Cytologic screening programs to detect cervical intraepithelial

Transcription

1 Combining Human Papillomavirus Testing or Cervicography With Cytology to Detect Cervical Neoplasia Michelle Howard, MSc; John W. Sellors, MD, MSc; Alice Lytwyn, MD, MSc; Paula Roth, MD; James B. Mahony, PhD Context. Cervicography and oncogenic human papillomavirus (HPV) testing have been proposed for improving the accuracy of cervical cancer screening. Objective. To examine whether cervicography and HPV testing can improve beyond chance the detection of cervical intraepithelial neoplasia (CIN) 2 or 3 in women with atypical cells of undetermined significance or lowgrade squamous intraepithelial lesions on cytology. Design. Cross-sectional analysis. Oncogenic HPV testing by Hybrid Capture II assay or cervicography combined with cytology was compared with the reference standard of colposcopy with directed biopsy. Setting. Community family practices. Participants. Three hundred four women with lowgrade cytologic abnormality. Main Outcome Measures. The gain in accuracy for detecting histologic CIN 2 or 3 or carcinoma. Because an adjunct test may improve sensitivity by chance alone, the sensitivity or specificity if the second test performed randomly was estimated. Results. Cervical intraepithelial neoplasia 2 or 3 was found in 11.8% (36/304) of the women and invasive squamous cell carcinoma in 0.3% (1/304). The sensitivity of cytology for detecting CIN 2 or 3 was 73.0% and increased by 21.6% to 94.6% with the addition of a cervigram showing a low-grade lesion or higher or a positive HPV test result. These gains were reduced to 8.1% and 10.8% above the sensitivities expected if the additional tests performed randomly. The corresponding specificities decreased from 49.1% to 32.2% and 33.0%. There was insufficient power to determine whether observed sensitivities were statistically significantly higher than the expected sensitivities. Conclusion. Adjunctive HPV testing or cervicography may provide similar gains in sensitivity, but they can appear misleadingly large if chance increases are not taken into account. (Arch Pathol Lab Med. 2004;128: ) Cytologic screening programs to detect cervical intraepithelial neoplasia (CIN) have led to reductions in morbidity and mortality from cervical cancer in several countries. 1 These successes notwithstanding, rigorously conducted meta-analyses have shown that the sensitivity of a cytology smear is low (30% 87%), in contrast to its specificity of 86% to 100% for high-grade (CIN 2 or 3) cervical lesions. 2 Among women with a low-grade squamous intraepithelial lesion (LSIL) or atypical squamous cells of undetermined significance (ASCUS) on screening Accepted for publication June 11, From the Departments of Family Medicine (Ms Howard and Dr Sellors) and Pathology and Molecular Medicine (Dr Mahony) and Colposcopy Clinic (Dr Roth), McMaster University, Hamilton, Ontario; the Program for Appropriate Technology in Health, Seattle, Wash (Dr Sellors); and the Department of Pathology, Sunnybrook and Women s College of Health Sciences Centre, Toronto, Ontario (Dr Lytwyn). Dr Lytwyn is now affiliated with the Department of Laboratory Medicine, McMaster University. The Hybrid Capture II test assays used for detection of HPV and the use of a luminometer were provided free of charge by Digene Corporation. The authors have no relevant financial interest in the products or companies described in this article. Reprints: Michelle Howard, MSc, Department of Family Medicine, McMaster University, 1200 Main St W, Room HSC-2V10, Hamilton, Ontario, Canada L8N 3Z5 ( mhoward@mcmaster.ca). cytology, the correlation with histologic findings has caused concern for conservative management, because approximately 10% to 20% of these women will have CIN 2 or 3 on histologic examination. 3 7 Cervicography and human papillomavirus (HPV) testing are adjunctive methods that may allow effective triage of women with lowgrade lesions on screening cytology. Cervicography is a procedure in which a projected photographic slide of the cervix simulates colposcopic magnification of 16 times. 8 Studies 6,9 13 that have evaluated cervicography in women with low-grade lesions on cytology, without verification bias, have found that the sensitivity and specificity for the detection of CIN 2 or 3 range from 54.5% to 90.0% and from 30.0% to 90.2%, respectively. Testing for oncogenic types of HPV in cervical specimens is another possible method to triage women with low-grade abnormalities on cytologic screening and has been found to have sensitivity and specificity ranging from 65% to 98% and from 52% to 74%, respectively, for detecting high-grade lesions. 4,14 17 The use of a second test in parallel with the first test (ie, cytology) must result in the same or increased sensitivity and the same or decreased specificity. These increases in sensitivity may be misleading because improvements in sensitivity would be expected by adding a second test, even if the second test performed randomly with respect Arch Pathol Lab Med Vol 128, November 2004 Cervicography and HPV Testing Howard et al 1257

2 Flow diagram of the studies from which women were analyzed. ASCUS indicates atypical squamous cells of undetermined significance; LSIL, low-grade squamous intraepithelial lesion; HPV, human papillomavirus; R, randomization; asterisk, colposcopy, cervicography, and cytology (at colposcopy clinic); and dagger, referral to colposcopy as soon as HPV detected. to disease identification. For this reason, the method described by Franco 18 was used to determine if any gain in test performance from the addition of the second test was greater than that expected if the second test offered no diagnostic information. This study assessed whether there was a gain in diagnostic accuracy over cervical cytology, above what would be expected by chance, by adding cervicography or HPV testing in a group of women with low-grade cytologic abnormalities. A cross-sectional analysis was performed using the data from 2 randomized controlled trials of cytology and HPV testing. 3,17 MATERIALS AND METHODS Women were eligible if they were aged 16 to 50 years and received a result of LSIL or ASCUS on routine cytologic screening. Women were recruited consecutively from 118 community family practices and 1 university student health clinic in Ontario between May 1995 and October Written informed consent was obtained at the physicians offices. The studies were approved by the Hamilton Health Sciences Research Ethics Board. The objective of the study was to determine the efficacy of HPV testing compared with repeat cytology for detecting CIN 2 or 3 or cancer, by randomizing women to repeat cytology alone (HPV testing was done in these groups but not used in the randomized trials) or to repeat cytology with HPV testing to direct referral to colposcopy (Figure). The maximum follow-up was 24 months, or sooner if a high-grade squamous epithelial lesion was found on a repeat cytologic examination for any woman, or if the HPV test result was positive in a woman randomized to the HPV testing strategy. The cytology and cervicography results presented in this article are those obtained at the time of colposcopy (ie, study exit). At enrollment, a written questionnaire on demographics and sexual health history was self-completed by the women, as previously described. 3,17 The HPV test findings used in the present analysis were obtained by the family physician from the area of the cervical os and the transformation zone with a Dacron swab at the last follow-up visit before colposcopy referral. The swab was placed into specimen transport medium for Hybrid Capture II (Digene Corporation, Gaithersburg, Md) testing. At colposcopy, the smear for cervical cytology was obtained by a colposcopy clinic nurse, using an extended-tip Ayre spatula and, as required, an endocervical brush. Cervicography was then performed by 1 of 2 trained nurses, after washing the cervix with 5% acetic acid. Finally, a colposcopic examination with directed biopsy or endocervical curettage (if deemed necessary) was performed. The physicians performing colposcopy were blinded to the cervicographic findings and the HPV results. Colposcopy was done at the McMaster University Colposcopy Clinic. The cervigrams were read by National Testing Laboratories (Fenton, Mo). Cervigram results were reported by the laboratories as (1) negative if (a) components of the transformation zone were visible, with no abnormal lesion or (b) components of the transformation zone were not visible; (2) atypical, recommending repeat cervigram and Papanicolaou test in 6 to 12 months, if (a) a lesion of doubtful significance was inside the zone or (b) alesion of doubtful significance was outside the zone; (3) positive, with colposcopy recommended, if the cervigram indicated (a) a probable normal variant and colposcopy was warranted to exclude significant disease, (b) results compatible with a low-grade lesion, (c) results compatible with a high-grade lesion, or (d) results compatible with cancer; or (4) technically defective if (a) the view of the cervix was obscured by mucus, blood, or an incorrect position, (b) there was insufficient acetic acid reaction, or (c) the cervigram was out of focus, overexposed, or underexposed. For the purpose of analysis, cervigram interpretations that were normal, revealed lesions of doubtful significance, or were judged as a probable normal variant were considered negative. The 2 positive cutoffs for the cervigram interpretation were (1) results compatible with a low-grade or high-grade lesion or carcinoma (categories 3b, c, or d in the previous paragraph) and (2) results compatible with a high-grade lesion or carcinoma (categories 3c or d in the previous paragraph). Cervigrams in which 1258 Arch Pathol Lab Med Vol 128, November 2004 Cervicography and HPV Testing Howard et al

3 Table 1. Demographic Characteristics and History of the Women in the 2 Randomized Controlled Trials* Characteristic Two-Year Study (n 150) Six-Month Study (n 154) Age, mean (SD), y Age at first intercourse, mean (SD), y No. of lifetime sexual partners, median or interquartile range Married or common law Previous pregnancy Ever used oral contraceptive Current smoker 34.3 (9.2) 17.9 (3.2) 4 (6) 57.4 (81/141) 65.5 (93/142) 91.9 (125/136) 36.6 (52/142) * Data are given as percentage (numerator divided by denominator) unless otherwise indicated (8.8) 17.3 (2.7) 4 (5) 49.6 (69/139) 55.8 (77/138) 87.0 (120/138) 35.3 (49/139) Table 2. The Performance of Cervical Cytology, Cervicography, and the Cervical Swab for Hybrid Capture II Human Papillomavirus (HPV) Testing for the Detection of 37 Women With a Biopsy-Proven Cervical Intraepithelial Neoplasia (CIN) 2 or 3 or Invasive Squamous Cell Carcinoma* Triage Test Triage Positive Threshold Sensitivity Specificity Positive Predictive Value Negative Predictive Value Cytology alone Atypical squamous cells of 73.0 ( ) 49.1 ( ) 16.6 ( ) 92.9 ( ) undetermined significance Cervigram alone Low grade or higher High grade or carcinoma 83.8 ( ) 24.3 ( ) 65.5 ( ) 95.9 ( ) 25.4 ( ) 45.0 ( ) 96.2 ( ) 90.1 ( ) HPV DNA test alone Positive for oncogenic type 86.5 ( ) 59.2 ( ) 22.7 ( ) 96.9 ( ) * Data are given as percentage (95% confidence interval). the view of the cervix was obscured and those that were technically defective were classified as negative. The Hybrid Capture II assay detects 13 of the most common oncogenic HPV types, and testing was carried out as previously described. 14 Specimens with relative light unit readings equal to or greater than the mean relative light unit reading of triplicate positive controls containing 1 pg/ml of HPV DNA were considered positive. Testing was performed at the McMaster University Regional Virology and Chlamydiology Laboratory. Tissue specimens were processed in the Department of Pathology, Hamilton Health Sciences. Results were reported using CIN terminology. Human papillomavirus effects were categorized as CIN 1 if no greater degree of dysplasia was present. Two expert gynecologic pathologists independently reviewed all biopsy and curettage material ( 0.88 for high-grade dysplasia and 0.83 for any grade of dysplasia, P.001 for both) and achieved consensus on any discordant specimens. For cases with more than 1 histologic specimen available, the most severe diagnosis was recorded. A colposcopic diagnosis of normal was recorded if the examination revealed no lesion and no biopsy was taken, or if the tissue obtained showed no CIN or cancer. Cervical cytology specimens taken at the colposcopy clinic were read in the Department of Pathology using Bethesda system terminology. The results of colposcopic examination and biopsy were used as the reference standard for the other tests. Sensitivity, specificity, and positive and negative predictive values were calculated for cervical cytology (taken at the time of colposcopy), cervicography, and HPV testing. To estimate the magnitude of the gain in sensitivity with the addition of cervicography or HPV testing, the proportion of additional true positives identified by the second test (cervicography or HPV testing) that were missed by cytology alone was calculated. To estimate the loss of specificity when adding a second test, the proportion of additional false positives identified by the second test was calculated. It was expected that these proportions would be significantly greater than zero. The 95% confidence intervals (CIs) were calculated using the normal approximation to the binomial distribution. 19 Differences between the observed sensitivities and specificities and the actual sensitivities and specificities (those calculated after adjusting for the chance increases expected if the second test performed randomly using the method by Franco 18 ) were compared using the normal approximation to the binomial distribution. 20 Differences were considered statistically significant at 0.05 (2-sided). RESULTS Of the 469 women enrolled, complete data on cervicography, HPV testing, cervical cytology, and the end point assessment of colposcopy were available for 304 (64.8%). The mean SD number of days between the HPV test and colposcopy was days. The demographic characteristics of the women are shown separately for the 2 trials in Table 1. Cervical cytology performed at colposcopy showed 46.4% (141/304) of women had normal cytology, 31.9% (97/304) had ASCUS, 16.1% (49/304) had LSIL, 5.3% (16/ 304) had a high-grade squamous epithelial lesion, and 0.3% (1/304) had carcinoma. There was no statistically significant association between the time since study entry and the cytology result (negative vs others) (data not shown). Cervical biopsies were performed on 55.9% (170/ 304) of the women at colposcopy. On colposcopy, CIN 2 or 3 was found in 36 (11.8%) of the 304 women, CIN 1 was found in 29 (9.5%), carcinoma was found in 1 (0.3%), and the cervix was normal in 238 (78.3%). Cervigram results were consistent with a low-grade lesion (category 3b) in 33.6% (102/304), a high-grade lesion (category 3c) in 5.9% (18/304), and carcinoma (category 3d) in 0.7% (2/ 304) of the women. Findings were negative or atypical but not requiring colposcopy in the remaining 59.9% (182/ 304) of the women, including 0.7% (2/304) and 0.3% (1/ 304) that were obscured views and technically defective, respectively. The prevalence of HPV infection was 46.4% (141/304); the prevalence of HPV in LISL was 65.3% (64/ 98) and 37.4% (77/206) in ASCUS. Table 2 shows the sensitivities, specificities, and positive and negative predictive values of cervicography and cervical HPV testing alone. The sensitivities of cytology and HPV testing for detecting CIN 2 or higher were 73.0% and 86.5% respectively. Sensitivity was low for cervicography when using the high-grade lesion or carcinoma cutoff Arch Pathol Lab Med Vol 128, November 2004 Cervicography and HPV Testing Howard et al 1259

4 Table 3. Comparison of the Observed Sensitivities and Specificities for the Combined Tests for Detecting Histologically Confirmed Cervical Intraepithelial Neoplasia 2 or 3 or Invasive Squamous Cell Carcinoma, in the 304 Women With Low-Grade Cervical Cytology at Study Entry* Sensitivity After Reduction Because of Random Test Observed P Value Cervical cytology ( atypical squamous cells of undetermined significance) or Cervigram assessment of low-grade lesion or higher ( ).06 Cervigram assessment of highgrade lesion or carcinoma ( ).97 Human papillomavirus DNA test positive ( ).14 Specificity After Reduction Because of Random Test Observed P Value ( ) 47.2 ( ) 33.0 ( ) * Using the method of Franco and Ferenczy, 25 which assumes that the second test is random with respect to disease detection. Data are given as percentage (95% confidence interval) unless otherwise indicated (24.3%) but improved when using the low-grade lesion or higher cutoff (83.8%). The cervigram assessment using the high-grade lesion or carcinoma cutoff identified 1 additional case of CIN 2 or 3, or true positive (1.50% [95% CI, 0.04% 8.20%]), increasing sensitivity from 73.0% to 75.7%. The cervigram assessment of a low-grade lesion or higher and the HPV test identified 8 additional true positives (21.6% [95% CI, 9.8% 38.2%]), increasing sensitivity to 94.6%. The addition of the cervigram using the high-grade lesion or carcinoma cutoff resulted in 5 additional false positives (1.8% [95% CI, 0.7% 4.6%]), reducing specificity from 52.5% for cytology alone to 47.2%. The addition of the cervigram using the low-grade lesion or higher cutoff and the HPV test identified 45 (16.9% [95% CI, 12.0% 22.7%]) and 43 (16.1% [95% CI, 12.0% 21.2%]) additional false positives, reducing specificity to 32.2% and 33.0% respectively. Actual sensitivities after adjusting for the increased sensitivity that would be expected if the second test performed randomly were 83.8% (vs 94.6% without adjustment, P.06) for the cervigram assessment and 86.5% (vs 94.6% without adjustment, P.14) for the HPV test, resulting in absolute decreases of 10.8% and 8.1% respectively, compared with the observed sensitivities (Table 3). The specificities for the combined tests were lower than for cytology alone, but were not as low as those calculated assuming the second test was random. For the HPV test, the observed specificities were statistically significantly better than the actual specificities after adjustment for chance increases in sensitivity. Classification of the 26 women with probable normal variants or lesions of doubtful significance as negative (including 2 women with histologic CIN 1 and 2 women with CIN 2 or 3) resulted in overestimates of specificities by approximately 10% (data not shown), but had little effect on sensitivity. All 3 women with technically defective cervigrams had normal histologic findings. COMMENT This study found that the use of a cervigram assessment threshold of a low-grade lesion or the use of a positive HPV test (Hybrid Capture II) result appeared to increase the sensitivity of cytology for detecting histologically confirmed CIN 2 or 3 by an increment of 21.6% in women with recent low-grade cytologic abnormalities. However, after adjusting the sensitivities of the combined tests to take into account increases in disease detection because of chance, these gains were reduced to 10.8% for cervicography and to 8.1% for HPV testing. None of the differences between the observed and actual sensitivities after adjustment were statistically significant; however, our power to detect these differences was small (20% 30%). Studies examining gains in the diagnostic use of combined tests should be interpreted with caution. Six studies 4,5,21 24 of combined cervical cytology and HPV testing for the detection of histologically confirmed cervical abnormalities were summarized by Franco and Ferenczy. 25 Five of the studies 4,5,21,23,24 demonstrated statistically significant gains in sensitivity for the combined method; however, after using an adjustment to take into account the chance increase in sensitivity from adding a second test, only 2 of the 5 studies 4,23 were found to have statistically significantly improved sensitivity. Previous studies 6,9,26 examining the usefulness of cervicography as an adjunct to cytology for women with low-grade cytologic abnormalities have reported increases in sensitivity of 25% to 45% above cytology alone. These studies did not take into account chance increases in sensitivity and may have overestimated the benefits of combined testing. Although we found that cervicography using a highgrade lesion interpretation to indicate a positive result had poor sensitivity, the use of a suspicious or atypical cervigram assessment has shown moderate to good sensitivity (61% 92%) in women with ASCUS or LSIL on cytology. 6,11,12,26 29 The sensitivity of HPV testing alone for detecting CIN 2 or 3 in the present study was 86.5%, but specificity was low (59.2%). These findings are similar to those of previous studies 7,14,15,30 using the Hybrid Capture II assay, in which sensitivity ranged from 89% to 98% and specificity from 47% to 73.9% among women referred with cytologic abnormalities. Some previous studies describing the diagnostic accuracy of cervicography 6,11,12,28,31 35 and HPV testing 31,36,37 were unable to verify disease status in women with negative test results, which can result in verification bias if corrections are not made. In screening studies in which disease prevalence is low, it is often not feasible to subject all women with negative test results to colposcopy. In such cases, a randomly selected sample of women with negative results can be examined by colposcopy to esti Arch Pathol Lab Med Vol 128, November 2004 Cervicography and HPV Testing Howard et al

5 mate the true performance of a screening test. 18 In the present study, all women underwent colposcopy. It is recognized that colposcopy with directed biopsy has imperfect sensitivity, 38,39 but it is the best tool available. An advantage of using combined tests, in addition to the potential for improved sensitivity, is the potential improvement in negative predictive value or the ability to rule out disease. This reassurance may be especially important in litigious societies or in settings where patient adherence with follow-up is problematic. It has been suggested that an additional benefit of improved sensitivity of combined testing is the possibility of increasing the cytologic screening interval, because the chances of missing a case of high-grade disease will be reduced. 40 However, improvements in sensitivity and negative predictive value when adding a second test may be undermined by the loss of specificity and positive predictive value, resulting in more false-positive referrals. 18 In this sample, the prevalence of HPV (46.4%) and the prevalence of a cervigram interpretation of a lesion (40.2%) were high, resulting in a loss of specificity for the addition of the HPV test, as a result of the large numbers of false-positive results for these tests. Although false-positive referrals would be generated, the combined testing was able to rule out disease in 97.7% of cases in which both tests were negative. The results of the present study should only be generalized to women with a recent abnormality on cervical cytology. The prevalence of CIN 2 or 3 is much lower ( 1% 41 ) in the general population with access to screening programs than in our sample (11.8%). Positive predictive value varies with the prevalence of the disease in a given setting, such that it is higher in higher prevalence settings because of the increased prior probability of disease, even if sensitivity of the diagnostic test is constant. The use of cervicography and HPV testing as adjuncts to cytology in lower disease prevalence screening settings would need to be evaluated in those settings. Preliminary results of the ASCUS/LSIL Triage Study 30 have suggested that HPV testing for triage of women with LSIL is not useful, because of the high prevalence of infection in these lesions, and that women with LSIL should be referred immediately to colposcopy. We did not have adequate sample size to stratify our analysis; however, the prevalence of HPV in LSIL in the present study was somewhat lower (65.3%) than in the ASCUS/LSIL Triage Study (83%). The HPV test result in our analysis was obtained at least 6 months after the initial LSIL diagnosis when women became eligible for the studies, and some regression of infection may have occurred. Because the study looked at short-term outcomes, it is not possible to know if longer-term outcomes, such as the incidence and mortality from cervical cancer in the context of a cervical cancer prevention program, would differ depending on whether HPV testing or cervicography was used. Therefore, adjustment for chance increases in the sensitivity of tests for long-term outcomes may also differ from results based on the short-term outcomes in this study. This study attempted to illustrate how the addition of a second test to cytology can increase sensitivity in part because of chance increases in detection. We did not have sufficient power to rule out the possibility that the 2 estimates were equivalent; however, the trends of decreasing sensitivity after adjustment were consistent for HPV testing and cervicography. Randomized trials comparing different management strategies will be informative, and adjustments to sensitivity estimates, if used, may alter the expected improvements in cost-effectiveness. The net benefits of improved sensitivity and negative predictive value of combined testing must be considered in relation to the availability of adjunctive testing and colposcopy resources and to patients adherence with follow-up in a given setting. This study was funded by grants MT13287 from the Medical Research Council of Canada, Ottawa, and from the Physicians Services Incorporated Foundation, North York, Ontario. We thank Stephen Walter for advice on data analysis and presentation. References 1. Parkin DM, Pisani P, Ferlay J. Global cancer statistics. CA Cancer J Clin. 1999;49: Nanda K, McCrory DC, Myers ER, et al. Accuracy of the Papanicolaou test in screening for and follow-up of cervical cytologic abnormalities: a systematic review. Ann Intern Med. 2000;132: Lytwyn A, Sellors JW, Mahony JB, et al. HPV Effectiveness in Lowgrade Paps (HELP) Study No. 1 Group. Comparison of human papillomavirus DNA testing and repeat Papanicolaou test in women with low-grade cervical cytologic abnormalities: a randomized trial. CMAJ. 2000;163: Cox JT, Lorincz AT, Schiffman MH, Sherman ME, Cullen A, Kurman RJ. Human papillomavirus testing by hybrid capture appears to be useful in triaging women with a cytologic diagnosis of atypical squamous cells of undetermined significance. Am J Obstet Gynecol. 1995;172: Wright TC, Sun XW, Koulos J. Comparison of management algorithms for the evaluation of women with low-grade cytologic abnormalities. Obstet Gynecol. 1995;85: Ferris DG, Payne P, Frisch LE, Milner FH, dipaola FM, Petry LJ. Cervicography: adjunctive cervical cancer screening by primary care clinicians. J Fam Pract. 1993;37: Manos MM, Kinney WK, Hurley LB, et al. Identifying women with cervical neoplasia using human papillomavirus DNA testing for equivocal Papanicolaou results. JAMA. 1999;281: Stafl A. Cervicography: a new method for cervical cancer detection. Am J Obstet Gynecol. 1981;139: Eskridge C, Begneaud WP, Landwehr C. Cervicography combined with repeat Papanicolaou test as triage for low-grade cytologic abnormalities. Obstet Gynecol. 1998;92: Jones MH, Jenkins D, Cuzick J, et al. Mild cervical dyskariosis: safety of cytologic surveillance. Lancet. 1992;339: Spitzer M, Krumholz BA, Chernys AE, Seltzer V, Lightman AR. Comparative utility of repeat Papanicolaou smears, cervicography, and colposcopy in the evaluation of atypical Papanicolaou smears. Obstet Gynecol. 1987;69: Schauberger CW, Rowe N, Gundersen JH, Jensen DP, Chadbourn M. Cervical screening with cervicography and the Papanicolaou smear in women with genital condylomata. J Reprod Med. 1991;36: Kesic VI, Soutter WP, Sulovic V, Juznic N, Aleksic M, Ljubic A. A comparison of cytology and cervicography in cervical screening. Int J Gynecol Cancer. 1993;3: Sellors JW, Lorincz AT, Mahony JB, et al. Comparison of self-collected vaginal, vulvar and urine samples with physician-collected cervical samples for human papillomavirus testing to detect high-grade squamous intraepithelial lesions. CMAJ. 2000;163: Shlay JC, Dunn T, Byers T, Baron AE, Douglas JM Jr. Prediction of cervical intraepithelial neoplasia grade 2 3 using risk assessment and human papillomavirus testing in women with atypia on Papanicolaou smears. Obstet Gynecol. 2000;96: ASCUS-LSIL Triage Study (ALTS) Group. Results of a randomized trial on the management of cytology interpretations of atypical squamous cells of undetermined significance. Am J Obstet Gynecol. 2003;188: Lytwyn A, Sellors JW, Mahony JB, et al. Adjunctive human papillomavirus testing in the 2-year follow-up of women with low-grade cervical cytologic abnormalities: a randomized trial and economic evaluation. Arch Pathol Lab Med. 2003;127: Franco EL. Statistical issues in human papillomavirus testing and screening. Clin Lab Med. 2000;20: Fleiss JL. Statistical Methods for Rates and Proportions. New York, NY: John Wiley & Sons; Uitenbroek DG. Binomial [SISA Web site] Available at: home.clara.net/sisa/binomial.htm. Accessed January 1, Hatch KD, Schneider A, Abdel-Nour WA. An evaluation of human papillomavirus testing for intermediate- and high-risk types as triage before colposcopy. Am J Obstet Gynecol. 1995;174: Ferenczy A, Franco E, Arseneau J, Wright TC, Richart RM. Diagnostic performance of Hybrid Capture human papillomavirus deoxyribonucleic acid assay combined with liquid-based cytologic study. Am J Obstet Gynecol. 1996;175: Arch Pathol Lab Med Vol 128, November 2004 Cervicography and HPV Testing Howard et al 1261

6 23. Cox JT, Schiffman MH, Winzelberg AJ, Patterson JM. An evaluation of human papillomavirus testing as part of referral to colposcopy clinics. Obstet Gynecol. 1992;80(pt 1): Kaufman RH, Adam E, Icenogle J, Reeves WC. Human papillomavirus testing as triage for typical squamous cells of undetermined significance and lowgrade squamous intraepithelial lesions: sensitivity, specificity, and cost-effectiveness. Am J Obstet Gynecol. 1997;177: Franco EL, Ferenczy A. Assessing gains in diagnostic utility when human papillomavirus testing is used as an adjunct to Papanicolaou smear in the triage of women with cervical cytologic abnormalities. Am J Obstet Gynecol. 1999; 181: Ferris DG, Schiffman M, Litaker MS. Cervicography for triage of women with mildly abnormal cervical cytology results. Am J Obstet Gynecol. 2001;185: McKinnon KJ, Ford RM, Hunter JC. Comparison of cytology and cervicography in screening a high risk Australian population for cervical human papillomavirus and cervical intraepithelial neoplasia. Aust NZ J Obstet Gynecol. 1993; 33: Mould TA, Singer A, Mansell ME, Gallivan S. Cervicography to triage women with borderline or mild dyskaryotic cervical PAP smears. Eur J Gynaecol Oncol. 2000;21: 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: Solomon D, Schiffman M, Tarone R, ALTS Study Group. 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: Schneider A, Zahm DM, Kirchmayr R, Schneider VL. Screening for cervical intraepithelial neoplasia grade 2/3: validity of cytologic study, cervicography, and human papillomavirus detection. Am J Obstet Gynecol. 1996;174: Schneider DL, Herrero R, Bratti C, et al. Cervicography screening for cervical cancer among 8460 women in a high-risk population. Am J Obstet Gynecol. 1999;180(pt 1): Denny L, Kuhn L, Risi L, et al. Two-stage cervical cancer screening: an alternative for resource-poor settings. Am J Obstet Gynecol. 2000;183: De Sutter P, Coibion M, Vosse M, et al. A multicentre study comparing cervicography and cytology in the detection of cervical intraepithelial neoplasia. Br J Obstet Gynaecol. 1998;105: Cecchini S, Bonardi R, Mazzotta A, Grazzini G, Iossa A, Ciatto S. Testing cervicography and cervicoscopy as screening tests for cervical cancer. Tumori. 1992;79: Kuhn L, Denny L, Pollack A, Lorincz A, Richard RM, Wright TC. Human papillomavirus DNA testing for cervical cancer screening in low-resource settings. J Natl Cancer Inst. 2000;92: Schiffman M, Herrero R, Hildesheim A, et al. HPV DNA testing in cervical cancer screening: results from women in a high-risk province of Costa Rica. JAMA. 2000;238: Massad LS, Halperin CJ, Bitterman P. Correlation between colposcopically directed biopsy and cervical loop excision. Gynecol Oncol. 1997;60: Belinson J, Qiao YL, Pretorius R, et al. Shanxi province cervical cancer screening study: a cross-sectional comparative trial of multiple techniques to detect cervical neoplasia. Gynecol Oncol. 2001;83: Kraemer HC. Evaluating Medical Tests: Objective and Quantitative Guidelines. Newbury Park, Calif: Sage Publications; Mitchell MF, Tortolero-Luna G, Wright T, et al. Cervical human papillomavirus infection and intraepithelial neoplasia: a review. J Natl Cancer Inst Monogr. 1996;21: Arch Pathol Lab Med Vol 128, November 2004 Cervicography and HPV Testing Howard et al