ANTICANCER RESEARCH 25: 3469-3480 (2005)

Comparing PAP Smear Cytology, Aided Visual Inspection, Screening , and HPV Testing as Optional Screening Tools in Latin America. Study Design and Baseline Data of the LAMS Study*

K. SYRJÄNEN1, P. NAUD2, S. DERCHAIN3, C. ROTELI-MARTINS4, A. LONGATTO-FILHO5, S.TATTI6, M. BRANCA7, M. ERZEN8, L.S. HAMMES2, J. MATOS2, R. GONTIJO3, L. SARIAN3, J. BRAGANCA3, F.C. ARLINDO4, M.Y.S. MAEDA5, A. LÖRINCZ9, G.B. DORES10, S. COSTA11 and S. SYRJÄNEN12

1Department of Oncology and Radiotherapy, Turku University Central Hospital, Turku, Finland; 2Hospital de Clinicas de Porto Alegre; 3Universidade Estadual de Campinas, Campinas; 4Hospital Leonor M de Barros, Sao Paulo, Brazil; 5Instituto Adolfo Lutz, Sao Paulo, Brazil and University of Minho, Braga, Portugal; 6First Chair, Gynecology Hospital de Clinicas, Buenos Aires, Argentina; 7Unit of , National Centre of Epidemiology, Surveillance and Promotion of Health, National Institute of Health (ISS), Rome, Italy; 8SIZE Diagnostic Center, Ljubljana, Slovenia; 9Digene Corp., Maryland, U.S.A.; 10Digene Brazil, Sao Paulo, Brazil; 11Department of Obstetrics and Gynecology, S. Orsola-Malpighi Hospital, Bologna, Italy; 12Department of Oral Pathology, Institute of Dentistry, University of Turku, Finland

Abstract. Objectives: This is a European Commission (EC)- positive in any test and for 5% of women with baseline PAP- funded ongoing study known as the LAMS (Latin American negative and 20% of HCII-negatives. All high-grade lesions Screening) study, where PAP smear/liquid-based cytology and (CIN2/3) were treated, and low-grade CIN are prospectively screening colposcopy were compared with i) three optional followed-up. Results: Of the 12,107 women, the following screening tools [visual inspection with acetic acid (VIA), or baseline data are available: epidemiological data (n=11,996), Lugol’s iodine (VILI), cervicography] and with ii) Hybrid conventional PAP smears (n=10,363), LBC, SurePATH Capture II from a) conventional samples and from b) self- (n=320), LBC, DNA-Citoliq (n=1,346), VIA (n=12.067), samples, in women at different risk for in Brazil VILI (n=3,061), cervicography (n=279), screening colposcopy and Argentina. Study Design: During 2002-2003, a cohort of (n=3,437), HCII conventional (n=4,710), HCII self-sampling 12,107 women attending four clinics: Campinas (CA), Sao (n=246) and cervical biopsies (n=1,524). The four sub- Paulo (SP), Porto Alegre (PA) and Buenos Aires (BA), were cohorts differ significantly in all their baseline data on the interviewed for risk factors, and examined using the 8 implicated risk factors of cervical cancer, consonant with their diagnostic arms. Colposcopy was performed for women origin from regions with different cancer incidence. Around 95% of all PAP smears were negative, with slight variations in the prevalence of LSIL and HSIL between the four centers. Significant differences were found in the detection rates of *LAMS: Latin American Screening Study, funded by the European abnormal findings in VIA, VILI and colposcopy between the Commission, INCO-DEV Contract # ICA4-CT-2001-10013 four centers (p=0.0001). The prevalence of HPV was practically identical (16.5-18.8%) in all four cohorts Correspondence to: Prof. Kari Syrjänen, MD, Ph.D., FIAC. SMW (p=0.486), with no differences in the relative viral loads. Consultants Ltd., Kylliäisentie 9, FIN-21620 Kuusisto, Finland. Biopsy results were different depending on whether the women Tel: +358-2-2557145, Fax: +358-2-2557178, e-mail: kari.syrjanen@ smwconsultants.fi underwent screening colposcopy (BA) or elective colposcopy (others). Conclusion: Four cohorts with significantly different Key Words: Screening, HPV, HPV testing, colposcopy, liquid-based baseline data are available, and prospective follow-up of these cytology, cervicography. women permits analysis of whether variations in cervical

0250-7005/2005 $2.00+.40 3469 ANTICANCER RESEARCH 25: 3469-3480 (2005)

Figure 1. Flowchart of patient examination, treatment and follow-up.

cancer incidence in these regions is due to i) different natural solutions to cope with this increasing problem (5-8, 10, 13). history of the precursor lesions, or ii) due to different levels of A variety of optional screening tools have been introduced exposure to the known risk factors. to be used in conjunction with, or independently of, the (7, 8, 10, 12, 13). These include: visual inspection Cervical cancer has an uneven geographic distribution, with with acetic acid (VIA) or Lugol’s iodine (VILI), the vast majority of cases being confined to regions where cervicography, , colposcopy, liquid-based- and the resources to combat the disease are the most meager, automated cytology, all under rigorous testing in different i.e., in developing countries (1-4). There is no argument that settings (7, 8, 10, 12-18). the declining trends in incidence and mortality rates Testing for Human papillomavirus (HPV) by different witnessed in developed countries over the past four decades molecular tools (Hybrid Capture, PCR) has been proposed are mainly attributable to the implementation of organized as an adjunct or independent screening tool, with several screening programs based on the cervical Pap smear (5-8). potential advantages (3, 7, 8, 12, 17, 19-24). Testing for the The best examples are the Nordic Countries, where etiological agent of cervical cancer (3, 4, 19, 25-27) offers organized screening has resulted in up to an 80% reduction an opportunity to detect the women at increased risk of in cervical cancer incidence since the early 1960’s (7-12). cervical cancer at the stage of latent and subclinical HPV Unfortunately, these highly effective organized screening infection, preceding (by several months to years) the clinical programs exist only in a few countries, and the prospects for stages (SIL, CIN) detectable by the PAP test, which makes effective PAP smear screening in the majority of the cervical cancer unique among all human malignancies (3, 4, developing countries seem gloomy, if not entirely 7, 8, 12, 13, 19 ,24 ,25, 28, 29). pessimistic, in the foreseeable future (7, 8, 10, 12, 13). This On the global scale, the countries of Latin America are fact has been well recognized among the scientific among those with the highest incidence and mortality rates community, emphasizing the necessity to find other of cervical cancer (1, 2, 7, 8, 10). Despite some local efforts

3470 Syrjänen et al: Testing Optional Screening Tools in Latin America to establish cervical cancer screening, no organised national Second visit. Women testing positive with any of these screening programs exist in any of these countries, techniques were examined by colposcopy at the second visit. The including Brazil and Argentina (7, 8, 17, 23, 30). Thus, cytology cut-off for referral included women with ASC-US, "abnormal" being the referral cut-off for VIA, VILI, cervicography there is no demonstrable trend to decrease the cancer and screening colposcopy. In addition, a 5% random sample of all morbidity in these two countries (1, 2, 30). These test-negative (PAP, VIA, VILI) women will be submitted to conditions prompted us to design the present study to test colposcopy, as will be 20% of those testing negative with HCII (at eight different screening tools in a large cohort of women 24 months), to assess the incidence rates of PAP smear enrolled in four clinics in regions with different incidences abnormalities and HPV infections. When subjected to colposcopy, of cervical cancer (30, 31). these baseline negative women will contribute to correction of the This multi-center trial has two main aims: i) to evaluate verification bias, otherwise inevitable in this type of study design, where only a minority of the patients are verified by the gold the feasibility of eight different diagnostic tests as screening standard (=colposcopic biopsy). tools in these settings, and ii) to test the hypothesis that the On colposcopy, all abnormal findings were confirmed by different incidence of cervical cancer in these regions directed punch biopsies. The result of the punch biopsy was used as depends on a) the different clinical course of cancer the gold standard for the cervical pathology, against which all the precursors, or b) on the different exposure of these women other diagnostic tests will be compared while calculating their to the known risk factors, e.g. HPV. The former is necessary performance (sensitivity, specificity, negative- and positive- to find out the cost-effective tools for cervical cancer predictive values). The women with biopsy-confirmed low-grade screening in these low-resource settings, while the latter CIN (HPV– or HPV+) comprise the cohort to be prospectively followed-up for a minimum of 24 months, to elucidate the disease offers the possibility of further elucidating the natural outcome. All high-grade lesions were promptly treated and history of cervical cancer and its precursors in women at followed-up for the same period, using repeated PAP test, VIA and different risk for cancer. HCII assay at 6-month intervals, and were subsequently The present communication describes the study design colposcopied if any of these tests were positive. and the baseline data of this multi-center trial, known as the Latin American Screening study (the LAMS study). Patients. In the first phase, the four clinics examined a total of 12,107 women, between February 2002 and June 2003, comprising Patients and Methods the LAMS Study cohort, from which all the baseline data are derived. The mean age of the women was 37.9 years (range 14-67; General study design. The ongoing LAMS (Latin American Median 37.7 yrs). Altogether, 74% of the women were Caucasian, Screening) study is a multi-center screening trial targeted at female 9% were colored, 16% mixed, and less than 1% of other origin. populations at different risk for cervical cancer in two Latin The key clinical characteristics and epidemiological data recorded American countries, Brazil and Argentina (32). In this study, from these patients at their first clinical visit are provided in Table funded by the INCO-DEV Programme of the European I. The institutional ethics committees of all four clinics separately Commission (EC) (ICA4-CT2001-10013), cervical cytology approved the study design and all the examination protocols. All (conventional and Liquid Based Cytology, LBC) is compared with patients gave their written consent to participate in the study. The 1) four optional screening tools suggested for low-resource settings: original intention was to enroll four cohorts of women, who a) visual inspection with acetic acid (VIA), b) visual inspection with differed in their baseline data, since they derived from populations Lugol’s iodine (VILI), c) cervicography and d) screening at different risk for cervical cancer. The special features of these colposcopy); and 2) with the new molecular diagnostic tools (HPV four clinics and their target populations are briefly described. testing by Hybrid Capture II), performed a) from physician- Hospital de Clinicas de Porto Alegre. This department is located in collected samples and b) from those taken by self-sampling devices. one of the major hospitals in the south of Brazil, linked to the Thus, eight different diagnostic tests are compared in a cohort of Federal University of Rio Grande do Sul (RS). RS State has the over 12,000 women, enrolled as detailed below. highest quality of life in Brazil, but still a relatively high incidence of First visit. The LAMS study is a combination of a population- cervical cancer (23/105) (30, 31). This department runs an outpatient based, cross-sectional (prevalence) study and a prospective cohort clinic with an average of 30,000 patients from the surrounding study of women enrolled in regions with different (low, community attending each year with all gynecological and obstetric intermediate, high) incidence of cervical cancer in these two indications. In the late 1990’s, a programme for cervical cancer countries. The patient flow-chart is illustrated in Figure 1. In the prevention was started in the region, and this clinic has become a first phase, consecutive series of women at their first visit to the reference center recognised by the State Ministry of Health and the four clinics: Campinas (CA), Sao Paulo (SP), Porto Alegre (PA) national authorities. Before the start of the LAMS study, almost and Buenos Aires (BA), were screened for cervical HPV infections 3,000 patients had been evaluated in these pilot studies. and CIN, using eight different diagnostic tools: 1) conventional Hospital Maternidade Leonor Mendes de Barros. This is a public PAP, 2) LBC, 3) VIA, 4) VILI, 5) cervicography, 6) colposcopy, 7) hospital located in the east side of downtown Sao Paulo, Brazil. HCII conventional and 8) HCII self-sampling. All four clinics used The hospital has a dual role; a) assistance of patients with the conventional PAP test, VIA and HCII. VILI was done in PA gynecological and obstetrical problems, and b) serves as a training only, while BA is the only clinic performing screening colposcopy. unit for residents in this speciality. The patients (n=50,000 per LBC and HCII self-sampling are done exclusively in SP, while year) coming from the surrounding poor areas are considered as cervicography is only tested in CA. medium-risk women for cervical cancer, well above the ASIR of

3471 ANTICANCER RESEARCH 25: 3469-3480 (2005)

Table I. Baseline characteristics of the study population enrolled in the four clinics.

Clinics Clinics Characteristics Porto Sao CampinasBuenos Significance Characteristics Porto Sao CampinasBuenos Significance Alegre Paulo Aires Alegre Paulo Aires

Number Years of enrolled 3.043 3.000 2.627 3.437 hormonal Age (M±SD) 41.3±10.7 36.9±9.9 35.4±11.8 37.5±11.5 p=0.0001 contraception 12.2±7.2 5.7±4.9 5.3±4.7 4.0±4.0 p=0.0001 Marital status: p=0.0001 Single 23.4% 28.4% 29.7% 43.1% Previous STD: p=0.0001 With partner 76.6% 71.6% 70.3% 56.9% No 85.4% 93.9% 90.9% 88.9% Yes 13.0% 5.7% 7.8% 4.7% Years of Not applicable 1.6% 0.4% 1.4% 6.4% Education 8.1±3.6 7.3±3.5 7.6±4.2 11.1±4.1 p=0.0001 Previous Race: p=0.0001 PAP smear: p=0.0001 Caucasian (white)73.4% 67.8% 67.2% 87.3% Yes 90.7% 93.7% 89.0% 85.4% Afro (black) 15.0% 11.0% 9.1% 0.6% Never 9.3% 6.3% 11.0% 14.6% Mixed 11.6% 18.8% 23.1% 11.9% Other 0.1% 2.4% 0.5% 0.8% No. of life-time PAP 7.6±6.8 6.3±5.0 7.9±6.6 7.0±6.7 p=0.0001 Age at sexual debut 18.7±4.1 18.5±4.0 18.2±4.0 18.4±3.8 p=0.0001 Time from previous Pregnant: p=0.0001 PAP 28.0±28.7 20.4±23.119.8±21.9 25.3±26.3 p=0.0001 No 98.1% 99.3% 92.5% 94.5% Yes 0.3% 0.1% 4.8% 3.2% History of previous Not applicable 1.6% 0.6% 2.7% 2.3% HPV-related pathology: p=0.0001 No. of Vulvar 3.4% 1.1% 2.3% 0.4% pregnancies 2.6±2.0 2.7±2.1 2.4±2.3 2.1±2.1 p=0.0001 Anal warts 0.5% 0.1% 0.3% 0% Oral warts 0.6% 0% 0% 0% No. of CIN 2.0% 1.0% 2.7% 0.1% deliveries 1.5±1.6 1.6±1.9 1.5±2.2 1.2±1.6 p=0.0001 Carcinoma 0.1% 0% 0% 0% No history 93.6% 97.8% 94.7% 99.5% No. of C. sections 0.5±0.8 0.7±0.9 0.6±0.9 0.2±0.7 p=0.0001 Smoking history: p=0.0001 Never 59.7% 64.5% 65.3% 59.6% No. of Yes 21.0% 20.4% 18.3% 28.4% abortions 0.6±1.0 0.4±0.9 0.3±0.8 0.6±1.0 p=0.0001 Yes, in the past 19.3% 15.1% 16.4% 12.0%

No. of life-time Years of sexual partners 2.9±4.1 2.6±4.1 2.4±2.9 3.0±4.3 p=0.0001 current smoking 18.7±10.4 17.0±9.4 16.3±10.6 13.2±10.0 p=0.0001 No. of partners during previous 12 months 0.9±0.5 1.0±0.5 1.0±0.6 1.1±2.4 p=0.001 No. of cigarettes/day 12.1±8.0 11.9±9.0 11.9±7.8 10.1±8.5 p=0.0001 Any partner with STD: p=0.0001 Years of Yes 14.8% 7.4% 7.9% 2.6% past smoking 10.9± 9.1 10.7±8.6 10.4±9.0 8.8±7.9 p=0.001 No 66.7% 90.6% 86.2% 82.5% Not applicable 18.5% 2.0% 5.9% 15.0% Past, cigarettes/day 10.9±9.8 11.1±10.711.7±11.1 11.1±11.0 p=0.658 Mode of contraception: p=0.0001 Years since No 7.3% 28.8% 19.0% 31.1% stopped smoking9.5± 8.2 7.8±6.9 9.2±8.2 6.3±7.0 p=0.0001 Hormonal 54.3% 25.7% 26.9% 13.9% Condom 11.8% 16.1% 10.3% 33.2% History of IUD 4.6% 6.7% 5.0% 7.2% drug abuse Sterilization 9.5% 18.3% 19.7% 1.4% Yes (n=109) 1.4% 0.3% 2.2% 0% p=0.0001 Other 12.5% 4.5% 19.0% 13.1% No 98.6% 99.7% 97.8% 100%

3472 Syrjänen et al: Testing Optional Screening Tools in Latin America

Table II. Results of the PAP smears at the first visit. was started in the Campinas area as an initiative of the Obstetrics and Gynecology Department, UNICAMP. The program has Clinics currently expanded to cover 47 cities in the Campinas area, integrating several centers at the secondary and tertiary level of PAP results Porto Sao Campinas Buenos Signi- health care. The Laboratory of Cytology at CAISM-UNICAMP Alegre Paulo Aires ficance receives well over 200,000 PAP smears every year, and the women N=3.005 N=1.372 N=2.552 N=3.434 attend the Unit of Cervical Pathology for additional examinations. Primera Catedra de Ginecologia, Hospital de Clinicas Jose de San Conventional PAP: p=0.0001 Martin. This clinic represents a colposcopy reference center Inadequate 0.0% 1.3% 0.9% 2.0% included in the Department of Obstetrics and Gynecology, at the Normal 91.7% 87.1% 85.9% 85.0% Buenos Aires University Hospital (Hospital de Clinicas). The clinic N-WEC 4.0% 6.9% 7.7% 8.9% ASCUS 2.1% 2.6% 3.8% 1.5% currently holds the leading position in the whole of Argentina, and LSIL 1.1% 0.9% 1.1% 1.2% is the referral center for the treatment of lower genital tract diseases HSIL 1.0% 0.4% 0.5% 0.9% in the Buenos Aires metropolitan area. The clinic operates on an AGUS 0% 0.9% 0% 0.1% outpatient basis, with a regular daily workload of approximately 30 SCC 0.1% 0% 0.1% 0.3% colposcopies every morning. The daily routine includes examination AC 0% 0.1% 0% 0% and triage of new patients referred to by other clinics. It is among this patient material that the present cohort was enrolled in BA, SurePATH: N=320 supplemented by the series of (n=850) women enrolled during a Inadequate 0.6% directed screening campaign in the last months of 2002. Normal 79.4% N-WEC 15.9% Epidemiological data. At the first visit, epidemiological data were ASCUS 1.6% collected by questionnaires focused on the known and suspected LSIL 1.9% risk factors of HPV, CIN and cervical cancer. The data recorded HSIL 0.6% by this questionnaire in the four clinics are summarized in Table I. AGUS 0% SCC 0% PAP smear. Cervical cytology was tested by three modes; AC 0% conventional and two different LBC techniques. Conventional PAP smears were taken by all centers, while LBC was tested in only one N=33 N=1.308 N=2 N=3 clinic (SP). The technique followed the conventional procedures DNA-Citoliq: p=0.987 for smear taking, fixation and staining. Interpretation of the smears Inadequate 0% 0.6% followed the Bethesda 2001 system (33). Cytology was subjected to Normal 90.9% 82.9% 100% 100% external quality control (QC), organised by one of the EC partners N-WEC 9.1% 6.5% (to be reported separately). ASCUS 0% 6.7% LSIL 0% 1.8% Liquid-based cytology (LBC). Two different systems of liquid-based HSIL 0% 0.8% cytology (LBC) were tested: DNA-Citoliq® (Digene Brazil, Sao AGUS 0% 0.6% Paulo, Brazil) and SurePath® (formerly AutoCytePREP®) SCC 0% 0.1% (TriPath, Durham, NC, USA). Both methods were designed for AC 0% 0% cytological screening and tested in previous studies (34, 35). SurePath®. The samples were collected with the SurePath® LBC N-WEC, Normal without endocervical cells; SCC, squamous cell system using the brush and transported into a vial of liquid carcinoma; AC, adenocarcinoma. preservation solution. The thin-layer slides were prepared following the Autocyte® protocol and stained by Papanicolaou method (34). The residual liquid-preserved samples were stored and used for analysis of DNA ploidy (35). cervical cancer in South-East Brazil, which is lowest in the country DNA-Citoliq®. The samples were collected using the brush of (18/105) (30). The Outpatient Clinic also includes the Cervical the DNA-Citoliq® System and immersed in the UCM (Universal Pathology Unit, attended by some 4,000 women per year, with Collecting Medium, Digene, US) vials. The sample processing steadily increasing numbers as a result of an organised campaign followed the manufacturer’s instructions. In brief, the samples in addressed to women with no access to at least one PAP test every the tube were homogenized in high-speed vortex for 20 seconds, three years. The laboratory tests of the SP partner were performed and a 200-Ìl aliquot was placed on a polycarbonate membrane, by the Pathology Division of Adolfo Lutz Institute. 25 mm in diameter and 5 Ìm of porosity, with uniform distribution Centre of Integral Care of Women’s Health. This hospital over the total area of the membrane. The slides and membranes (CAISM) is connected with the State University of Campinas were placed in a system with the capacity to process 12 samples (UNICAMP), located at the outskirts of the city of Campinas, simultaneously (Lamigene®, Digene Brazil) (35). The slides, Brazil. This hospital offers all diagnostic and therapeutic services mounted in the Lamigene, were placed in a metal box (Prepgene) for cervical cancer and its precursor lesions as well as for other and the cover is closed for 10 seconds with a constant pressure. The gynecological malignancies. An average of 250 new cases of pre- end result was a slide with a homogeneous "imprint" of the sample. invasive lesions and 100 new cervical carcinomas are detected and The slides were fixed in absolute ethanol and stained with the treated every year. In 1968, a cervical cancer screening program conventional Papanicolaou method (35).

3473 ANTICANCER RESEARCH 25: 3469-3480 (2005)

Table III. Results of the visual inspection at the first visit in the four Table V. Results of Hybrid Capture II assay in the four clinics. clinics. Clinics Clinics Results of Porto Sao Campinas Buenos Signi- Results of Porto Sao Campinas Buenos Signi- Hybrid Alegre Paulo Aires ficance VIA & VILI Alegre Paulo Aires ficance Capture 2 N=3.039 N=2.999 N=2.592 N=3.437 (HCII) assay (2 pg/ml) VIA: p=0.0001 N=1.104 N=807 N=1.440 N=940 Normal 83.7% 87.0% 93.3% 90.3% HC II assay Abnormal 16.1% 12.8% 6.6% 9.4% at 1st visit: p=0.154

Suggesting cancer 0.1% 0.1% 0.2% 0.3% Positive 15.4% 16.5% 18.8% 16.9% Negative 84.6% 83.5% 81.2% 83.1% N=3.039 N=0 N=22 N=0 Viral Load 29.6± 38.7± 33.8± 48.4± p=0.223 VILI: p=0.038 (RLU/CO) 170.9 244.7 183.5 267.3 (M±SD) Normal 77.1% 100% N=263 N=0 N=26 N=0 Abnormal 22.8% HCII assay at colposcopy: p=0.262 Suggesting cancer 0.2% Positive 30.4% 38.5% Negative 69.6% 61.5% Viral load (RLU/CO) Table IV. Results of colposcopy in the four clinics (M±SD) 107± 175± p=0.336 325 473 Clinics

Results of Porto Sao Campinas Buenos Signi- Colposcopy Alegre Paulo Aires ficance

Screening acid and, after 1 minute, visualized with a 100W light. At the second Colposcopy: N=3.437 step, the cervix is painted with Lugol’s iodine. A positive test is recorded in cases with altered acetic acid (VIA) and/or Lugol’s (VILI) Normal 65.4% staining. The objective is not to establish the diagnosis, but to Abnormal 34.6% distinguish normal from abnormal. Here, we distinguished three categories: normal cervix, abnormal cervix and cervix with suspected Significant cancer (36). All positive tests were controlled by colposcopy and biopsy. abnormality* 18.6% Cervicography. Cervicography, as an independent diagnostic N=1.038 N=594 N=1.068 technique, was described in 1981 by Stafl (37). This test basically Elective involves the examination of magnified photographic documentation Colposcopy: p=0.0001 of the cervix impregnated with acetic acid (37, 38). Preparation of the cervix is done in a similar way as for colposcopy. The cervix is Normal 72.8% 64.0% 80.1% visualized in a self-retaining speculum, and a specifically designed Abnormal 27.2% 36.0% 19.9% hand-held camera (Cerviscope® , National Testing Laboratories, NTL, MO, USA) is used to take the photographs (cervigrams), Significant using a special film. An electronic data bank permits accurate abnormality* 5.3% 6.9% 7.3% p=0.0001 identification of the slides (cervigrams) of each patient, classifying *Dense acetowhite lesion, coarse mosaic, coarse punctuation. them as technically defective (no diagnosis can be defined), negative, atypical and positive (38). In this study, all cervigrams were interpreted by one of the authors, who is an NTL-authorized reader of cervigrams (S. Costa, Bologna, Italy). These results are still pending at the time of writing. Visual inspection (VIA, VILI). Originally introduced for purposes of down-staging as unaided visual inspection (UVI), the technique has Screening colposcopy. The value of colposcopy as a screening tool been developed to include the application of acetic acid, known as VIA has long been disputed (39). Because of its lower specificity, (36). In completing VIA, the uterine cervix is painted with 5% acetic colposcopic screening cannot, in most cases, compete with the

3474 Syrjänen et al: Testing Optional Screening Tools in Latin America

Table VI. Results of the first cervical biopsies taken in the four clinics. and diagnosed using the standard CIN nomenclature. For the study purposes, the pathologists were also asked to notify the Clinics morphological changes suggestive for the presence of HPV in cases with no CIN, i.e., HPV-NCIN (=flat condyloma). Results of Porto Sao Campinas Buenos Signi- The slides from two of these centers (CA, SP) have been Cervical Biopsies Alegre Paulo Aires ficance subjected to re-examination by a panel of pathologists from EC countries (ME, KS). The consensus diagnosis of the panel was N=278 N=243 N=214 **N=789 N=1.524 considered to be the final diagnosis, and also comprised the specific diagnostic categories used in classifying cervical pathology. Normal 67.3% 31.3% 43.5% 72.4% All laboratories were also instructed to prepare special slides for HPV-NCIN 4.0% 5.3% 4.7% 15.6% immunohistochemical (IHC) analysis of specific biomarkers, to be Condyloma 2.2% 0.4% 0.9% 0.1% CIN I 12.6% 10.3% 17.3% 2.0% completed for all CIN lesions and carcinomas. From these patients, CIN II 4.3% 4.1% 7.9% 1.6% scrapings for RNA extraction were also collected and stored for CIN III 8.6% 4.5% 4.2% 5.2% future analysis by cDNA-microarrays for the global gene expression Invasive SCC <5 mm 0.4% 1.2% 0.9% 0.0% patterns (to be reported separately). Invasive SCC >5 mm 0.7% 1.2% 0.5% 2.8% VAIN 0.0% 0.0% 0.0% 0.3% Statistical analysis. Statistical analyses were performed using the Metaplasia *0.0% 22.6% 16.4% *0.0% SPSS® computer program package (SPSS for Windows, version Reactive atypia 0.0% 3.7% 0.5% 0.0% 11.5). In examining these baseline data, frequency tables were Microglandular analyzed using the Chi-square test, with the likelihood ratio (LR) hyperplasia 0.0% 1.2% 0.5% 0.0% significance test between the categorical variables. OR and 95% Subclinical HPV confidence intervals (95% CI) were calculated where appropriate. or HPV suspicion 0.0% 4.5% 2.8% 0.0% Differences in the means of continuous variables between the Chronic inflammation 0.0% 9.1% 0.0% 0.0% groups were analyzed using non-parametric tests (Mann-Whitney, CINIII & AIS 0.0% 0.4% 0.0% 0.0% Kruskal-Wallis) or ANOVA, after careful control of the normal distribution (Kolmogorov-Smirnov test with Lilliefors correction). Total 100.0% 100.0% 100.0% 100.0% p=0.0001 In this report, logistic regression models were used to analyze the power of different variables as predictors of some of the outcome *Not recorded as a separate entity; **Screening colposcopies variables both in univariate (crude OR and 95% CI) analysis. In all tests, the values p<0.05 were regarded statistically significant.

Results organised cytological screening in costs. The costs of colposcopy, however, vary significantly from country to country, and the cost- The baseline data collected from the 12,107 women, effectiveness of colposcopy needs to be evaluated separately in stratified according to the four centers, are summarized in each individual setting. Thus, screening colposcopy was included Table I. These four cohorts (ranging from 2,627 up to 3,437 among the diagnostic tools tested in the LAMS study. Colposcopy women) statistically differed from each other significantly in was performed systematically for all women examined at the BA most recorded variables. Women examined in PA were clinic. In classifying the colposcopic patterns, the terminology of significantly older than the others. The frequency of being the International Federation of Cervical Pathology and Colposcopy single was highest among the BA women, who were also (IFCPC) was followed (40). better educated (measured by years of education). Due to HPV testing. HPV testing was carried out by the Hybrid Capture 2 the geographic differences, the distribution of races was also (HCII) assay, using cervical swabs (collected by a physician) and different in the four cohorts (p=0.0001). The age of onset self-sampling devices (tampons), as described previously (41). The of sexual activity was significantly different (p=0.0001). The HCII assay was performed using the automated HCII test system, highest prevalence of pregnant women was encountered according to the manufacturer’s protocol (24, 29). The samples among those in CA, followed by BA. The four cohorts also were analysed only for the presence of high-risk HPV types 16, 18, differed in the number of pregnancies, deliveries, Cesarean 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68. The usual limit of 1 pg/ml of HPV16 DNA was used as the positive control (CO). sections and abortions. Samples were classified as high-risk HPV-positive, if the relative Women in BA reported the highest mean number of life- light unit (RLU) reading of the luminometer was equal to or time sexual partners (LSP), which probably was due to their greater than the mean of CO values, i.e., RLU/CO ≥1.0 pg/ml older mean age rather than promiscuity, because age and being the cut-off for test positivity (24, 29). LSP were significantly correlated up to the age 35 (Spearman rho 0.072, p=0.001; linear regression analysis Cervical biopsies. Directed punch biopsies (and cone biopsies) were p=0.0001). The frequency of partners with diagnosed STD fixed in formalin, embedded in paraffin and processed into 5-Ìm- thick Hematoxylin-eosin (HE)-stained sections for light microscopy, was highest in PA (14.8%), but much less elsewhere, and following the routine procedures. All biopsies were examined on a indeed very low in the BA cohort. The same applies to the daily routine basis in the Pathology Departments of the four clinics, past history of HPV-related pathology.

3475 ANTICANCER RESEARCH 25: 3469-3480 (2005)

The modes of contraception were different in these four referred due to another positive test; PAP, VIA, VILI, cohorts as was the number of years with oral contraception. cervicography). No difference was found between the two In both aspects, PA differed from the other centers most clinics (PA, CA) where such samples were available. markedly, since over 50% of the women used hormonal The results of the baseline cervical biopsies are given in contraception, and the mean number of years used was Table VI. The differences reflect different practices in more than twice higher than elsewhere. The frequency of categorizing the lesions rather than marked true differences previous STDs was also highest among PA women. The four in the prevalence of significant pathology. Detection of series also differed significantly as to the previous PAP high-grade CIN was quite similar in all clinics, as was CIN II smear taken, the life-time number of PAP smears and the and CIN I, by taking into account that BA systematically time elapsed since the last smear. seemed to grade these lesions one grade lower, i.e., CIN I Concerning smoking history, the highest proportion of with HPV as HPV-NCIN etc. Similarly, moving metaplasia current smokers was found among women in BA (28.4%), (not recorded in PA and BA) into the category of normal and the lowest in CA (18.3%). The mean number of years cervix would render the figures of that category very similar. of current smoking was lowest in BA, however. The other smoking-related variables, with the exception of the number Discussion of cigarettes daily in ex-smokers, were also significantly different in the four cohorts. In the entire series, there were The necessity to develop optional diagnostic tools for 109 drug abusers. Drug abuse was significantly associated cervical cancer screening, particularly in low-resource with current and past smoking (48.6%; 32.1%)(p=0.0001), settings, is widely recognized (1, 4, 6-8, 10, 12, 13, 16, 17). with a previous history of STD (OR 3.45, 95% CI=2.20- Such potential optional screening tools include VIA, VILI, 5.42), as well as with HCII-positive test (OR 2.94, 95% HPV testing, cervicography and, possibly, screening CI=1.73-5.00). colposcopy (16, 17, 19-23, 30). There is no argument that The results of the first PAP smear are summarized in organized cytology screening is the only cost-effective means Table II, stratified according to the clinics, and shown of cervical cancer control, and should be used as the gold separately for conventional PAP and LBC. In conventional standard to which the other screening technology should be PAP smears, the distribution of the different diagnostic compared (7-9). It is also clear that a direct one-to-one categories was significantly different in the four series, most transfer of a screening program, that works well in one notably due to different rates of ASCUS and HSIL and, to country, to a completely different setting in some (low- a lesser extent, due to inadequate smears and those with resource) country is not a realistic mode of action (6-8, 10, AGUS. LBC was tested in SP only, and the comparison of 13, 17, 28). Instead, the only feasible option to establish a the results with the conventional PAP showed a significant cost-effective screening program in an individual country is difference (p=0.0001; data not shown). through extensive comparison of the optional tests in these The results of VIA and VILI in the four clinics, as individual settings (7, 8, 10, 12, 13, 17, 28, 30). Under such depicted in Table III, indicate that VIA was practiced by all circumstances, the effectiveness of e.g. VIA, cervicography, clinics, whereas VILI was done only in PA. The prevalence colposcopy, HPV testing or LBC, as compared with the of abnormal findings in VIA was almost three-fold higher conventional PAP test screening, will be determined by a in PA as compared to SP. The proportion of "suggesting large number of factors, including a) the prevalence of HPV cancer" category varied between 0.1 and 0.3%. infections and CIN in the target population, b) the accuracy The results of colposcopy are summarized in Table IV, of cytological diagnosis, as well as c) the costs and separately for the screening mode and elective mode. There applicability of these diagnostic tests (7, 8). was a significant difference between the three clinics where Several multi-center trials, testing these alternative elective colposcopy was done (p=0.0001), in that the screening tools in different settings, are currently ongoing. frequency of abnormal findings was almost twice as high in These include the ALTS study (42), the POBASCAM study SP than in CA, at 36% and 19.9%, respectively. The results (43), the HART study (44), the TOMBOLA project (45), of screening colposcopy in BA were almost identical with the French Cohort studies (46) and the recently started those of elective colposcopy performed in SP (p=0.259). Canadian Cervical Cancer Screening Trial (CCCaST) (47). HPV detection data with HCII in the four centers are We have recently concluded another multi-center trial shown in Table V. The prevalence of HPV was surprisingly testing HCII, PCR and PAP smear as screening tools in constant in these four cities, ranging from 15.4% to 18.8%, three New Independent States of the former Soviet Union with no statistical significance (p=0.154). The same is true (the NIS Cohort Study) (48), from which new data are of the mean viral loads, measured by the RLU/CO values. currently emerging (28). The design of all these major The HPV detection rates were higher when HCII testing studies is principally quite similar; population-based cohorts was done on the occasion of elective colposcopy (for women of women, in whom usually 2-3 different diagnostic tests are

3476 Syrjänen et al: Testing Optional Screening Tools in Latin America compared in a screening setting. The cohort size in many of baseline data on cervicography and HPV testing with self- them is also surprisingly similar, i.e., around 12,000 women sampling are not available yet. In the first PAP smear, some as in our LAMS study (32, 42, 44, 47). The LAMS study is 95% of the samples were negative in all four clinics (Table unique, however, in several respects: i) it compares eight II). There were slight differences in ASCUS and HSIL different diagnostic tests instead of two to three as in the frequencies, whereas that of LSIL was practically constant. other studies, ii) the target women are derived from These figures are similar to those reported in other geographic areas with a different incidence of cervical population-based studies (30, 42-48). There were some cancer, and iii) the study has a prospective component differences in the ASCUS, LSIL and HSIL rates between testing the value of several new biomarkers as predictors of conventional PAP and LBC performed in one of the clinics, the natural history of cancer precursors. as discussed in a separate report (50). While designing the LAMS study, the rationale was to build The results of VIA evaluation were different in the four up a series that would enable testing of the hypothesis, sub-cohorts, with the highest frequency (16%) of abnormal whether the differences in cervical cancer incidence between findings reported in PA (Table III). The category the distinct regions of Brazil and Argentina could be explained "suggesting cancer" was rare (0.1-0.3%) in all clinics. As by a) the different exposure of the women to the known risk compared with other studies using VIA, these detection factors (e.g. HPV), or b) by the divergent natural history of rates represent a good average between the reported the cervical precursor lesions. Accordingly, the Brazilian extremes (17, 18, 36). Of note is the increase in abnormal partners represent two different regions, South Brazil (Porto findings using VILI as compared to VIA in the same Alegre) and South-East Brazil (Sao Paulo, Campinas) (30, patients, 22.8% and 16.1%, respectively. The detailed data 31). The Argentine partner clinic is from downtown Buenos of VIA and VILI as well as their performance Aires (i.e., Entre Rios Province), where the national registry characteristics in detection of biopsy-confirmed high-grade reports cervical cancer incidence of 32/105 in year 2000. These lesions will be reported separately. figures are above the country average, because increasing The role of colposcopy as a screening tool has been numbers of a low-income population have moved to the city heavily disputed (39, 40). The opinion among the leading during the past several years (49). colposcopists favors the idea that colposcopy should not be Between February 2002 and June 2003, a total of 12,107 used as a screening tool. The technique suffers from women were examined by the four clinics (32), fulfilling the inherent low specificity, and colposcopic screening cannot original goal of having four sub-cohorts of 3,000 women compete in costs with organized cytological screening in enrolled from each region. As shown by the baseline data most settings. Naturally, this depends on the cost of (Table I), these four sub-cohorts differ very significantly colposcopic examination, which, in most developed (p=0.0001) or significantly (p=0.001) in most key countries, is the cost-critical diagnostic tool. It is well characteristics recorded as known or implicated risk factors known, however, that the costs of colposcopy vary of HPV, CIN and cervical cancer. These included their significantly from country to country (e.g. 2 USD in Brazil, marital status, racial distribution, level of education, sexual but 100-200 USD in USA and Europe) (30, 31, 40). Thus, if behavior and obstetric history, modes of contraception, PAP the performance of colposcopy proves to be even close to smear history, history of STDs and HPV-linked pathology, that of other techniques, it is possible that screening as well as smoking history. In addition, the cohort also colposcopy might be a viable and cost-effective screening comprised enough drug abusers (n=109) to calculate tool in these settings. meaningful statistics, e.g. by using a nested case-control Several studies in Brazil and Argentina confirm the high study with 1:4 age-matched case-control design (to be prevalence of HPV in cervical carcinomas and, like reported). Considering the patients from SP and CA to elsewhere, high-risk HPV types 16 and 18 predominate in represent a "low-incidence" area in this parameter, those cervical, vulvar and penile carcinomas (49, 51-53). A large- from Porto Alegre an "intermediate-risk" area, and women scale prospective cohort study, run in collaboration between from BA deriving from a "high-incidence" area, these data Ludwig Institute (Sao Paulo) and McGill University can be stratified according to the known cancer risk (to be (Montreal), has recently provided important data on the reported separately). This type of analysis, combined with natural history of HPV infections (51). Depending on the the prospective follow-up of all the low-grade precursor target populations, the HPV prevalence in these two lesions, should enable us to define whether any differences countries varies, but the figures in screening populations in the natural history could be disclosed in these women at seem to be surprisingly similar to the baseline HPV data of different risk for cervical cancer. the LAMS study (Table V). Interestingly, these prevalence Here, only the findings in the first PAP smear, VIA and rates (16-18%) are lower than those (25%-35%) detected VILI, colposcopy, cervical biopsy and HCII testing are among the screening populations in Russia, Belarus and summarized, without any further analysis of the data. The Latvia, by our NIS Cohort study (28, 48). In all future

3477 ANTICANCER RESEARCH 25: 3469-3480 (2005) calculations of the test performance characteristics, the 7 Miller AB, Nazeer S, Fonn S, Brandup-Lukanow A, Rehman R, baseline cervical biopsies (Table VI) will be used as the gold Cronje H, Sankaranarayanan R, Koroltchouk V, Syrjänen K, Singer standard. To correct the verification bias, 20% of all test- A and Onsrud M: Report on Consensus conference on Cervical Cancer Screening and Management. Int J Cancer 86: 440-447, 2000. negative women will be invited to a second HCII after 2 8 Franco E, Syrjanen K, de-Wolf C, Patnick J, Ferenczy A, years, while 5% of all negative women were directly referred McGoogan E, Bosch X, Singer A, Munoz N, Meheus A et al: to colposcopy as a quality-control procedure (Figure 1). New developments in cervical cancer screening and prevention. In conclusion, the ongoing LAMS study is another multi- Geneva, Switzerland, June 17-19, 1996. Workshop. Cancer center trial testing eight different diagnostic tests as potential Epidemiol Biomarkers Prev 5: 853-856, 1996. screening tools in a cohort of over 12,000 women enrolled in 9 Hakama M: Screening for cervical cancer: Experience from the two Latin American countries, Argentina and Brazil (32). Nordic Countries. In: Franco E, Monsonego J (eds.): New The design of the LAMS study permits a detailed analysis of Developments in Cervical Cancer Screening and Prevention. Blackwell Science, Oxford, pp. 190-199, 1997. a number of important issues, mandatory for designing new 10 Ponten J, Adami H-O, Bergström R, Dillner J, Friberg, L-G, strategies for a cost-effective programme to control cervical Gustafsson L, Miller AB, Parkin DM, Sparen P and Trichopoulos cancer (6-8, 10-13). In addition, the LAMS study design D: Strategies for global control of cervical cancer. Int J Cancer provides an opportunity to follow-up the patients with low- 60: 1-26, 1995. grade lesions and obtain important data on the natural 11 Sigurdsson K: The Icelandic and Nordic cervical screening history of HPV infections in women from distinct regions of programs: trends in incidence and mortality rates through 1995. the continent, with different incidence of cervical cancer. Acta Obstet Gynecol Scand 78: 478-485, 1999. 12 Syrjänen K: Early detection of CIN, HPV and prevention of Combined with the modern molecular diagnostic tools cervical cancer. In: Syrjänen K and Syrjänen S. Papillomavirus applied to the biopsies, a unique opportunity is offered to Infections in Human Pathology. Chapter 8. J. Wiley & Sons, examine the pathogenetic mechanisms and biological New York, pp. 252-280, 2000a. behavior of HPV-associated cervical cancer precursors in 13 Syrjänen K, Erzen M and Syrjänen S: Cervical cancer control these women. These follow-up data should also permit by organised screening. Issues to be considered in designing a identification of the low-risk patients, to be followed-up only, national programme. Kolposkopia 2: 95-116, 2001. which should have a major impact on the health economy 14 McGoogan E: Liquid-based cytology: the new screening test for issues in these countries. Even more importantly, this cervical cancer control. J Fam Plann Reprod Health Care 30: 123-125, 2004. approach enables us to elucidate whether the different 15 Confortini M, Bonardi L, Bulgaresi P, Cariaggi MP, Cecchini cervical cancer incidence in these regions is due to i) a S, Ciatto S et al: A feasibility study of the use of the AutoPap different natural history of the precursor lesions, or ii) a screening system as a primary screening and location-guided different level of exposure to the known risk factors. rescreening device. Cancer 99: 129-134, 2003. 16 Cronje HS, Parham GP, Cooreman BF, de Beer A, Divall P Acknowledgements and Bam RH: A comparison of four screening methods for cervical neoplasia in a developing country. Am J Obstet This study has been supported by the European Commission, Gynecol 188(2): 395-400, 2003. INCO-DEV Programme (Contract# ICA4-CT-2001-10013). The 17 Miller AB, Sankaranarayanan R, Bosch FX and Sepulveda C: generous contribution of DIGENE Inc. (USA) in donating the Can screening for cervical cancer be improved, especially in HCII tests at our disposal is gratefully acknowledged. developing countries? Int J Cancer 107(3): 337-340, 2003. 18 Ngelangel CA, Limson GM, Cordero CP, Abelardo AD, Avila References JM and Festin MR: Acetic-acid guided visual inspection vs. cytology-based screening for cervical cancer in the Philippines. 1 Ferlay J, Bray F, Pisani P and Parkin DM: GLOBOCAN 2000: Int J Gynaecol Obstet 83(2): 141-150, 2003. Cancer Incidence, Mortality and Prevalence Worldwide, 19 IARC Monographs on the Evaluation of Carcinogenic Risks to Version 1.0. IARC Cancer Base No. 5. Lyon, IARC Press, 2001. Humans. Vol 64. Papillomaviruses. Lyon. IARC: pp. 1-409, 1995. 2 Parkin DM, Pisani P and Ferlay J: Estimates of the worldwide 20 Petry KU, Menton S, Menton M, van Loenen-Frosch F, de incidence of 25 major cancers in 1990. Int J Cancer 80: 827-841, Carvalho Gomes H, Holz B et al: Inclusion of HPV testing in 1999. routine cervical cancer screening for women above 29 years in 3 Syrjänen K and Syrjänen S: Papillomavirus Infections in Human Germany: results for 8466 patients. Br J Cancer 88: 1570-1577, Pathology. J. Wiley & Sons, Chichester, pp. 1-615, 2000. 2003. 4 Aubin F, Pretet JL and Mougin C (eds.): Papillomavirus 21 Cuzick J, Sasieni P, Davies P, Adams J, Normand C, Frater A, Humains. Biologie et Pathologie Tumorale. Editions TEC & van Ballegooijen M and van den Akker E: A systematic review DOC, Paris pp. 1-759, 2003. of the role of human papillomavirus testing within a cervical 5 Miller AB: Evaluation of the impact of screening for cancer of screening programme. Health Technol Assess 3: 1-204, 1999. the cervix. IARC Sci Publ, pp. 149-160, 1986. 22 Kulmala SM, Syrjanen S, Shabalova I, Petrovichev N, 6 Miller AB and Robles SC: Workshop on screening for cancer Kozachenko V, Podistov J et al: Human papillomavirus testing of the uterine cervix in Central America. Bull Pan Am Health with the hybrid capture 2 assay and PCR as screening tools. J Organ 30: 397-408, 1996. Clin Microbiol 42: 2470-2475, 2004.

3478 Syrjänen et al: Testing Optional Screening Tools in Latin America

23 Franco EL: Chapter 13: Primary screening of cervical cancer 35 Longatto Filho A, Utagawa ML, Miranda Pereira SM, Pittoli with human papillomavirus tests. J Natl Cancer Inst Monogr 31: JE, Aguiari LS, Maeda MYS, di Loreto C, Roteli-Martins C, 89-96, 2003. Lima TP, Galvane JO and Syrjänen KJ: Low- and high-risk 24 Lörincz AT and Richart RM: Human papillomavirus DNA HPV types detected with Hybrid Capture II in liquid-based testing as an adjunct to cytology in cervical screening programs. cytology (DNA CITOLIQ®) system: experience from the Arch Pathol Lab Med 127(8): 959-968, 2003. LAMS (Latin American Screening) study. J Bras Doencas Sex 25 zur Hausen H: Papillomaviruses and cancer: from basic studies Transm 16: 17-20, 2004. to clinical application. Nature Rev Cancer 2: 342-350, 2002. 36 Sankaranarayanan R, Syamalakumary B, Wesley R, Amma NS, 26 Bosch FX and de Sanjose S: Chapter 1: Human papillomavirus Parkin DM and Nair MK: Visual inspection with acetic acid in and cervical cancer – burden and assessment of causality. J Natl the early detection of cervical cancer and precursors. Int J Cancer Inst Monogr 31: 3-13, 2003. Cancer 80: 161-163, 1999. 27 Munoz N, Bosch FX, de Sanjose S, Herrero R, Castellsague X, 37 Stafl A: Cervicography: a new method for cervical cancer Shah KV et al: Epidemiologic classification of human detection. Am J Obstet Gynecol 139: 815-821, 1981. papillomavirus types associated with cervical cancer. N Engl J 38 Costa S, Sideri M, Syrjänen K, Terzano P, De Nuzzo M, De Med 348: 518-527, 2003. Simone P, Cristiani P, Finarelli C, Bovicelli A, Zamparelli A 28 Syrjänen S, Shabalova IP, Petrovichev N, Kozachenko VP, and Bovicelli L: Combined Pap smear, cervicography and HPV Zakharova T, Pajanidi A, Podistov JI, Chemeris G, Soazeva DNA testing in the detection of cervical intraepithelial LG, Lipova EV, Tsidaeva I, Ivanchenko O, Pshepurko G, neoplasia and cancer. Acta Cytol 44: 310-318, 2000. Zakharenko S, Nerovjna R, Kljukina LB, Erokhina OA, 39 Parham GP, Andrews NR and Lee ML: Comparison of Branovskaja MF, Nikitina M, Grunberga V, Grunberg A, immediate and deferred colposcopy in a cervical screening Juschenko A, Tosi P, Cintorino M, Santopietro R and Syrjänen program. Obstet Gynecol 95: 340-344, 2000. KJ: Acquisition of incident high-risk human papillomavirus 40 Walker P, Dexeus S, De Palo G, Barrasso R, Campion M, (HPV) infections and PAP smear abnormalities in a cohort of Girardi F, Jakob C and Roy M: Nomenclature Committee of women subjected to HPV screening in the New Independent the International Federation for Cervical Pathology and States (NIS) of the former Soviet Union. J Clin Microbiol 42: Colposcopy. International terminology of colposcopy: an 505-511, 2004. updated report from the International Federation for 29 Lörincz AT: Screening for cervical cancer: new alternatives and Cervical Pathology and Colposcopy. Obstet Gynecol 101: research. Salud Publica Mex 45(Suppl 3):376-387, 2003. 175-177, 2003. 30 Naud P, Camozzato A, Hammes L, Prolla J, Catalan F, 41 Belinson JL, Qiao YL, Pretorius RG, Zhang WH, Rong SD, Larangeira M, Matos J, Milovanovic P, Schwartsmann G, Huang MN, Zhao FH, Wu LY, Ren SD, Huang RD, Vetorazzi J, Quadros C, Gomes T, Lemos A, D’Ávila A, Washington MF, Pan QJ, Li L and Fife D: Shanxi Province Manfrói A, Marc C, Berti C, Born C, Rigol C, Vettori D, Santos cervical cancer screening study II: self-sampling for high-risk D, Dias E, Cislaghi G, Di Domenico K, Costa M, Santos R, human papillomavirus compared to direct sampling for human Marroni R, Magno V and Syrjänen K: Cervical cancer screening papillomavirus and liquid based cervical cytology. Int J Gynecol in Porto Alegre, Brazil: alternative methods for detecting cancer Cancer 13: 819-826, 2003. precursors in a developing country. J Lower Gen Tract Dis 5: 24- 42 Human papillomavirus testing for triage of women with 28, 2001. cytologic evidence of low-grade squamous intraepithelial 31 Naud P, Bozzetti MC and Prolla JC: Screening in cervical lesions: baseline data from a randomized trial. The Atypical cancer prevention in Porto Alegre, Brazil: the experience of a Squamous Cells of Undetermined Significance/Low-Grade programme in a developing country. In: Franco E, Monsonego Squamous Intraepithelial Lesions Triage Study (ALTS) Group. J (eds.). New Developments in Cervical Cancer Screening and J Natl Cancer Inst 92: 397-402, 2000. Prevention. Blackwell Science, Oxford, pp. 250-262, 1997. 43 Bulkmans NW, Rozendaal L, Snijders PJ, Voorhorst FJ, Boeke 32 Syrjänen K, Naud P, Derchain SM, Roteli-Martins C, Longatto- AJ, Zandwijken GR et al: POBASCAM, a population-based Filho A, Tatti S, Branca M, Erzen M, Serpa-Hammes L, Gontijo randomized controlled trial for implementation of high-risk R, Lima TP, Maeda MYS, Lorincz A, Dores GB, Costa S and HPV testing in cervical screening: design, methods and baseline Syrjänen S: Comparing PAP smear cytology, aided visual data of 44,102 women. Int J Cancer 110: 94-101, 2004. inspection (VIA), screening colposcopy, cervicography and HPV 44 Cuzick J, Szarewski A, Cubie H, Hulman G , Kitchener H, testing by HCII (normal and self-sampling) as optional screening Luesley D et al: Management of women who test positive for tools in Latin America. Experience from the LAMS study. In: high-risk types of human papillomavirus: the HART study. Monsonego J (ed.). 5th International Multidisciplinary Congress Lancet 362: 1871-1876, 2003. EUROGIN 2003. Monduzi Editore, Milano pp. 65-73, 2003. 45 Little J: Human papillomavirus testing. Effectiveness of testing 33 Solomon D, Davey DD, Kurman R et al: for the Forum Group for high risk HPV for triage of low grade abnormal smears is Members and the Bethesda 2001 Workshop. The 2001 Bethesda being assessed in TOMBOLA trial. BMJ 323: 109, 2001. System: terminology for reporting results of cervical cytology. 46 Clavel C, Masure M, Bory JP, Putaud I, Mangeonjean C, JAMA 287: 2114-2119, 2002. Lorenzato M et al: Human papillomavirus testing in primary 34 Bishop JW, Bigner SH, Colgan TJ, Husain M, Howell LP, screening for the detection of high-grade cervical lesions: a McIntosh KM, Taylor DA and Sadeghi MH: Multicenter study of 7932 women. Br J Cancer 84: 1616-1623, 2001. masked evaluation of AutoCytePREP thin layers with matched 47 Franco EL: Randomized controlled trials of HPV testing and conventional smears: including initial biopsy results. Acta Cytol Pap cytology: toward evidence-based cervical cancer prevention. 42: 189-197, 1998. Int J Cancer 110: 1-2, 2004.

3479 ANTICANCER RESEARCH 25: 3469-3480 (2005)

48 Syrjänen S, Shabalova IP, Petrovichev N, Kozachenko VP, 51 Franco E, Villa L, Rohan T, Ferenczy A, Petzl-Erler M and Zakharova T, Pajanidi A, Podistov JI, Chemeris G, Soazeva Matlashewski G: Design and methods of the Ludwig-McGill LG, Lipova EV, Tsidaeva I, Ivanchenko O, Pshepurko G, longitudinal study of the natural history of human Zakharenko S, Nerovjna R, Kljukina LB, Erokhina OA, papillomavirus infection and cervical neoplasia in Brazil. Branovskaja MF, Nikitina M, Grunberga V, Grunberg A, Ludwig-McGill Study Group. Rev Panam Salud Publica 6:223- Juschenko A, Tosi P, Cintorino M, Santopietro R and Syrjänen 233, 1999. KJ: Human papillomavirus testing and conventional PAP smear 52 Carvalho MO, Almeida RW, Leite FM, Fellows IB, Teixeira cytology as optional screening tools of women at different risk MH, Oliveira LH et al: Detection of human papillomavirus for cervical cancer in countries of former Soviet Union. J Lower DNA by the hybrid capture assay. Braz J Infect Dis 7: 121- Genital Tract Dis 6: 97-110, 2002. 125, 2003. 49 Matos E, Loria D, Amestoy GM, Herrera L , Prince MA, 53 Cruz MR, Cerqueira DM, Cruz WB, Camara GN, Brigido MM, Moreno J et al: Prevalence of human papillomavirus infection Silva EO et al: Prevalence of human papillomavirus type 16 among women in Concordia, Argentina: a population-based variants in the Federal District, Central Brazil. Mem Inst study. Sex Transm Dis 30: 593-599, 2003. Oswaldo Cruz 99: 281-282, 2004. 50 Longatto-Filho A, Maeda MYS, Erzen M, Branca M, Roteli- Martins C, Naud P, Derchain SM, Serpa-Hammes L, Matos J, Gontijo R, Sarian L, Lima TP, Tatti S, Syrjänen S and Syrjänen K: Conventional PAP smear and liquid-based cytology (LBC) as optional screening tools in low-resource settings of Latin America. Experience from the LAMS Study. Acta Cytologica, Received February 22, 2005 in press, 2005. Accepted May 30, 2005

3480