sensitivity and specificity have proven elusive because strategies that maximize sensitivity have typically produced relatively poor speci-ficity. The development of invasive cer-vical cancer is typically a slow process that takes decades rather than years to
occur. 8 This fact reduces the relative
benefits of achieving maximum sensitiv-ity if poor specificsensitiv-ity can lead to potential harm.
The recent ACS/ASCCP/ASCP cervi-cal cancer screening guidelines specifi-cally acknowledge that the benefits of screening should be balanced against the potential harms, with the total number of colposcopies serving as the primary
measure ofharm.2 Colposcopy was
cho-sen as the measure of harm because it has been shown to be associated with consid-erable psychological distress, physical discomfort during the procedure, and the potential to lead to more invasive treatments with short- and long-term risks:: The baseline data from the ATHENA study provide us with an ex-ceptional opportunity to evaluate both the benefit and potential harms that would be produced during a single
r
Clinical outcomes of different strategies for detection of CIN2 or more severe
l
Tests
Strategy number and performed, Colposcopies
name n
Cytology with reflex 35,546
HPV (ASC-US triage)
...
Cytology alone 34,254
3 Cotesting with reflex 68,508
for ASC-US 4 Cotesting with 68,508 genotyping and cytology triage: HPV 16/HPV 18 and ASC-US HPV-positive threshold 5 Cotesting with 68,508 genotyplng and cytology triage: HPV 16/HPV 18 and LSIL threshold 6 HPV alone 7 HPV with reflex to cytology 34,254 37,126 8 HPV with genotyping 34,254 9 HPV with genotyping 36,423
and reflex cytology: ASC-US threshold
10 HPV with genotyping 36,423
and reflex cytology: LSIL threshold performed, n 816 ··· 1644 816 1202 1030 2341 596 580 982 810 Colposcopies to detect 1 CIN2 or more severe, n 5.7 ... 11.0 5.7 6.4 6.0 9.7 4.5 4.8 5.5 5.0 CIN2 or more severe cases Identified, n 144 . ... 149 144 189 173 242 133 122 178 162 Cases identified for 12 month follow-up Sensitivity, (estimated), n % 0 51.4 ··· 0 53.2 109 51.4 64 67.5 80 61.8 0 86.4 109 47.5 120 43.6 64 63.6 80 57.9 Sensitivity relative to False-ASC-US positive triage rate, % 1.00 12.0 1.03 26.6 1.00 12.0 1.31 18.0 1.20 15.2 1.68 37.3 0.92 8.2 0.85 8.1 1.24 14.3 1.13 11.5 Specificity relative to ASC-US triage 1.00 0.83 1.00 0.93 0.96 0.71 1.04 1.04 0.97 1.00
ASC-US, atypical squamous cells of undetermined significance; GIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; LS/L, low-grade squamous intraepitheliallesion.
Cox. Cervical cancer screening strategies: evaluation of results from the ATHENA HPV study. Am J Obstet Gynecol2012.
screening strategies based on various combinations of cytology, pooled testing for 12 high-risk HPV types, and geno-typing for HPV 16 and HPV 18.
To evaluate trade-offs between sensi-tivity and potential harms of the differ-ent screening strategies, we used a scat-terplot of each strategy's sensitivity for CIN3 or more severe vs the number of colposcopies required. The scatterplot delineates 3 tiers in the balance between sensitivity for CIN3 or more severe, or benefit of each screening strategy, and the number of colposcopies, a measure of potential harm.
The most sensitive strategy, HPV
alone with referral of all HPV-positive
women to colposcopy (strategy 6), does not appear to offer the right balance be-tween maximum detection of CIN3 or more severe and potential for harm
be-cause it results in 2-3 times more
colpos-copies than the other strategies.
Like-wise, the 5 strategies in the lowest tier
that are the least sensitive for detection of CIN3 or more severe would appear to be less attractive options for clinicians and policy makers, despite the fact that all but 1 strategy (cytology alone) offer the least potential for harm by requiring the lowest number of colposcopies.
How-ever, the low sensitivity of these
strate-gies for detection of CIN3 or more severe at baseline increases the burden of cases that are missed at baseline and that bur-den would depend on potential ibur-dentifi- identifi-cation for those strategies that have 12 month follow-up.
The most attractive strategies from the
perspective of a benefit-vs-harm analysis appear to be the 4 strategies occupying the middle portion of the scatterplot. These strategies all combine what we consider to be a reasonable sensitivity for CIN3 or more severe with only a modest (less than 2-fold) increase in the required
number of colposcopies compared with
the less sensitive strategies. All 4 of the
strategies that occupy the middle portion of the scatterplot use HPV testing with
genotyping for HPV 16/HPV 18, and 2
are cotesting strategies. Of these, cotest-ing with genotypcotest-ing triage (strategy 4) is the most sensitive strategy but re-quires the most colposcopies and twice the initial tests as the non-cotesting options in this group, strategies 9 and 10. HPV with genotyping and cytology
triage with LSIL (strategy 10) requires
the fewest colposcopies and therefore would be most applicable in settings in which potential harm from colposcopy is of greater concern than benefit.
The 2 strategies within this tier that seem to optimize the balance between sensitiyity and specificity are cotesting with genotyping and cytology triage with LSIL (strategy 5) and HPV with genotyp-ing and cytology triage with ASC-US (strategy 9). The latter strategy results in a 50o/o reduction in the number of re-quired screening tests and is also slightly
r
L J
Clinical outcomes of different strategies for the detection of CIN3 or more severe
l
Tests Strategy number and performed,
name n
Cytology with reflex 35,546 HPV (ASC-US
triage}
2 Cytology alone
3 Cotesting with reflex for ASC-US
34,254 68,508 4 Cotesting with 68,508 genotyping and cytology triage: HPV 16/HPV 18 and ASC-US HPV-positive threshold 5 Cotesting with 68,508 genotyping and cytology triage: HPV 16/HPV 18 and LSIL threshold 6 HPV alone 7 HPV with reflex to cytology 8 HPVwith genotyping 9 HPV with genotyping and reflex cytology: ASC-US threshold 10 HPVwith genotyping and reflex cytology: LSIL threshold 34,254 37,126 34,254 36,423 36,423 Colposcopies to detect 1 Colposcopies CIN3 or more performed, n severe, n 816 7.7 1644 15.1 816 7.7 1202 8.3 1030 7.7 2341 13.8 596 6.1 580 5.7 982 7.2 810 6.4
Cases identified Sensitivity Specificity CIN3 or more for 12 month relative to False- relative to severe cases follow-up Sensitivity, ASC-US positive ASC-US identified, n (estimated), n % triage rate,% triage
106 0 56.1 1.00 12.4 1.00 109 0 57.7 1.03 26.8 0.84 106 72 56.1 1.00 12.4 1.00 144 34 76.2 1.36 18.5 0.93 134 44 70.9 1.26 15.7 0.96 170 0 89.9 1.60 38.0 0.71 98 72 51.9 0.92 8.7 1.04 101 69 53.4 0.95 8.4 1.05 136 34 72.0 1.28 14.8 0.97 126 44 66.7 1.19 12.0 1.01
ASC-US, atypical squamous cells of undetermined significance; GIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; LSIL, low-grade squamous intraepitheliallesion. Cox. Cervical cancer screening strategies: evaluation of results from the ATHENA HPV study. Am JOb stet Gynecol2012.
more sensitive and requires slightly fewer colposcopies to detect 1 CIN3 or more severe case.
An obvious limitation of this analysis is that it uses only the baseline data from the ATHENA trial and does not include the results of those women identified as needing 12 month follow-up. Because the design of the ATHENA study re-ferred all women with abnormal screen-ing test results at enrollment for colpos-copy, many women with CIN2 or more severe lesions missed by a given screen-ing test were identified by the other screening test, and their CIN2 or more severe lesions were treated. Therefore, it is impossible to know what percentage of the CIN2 or more severe lesions missed at enrollment by a given screening test would be detected on subsequent screen-ing or follow-up.
In a setting in which it would be
consid-ered unethical to follow up women with known CIN3 or more severe lesions, a ran-domized trial in which women are as-signed to each of the different screening strategies would be required to determine
exactly how subsequent rounds of fol-low-up or screening would perform. We therefore consider it reassuring that our approach of comparing benefits and harms of different screening strategies ar-rived at many of the same conclusions as have analyses incorporating multiple rounds of screening and mathematical modeling studies. For example, in this analysis, a strategy of cytology with HPV triage of ASC-US (strategy 1) is clearly su-perior to cytology alone (strategy 2) be-cause both have similar sensitivities but
with cytology alone almost doubling the number of required colposcopies.
1.e8 American Journal of Obstetrics & Gynecology MONTH 2012
Multiple cost-effectiveness analyses and metaanalyses have previously docu-mented the attractiveness of cytology with triage of ASC-US by HPV over a strategy of cytology with referral of all
women with abnormal cytology results to colposcopy.9-12 Likewise, the group of
screening strategies that appear to be the most attractive after comparing benefits and harms includes the strategy recently classified by the ACS as the preferred strategy when screening women aged 30 years or older (cotesting).2
However, of the 2 cotesting strategies recommended by the ACS/ ASCCP! ASCP, cotest.illg with genotyping triage (strategy 4) detects 26.4% more CIN3 or more se-vere than cotesting with 12 month fol-low-up of all cytology-negative/HPV-posi-tive women (strategy 3) but requires nearly 50% more colposcopies at the
ini-General Gynecology
H
Sensitivity for CIN3 or more severe and number of colposcopies
Primary Screening Strategies For CIN3+ Endpoint
100.0
.
.
90.0 .-
--
a.;.:;t~-.-,.::gy 80.0-
.
-
-
.
-• Sir tegy 4•
I"Jt 11 70.0--
-
-
•
lrn!---
--
-
--
-
. .
•
I~ 11-
,o 6C.O ... -.,..-
-.
+
Stratagy 1 or 3 • :3::r':',~~~\;Y -~.?."
·
s:
50.0 +,.sgategyB trategy 7 :0:0·
u;
c:
(!} 40.0m
Colposcopies. Initial testsStrategy n Sensitlvtty, % performed, n
Strategy 1 Cyiolc.gy wtlh reflex HPV (ASC-US Tnage) 816 561 35.546
30.0 S.:_[-:.-.h:·q_'{·-1 Cytolc.gy alone 11344 57 7 34254
Strategy 3 Cotesttng •.vith reflex for ASC-US 816 56.1 6<3.503
,. Cotesltng wtth genotypmg and c>1ology triage HPV 16/18 and ASCUS HPV+ threshold 1202 7'62 d3,5(l£1
Zii.O Cotes·ltng wrth genotyptng and cytology tnage HPV 16/1C and LSIL threshold 10.30 71) 9 61l.50g •.~ u ";j !l:.~.J !l -:. HPValons 2341 89 9 34,254
Strategy 7 HPV wtlh c'Y!ology \nage 596 51 9 37.126
10.0 Slr<!tegy C HPV wrth genotyping tnage 580 534 34.254
HPV wtlh genotyptng and refleJ< cytology ASC-US thteshold 98< 72 0 36423
,,
HP\/ Wtlh genotyptng and cytology (LSIL cut-off) tnage 810 GB 7 '31'\.42~o.c
0 500 1000 1500 2000 2500
Number of Colposcopies
Scatterplot of sensitivity for CIN3 or more severe and number of colposcopies for each screening strategy. Clear outliers as measured by numbers of colposcopies are in red. Strategies with the lowest sensitivity but also the lowest number of initial colposcopies are in black and those with intermediate sensitivity and number of colposcopies are in blue. Strategies that are presently recommended in the American Society for Colposcopy and Cervical Pathology guidelines are in bold.
ASC·US, atypical squamous cells of undetermined significance; GIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; LS/L, low-grade squamous intraepitheliallesion.
Cox. Cervical cancer screening strategies: evaluation of results from theATHENA HPV study. Am f Obstet Gynecol2012.
tial round of screening. Although this would appear to significantly increase potential harm when compared with cotesting without genotyping, this op-tion obviates the need for repeat cotest-ing in 12 months for women positive for HPV 16/18 and reduces the risk of de-layed diagnosis and or loss to follow-up of women with significant disease. 13•14
Moreover, approximately half of the women who are followed up and un-dergo cotesting 12 months later will re
-quire colposcopy because of persistent high-risk HPV infections.15
Previous studies have shown other
screening strategies to be effective. 14 •16-21
Analysis of data from a Finnish screening trial demonstrated that triage of HPV-positive women by cytology (strategy 7), as in the ATHENA trial, was more spe-cific than conventional cytology screen-ing and decreased colposcopy referrals but, in contrast to our results, was also more sensitive. 17 N aucler et al18 evaluated
11 screening strategies using data from a Swedish screening trial and found that HPV with cytology triage and repeat HPV testing at 1 year of cytology-negative/ HPV-positive women was the most feasi-ble of !he screening strategies. Similarly, a large population-based screening trial in The Netherlands (VUSA-Screen Study) also found HPV with cytology triage and 1 year follow- up by repeat cytology ofHPV-positive women to be a feasible alternative to cytology screening. H
R
I:~I·
,
R
c
I J
General Gynecology
It is difficult to compare these analyses with ours because they are based on 1-5 year follow-up data, and each of these studies had quite different study designs compared with ATHENA. Moreover, neither the Finnish nor The Netherlands trials evaluated strategies that included HPV 16/HPV 18 genotyping. It should be stressed, however, that the ATHENA base-line data do not support a conclusion that HPV with cytology triage is competitive with any of the HPV 16/HPV 18 genotyp-ing strategies that include cytology, unless an absolute reduction in baseline colpos-copies is the primary goal. In addition, an-other recent report from a large screening population in The Netherlands (Popula-tion-Based Screening Study Amsterdam [POBASCAM)) has documented that early detection of CIN3 or more severe as-sociated with HPV 16 was a major compo-nent of the benefit of testing for HPV. 19
Other more complex HPV genotype-based approaches may further improve upon these strategies. In particular, sev-eral studies, including ATHENA, have shown that HPV 16 has a much higher baseline positive predictive value (PPV) than the other high-risk types and that, at least cross-sectionally, HPV 18 has a similar PPV to the pool of 12 other high-risk types:lS,/ This raises the question as to whether HPV 18-positive women would be more efficiently managed by short-term (6-12 months) repeat testing for HPV 18 to establish persistence be-fore referral to colposcopy, similar to the management of women positive for one of the other 12 high-risk HPV types. However, HPV 18 is associated more with cancer than these other types22 and
also is associated with endocervical le-sions that are difficult to detect.
In ATHENA, 3 of the 6 cancers de-tected at baseline were HPV 18 positive, as were 8 of the 16 adenocarcinoma in situ cases and 1 cancer detected in fol-low-up.6 So although the PPV for CIN3 or more severe of a positive HPV 18 test at baseline is lower than typically recom-mended for immediate colposcopy, the greater severity of the disease burden substantiates this approach. Other ques-tions related to the behavior of different HPV types also need to be addressed, such as whether HPV 45, which shows a
similar predilection for adenocarcinoma as HPV 18, should be managed in the same way as HPV 18.
The strategies with 12 month follow-up for women at intermediate risk will
in-crease both the ultimate sensitivity of each of these strategies as well as the total num-ber of colposcopies and must also take into account loss to follow-up. Hence, any strategy selected should consider the trade-offs of sensitivity and specificity in the con-text of real-world clinical practice. As illus-trated by Kitchener et al,13 poor follow-up of HPV -positive women not sent immedi-ately to colposcopy could negate some of the benefits of HPV testing to identify women at risk for CIN2 or more severe while leaving unaffected the added safety of a negative HPV test. Within the context of poor follow-up of screen-positive pa-tients, more sensitive methods of manag-ing HPV-positive women may be pre-ferred over methods with increased specificity and PPV, so that the opportu-nity for immediate detection and treat-ment of precancerous lesions will not be missed.
In conclusion, this analysis demon-strates that multiple cervical cancer screening strategies are more effective than the present standard of cytology screening with ASC-US triage. Strategies that maximize early detection of CIN3 or more severe without excessive increases in initial screening tests and colposco-pies, yet also identify women at interme-diate risk in need of 12 month follow- up, would appear to provide optimal bal-ance between benefit and harms. Of these options, strategies that incorporate initial screening with HPV and triage of HPV-positive women by a combination of genotyping for HPV 16/HPV 18 and cytology may best fulfill these require-ments for more balanced screening, al-though other options may also be com-pelling in different settings. When the 3 year follow-up data from ATHENA be-come available, formal cost -effectiveness modeling can be performed to better de-termine the benefits of cervical cancer screening strategies that incorporate HPV with genotyping for HPV 16/HPV 18 and cytology.
1.e1 0 American Journal of Obstetrics & Gynecology MONTH 2012
ACKNOWLEDGMENTS
We thank Teresa Wright, MD, for her valuable contributions to the design and execution of the ATHENA HPV Study and for assistance in the analysis of data. The ATHENA study testing sites and participants include the following:
United States, Comprehensive Clinical Trials, West Palm Beach, FL; Green Clinic, Ruston, LA; Philadelphia Clinical Research, Philadel-phia, PA; Visions Clinical Research, Boynton Beach, FL; Women's Health Specialist, Costa Mesa, CA; Mount Vernon Clinical Research, At-lanta, GA; Tennessee Women's Care, Nash-ville, TN; Chattanooga Medical Research, Chat-tanooga, TN; 08/GYN Specialists of the Palm Beaches, West Palm Beach, FL; Segallnstitute for Clinical Research, North Miami, FL; South Carolina Clinical Research Center, Columbia, SC; Bluegrass Clinical Research, Louisville, KY;
Delaware Valley OB-GYN and Infertility Group, Plainsboro, NJ; Advanced Research Associ-ates, Corpus Christi, TX; Advanced Clinical Concepts, West Reading, PA; Miami Research Associates, Miami, FL; Center for Women's Health of Lansdale, Lansdale, PA; Blue Skies Center for Women, Colorado Springs, CO; Vi-sions Clinical Research, Tucson, AZ; Impact Clinical Trials, Las Vegas, NV; Physicians' Re-search Options, Lakewood, CO; Four Rivers Clinical Research, Paducah, KY; Medical Net-work for Education and Research, Decatur, GA;
Women's Health Research, Phoenix, AZ; Im-pact Clinical Trials, Los Angeles, CA; HWC Women's Research Center, Englewood, OH; Texas Medical Center, Houston, TX; Mobile 08/GYN, Mobile, AL; Altus Research, Lake Worth, FL; Tacoma Women's Specialist, Ta-coma, WA; Phoenix OB-GYN Association, Moorestown, NJ; The Woman's Clinic, Boise, ID; Impact Clinical Trials, Los Angeles, CA; eCast Corp, North Charleston, SC; State of Franklin Healthcare Associates Research,
Johnson City, TN; Quality of Ufe Medical and Research Center, Tucson, AZ; Eastern Carolina Women's Canter, New Bern, NC; Tidewater Clinical Research, Virginia Beach, VA; St John's Center for Clinical Research, Jacksonville, FL, R. Myers; M and 0 Clinical Research, Ft. Lau-derdale, FL; Lyndhurst Gynecologic Associ-ates, PA, Winston-Salem, NC; Enterprise Women's Center, Enterprise, AL; Salt Lake Re-search, Salt Lake City, UT; Women's Health Care at Frost Street, San Diego, CA; Atlanta North Gynecology Center for Research, Ro-swell, GA; Women's Clinical Research, New-burgh, IN; Jacksonville Canter for Clinical Re-search, Jacksonville, FL; Women's OB-GYN,
Saginaw, Ml; Clinical Research Consultants, Hoover, ALr Edinger Medical Group Research Center, Fountain Valley, CA; Health Awareness, Jupiter, FL; Physician Care Clinical Research,
Sarasota, FL; Woman's Health Practice, Cham-paign, IL; Clinical Trials Management, Caving-ton, LA; Advanced Research Associates, Dal-las, TX; Fellows Research Alliance, Savannah, GA; Fellows Research Alliance, Hilton Head,
" w.AJO<
,,orgSC; Women's Care Florida, Tampa, FL; Ad-vanced Research Associates, McAIIen, TX; Pre -cision Trials, Phoenix, AZ; and Yassear Clinical Research, Carmichael, CA.
REFERENCES
1. Moyer VA, on behalf of the US Preventive Services Task Force. Screening for cervical cancer: US Preventive Services Task Force Recommendation Statement. Ann Intern Med 2012;156:880-91, W312.
2. Saslow D, Solomon D, Lawson HW, et al. Amer
-ican Canoer Society, American Society for Col pos-copy and Cervical Pathology, and American Society for Clinical Pathology screening guidelines for the prevention and early detection of cervical cancer. Am J Clin Pathol2012;137:516-42.
3. American College of Obstetrics and Gyne-cology. Cervical cancer recommendations. Available at: http:/twww.aoog.org/About_ACOGI AnnouncementSINew_Cervicai_Canc:er_Screening_ recommendations. Accessed June 1, 2012.
4. Stoler MH, Wright TC Jr, Sharma A. et al. High-risk human papillomavirus testing in women with ASC-US cytology: results from the ATHENA HPV study. Am J Clin Pathol 2011; 135:468-75.
5. Castle PE, Stoler MH, Wright TC Jr, Sharma
A. WrightTL, Behrens CM. Performance of car -cinogenic human papillomavirus (HPV) testing and HPV16 or HPV18 genotyping for cervical cancer screening of women aged 25 years and older: a subanalysis of the A THENA study. Lan -cet Oncol2011 ;12:880-9.
6. Wright TC Jr, Stoler MH, Behrens CM, Apple R, Derion T, Wright TL. The ATHENA human papillomavirus study: design, methods, and baseline results. Am J Obstet Gynecol 2012; 206:46.e1-11 .
7. Wright TC Jr, Stoler MH, Sharma A, et al. Evaluation of HPV-16 and HPV-18 genotyping
General Gynecology
for the triage of women with high-risk HPV+ cytology-negative results. Am J Clin Pathol 2011; 136:578-86.
8. McCredie MR, Sharples KJ, Paul C, et al. Natural history of cervical neoplasia and risk of invasive cancer in women with cervical intraepi-thelial neoplasia 3: a retrospective cohort study. Lancet Oncol 2008;9:425-34.
9. Kim JJ, Wright TC, Goldie SJ. Cost-effective-ness of alternative triage strategies for atypical squamous cells of undetermined significance.
JAMA 2002;287:2382-90.
10. Kim JJ, Wright TC, Goldie SJ. Cost-effec-tiveness of human papillomavirus DNA testing in the United Kingdom, The Netherlands, France, and Italy. J Natl Cancer lnst 2005; 97:888-95.
11. Kulasingam SL, Kim JJ, Lawrence WF, et al. Cost -effectiveness analysis based on the atypical squamous cells of undetermined signif-icance/low-grade squamous intraepithelial l e-sion Triage Study (AL TS). J Natl Cancer lnst 2006;98:92-1 00.
12. Arbyn M, Sasieni P, Meijer CJ, Clavel C, Koliopoulos G, Dillner J. Chapter 9: clinical ap-plications of HPV testing: a summary of meta-analyses. Vaccine 2006;24(Suppl 3):S3/78, S3/89.
13. Kinney W, Fetterman B, Cox JT, Lorey T, Flanagan T, Castle PE. Characteristics of 44 cervical cancers diagnosed following Pap-neg-ative, high risk HPV-positive screening in routine clinical practice. Gynecol Oncol 2011 ;121: 309-13.
14. Rijkaart DC, Berkhof J, van Kemenade FJ, et al. HPV DNA testing in population-based cer-vical screening (VUSA-Screen study): results and implications. Br J Cancer 2012; 106: 975-81.
15. Rodriguez AC, Schiffman M, Herrero R, et
al. Rapid clearance of human papillomavirus
and implications for clinical focus on persistent infections. J Natl Cancer lnst 2008;100:513-7. 16. Cuzick J, Szarewski A, Mesher D, et al. Long-term follow-up of cervical abnormalities among women screened by HPV testing and cytology-Results from the Hammersmith study. lnt J Cancer 2008;122:2294-300.
17. Leinonen M, Nieminen P, Kotaniemi- Tal-onen L, et al. Age-specific evaluation of primary human papillomavirus screening vs conven-tional cytology in a randomized setting. J Natl Cancer lnst 2009; 101 :1612-23.
18. Naucler P, Ryd W, Tornberg S, et al. Effi-cacy of HPV DNA testing with cytology triage and/or repeat HPV DNA testing in primary cer-vical cancer screening. J Natl Cancer lnst 2009;1 01:88-99.
19. Rijkaart DC, Berkhof J, Rozendaal L, et al. Human papillomavirus testing for the detection of high-grade cervical intraepithelial neoplasia and cancer: final results of the POBASCAM r an-domised controlled trial. Lancet Oncol 2012; 13:78-88.
20. Ronco G, Giorgi-Rossi P, Carozzi F, et al. Efficacy of human papillomavirus testing for the detection of invasive cervical cancers and cervical intraepithelial neoplasia: a ran-domised controlled trial. Lancet Oncol 201 0; 11:249-57.
21. Sankaranarayanan R, Nene BM, Shastri
SS, et al. HPV screening for cervical cancer in rural India. N Engl J Med 2009;360:1385-94. 22. de Sanjose S, Quint WG, Alemany L, et al. Human papillomavirus genotype attribution in invasive cervical cancer: a retrospective cross-sectional worldwide study. Lancet Oncol 201 0; 11 :1048-56.
23. Kitchener HC, Almonte M, Thomson C, et al. HPV testing in combination with liquid-based cytology in primary cervical screening (ARTIS-TIC): a randomised controlled trial. Lancet On-col 2009; 10:672-8.
ELSEVIER
06977324001 Elsevier GmbH Professional Solutions Hackerbrucke 6 · D-80335 Munchen Tel: +49-(0) 89-5383-704 Fax: +49-(0) 89-5383-725 e-mail: [email protected] www.elsevier.de/professional-solutions 951001234No responsibility is assumed by Elsevier, its licensors or associates for any injury
and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages sho~ld be made.
This article reprint is distributed with the support of: Roche Diagnostics GmbH.
