Efficiency of the Hybrid Capture 2 HPV DNA Test in Cervical Cancer Screening A Study by the French Society of Clinical Cytology

Anatomic Pathology / HPV DNA TEST IN CERVICAL CANCER SCREENING Efficiency of the Hybrid Capture 2 HPV DNA Test in Cervical Cancer Screening A Study by the French Society of Clinical Cytology Patricia de Cremoux, MD, PhD, 1 Joël Coste, MD, PhD, 2 Xavier Sastre-Garau, MD, PhD, 1 Martine Thioux, 1 Christelle Bouillac, 1 Sylvain Labbé, MD, 3 Isabelle Cartier, MD, 4 Marianne Ziol, MD, 5 Anne Dosda, MD, 6 Catherine Le Galès, PhD, 7 Vincent Molinié, MD, 8 Marie-Cécile Vacher-Lavenu, MD, PhD, 9 Béatrix Cochand-Priollet, MD, PhD, 10 Philippe Vielh, MD, PhD, 1 and Henri Magdelénat, PhD, 1 for the French Society of Clinical Cytology Study Group * Key Words: Cervical cancer; Screening; Human papillomavirus; Hybrid Capture 2 test Abstract The aim of this study was to determine the efficiency of the Hybrid Capture 2 (HC2; Digene, Gaithersburg, MD) human papillomavirus (HPV) assay for the detection of cervical neoplasia. Of the 1,785 patients recruited, 462 (25.88%) were referred for colposcopy owing to previously detected cytologic abnormalities, and 1,323 (74.12%) were voluntary candidates for screening. For all patients, a Papanicolaou smear and a monolayer smear (ThinPrep, Cytyc, Boxborough, MA) were done. HPV DNA was detected on the residual liquid-based material. False-positive results were observed in 111 cases and comprised 34 cross-reactions (1.90%) and 77 false-positive cases (4.31%) owing to a contiguous strong chemiluminescence signal. Interestingly, all these samples had a relative light units value of 1 to 3 and were contiguous to a sample with a very high HPV DNA load. The final results showed that high-risk and low-risk HPV DNA were detected in 480 samples (26.89%) and 135 samples (7.56%), respectively. Although HC2 can be considered a reliable and sensitive test for HPV DNA detection, we do not advocate its use for large-scale screening for cervical neoplasia. Cervical cancer represents the second most common cancer in women, with 465,000 new cases and 200,000 deaths worldwide every year. 1 The universally accepted screening method for preventing cervical cancer is cytologic examination of cervical smears with Papanicolaou staining (Pap smear). The Pap smear allows the identification of lesions that are precancerous or cancerous. Well-organized, cytology-based screening programs for cervical cancer have been effective in reducing cancer incidence and preventing premature deaths. However, even in some countries in which screening programs have been pursued extensively for many years, cervical cancer has not been eradicated. 2,3 One reason is the limited access of women to screening, and another involves the shortcomings of the technique itself, ie, the frequency of false-negative results 4 and sampling error, when abnormal cells are not recovered on the smear. Sensitivities of only 40% to 80% for high-grade cervical intraepithelial neoplasia (CIN 2 or CIN 3) have been reported 5,6 and reviewed by Cuzick et al. 7 New technologies have been developed that, although initially expensive, ultimately might make screening possible at a much lower cost. 8 The main risk factor for the transformation of precancerous lesions into invasive lesions is sexually transmitted infection by the human papillomavirus (HPV). More than 85 HPV types have been described. Only a limited number of the approximately 40 types infecting the anogenital tract can be found in anogenital cancer biopsy specimens. Specifically, these types possess cell-immortalizing and cell-transferring properties. The main representatives are HPV types 16, 18, 31, 33, 39, 45, 52, 58, and 69. They can be considered high- and 492 Am J Clin Pathol 2003;120:492-499 Downloaded 492 from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022

Anatomic Pathology / ORIGINAL ARTICLE intermediate-risk types. Other HPV types are considered lowrisk types, such as the 2 main types found in nonmalignant genital tract lesions (condylomata acuminata), HPV-6 and HPV-11, but also HPV-42, HPV-43, and HPV-44. They are detected rarely in malignant lesions. 9,10 In addition, some studies indicate that the risk of progression of low-grade intraepithelial neoplasia is linked to the persistence of HPV, mainly high-risk HPV with a high viral load. 11,12 HPV can be detected reliably only by DNA-based tests. The most reliable tests are consensus primer polymerase chain reaction (PCR) assay or specific primer PCR and the Hybrid Capture 2 microtiter assay (HC2; Digene, Gaithersburg, MD). 6,13-15 Under these conditions, the sensitivity of the addition of high-risk HPV DNA tests to cytology is almost always higher for detecting CIN 2 or CIN 3 than cytology alone. These studies suggest that adding HPV testing to primary screening could increase the diagnosis of high-grade CIN by 30% to 100%. 16 Furthermore, some studies demonstrate that despite the high sensitivity of detection methods currently in use, no viral genome is detected in 5% to 10% of invasive cancers, 17-19 10% to 15% of highgrade intraepithelial neoplasias, and 20% to 30% of lowgrade intraepithelial neoplasias. 9,18,20 This may be due to methodological problems, the presence of viruses that have not yet been identified, or both. 21,22 In contrast, 10% of all women tested have a viral infection but no cytologic or histologic abnormalities. 17,18,22 In the present study, our primary objective was to perform a critical analysis of the HPV assay by the HC2 test to raise preliminary recommendations for its use in the context of large-scale screening. The secondary objective was to evaluate HPV detection by the HC2 test in a prospective study conducted by the French Society of Clinical Cytology (Paris), consisting of comparing the cost-effectiveness of monolayers (ThinPrep, Cytyc, Boxborough, MA) and HPV testing with that of conventional Pap smears for cervical cancer screening. The present work evaluates the addition of HPV detection to the monolayer smear to determine the sensitivity of the detection of cervical abnormality. Materials and Methods Study Population The study population comprised adult women candidates for cervical cancer screening. Patient accrual and data collection started in September 1999 and were completed in May 2000. Women older than 18 years, not pregnant, without any recent (<1 year) history of surgery or laser treatment of the cervix, whose cervix was visible by the physician, and who gave written informed consent were enrolled in the study. A total of 1,785 women were eligible for HPV DNA detection. Two populations were considered: group 1, 462 consecutive women (25.88%) referred for colposcopy owing to abnormalities detected on previous screening smears, and group 2, 1,323 consecutive women (74.12%) who were voluntary candidates for the screening of cervical lesions. Experienced gynecologists and cytopathologists at 4 independent French centers recruited women into the 2 groups: 2 private laboratories (Pathology Center, Besançon [1]; and Pathology Center [Dr R. Cartier], Paris [2]) and 2 university hospital pathology laboratories (Louis Mourier Hospital, Colombes [3]; and Jean Verdier Hospital, Bondy [4]). The Ethics Committee (Cochin Hospital, Paris) approved this protocol proposed by the French Society of Clinical Cytology in July 1998. The investigating gynecologists completed a standardized questionnaire for all patients, recording sociodemographic, medical, gynecologic, and colposcopic data, as previously described. 23 In addition, all women underwent colposcopic examination with directed biopsy. In the present study, we first performed a conventional Pap smear and then obtained a sample for the monolayer technique. We used cervical sampling equipment (Cervex- Brush, Rovers Medical Devices, Oss, the Netherlands; or appropriate brushes and spatulas) for Pap smears, and the sample was rinsed by shaking and rotation in PreservCyt (Cytyc, Boxborough, MA). The resulting suspension was used to prepare the monolayer slide. The residual sample was used to test for HPV DNA sequences. All women then underwent colposcopic examination. The cell suspensions were sent every 2 weeks to the laboratory at Institut Curie, Paris. The method for evaluating conventional Pap smears and monolayer smears has been described previously. 23,24 Lesions were classified according to the International Federation of Cervical Pathology and Colposcopy classification system. 25 Smears were classified cytologically into 5 categories (within normal, atypical squamous cells of undetermined significance [ASCUS] or atypical glandular cells of undetermined significance [AGUS], low-grade squamous intraepithelial lesion, high-grade squamous intraepithelial lesion, cancer), and biopsy specimens were classified histologically into 4 categories (normal, condylomata or CIN 1, CIN 2 or CIN 3, carcinoma). All cytologic and pathologic interpretations were performed under blinded conditions. An optimized consensus interpretation was performed for all discordant results and represented the gold standard for the examination. 23 HPV Detection HPV detection was performed in 1 center (Institut Curie) under blinded conditions. The residual cell suspension after monolayer preparation was used to test for HPV DNA Downloaded from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022 Am J Clin Pathol 2003;120:492-499 493 493 493

de Cremoux et al / HPV DNA TEST IN CERVICAL CANCER SCREENING sequences. All samples were treated less than 1 month after sampling, according to the manufacturer s instructions. If less than 2.5 ml remained in the vial, the HC2 test was not performed because of insufficient cellularity for analysis (25% of samples). A maximum volume of 10 ml was used for the HC2 test. Cells were centrifuged, and the pellet was suspended in Digene Specimen Transport Medium and frozen until analysis. The HC2 system was used to test for high-risk and lowrisk HPV DNA. 26,27 The HC2 test uses unlabeled full genomic HPV RNA probes that are hybridized with target DNA from specimens. Resulting RNA DNA hybrids are captured on microplate wells by an immobilized antibody that specifically recognizes RNA DNA hybrids. Hybrids are reacted with a second antibody conjugated to alkaline phosphatase. Bound alkaline phosphatase is revealed by addition of a chemiluminescent dioxetane-based substrate: the substrate is cleaved by the bound alkaline phosphate conjugate, and the emitted light is measured in a microplate luminometer (DML 2000, Digene). Three negative and 3 positive control samples provided by the manufacturer were included in each series. In addition, we used previously evaluated (using type-specific PCR) HPV-6, -11, and -16 positive samples (5 invasive cervical carcinomas) as additional positive control samples and HPVnegative samples (3 endometrial carcinomas) as additional negative control samples. Results were expressed in relative light units (RLUs), which represent the ratio of light emission of a sample to the average of 3 positive control samples provided by the manufacturer, containing 1 pg/ml of HPV-16 DNA. An RLU value of greater than or equal to 1 corresponding to 5,000 or more HPV DNA copies per test well was considered positive. 28 If the coefficient of variation of expected negative and positive controls was more than 25%, the outlying value was discarded and a new mean value was calculated using the remaining 2 validated positive or negative control samples. Two mixtures of RNA probes were used: 1 for low-risk HPV (probes complementary to the DNA of HPV types 6, 11, 42, 43, and 44) and 1 for high-risk HPV (probes complementary to the DNA of HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68). Each sample was analyzed once, according to the manufacturer s recommendations. All samples with a ratio between 1 and 2 were retested in an independent run, together with all positive samples contiguous to a positive sample. Statistical Analysis For comparison of the diagnostic performances of the 3 techniques, the optimized consensus diagnosis was considered to represent the best possible performance. The sensitivity (main assessment criterion) and the specificity of each of the 3 methods were compared using paired samples. To adjust for variations in stratification (group), center, and all other identified confounding factors, a conditional logistic regression model was used. Pairwise tests (McNemar χ 2 and 2-tailed Wilcoxon test) were used to compare other characteristics of interest between the methods, eg, proportion of slides that were not interpretable and frequency of different types or abnormality. Statistical analysis was performed using SAS version 8.01 statistical software (SAS Institute, Cary, NC). Data are given as mean ± SD unless otherwise indicated. Results Study Population The final cohort comprised 1,785 women. Table 1 shows the characteristics of the population tested for HPV DNA at study entry. The age was 34.5 ± 11.5 years; the age at puberty was 12.9 ± 1.4 years; the age at first intercourse was 18.4 ± 2.2 years; and the number of pregnancies was 1.4 ± 1.7. Viral Typing Of the 649 low-risk and/or high-risk HPV-positive samples, 159 were retested, either because they were contiguous with a sample with a very high viral load (risk of false-positive reading) or because they had a very high viral load of a high-risk or low-risk HPV (risk of cross-reaction), and 111 samples (69.8%) were finally found to be HPVnegative. These samples corresponded to 34 cross-reactions Table 1 Characteristics of 1,785 Patients Characteristic No. (%) Educational level No schooling or primary schooling only 89 (4.99) Secondary 959 (53.73) Higher 710 (39.78) Current smoking 604 (33.84) Postmenopausal 175 (9.80) First smear 162 (9.08) Previous abnormal smears 392 (21.96) Contraception Combined estrogen-progestin pill 907 (50.81) Intrauterine device 181 (10.14) Condom 51 (2.86) Colposcopic findings* Normal colposcopy 1,180 (66.11) Acetowhite epithelium 324 (18.15) Iodine-negative epithelium 565 (31.65) * Patients may have more than 1 abnormal finding, thus the total exceeds the total number of patients. 494 Am J Clin Pathol 2003;120:492-499 Downloaded 494 from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022

Anatomic Pathology / ORIGINAL ARTICLE (1.90% of the whole population) and 77 false-positive readings (4.31% of the whole population) due to the intensity of the neighboring chemiluminescence. All false-positive samples had RLU values between 1 and 3, and all samples with an RLU value of more than 3 remained positive. Following retesting, we finally observed 135 samples positive for low-risk HPV (7.56%) and 480 samples positive for high-risk HPV (26.89%) with the following distribution: 23% with high-risk HPV positivity alone, 3% with low-risk HPV positivity alone, and 4% with simultaneous low-risk and high-risk HPV positivity Table 2. High-risk HPV DNA was detected in 56.9% of group 1 patients and 16.02% of group 2 patients. Low-risk HPV DNA was detected in 11.9% of group 1 patients and 5.97% of group 2 patients. Relationship Between Viral Typing and Pathologic Characteristics In an attempt to evaluate the relationship between HPV viral load and histologic features, the HPV DNA viral load first was analyzed in the whole positive population and in the various groups of high-risk, low-risk, and combined highrisk and low-risk HPV. The mean ± SD HPV load was higher in samples positive for both high-risk and low-risk HPV (RLUs of 416 ± 721 for low-risk HPV and 471 ± 720 for high-risk HPV) than for samples positive for only highrisk HPV DNA (RLUs of 384 ± 644). The lowest load was observed for samples positive only for low-risk HPV DNA (RLUs of 94 ± 195). HPV positivity and viral DNA load then were found to be increased as a function of histologic grade Table 3. The highest viral load category (RLUs 100) was found in 71% of invasive carcinomas (10/14) compared with only 4.13% of samples without histologic abnormalities. Of the highest viral loads, 40.0% and 47.3% were found in CIN 1 and CIN 2 or CIN 3 lesions, respectively. No HPV DNA sequence was detected (neither highrisk nor low-risk) in 85.30% of the samples with no histologic lesion, in 30.9% of condylomata or CIN 1, in 17.4% of CIN 2 or CIN 3, and in 14% of carcinomas (2/14). We also analyzed the relationship between high-risk HPV DNA positivity and cytologic findings Table 4 to determine whether the sensitivity of morphologic examination was improved by HPV DNA detection. The frequency of high-risk HPV DNA positivity increased as a function of cytologic abnormalities (from 42.6% in the ASCUS/AGUS group to 92% [1/13] in the carcinoma group). High-risk HPV DNA positivity also was found in samples with no cytologic abnormalities (13.02%), which is close to the 14.70% of high-risk HPV DNA positivity found in samples with no histologic lesion. Relationship to Age The relationship between the distribution of HPV DNA positivity and age was studied in the whole population. No statistically significant difference was observed in the distribution of low-risk and high-risk HPV DNA. However, younger women (younger than 25 years) tended to have a Table 2 Distribution of Positivity for Human Papillomavirus (HPV) DNA Among Centers * HPV Group Center HPV-Positive Samples Low-Risk High-Risk Low- and High-Risk 1 201 (23.48) 29 (3.39) 146 (17.06) 26 (3.04) 2 241 (38.44) 18 (2.87) 183 (29.18) 40 (6.38) 3 75 (36.59) 8 (3.90) 60 (29.27) 7 (3.41) 4 21 (21.65) 3 (3.09) 14 (14.43) 4 (4.12) Total 538 (30.14) 58 (3.25) 403 (22.58) 77 (4.31) * Data are given as number (percentage). Percentages for the Total row are based on 1,785. Table 3 Relationship Between High-Risk HPV Load in Cytologic Specimens and Histologic Characteristics * Histologic Characteristics Invasive Viral load Normal Condyloma or CIN 1 CIN 2 or CIN 3 Carcinoma HPV-negative 1,219 (85.30) 54 (30.9) 29 (17.4) 2 (14) HPV-positive RLU value, 1-9 78 (5.46) 18 (10.3) 20 (12.0) 2 (14 RLU value, 10-99 73 (5.11) 33 (18.9) 39 (23.2) 0 (0) RLU value, 100 59 (4.13) 70 (40.0) 79 (47.3) 10 (71) Total 1,429 175 167 14 CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; RLU, relative light units. Downloaded from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022 Am J Clin Pathol 2003;120:492-499 495 495 495

de Cremoux et al / HPV DNA TEST IN CERVICAL CANCER SCREENING higher rate of low-risk HPV DNA (10% vs 7%). Sensitivity and Specificity The sensitivity and specificity of the HC2 test alone and combined with optimized reading of monolayer smears were compared with those of various optimized cytologic techniques. 23 Table 5 shows the comparison of cytologic methods and the HPV DNA test for the detection of histologic lesions equal to or higher than CIN 1. The HPV DNA test alone (high-risk or high- and low-risk) was less sensitive and less specific than all of the optimized cytologic techniques for groups 1 and 2 (P <.001). The sensitivity and specificity of the monolayer smear HPV DNA test combination were higher than for the HPV DNA test alone, but they were not significantly superior to cytologic optimized interpretations of lesions equal to or higher than ASCUS or AGUS in groups 1 and 2 (P =.06). Table 6 shows the comparison of the same methods for the detection of lesions CIN 2 or higher. When considering lesions higher than high-grade squamous intraepithelial lesion, the sensitivity of the HPV DNA test alone (high-risk or high/low-risk) and the sensitivity and specificity of the monolayer smear HPV DNA test combination was not statistically different from those of the optimized cytologic techniques in group 1 (P =.66), whereas the specificity of the HPV DNA test alone was clearly lower for group 1 (P =.001). The sensitivity of the HPV DNA test alone and the monolayer smear HPV DNA test combination was increased but was not superior to that of optimized reading of cytologic techniques in groups 1 and 2 (P =.73) The specificity of the HPV DNA test alone was much lower than that of optimized reading of cytologic techniques (P <.001) in group 2, whereas the specificity of the monolayer smear HPV DNA test combination was not statistically different from that of optimized cytologic techniques for group 2. Discussion This study was designed to evaluate HPV DNA by using the HC2 assay from the residual sample of a liquid-based Pap test (ThinPrep). HC2 is a second-generation commercial HPV detection test, designed to detect HPV types divided into high-risk and low-risk groups. It has been correlated with detection of HPV DNA sequences by type-specific or consensus PCR in experienced laboratories. A good correlation has been reported between the HC2 test and Southern blot and PCR. 29-33 The published results showed that HC2 and PCR with GP5+/GP6+ or MY09/MY11 L1 consensus primers gave very similar results in terms of HPV detection Table 4 High-Risk HPV Status According to Cytologic Features * High-Risk HPV Status Normal ASCUS or AGUS LSIL HSIL Carcinoma HPV-positive 173 (13.02) 46 (42.6) 115 (68.0) 133 (82.6) 12 (92) HPV-negative 1,156 (86.98) 62 (57.4) 54 (32.0) 28 (17.4) 1 (8) Total 1,329 108 169 161 13 AGUS, atypical glandular cells of undetermined significance; ASCUS, atypical squamous cells of undetermined significance; HSIL, high-grade squamous intraepithelial lesion; HPV, human papillomavirus; LSIL, low-grade squamous intraepithelial lesion. * Data are given as number (percentage). Table 5 Sensitivity and Specificity of Conventional Pap Smear, ThinPrep Monolayer, and HPV DNA Test for the Detection of Histologic Lesions CIN 1 or Higher * Group 1 Group 2 Sensitivity Specificity Sensitivity Specificity Cytology results, ASCUS or AGUS or higher Conventional Pap optimized interpretation 92 80 74 91 ThinPrep monolayer optimized interpretation 89 76 73 90 HPV DNA test High-risk HPV-positive 79 77 64 86 High- and low-risk HPV-positive 82 74 69 83 ThinPrep monolayer and high-risk HPV-positive 85 82 67 94 when ASCUS or AGUS AGUS, atypical glandular cells of undetermined significance; ASCUS, atypical squamous cells of undetermined significance; CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; Pap, Papanicolaou. * Group 1, consecutive women referred for colposcopy owing to abnormalities detected on previous screening smears; group 2, consecutive women who were voluntary candidates for screening of cervical lesions. Data are given as percentages. For proprietary information, see the text. 496 Am J Clin Pathol 2003;120:492-499 Downloaded 496 from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022

Anatomic Pathology / ORIGINAL ARTICLE Table 6 Sensitivity and Specificity of Conventional Pap Smear, ThinPrep Monolayer, and HPV DNA Test for the Detection of Histologic Lesions CIN 2 or Higher * Group 1 Group 2 Sensitivity Specificity Sensitivity Specificity Cytology results of HSIL or higher Conventional Pap optimized interpretation 85 92 60 99 ThinPrep monolayer optimized interpretation 78 94 65 99 HPV DNA test High-risk HPV-positive 80 54 96 85 High- and low-risk HPV-positive 81 50 96 82 ThinPrep monolayer and high-risk HPV-positive 80 93 76 97 when ASCUS or AGUS AGUS, atypical glandular cells of undetermined significance; ASCUS, atypical squamous cells of undetermined significance; CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; Pap, Papanicolaou. * Group 1, consecutive women referred for colposcopy owing to abnormalities detected on previous screening smears; group 2, consecutive women who were voluntary candidates for screening of cervical lesions. Data are given as percentages. For proprietary information, see the text. in large series of human cervical samples. Peyton et al 31 showed more than 90% agreement between HC2-positive results with both Digene Specimen Transport Medium and PreservCyt medium. Moreover, Clavel et al 30 argued that the HC2 assay is a more convenient and easier test than the PCR assay for routine use. The first-generation assay (HC) had a high rate of false-positive results (18%), 16 but the secondgeneration assay (HC2) is more accurate and provides better sensitivity. This is due mainly to a reformulation of hybridization reagents and, to a lesser extent, the addition of new HPV probes (types 58 and 68). 28 Although the HC2 assay is convenient, its reliability depends on critical analysis of the results, as each sample is examined only once and not in duplicate owing to the high cost of the test and in accordance with the manufacturer s instructions. Therefore, 2 problems may be encountered, leading to 6.2% false-positive results. First, cross-hybridization with other HPV types may occur. Although the specificity of the test has been demonstrated, we observed that a very high load of low-risk HPV DNA induced cross-hybridization with high-risk HPV DNA, leading to the detection of a significant viral load of highrisk HPV. Similar results were observed with a high load of high-risk HPV DNA. This could be analyzed in our series because high- and low-risk HPV DNA were quantified further, but in the majority of studies and in routine use, only the detection of high-risk HPV DNA actually is performed. We observed a 1.90% cross-hybridization rate. The resulting risk is that of a false-positive result of high-risk type HPV in a sample with low-risk HPV. The cross-reaction is not constant and unpredictable. This has been described by Vernon et al. 33 Cuzick et al 34 analyzed the positive predictive value of the HC2 test in primary screening of older women, by comparing different cutoff values for HPV DNA positivity (1, 2, and 4 pg/ml). Under these conditions, all HPVpositive cases of CIN 2 or CIN 3 had HPV DNA levels greater than 4 pg/ml, giving 100% sensitivity for CIN 3 lesions and 95.2% sensitivity for all high-grade lesions. The false-positive rate also was reduced (2.3% at 2 pg/ml and 2.1% at 4 pg/ml vs 4.9% at 1 pg/ml), without any loss of sensitivity. The positive predictive value for HC2 was equal to 17%, 27%, and 28% at the 1, 2, and 4 pg/ml cutoff values, respectively. 34 Second, the chemiluminescent emission of high-load HPV DNA of a sample may lead to false-positive reactions in contiguous samples. This was observed in 4.31% of samples in our series. Critical analysis of the results requires detection and retesting of all samples with doubtful results. In the present study, we detected more high-risk than low-risk HPV DNA (25% vs 7%). High-risk HPV DNA was detected mainly in the patients referred for colposcopy owing to previously detected abnormalities such as ASCUS or AGUS (group 1; 57%), whereas 16% of high-risk HPV DNA cases were detected in the group systematically screened for cervical cancer (group 2). Two of the 4 centers had a higher incidence of high-risk HPV DNA (centers 2 and 3). This could be related to the distribution of their recruited populations, which included more women from group 1 than from group 2. The distribution of high-risk HPV DNA in our population is in agreement with previously published series 16,35-39 analyzed by the HC2 test and/or consensus GP5+/GP6+ PCR 38 or MY09/MY11 PCR. 37 Transient infections are much more common in younger women and are not always of clinical significance in women younger than 30 years. 16,40 In our series, no significant difference in the distribution of HPV positivity according to age was observed. However, there was a trend toward an increased incidence of low-risk HPV infection in women younger than 25 years. We cannot evaluate the incidence of Downloaded from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022 Am J Clin Pathol 2003;120:492-499 497 497 497

de Cremoux et al / HPV DNA TEST IN CERVICAL CANCER SCREENING transient HPV infection, as no follow-up of our population was performed. Woodman et al, 38 in a follow-up study of a cohort comprising 1,075 very young females (mean age, 18 years) who had cytologically normal and HPV DNA negative results at presentation, demonstrated that the period during which the virus can be detected sequentially usually is short. This cannot be extrapolated to larger populations, but these very young females had a high incidence of HPV DNA positivity because they are more likely to have HPV infection; however, they less frequently develop severe disease. In addition, transient HPV infection, as frequently observed in very young women, cannot be analyzed by single testing, which constitutes a limitation of our study. Repeated testing should be performed after 1 year, especially for young women with HPV DNA positive and cytology-negative results, to avoid false-positive results. 40 We used a semiquantitative method to distinguish between low-viral-load and high-viral-load samples, ie, an arbitrary stratification (ratio of RLU/positive control samples of <10, <100, and 100). Viral load is known to wax and wane during the course of an HPV-positive episode. Lowviral load usually appears at the beginning and at the end of the infection. 38 It also has been demonstrated that high-risk HPV clearance precedes regression of cervical lesions by an average of 3 months. 41 The finding of a low-viral-load sample in a normal smear, therefore, provides little prognostic information. In the present study, the viral load was correlated with histologic grade. Moreover, patients with abnormal smears had the highest HPV viral load. This is observed particularly in cervical carcinoma, in which 71% of samples (10/14) had the highest viral load. In the present study, the HPV DNA test was assessed for cervical cancer screening in comparison with 2 cytologic techniques (conventional Pap smear and liquid-based smear) in 2 contexts: systematic testing and diagnosis of ASCUS or AGUS. In both contexts, our results showed that the sensitivity of the HPV DNA test alone was lower than that of optimized reading of conventional cytologic techniques. This finding is not in accordance with the results of some studies, 35,36 but it is similar to those of others. 39,42 In view of the natural history of cervical HPV infections, systematic HPV DNA detection has been proposed for cervical cancer screening but has been considered a complementary test to the smear, especially for patients with ASCUS or AGUS. 26,42 The HC2 assay is a sensitive test to detect HPV DNA sequences in experienced laboratories; however, its use as the only primary test for large-scale screening of cervical neoplasia and, therefore, the management of patients with ASCUS or AGUS, should not be advocated. From the 1 Department of Tumor Biology, Institut Curie, Paris; 2 Department of Biostatistics, Cochin Hospital, AP-HP and Paris V University, Paris; 3 Pathology Center, La Rochelle; 4 Pathology Center (Dr R. Cartier), Paris; 5 Department of Pathology, Jean Verdier Hospital, Paris XIII University, Bondy; 6 Department of Pathology, Louis Mourier Hospital, Colombes; 7 Center for Health Economics and Administration Research, INSERM U537, CNRS UPRESA 80-52, Le Kremlin Bicêtre; 8 Department of Pathology, Foch Hospital, Suresnes; 9 Department of Pathology, Cochin Hospital, AP-HP and Paris V University, Paris; and 10 Department of Pathology, Lariboisière Hospital, AP-HP, Paris, France. Supported by grant 9099 from the ARC (Research Association Against Cancer); and grant AOM 98010 from the DGS (National Health Board) and PHRC (Clinical Research Program for Hospitals), French Ministry of Health, Paris. Address reprint requests to Dr de Cremoux: Dept Biologie des Tumeurs, Institut Curie, 26 rue d Ulm, 75248 Paris Cedex, France. * S. Arkwright, A. Biaggi, C. Besançon-Roux, L. Carbillon, I. Cartier, B. Cochand-Priollet (clinical coordinator), J. Coste (methodological coordinator), P. Dauvergne, P. de Cremoux, A. Dosda, E. Foucher, I. Gouget, B. Karkouche, S. Labbé, C. Le Galès, H. Magdelenat, E. Merea, V. Molinié, A. Petitjean, P. Piquet, X. Sastre-Garau, N. Seince, M.C. Vacher-Lavenu, P. Vielh, M. Ziol. References 1. Herrero R. Epidemiology of cervical cancer. J Natl Cancer Inst Monogr. 1996;21:1-6. 2. Koss LG. The Papanicolaou test for cervical cancer detection: a triumph and a tragedy. JAMA. 1989;261:737-743. 3. Sasieni PD, Cuzick J, Lynch-Farmery E, and the NCN Working Group. Estimating the efficacy of screening by auditing smear histories of women with and without cervical cancer. Br J Cancer. 1996;73:1001-1005. 4. Janerich DT, Hadjimichael O, Schwartz PE, et al. The screening histories of women with invasive cervical cancer. Am J Public Health. 1995;85:791-794. 5. Cox JT, Lorincz AT, Schiffman MH, et al. Human papillomavirus testing by hybrid capture appears to be useful in aging women with a cytologic diagnosis of atypical squamous cells of undetermined significance. Am J Obstet Gynecol. 1995;172:946-954. 6. Cuzick J, Szarewski A, Terry G, et al. Human papillomavirus testing in primary cervical screening. Lancet. 1995;345:1533-1537. 7. Cuzick J, Sasieni P, Davies P, et al. A systemic review of the role of human papillomavirus (HPV) testing within a cervical screening programme: summary and conclusions. Br J Cancer. 2000;83:561-585. 8. Méréa E, Le Galès C, Cochand-Priollet B, et al. The cost of screening for cancerous and precancerous lesions of the cervix. Diagn Cytopathol. 2002;27:251-257. 9. Bergeron C, Barasso R, Beaudenon S, et al. Human papillomavirus associated with cervical intraepithelial neoplasia: great diversity and distinct distribution in low- and high-risk lesions. Am J Surg Pathol. 1992;16:641-649. 10. Zur Hausen H. Papillomavirus causing cancer: evasion from host-cell control in early events of carcinogenesis. J Natl Cancer Inst. 2000;92:690-698. 498 Am J Clin Pathol 2003;120:492-499 Downloaded 498 from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022

Anatomic Pathology / ORIGINAL ARTICLE 11. Remmink AJ, Walboomers JMM, Helmerhorst TJM, et al. The presence of persistent high-risk HPV genotypes in dysplastic cervical lesions is associated with progressive disease: natural history up to 36 months. Int J Cancer. 1995;61:306-311. 12. Ho GYF, Burk RD, Klein S, et al. Persistent genital human papillomavirus infection as a risk factor for persistent cervical dysplasia. J Natl Cancer Inst. 1995;87:1365-1371. 13. Reid R, Greenberg MD, Lorincz A, et al. Should cervical cytologic testing be augmented by cervicography or human papillomavirus deoxyribonucleic acid detection? Am J Obstet Gynecol. 1991;164:1461-1471. 14. Schneider A, Zahm DM, Kirchmayr R, et al. Screening for cervical intraepithelial neoplasia grade 2/3: validity of cytologic study, cervicography, and human papillomavirus detection. Am J Obstet Gynecol. 1996;174:1534-1541. 15. Ratnam S, Prafull G, Franco E, et al. Utility of HPV testing in combination with Papanicolaou smear in primary cervical screening. Presented at the 17th International Papillomavirus Conference; January 9-15, 1999; Charleston, SC. 16. Cuzick J. Screening for cancer: future potential. Eur J Cancer. 1999;35:1925-1932. 17. Bosh FX, Manos M, Munoz N, et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. J Natl Cancer Inst. 1995;87:796-802. 18. Lombard I, Vincent-Salomon A, Validire P, et al. Human papillomavirus genotype as a major determinant of the course of cervical cancer. J Clin Oncol. 1998;16:2613-2619. 19. zur Hausen H. Papillomaviruses and cancer: from basic studies to clinical applications. Nat Rev Cancer. 2002;2:342-350. 20. Lörincz AT, Reid R, Jenson B, et al. Human papillomavirus infection of the cervix: relative risk associations of 15 common anogenital types. Obstet Gynecol. 1992;79:328-337. 21. de Villiers EM. Human pathogenic papillomavirus types: an update. Curr Top Microbiol Immunol. 1994;186:1-12. 22. Walboomers JMM, Meijer CJLM. Do HPV-negative cervical carcinomas exist? J Pathol. 1997;181:253-254. 23. Cochand-Priollet B, Le Gales C, de Cremoux P, et al, for the 20 Monolayers French Society of Clinical Cytology Study Group Diagnostic. Cost-effectiveness of monolayers and human papillomavirus testing compared to that of conventional Papanicolaou smears for cervical cancer screening: protocol of the study of the French Society of Clinical Cytology. Diagn Cytopathol. 2001;24:412-420. 24. Coste J, Cochand-Priollet B, de Cremoux P, et al, and the French Society of Clinical Cytology. Cross sectional study of conventional cervical smears, monolayer cytology, and human papillomavirus DNA testing for cervical cancer screening. BMJ. 2003;326:733-737. 25. Stafl A, Wilbanks GD. An international terminology of colposcopy: report of the Nomenclature Committee of the International Federation of Cervical Pathology and Colposcopy. Obstet Gynecol. 1991;77:313-314. 26. Sherman ME, Mendoza M, Lee KR, et al. Performance of liquid-based thin-layer cervical cytology: correlation with reference diagnoses and human papillomavirus testing. Mod Pathol. 1998;11:837-843. 27. Manos MM, Kinney WK, Hurley LB, et al. Identifying women with cervical neoplasia using human papillomavirus DNA testing for equivocal Papanicolaou tests. JAMA. 1999;281:1605-1610. 28. Poljack M, Brencic A, Seme K, et al. Comparative evaluation of first- and second-generation Digene Hybrid Capture assays for detection of human papillomavirus associated with high or intermediate risk for cervical cancer. J Clin Microbiol. 1999;37:796-797. 29. Lörincz A. Hybrid Capture method for detection of human papillomavirus DNA in clinical specimens. Papillomavirus Rep. 1996;7:1-5. 30. Clavel C, Masure M, Putaud I, et al. Hybrid Capture II, a new sensitive test for human papillomavirus detection: comparison with Hybrid Capture I and PCR results in cervical lesions. J Clin Pathol. 1998;51:737-740. 31. Peyton CL, Schiffman M, Lörincz AT, et al. Comparison of PCR- and Hybrid Capture based-human papillomavirus detection systems using multiple cervical specimen collection strategies [published correction appears in J Clin Microbiol. 1999;37:478]. J Clin Microbiol. 1998;36:3248-3254. 32. Bozzetti M, Nonnenmacher B, Mielzinska II, et al. Comparison between Hybrid Capture II and polymerase chain reaction results among women at low-risk for cervical cancer. Ann Epidemiol. 2000;10:466. 33. Vernon SD, Unger ER, William D. Comparison of human papillomavirus detection and typing by cycle; sequencing, line blotting, and hybrid capture. J Clin Microbiol. 2000;38:651-655. 34. Cuzick J, Beverley E, Ho L, et al. HPV testing in primary screening of older women. Br J Cancer. 1999b;81:554-558. 35. Solomon D, Schiffman M, Tarone R. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: baseline results from a randomized trial. J Natl Cancer Inst. 2001;93:293-299. 36. Clavel C, Masure M, Bory JP, et al. Human papillomavirus testing in primary screening for the detection of high-grade cervical lesions: a study of 7932 women. Br J Cancer. 2001;89:1616-1623. 37. The Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesions Triage Study (ATLS) Group. Human papillomavirus testing for triage of women with cytologic evidence of low-grade squamous intraepithelial lesions: baseline data from a randomized trial. J Natl Cancer Inst. 2000;92:397-402. 38. Woodman CBJ, Collins S, Winter H, et al. Natural history of cervical human papillomavirus infection in young women: a longitudinal cohort study. Lancet. 2001;357:1831-1836. 39. Zuna RE, Moore W, Dunn ST. HPV DNA testing of the residual sample of liquid-based Pap test: utility as a quality assurance monitor. Mod Pathol. 2001;14:147-151. 40. Cuzick J, Woodman CBJ, Collins S, et al. Natural history of cervical human papillomavirus infection in young women: a longitudinal cohort study. Lancet. 2001;357:1831-1836. 41. Nobbenhuis MAE, Helmerhost JM, Van den Brule AJC, et al. Cytological regression and clearance of high-risk human papillomavirus in women with an abnormal cervical smear. Lancet. 2001;358:1782-1783. 42. Lee NW, Kim D, Park JT, et al. Is the human papillomavirus test in combination with the Papanicolaou test useful for management of patients with diagnoses of atypical squamous cells of undetermined significance/low-grade squamous intraepithelial lesions? Arch Pathol Lab Med. 2001;125:1453-1457. Downloaded from https://academic.oup.com/ajcp/article-abstract/120/4/492/1759022 Am J Clin Pathol 2003;120:492-499 499 499 499