| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Division of Epidemiology and Biostatistics, European Institute of Oncology, Milan, Italy; and the Department of Surgery, Jules Bordet Institute, the Department of Gynecology, Akademisch Ziekenhuis Vrije Universiteit Brussels, and the Prevention and Screening Unit, Public Health School, Catholic University of Louvain, Brussels, Belgium.
Address reprint requests to: Michel Coibion, MD, Jules Bordet Institute, Héger-Bordet Strasse 1, Brussels 1000, Belgium, E-mail: epica{at}ulb.ac.be
 |
Abstract |
|---|
Top
 Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Methods: Five thousand five hundred fifty women 18–91 years old were randomized to cytology plus cervicography or cytology alone. One year later, women were rescreened using both cytology and cervicography. All women with positive lesions were referred for colposcopically directed biopsies and treatment as indicated. We expected to decrease by half the number of cervical intraepithelial neoplasia (CIN) lesions at 1-year screening by adding cervicography to cytology at initial screening.
Results: Compared with cytology screening alone, screening with cytology plus cervicography showed a 30% reduction in CIN I-II-III cervical lesions at 1-year screening round (relative risk [RR] 0.70; 95% confidence interval [CI] 0.32, 1.55, P = .35) and a 43% reduction in CIN II-III cervical lesions (RR 0.57; CI 0.14, 2.16, P = .36). Most lesions detected by one test were not detected by the other. In a retrospective search, nine biopsies (one normal and eight revealing CIN I or more on histopathology) were found to be positive for the human papillomavirus; five of the nine biopsies were found by cytology, three by cervicography, and one by both tests.
Conclusion: The addition of cervicography to cytology in initial screening did not significantly decrease the number of premalignant cervical lesions detected 1 year later, probably because of the transient nature of most of these lesions. Because cytology and cervicography seemed to detect different premalignant lesions, it is possible that cervicography could detect lesions that do not express the cellular abnormalities necessary for detection by cytology.
There is compelling evidence that cervical cytology screening can lead to substantial decreases in incidence and mortality of cervix cancer.1 However, a number of studies have shown that cervical cytology can fail to detect invasive cervix cancers, with negative cytology reports issued on 20–40% of all invasive cervix cancer.2–4 Therefore, other screening methods have been devised, including cervicography,5 a technique often presented as capable of increasing the sensitivity of cervix cancer screening. Cervicography consists of photography of the cervix with a special reflex camera after application of a 5% solution of acetic acid. The pictures (cervicograms)6 are magnified and projected onto a screen. Abnormalities are then sought as with colposcopy.
Cervicography can detect premalignant lesions other than those detected by cytology.7–10 Therefore cervicography might capture some of the false-negative cytology results. We compared the ability of cytology plus cervicography with that of cytology alone to decrease the number of premalignant lesions detected in subsequent screening in a randomized trial.
|
Materials and Methods |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Women were randomized to initial screenings of cytology alone or cytology plus cervicography. Women whose cytology or cervicography was positive at initial screening were referred for colposcopy-biopsy. If the cervix was colposcopically normal, no biopsy was taken. When cervical intraepithelial neoplasia (CIN) was confirmed histopathologically, appropriate treatment was given. Women with CIN lesions at initial screening remained in their respective study arms.
One year after randomization, women were recalled for a second screening round comprising cytology and cervicography because combining the two techniques was the best strategy for a high detection rate. Women positive for one of the two screening tests were referred for colposcopy-biopsy.
The principal study endpoint was the rate of histopathologically confirmed CIN lesions (ie, the true positive lesions) 1 year after the initial screening. The protocol was accepted by ethical committees of the three participating institutions and complied with the Helsinki Declaration. Because the screening procedures used always included cytology and the study did not involve any other intervention than those routinely performed for cervical cancer screening, women were not told they were part of the study.
Previous records showed that frequency of histopathologically confirmed CIN lesions in that population was around 2%. Results from a previous study in a population of same age distribution9 showed that we could expect a 50% decrease in the number of histopathologically confirmed CIN lesions found 1 year after a cytology screening supplemented with the cervicography. So, for a two-sided 5% 
-error level and 80% power, 5666 women had to be randomized.
Statistical testing of differences was based on intent-to-treat. Uncorrected 
2 and Fisher exact test were used for univariate analysis. Relative risks with corresponding 95% confidence intervals (CIs) were calculated to evaluate the effect of supplementing cytology with cervicography. Because frequency of lesions detected was similar in the three participating centers, between-center variation was not calculated.
Randomization was done by opaque, sealed envelopes prepared by a person not involved in the screening procedures and opened consecutively in the consultation room just before screening tests were performed. The envelopes contained computer-generated random numbers that determined study arms. The selected code was stapled to the personal file of each woman.
Cytology and cervicography were performed by 20 general practitioners or gynecology fellows. Initial and second screenings generally were performed by different physicians. Assessments of cervicograms and cytology specimens were separate processes, without any mutual influence. Histopathologists, cytologists, and cervicogram readers were blinded to study arms.
Cervicogram readers were colposcopists trained to recognize cervical abnormalities using the classification system of the National Testing Laboratory Co. (ie, defective, negative, atypical, positive cervicograms). Our interpretation of cervicograms was based on visual impression of the reader and standardized criteria for image interpretation that were defined by consensus and reproducible between readers. These criteria will be the subject of a future publication. According to our reading procedure, a portion of acetowhite positive lesions were atypical and thus not referred for colposcopic evaluation. A cervicogram was defective if the reader could not see the cervical transformation zone correctly (eg, because of bleeding due to smear sampling or of lack acetic acid application).
Testing for human papillomavirus (HPV) was done by staff blinded to randomization arms and results of cytology, cervicography, or histologic examination. Papillomavirus DNA sequences in biopsy specimens were studied using the hot-start polymerase chain-reaction DNA amplification method on paraffin-embedded tissue, as previously described, with the L1 consensus primers MY11 (positive stand) and MYO9 (negative stand).11–13 Tissue specimens of condyloma acuminata of vulvar and cervical carcinoma, previously proved to contain HPV types 6, 11, 16, 18, and 33, served as positive controls. After amplification, to provide increased sensitivity of detection over ethidium bromide staining, we also used DNA enzyme immunoassay detection with specific biotinylated HPV DNA probes for HPV 6, 11, 16, 18, and 33, as described by Sangiuolo and coworkers.14
|
Results |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
. Two thousand seven hundred seventy-two women were randomized in the cytology-alone arm and 2778 in the cytology-plus-cervicography arm. The two study arms were well balanced for the major baseline covariates (Table 1). Because women included in the study were already part of the regular screening routines installed in the three participating institutions, there was no difference between groups in prior abnormal cytology results.
|
|
) yielded a nearly equivalent number of positive cytology results in the two arms (13 and 12). In the cytology-plus-cervicography arm, 101 (3.6%) women had positive cervicograms. No woman was positive for both cytology and cervicography. Missing work-up data in the initial screening round were due to colposcopy biopsies done in other health facilities from which, after reasonable efforts, we could not obtain follow-up results. Six times more histopathologically proven premalignant lesions were found in the cytology-plus-cervicography arm than in the cytology-alone arm because a notable number of CIN I lesions were detected by cervicography.
|
shows results of the second screening round. Age and parity distribution remained well-balanced between arms. Rescreening rates and average time between initial and second screening were equivalent. The second screening, using both cytology and cervicography, found 17 true positive lesions in the cytology-alone arm and 12 true positive lesions in the cytology-plus-cervicography arm. Adding cervicography to cytology 1 year later resulted in a statistically nonsignificant 30% reduction in the number of women with premalignant lesions of the cervix. The absolute reduction was lower for CIN I lesions than for CIN II-III lesions, 20% and 43%, respectively. No women found with CIN lesions at initial screening had CIN lesions at second screening because of treatment. A comparable number of lesions were detected 9–12.4 months or 12.5–33 months after initial screening (data not shown).
|
shows which tests identified the premalignant lesions found in the study. The greater ability of cervicography to detect CIN I lesions resulted in significantly more low-grade lesions in the cytology-plus-cervicography arm (27 versus 10 for both screening arms, P < .001). For the two screening rounds, only one true-positive lesion was detected by cytology and cervicography at second screening. With specific regard to CIN II-III lesions, the number of smear-detected lesions remained constant no matter which randomization arm or screening strategy was used (a total of seven in each arm for the two screening rounds). This indicates that at initial screening, cervicography did not affect smear-detectable lesions; each screening test seemed to work independently of the other, picking up different lesions.
|
that in the cytology only arm, at second screening cervicography alone detected four CIN II-III lesions that had not been identified by two successive cytology screenings 1-year apart. That is 36% (100 [4 ÷ 11]) of all CIN II-III lesions discovered in that arm.
In light of the estimated false-negative rates, we were interested in how screening tests compared in the detection of HPV infection. Therefore, we retrospectively conducted a search for HPV on biopsy materials from 71 women, irrespective of screening round. The number of biopsies is less than the number of positive screening tests because the work-up policy was to avoid random biopsies when colposcopy was normal and satisfactory. Also, some work-up procedures had been done outside the three participating institutions, and that biopsy material could not be recovered. Table 5 shows the results from HPV testing. Apart from one lesion positive for HPV 33, each screening test identified distinct HPV-positive lesions. Most biopsies without CIN lesion were the result of false-positive cervicographic screening. Only one negative biopsy tested positive for HPV.
|
, 3, and 4 show that during the study, the total number of false-positive results was 12 for cytology and 152 for cervicography (ie, all women who were positive for a test minus the true-positive lesions found by that test minus the missing work-up data). |
Discussion |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Adding cervicography to cytology at initial screening achieved a six-fold increase in detection of histopathologically proven CIN lesions (Table 2). Why such a higher detection rate did not translate to a more substantial reduction in the number of CIN lesions discovered 1 year later remains unknown (Table 3).
One explanation arises from the highly transient nature of CIN I lesions. Because cervicography has a much greater ability than cytology to detect CIN I lesions,7–10 the initial screening using both tests succeeded in identifying the majority of women carrying a CIN I lesion. However, some of the cervicographically detected CIN I lesions spontaneously regressed during the screening interval, while new CIN I lesions developed in other women. The same phenomenon could apply to the CIN II-III lesions, although they are less transient than the CIN I lesions.
Another explanation is that cytology and cervicography are complementary in finding CIN lesions. We expected that at initial screening, cervicography would detect premalignant lesions that were not yet detectable or missed by cytology but that would become detectable by cytology during further screening rounds. However, with one exception, lesions detected by one test were missed by the other, and in the cytology-alone arm, two successive screening rounds missed 36% of CIN II-III lesions. The retrospective assessment of HPV infection in biopsy materials (Table 5) strengthened our impression that cytology and cervicography were complementary methods for detecting clinically relevant premalignant lesions.
Although our findings should be verified in a group of women at higher risk to develop cervix cancer, we contemplated the hypothesis that cervicographically detectable premalignant lesions are of another phenotype that does not express the cellular abnormalities necessary for detection by cytology. False-negative problems of cytology would consist of sampling or reporting failures and, in some instances, of lesions with malignant potential that do not develop in a way easily detectable by cytology.
Our hypothesis is supported by the observation that in all retrospective audits of screening histories of women who suffered from cervix cancer, a number of good quality smears remained negative on rereading, and many smears reclassified as positive had abnormalities difficult to recognize. Further evidence comes from the age distribution of cervical cancers diagnosed after negative cytology. If some lesions with malignant potential are not recognized by cytology, the incidence of interval cervical cancers should increase with age. Appropriate data for examining this question were found in four studies.15–18 In all four studies, the incidence of interval cervical cancer was higher among older than younger women. In two studies,17,18 the highest rate of interval cancer in older women was noticeable among women who had two negative cytologies, compared with women who had had only one negative cytology, a detail that reinforces the likelihood of our hypothesis.
Cervicography might be able to detect the HPV-positive premalignant lesions of the cervix that are not detected by cervical cytology. Future studies in high-risk populations combining cytology, HPV testing, and cervicography are needed to better assess the extent to which cytology and cervicography are complementary in detecting HPV-positive cervical lesions with the potential to evolve into invasive cancer. Future efforts also must aim to reduce the false-positive results from cervicogram reading.
|
Footnotes |
|---|
Supported by a grant from the Europe Against Cancer Programme of the European Commission.
Received March 5, 1998. Received in revised form August 31, 1998. Accepted September 17, 1998.
|
References |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
2. Mitchell H, Giles G. Cancer diagnosis after a report of negative cervical cytology. Med J Aust 1996;164:270–3.[Medline]
3. Sasieni PD, Cuzick J, Lynch-Farmery E. Estimating the efficacy of screening by auditing smear histories of women with and without cervical cancer. Br J Cancer 1996;73:1001–5.[Medline]
4. Slater DN, Milner PC, Radley H. Audit of deaths from cervical cancer: Proposal for an essential component of the national screening programme. J Clin Pathol 1994;47:27–8.
5. Stafl A. Cervicography: A new method for cervical cancer detection. Am J Obstet Gynecol 1981;139:815–25.[Medline]
6. Coppleson M, Monaghan JM, Morrow CP, Tattersall MHN. Gynecologic oncology: Fundamental principles and practice. 2nd ed. London: Churchill Linvingstone, 1992.
7. Mitchell H. Is there a role for screening by cervicography in Australia? Med J Aust 1989;150:441–4.[Medline]
8. Szarewski A, Cuzick J, Edwards R, Butler B, Singer A. The use of cervicography in a primary screening service. Br J Obstet Gynaecol 1991;98:313–7.[Medline]
9. Coibion M, Autier P, Vandam P, Delobelle A, Huet F, Hertens D, et al. Is there a role for cervicography in the detection of the premalignant lesions of the cervix uteri? Br J Cancer 1994;70: 125–8.[Medline]
10. De Sutter P, Coibion M, Vosse M, Huet F, Wesling F, Wayenberg M, et al. A multicenter study comparing cervicography and cytology in the detection of cervical intraepithelial neoplasia. Br J Obstet Gynaecol 1998;105:613–20.[Medline]
11. Manos MM, Wright DK, Lewis AJ, Broker TR, Wolinsky SM. The use of polymerase-chain-reaction amplification for the detection of genital papillomaviruses. In: Furth M, Greaves M, eds. Molecular diagnostics of human cancer, cancer cells 7. New York: Cold Spring Harbor Laboratory, 1989: 209–4.
12. Nuovo G. PCR in situ hybridization: Protocol and applications. New York: Raven Press, 1992:63–81.
13. Ting Y, Manes MM. Detection and typing of genital human papillomaviruses. In: Innis MA, Gelfand DH, Sninsky JJ, White T eds. PCR protocols: A guide to methods and applications. San Diego: Academic Press, 1990:356–71.
14. Sangiuolo F, De Santis L, Cavicchini A, Angeloni U, Romanini C, Novelli G. A new method for direct analysis of PCR amplified human papillomavirus using DNA enzyme immunoessay (DEAI). Int J Clin Lab Res 1994;24:223–6.[Medline]
15. Mitchell H, Medley G, Giles G. Cervical cancers diagnosed after negative results on cervical cytology: Perspective in the 1980s. BMJ 1990;300:1622–6.
16. van Oortmarssen, Habbema JFD. Cervical cancer screening data from two cohorts in British Columbia. In: Screening for cancer of the uterine cervix. Hakama M, Miller AB, Day NE, eds. IARC Scientific Publication No. 76. Lyon, France: International Agency for Research on Cancer, 1986: 47–60.
17. Choi NW, Nelson NA. Results from a cervical cancer screening programme in Manitoba, Canada. In: Hakama M, Miller AB, Day NE, eds. Screening for cancer of the uterine cervix. IARC Scientific Publication No. 76. Lyon, France: International Agency for Research on Cancer, 1986: 61–8.
18. Pettersson F, Näslund I, Malker B. Evaluation of the effect of Papanicolaou screening in Sweden: Record linkage between a central screening registry and the National Cancer Registry. In: Hakama M, Miller AB, Day NE, eds. Screening for cancer of the uterine cervix. IARC Scientific Publication No. 76. Lyon, France: International Agency for Research on Cancer, 1986:91–108.
This article has been cited by other articles:
|
|
 |
|
  C. Ferreccio, M. C. Bratti, M. E. Sherman, R. Herrero, S. Wacholder, A. Hildesheim, R. D. Burk, M. Hutchinson, M. Alfaro, M. D. Greenberg, et al. A Comparison of Single and Combined Visual, Cytologic, and Virologic Tests as Screening Strategies in a Region at High Risk of Cervical Cancer Cancer Epidemiol. Biomarkers Prev., September 1, 2003; 12(9): 815 - 823. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
