Comparison of Human Papillomavirus Distribution in Cytologic Subgroups of Low-Grade Squamous Intraepithelial Lesion

Transcription

1 288 Comparison of Human Papillomavirus Distribution in Cytologic Subgroups of Low-Grade Squamous Intraepithelial Lesion Rosemary E. Zuna, MD 1 Sophia S. Wang, PhD 2 Mark Schiffman, MD, MPH 2 Diane Solomon, MD 3 for the Atypical Squamous Cells of Undetermined Significance/Low- Grade Squamous Intraepithelial Lesion Triage Study Group 1 Pathology Department, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma. 2 Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland. 3 Division of Cancer Prevention, National Cancer Institute, Bethesda, Maryland. BACKGROUND. Low-grade squamous intraepithelial lesion (LSIL) subsumes the formerly delineated cytologic categories of human papillomavirus (HPV)-associated cell changes (koilocytotic atypia) and low-grade dysplasia/cervical intraepithelial neoplasia (CIN) Grade 1 (CIN1). In this study, the objective was to determine whether these 2 morphologic subcategories are characterized by differences in risk for CIN3 and/or HPV type distribution. METHODS. Within the Atypical Squamous Cells of Undetermined Significance/ Low-Grade Squamous Intraepithelial Lesion Triage Study, all cytologic interpretations of HPV cellular changes and CIN1 rendered by any of the pathology reviewers (community laboratory, clinical center, or Pathology Quality-Control Group) on referral Papanicolaou (Pap) tests or enrollment ThinPrep 1 Pap tests were included for analysis. HPV testing was performed by Hybrid Capture TM 2 (HC2) and by polymerase chain reaction based reverse-line blot analysis for 27 HPV types. The absolute risks of cumulative detection of CIN3 or cancer (CIN3 þ) and CIN2 or worse (CIN2 þ) over 2 years of follow-up were calculated for the various cytologic interpretations. RESULTS. For each review group and cytology preparation, most LSIL interpretations (from approximately 2 of 3 interpretations to 3 of 4 interpretations) were subcategorized as CIN1 rather than HPV cellular changes. HPV type 16 (HPV-16) was the most common HPV type and was identified in 21% to 24% of CIN1 and in Supported by the National Cancer Institute, National Institutes of Health, Department of Health and Human Services (contracts CN-55153, CN , CN-55155, CN-55156, CN-55157, CN , CN-55159, and CN ) Some of the equipment and supplies that were used in this study were donated or provided at reduced cost by Digene Corporation (Gaithersburg, MD), Cytyc Corporation (Marlborough, MA), National Testing Laboratories (Fenton, MO), Denvu (Tucson, AZ), TriPath Imaging, Inc. (Burlington, NC), and Roche Molecular Systems Inc. (Alameda, CA). Affiliations of the Atypical Squamous Cells of Unknown Significance Low-Grade Squamous Intraepithelial Lesion Triage Study Group: National Cancer Institute, Bethesda MD (D. Solomon, Project Officer; and M. Schiffman, Co-Project Officer); Clinical Centers: University of Alabama at Birmingham, AL (E. E. Partridge, Principal Investigator; L. Kilgore, Co- Principal Investigator; and S. Hester, Study Manager); University of Oklahoma, Oklahoma City, OK (J. L. Walker, Principal Investigator; G. A. Johnson, Co- Principal Investigator; and A. Yadack, Study Manager); Magee-Womens Hospital of the University of Pittsburgh Medical Center Health System, Pittsburgh, PA (R. S. Guido, Principal Investigator; K. McIntyre-Seltman, Co-Principal Investigator; R. P. Edwards, Investigator; and J. Gruss, Study Manager); University of Washington, Seattle, WA (N. B. Kiviat, Co-Principal Investigator; L. Koutsky, Co-Principal Investigator; and C. Mao, Investigator); Colposcopy Quality Control Group (D. Ferris, Principal Investigator [Medical College of Georgia, Augusta, GA]; J. T. Cox, Co-Investigator [University of California at Santa Barbara, Santa Barbara, CA]; and L. Burke, Co-Investigator [Beth Israel Deaconess Medical Center Hospital, Boston, MA]); Human Papillomavirus Quality Control Group (C. M. Wheeler, Principal Investigator [University of New Mexico Health Sciences Center, Albuquerque, NM]; C. Peyton-Goodall, Laboratory Manager [University of New Mexico Health Sciences Center, Albuquerque, NM]; and M. M. Manos, Co-Investigator [Kaiser Permanente, Oakland, CA]); Pathology Quality Control Group (R. J. Kurman, Principal Investigator [Johns Hopkins Hospital, Baltimore, MD]; D. L. Rosenthal, Co-Investigator [Johns Hopkins Hospital, Baltimore, MD]; M. E. Sherman, Co-Investigator [The National Cancer Institute, Rockville, MD]; and M. H. Stoler, Co-Investigator [University of Virginia Health Science Center, Charlottesville, VA]); Westat (Coordinating Unit [Rockville, MD]: J. Rosenthal, Project Director; M. Dunn, Data Management Team Leader; J. Quarantillo, Senior Systems Analyst; and D. Robinson, Clinical Center Coordinator; Quality of Life Group: D. M. Harper, Dartmouth Medical School [Hanover, NH]; A. T. Lorincz, Senior Scientific Officer [Digene Corporation, Gaithersburg, MD]; and B. Kramer, Senior Programmer/Analyst [Information Management Services, Inc., Silver Spring, MD]. Address for reprints: Rosemary E. Zuna, MD, Pathology Department, BMSB 451, University of Oklahoma Health Sciences Center, Oklahoma City. OK 73104; Fax: (405) ; rosemary-zuna@ouhsc. edu Received May 30, 2006; accepted June 28, *This article is a US Government work and, as such, is in the public domain in the United States of America. Published 2006 by the American Cancer Society* DOI /cncr Published online 1 September 2006 in Wiley InterScience (

2 HPV Cellular Changes and CIN1/Zuna et al % to 18% of HPV cellular changes. Nononcogenic types were identified alone in from 9% to 11% of CIN1 compared with 17% to 20% of HPV cellular change. The absolute risks of CIN1 and HPV cellular changes for cumulatively detected CIN3 þ ranged from 12% to 16% for CIN1 and from 6% to 9% for HPV cellular changes. CONCLUSIONS. Both cytologic subcategories of LSIL were associated predominantly with oncogenic HPV types; however, the proportion of HPV types was lower and the absolute risks for CIN3 þ were higher for CIN1 compared with HPV cellular changes. The concordance in subcategorizing LSIL was low, and the authors concluded that the diagnostic distinction is of limited clinical utility for individual patient management. Cancer (Cancer Cytopathol) 2006;108: Published 2006 by the American Cancer Society.* KEYWORDS: low-grade squamous intraepithelial lesion, koilocytotic atypia, cervical intraepithelial neoplasia Grade 1, human papillomavirus, Papanicolaou test, cervical cancer screening. With their seminal descriptions of koilocytotic atypia in the squamous epithelium of the cervix, Koss and Durfee 1 and Ayre 2 suggested the possibility of a viral association with premalignant lesions of the cervix. Meisels et al. 3,4 further identified the balloon cell or koilocyte as the primary cytologic manifestation of the flat condylomata and stressed the importance in distinguishing those lesions from mild dysplasia (cervical intraepithelial neoplasia Grade 1 [CIN1]). Early terminology distinctions between koilocytic atypia and mild dysplasia were based on the belief that viral infection was distinct etiologically from carcinogenesis and that mild dysplasia was the true early precursor lesion to cervical cancer. Since then, however, a group of 13 to 15 human papillomaviruses (HPV) types have been recognized as oncogenic; and, today, it is well accepted that HPV is unequivocally the etiologic agent for cervical cancer and its precursors. 5 7 Although koilocytotic changes are accepted as HPV-induced abnormalities, 8,9 their risk relative to CIN1 for a subsequent diagnosis of cervical precancer still is debated. In some countries, the distinction still is made in daily practice. The morphologic characteristic of cellular changes of HPV (HPV changes) (also known as koilocytotic atypia ) is the perinuclear clearing or halo, which is devoid of organelles and is surrounded by a dense, cytoplasmic rim filled with tonofilaments Studies attempting to distinguish HPV changes from CIN1 have shown similar biology and lack of diagnostic reproducibility for separating these categories. 14,15 Consequently, the Bethesda System of cytologic classification, which was defined originally in ,17 and revised in 2001, 18 combined lesions that were described as HPV changes with lesions that were diagnosed as CIN1 into a single category known as lowgrade squamous intraepithelial lesion (LSIL). A significant amount of new information is now available regarding HPV types and their categorization by oncogenicity. 6,19 21 The objective of the current study, which was conducted within the Atypical Squamous Cells of Undetermined Significance (ASCUS)/ LSIL Triage Study (ALTS), a multicenter, randomized clinical trial that compared management strategies for low-grade and equivocal cytologic categories, was to investigate the distribution of HPV types comprehensively within the 2 morphologic patterns of LSIL- HPV changes and CIN1- and to identify any differences in the risk for CIN3 or HPV type distribution. MATERIALS AND METHODS Study Population The structure of ALTS has been described previously. 22 Briefly, enrollment was offered to women who were referred from the community with ASCUS or LSIL conventional Papanicolaou (Pap) smear interpretations from November 1996 through December 1998 at 4 clinical centers: the University of Alabama (Birmingham, AL), Magee-Womens Hospital of the University of Pittsburgh Medical Center Health System (Pittsburgh, PA), the University of Oklahoma (Oklahoma City, OK), and the University of Washington (Seattle, WA). Written informed consent was obtained from each participant. At enrollment, cervical specimens and complete questionnaire data, including demographic, hormone, and sexual histories, were collected. The current analysis included all women who were diagnosed with LSIL in referral and/or enrollment cytology specimens, as specified below. Women were followed every 6 months for 2 years for disease outcomes of histologic CIN2 or worse (CIN2 þ) and CIN3 or cancer (CIN3 þ). Because CIN2 is the community threshold for treatment, we considered the clinical center

3 290 CANCER (CANCER CYTOPATHOLOGY) October 25, 2006 / Volume 108 / Number 5 TABLE 1 Distribution of Cervical Intraepithelial Neoplasia Grade 1and Human Papillomavirus (HPV) Changes among Varying Definitions of HPV Status by 4 Distinct Definitions: for Low-Grade Intraepithelial Squamous Lesion: Original Referral Papanicolaou (Pap) Test (Conventional) Interpretation at the Clinical Site, Pathology Quality-Control Review Interpretation of the Original Pap Test, Clinical Center Interpretation of the enrollment ThinPrep 1 Pap Test, and Pathology Quality-Control Review of the Enrollment ThinPrep 1 Pap Test No. of patients (%) HC2-positive HC2-negative Diagnosis HPV16 non-hpv16 PCR-positive, HPV16 non-hpv16 PCR-positive, Chi-square P value Original referral Pap (conventional) interpretation at the clinical sites (n ¼ 1476)* CIN1 (n ¼ 1262) 273 (22) 628 (50) 95 (8) 71 (6) 6 (0) 17 (1) 37 (3) 135 (11) HPV y (n ¼ 178) 26 (15) 80 (45) 25 (14) 15 (8) 2 (1) 1 (1) 9 (5) 20 (11).009 Path QC review interpretation of the original Pap (n ¼ 1283) { CIN1 (n ¼ 1067) 228 (21) 546 (51) 88 (8) 53 (5) 4 (0) 18 (2) 35 (3) 95 (9) HPV y (n ¼ 212) 29 (14) 94 (44) 32 (15) 13 (6) 2 (1) 5 (2) 10 (5) 27 (13).003 Clinical center interpretation of the enrollment ThinPrep 1 Pap (n ¼ 1244) CIN1 (n ¼ 675) 163 (24) 372 (55) 45 (7) 28 (4) 0 (0) 3 (0) 12 (2) 52 (8) HPV y (n ¼ 412) 74 (18) 228 (55) 64 (16) 24 (6) 1 (0) 5 (1) 5 (1) 11 (3) <.0001 Path QC review of the enrollment ThinPrep 1 (n ¼ 1196) k CIN1 (n ¼ 931) 215 (23) 532 (57) 86 (9) 45 (5) 0 (0) 3 (0) 16 (2) 34 (4) HPV y (n ¼ 254) 42 (17) 141 (56) 41 (16) 13 (5) 0 (0) 7 (3) 5 (2) 5 (2) <.0001 HC2 indicates Hybrid Capture 2; PCR, polymerase chain reaction; HPV, human papillomavirus; HPV cellular changes, Pap, Papanicolaou; CIN1, cervical intraepithelial neoplasia Grade 1; Path QC, pathology quality control. * Low-grade squamous intraepithelial lesion, not otherwise specified (LSIL, NOS), n ¼ 36 women; missing, n ¼ 96 women. y HPV refers to HPV cellular changes. { LSIL, NOS in 4 women; data missing for 89 women. LSIL, NOS in 57 women; data missing for 98 women. k LSIL, NOS in 11 women; data missing for 94 women. diagnosis of CIN2 þ as 1 outcome, and the Pathology Quality-Control (QC) group review diagnosis for the more stringent outcome of CIN3 þ as a true precancer state. Definition of LSIL On average, women were enrolled into ALTS 2 months after the original abnormal Pap test result of ASCUS or LSIL ( referral Pap ). At enrollment, a liquid-based cytology specimen (ThinPrep 1 ; Cytyc Corporation, Marlborough, MA) was collected with a broom-type sampler (Papette TM broom; Wallach Surgical Devices, Incorporated, Orange, CT) and transferred to a PreservCyt 1 vial (Cytyc Corporation). A second cervical specimen was then collected using a Dacron swab andplacedintospecimentransportmedium TM (STM) (Digene Corporation, Gaithersburg, MD), which was used for the prototype linear-array polymerase chain reaction (PCR) HPV DNA typing (see below). The referral Pap smears were read by the community laboratory, and the enrollment ThinPrep Pap tests initially were interpreted by a clinical site pathologist; both the referral smears and the enrollment ThinPrep Pap tests were then reviewed by the Pathology QC Group. 22,23 Therefore, for each woman, there were 4 independently rendered cytologic interpretations: 1) the community laboratory interpretation of the original referral Pap smear (which was always ASCUS or LSIL, according to the enrollment eligibility criteria); 2) the Pathology QC Group masked review of the original referral Pap smear; 3) the clinical center interpretation of the enrollment ThinPrep test; and 4) the Pathology QC masked review of the enrollment Thin- Prep test. In total, 2546 women were identified with LSIL by any 1 or more of the 4 interpretation groups, including 2371 women for whom HPV data were available. 23 Those 2371 women constitute the population for this study, whereas the remaining 175 women without HPV data were excluded. Cytologic Definition of HPV Changes and CIN1 For research purposes and data collection, the broad category of LSIL was divided into 3 subcategories: 1) LSIL, not otherwise specified (NOS); 2) LSIL, cellular changes of HPV; and 3) LSIL, CIN1. For the current analysis, we considered that LSIL, cellular changes of HPV (HPV changes) was analogous to the previously used term koilocytotic atypia and that LSIL, CIN1

4 HPV Cellular Changes and CIN1/Zuna et al. 291 (CIN1) was analogous to mild dysplasia/cin1. Because only 4% of LSIL interpretations overall were NOS, as indicated in Table 1, this subcategory was not included in subsequent analyses. The clinical center and Pathology QC pathologists applied criteria that were based largely on the Bethesda System 17 to distinguish HPV changes from CIN1. Cells with well defined cytoplasmic halos and thickened cytoplasmic rims in association with only mild nuclear changes nuclear enlargement, hyperchromasia, and/or smudged chromatin were interpreted as HPV changes (Fig. 1). 4 In contrast, the cases that were categorized as CIN1 demonstrated some superficial squamous cells with nuclei that were at least 3-fold the size of a normal intermediate cell nucleus and hyperchromatic. Perinuclear halos could be present as long as the nuclear enlargement and hyperchromasia were sufficient for dysplasia. FIGURE 1. These photomicrographs of representative cells illustrate the diagnostic categories that were used for low-grade squamous intraepithelial lesions in the current study (modified Papanicolaou stain; original magnification 63). (A) Human papillomavirus cellular changes and (B) cervical intraepithelial neoplasia of Grade 1. HPV Testing Two types of HPV tests were used in this study. The PreservCyt sample was used for the Hybrid Capture 2 TM (HC2) test with the high-risk probe (Probe B), which targets the following oncogenic HPV types: HPV type (HPV-16), HPV-18, HPV-31, HPV-33, HPV- 35, HPV-39, HPV-45, HPV-51, HPV-52, HPV-56, HPV- 58, HPV-59, and HPV DNA isolated from the STM samples was analyzed for individual HPV genotype(s) by using a PCR-based reverse-line blot technique 25,26 (Roche Molecular Systems, Alameda, CA), which utilizes biotinylated PGMY09/PGMY11 consensus primers that amplify a 450-base pair fragment of HPV L1 open reading frame for a large number of HPV types. This approach can identify 1 or more of 27 individual HPV types in a single test. Included in this test are primers for oncogenic HPV types HPV-16, HPV-18, HPV-31, HPV-33, HPV-35, HPV-39, HPV-45, HPV-51, HPV-52, HPV-56, HPV-58, HPV-59, and HPV- 68; HPV types HPV-6, HPV-11, HPV-26, HPV-40, HPV-42, HPV-53, HPV-54, HPV-55, HPV-57, HPV-66, HPV-73, HPV-82, HPV-83, and HPV-84; and b-globin DNA as an internal standard. 24 Tests with multiple HPV types were considered to be oncogenic if at least 1 oncogenic HPV type was identified or if only types were identified. We evaluated viral load by using HC2 signal-strength measurements (in relative light units [RLU]), which were related linearly to the number of HPV copies and converted to pg/ml of HPV DNA. 27,28 All cytologic interpretations and HPV tests were performed independently without knowledge of the other results. Because it was proposed recently that HPV-66 is an oncogenic type, 20,21 we also conducted all analyses with HPV-66 included as an oncogenic type and observed minimal impact on our results. Similarly, the inclusion of HPV-26, HPV-73, and/or HPV-82 as oncogenic types did not alter our results. Statistical Methods Within each of the 4 interpretation groups identified above, we compared the distribution (numbers and percentages) of women who were diagnosed with CIN1 or HPV changes according to the different review groups by using k statistics and the McNemar test for asymmetry. We also compared the distribution of women who were diagnosed with CIN1 or HPV changes according to their HPV status. HPV status first was dichotomized as HC2-positive or HC2-negative; then, it was stratified further by the following risk categories based on PCR results: HPV-16-positive; any

5 292 CANCER (CANCER CYTOPATHOLOGY) October 25, 2006 / Volume 108 / Number 5 oncogenic HPV-positive, excluding HPV-16; oncogenic HPV-negative, HPV-positive; and HPV-negative by PCR. Differences in the distribution of HPV status between CIN1 and HPV changes were quantified by using the Fisher exact test for significance. P values (2-tailed).05 were considered statistically significant. This analysis was repeated for each of the 4 categories of cytologic interpretation. Oncogenic and HPV type distribution between CIN1 and HPV changes was delineated further. Absolute risks (or positive predictive value) for disease outcomes of CIN2 þ, as defined by the clinical center pathologists, and for CIN3 þ, as defined by the Pathology QC Group, were calculated. All results presented are stratified by the 4 cytologic interpretations and by HPV status. RESULTS We analyzed LSIL interpretations, subcategorized as HPV changes or CIN1, that were rendered by 4 independent reviews: 1) the community laboratory interpretation of the original referral Pap test; 2) the Pathology QC review of the original referral Pap test; 3) the clinical center interpretation of the enrollment ThinPrep Pap test; and 4) the Pathology QC review of the enrollment ThinPrep Pap test. These reviews define the 4 different definitions of LSIL. Most LSIL interpretations (from approximately 2 of 3 interpretations to 3 of 4 interpretations) were subcategorized as CIN1 rather than cellular changes of HPV. The percentages of LSIL categorized as CIN1 or HPV changes, however, varied within the 4 interpretation groups of LSIL (Table 1). Among women who had interpretations of LSIL on their original referral Pap test (n ¼ 1476 women), 86% and 12% were identified as CIN1 and HPV changes, respectively; among women who had interpretations of LSIL in the Pathology QC review of the referral Pap test (n ¼ 1283 women), 83% and 17% were identified as CIN1 and HPV changes, respectively; among women who had interpretations of LSIL in the clinical center review of their ThinPrep Pap test (n ¼ 1244 women), 65% and 33% were identified as CIN1 and HPV changes, respectively; and, among women who had interpretations of LSIL in the Pathology QC review of their ThinPrep Pap test (n ¼ 1196 women), 78% and 21% were identified as CIN1 and HPV changes, respectively. There was moderate agreement (77%) on the Pathology QC interpretations of CIN1 and HPV changes for the original referral Pap test compared with the paired ThinPrep Pap test from the same women (k ¼.2). However, agreement was poor (37%) between interpretations that were made by the community laboratory and the clinical center; and, in general, those interpretations were more severe than the interpretations made by the Pathology QC group (k <.05; McNemar P ¼.001). Therefore, in this report, we present the results from all 4 cytologic interpretations. In total, 2546 women were diagnosed with LSIL (HPV cellular changes or CIN1) by 1 ofthe4pathology interpretations. Of 827 women who were diagnosed with HPV changes by any of the 4 interpretations, only 4 women (<1%) had results that were identified as HPV changes according to 4 interpretations; 45 women (5%) had results that were identified as HPV changes according to 3 interpretations, 211 women (26%) had results that were identified as HPV changes according to 2 interpretations, and 567 women (69%) had results that were identified as HPV changes according to a single interpretation. Similarly, of 2312 women who were diagnosed with CIN1 by any of the 4 interpretations, 129 women (6%) had results that were identified as CIN1 according to all 4 interpretations; 359 women (16%) had results that were identified as CIN1 according to 3 interpretations, 788 women (34%) had results that were identified as CIN1 according to 2 interpretations, and 1036 women (45%) had results that were identified as CIN1 according to a single interpretation. For all interpretations of LSIL (i.e., Table 1), the HPV risk category (according to the HC2 and PCR results) varied significantly between CIN1 and HPV changes, as measured with the Fisher exact test. CIN1 was associated with a higher percentage of HC2-positive/PCR HPV-16-positive resutls (21 24%) compared with HPV changes (14 18%). CIN1 also showed a higher percent of HC2-positive/PCR oncogenic typepositive (HPV-16-negative) results compared with HPV changes, but only for the referral smear interpretations and not for the enrollment ThinPrep cytology. The percentage of HC2-positive women who had types identified by PCR was much lower than the percentage of HC2-positive women who had oncogenic types identified by PCR. The relative numbers differed notably for HPV changes and CIN1 diagnoses. The percentage of HPV-positive women ranged from 7% to 9% for women who had an interpretation of CIN1 and was nearly 2-fold higher for women who had HPV changes (14 16%) for all interpretation groups. The numbers of tests that had discordant HC2 results and PCR results were small and accounted for 6% to 15% of all CIN1 and HPV changes (Table 1). Specifically, there were a few women in all categories who had HC2-positive results but HPV-negative PCR results (4 8%) or HC2-negative results but HPV-positive PCR results for either oncogenic or types (2 8%). 29 The remaining 2% to 13% of women

6 HPV Cellular Changes and CIN1/Zuna et al. 293 TABLE 2 The Absolute Risk for a Diagnosis of Cervical Intraepithelial Neoplasia Grade 3 or Worse by Expert Pathology Review According to Cervical Intraepithelial Neoplasia Grade 1 and Human Papillomavirus (HPV) Changes among Various Definitions of HPV Status According to Distinct Definitions for Low-Grade Squamous Intraepithelial Lesion: Original Referral Papanicolaou (Pap) Test (Conventional) Interpretation at the Clinical Site, Pathology Quality-Control Review Interpretation of the Original Pap Test, Clinical Center Interpretation of the enrollment ThinPrep 1 Pap Test, and Pathology Quality-Control Review of the Enrollment ThinPrep 1 Pap Test Absolute risk (range), % HC2-positive HC2-negative Diagnosis Total HPV-16 non-hpv-16 PCRpositive, HPV-16 non-hpv-16 PCRpositive, Original referral Pap (conventional) interpretation at the clinical sites (n ¼ 1476) CIN1 (n ¼ 1262) 16.2 ( ) 39.9 ( ) 11.9 ( ) 4.2 ( ) 9.9 ( ) 33.3 ( ) 11.8 ( ) 2.7 ( ) 3.7 ( ) HPV y (n ¼ 178) 8.8 ( ) 34.6 ( ) 5.0 ( ) 4.0 ( ) 6.7 ( ) * * 0 (0 33.6) 0 (0 16.8) Path QC review interpretation of the original Pap (n ¼ 1283) CIN (n ¼ 1067) 13.9 ( ) 36.0 ( ) 9.3 ( ) 4.6 ( ) 13.2 ( ) * 5.6 ( ) 2.9 ( ) 2.1 ( ) HPV y (n ¼ 212) 6.1 ( ) 17.2 ( ) 5.3 ( ) 3.1 ( ) 7.7 ( ) * * 0 (0 30.8) 3.7 ( ) Clinical center interpretation of the enrollment ThinPrep 1 Pap (n ¼ 1244) CIN1 (n ¼ 675) 14.3 ( ) 30.1 ( ) 10.8 ( ) 4.4 ( ) 7.1 ( ) * * 8.3 ( ) 1.9 ( ) HPV y (n ¼ 412) 8.1 ( ) 23.0 ( ) 5.7 ( ) 1.6 ( ) 4.2 ( ) * * * 18.2 ( ) Path QC review of the enrollment ThinPrep 1 (n ¼ 1196) CIN1 (n ¼ 931) ( ) 28.4 ( ) 8.5 ( ) 3.5 ( ) 8.9 ( ) * * 6.3 ( ) 0 (0 10.3) HPV y (n ¼ 254) 7.4 ( ) 28.6 ( ) 2.8 ( ) 2.4 ( ) 0 (0 24.7) * 0 (0 41.0) * * PCR indicates polymerase chain reaction; HC2, Hybrid Capture 2; HPV, human papillomavirus; HPV cellular changes; Pap, Papanicolaou; CIN1, cervical intraepithelial neoplasia Grade 1; Path QC, pathology quality control. * Denominator 5. y HPV refers to HPV cellular changes. had HC2-negative results and PCR-negative results, as described in full previously. 23 Higher percentages of referral smear interpretations were HPV-negative according to both HC2 and PCR results compared with the ThinPrep interpretation, presumably because of the resolution of some HPV infections in the interval between the referral smear and the enrollment specimen collections (on average, 2 months later) used for ThinPrep and HPV testing. The absolute risks or positive predictive values for developing CIN3 þ within the 2-year follow-up are shown in Table 2. The absolute risk was highest for the women with HC2-positive/HPV-16-positive CIN1 results across all 4 cytologic interpretation groups (range, %). The absolute risk for women who had HPV changes and the same HPV risk status ranged from 17.2% to 34.6%. For women who had HC2- positive results and PCR-positive results for oncogenic HPV types other than HPV-16, the absolute risks for developing CIN3 þ were substantially lower for both CIN1 (range, %) and HPV changes (range, %) (Table 2). The absolute risks for women who had HC2-positive results and HPVpositive PCR results were lower still (CIN1: range, %; HPV changes: range, %). For women with CIN1 who had HC2-negative results, the absolute risks of CIN3 þ were very low; and reliable comparisons of CIN1 with HPV changes were not possible because of small numbers. In particular, the risk for CIN3 þ among women who had HPV-negative results according to both HC2 and PCR was very low ( %), except in 1 group with small numbers (2 of 11 women; 18%). The absolute risk for progression to CIN2 þ (Table 3) had a pattern that was similar to that for CIN3 þ. Again, the absolute risk for progression to CIN2 þ was highest for women who had HC2-positive/PCR HPV-16-positive results (CIN1: range, %; HPV changes: range, %). These risks were nearly double the absolute risks for CIN2 þ for women who had HC2-positive results and PCR-positive results for oncogenic HPV types other than HPV-16 (CIN1: range, %; HPV changes: range, %). Among women who had HC2-positive/PCR HPV-positive results, higher absolute risks for CIN2 þ were observed for women who had CIN1 (range: %) than for women who had HPV changes (range: %), although none of the comparisons were statistically significant, probably because of the small sample size in each strata. Women

7 294 CANCER (CANCER CYTOPATHOLOGY) October 25, 2006 / Volume 108 / Number 5 TABLE 3 The Absolute Risk for a Diagnosis of Cervical Intraepithelial Neoplasia Grade 2 or Worse by Clinical Center Review According to Cervical Intraepithelial Neoplasia Grade 1 and Human Papillomavirus (HPV) Changes among Various Definitions of HPV Status According to Distinct Definitions for Low-Grade Squamous Intraepithelial Lesion: Original Referral Papanicolaou (Pap) Test (Conventional) Interpretation at the Clinical Site, Pathology Quality-Control Review Interpretation of the Original Pap Test, Clinical Center Interpretation of the enrollment ThinPrep 1 Pap Test, and Pathology Quality-Control Review of the Enrollment ThinPrep 1 Pap Test Absolute risk (range), % HC2-positive HC2-Negative Diagnosis Total HPV-16 non-hpv-16 PCRpositive, HPV-16 non-hpv-16 PCRpositive, Original referral Pap (conventional) interpretation at the clinical sites (n ¼ 1476) CIN1 (n ¼ 1262) 27.2 ( ) 52.8 ( ) 24.8 ( ) 17.9 ( ) 16.9 ( ) 33.3 ( ) 11.8 ( ) 5.4 ( ) 9.6 ( ) HPV y (n ¼ 178) 12.9 ( ) 42.3 ( ) 12.5 ( ) 8.0 ( ) 0 (0 21.8) * * 0 (0 33.6) 0 (0 16.8) Path QC review interpretation of the original Pap (n ¼ 1283) CIN1 (n ¼ 1067) 25.3 ( ) 48.3 ( ) 23.3 ( ) 17.1 ( ) 20.8 ( ) * 5.6 ( ) 11.4 ( ) 6.3 ( ) HPV y (n ¼ 212) 14.8 ( ) 41.4 ( ) 12.8 ( ) 15.6 ( ) 0 (0 24.7) * * 0 (0 30.8) 7.4 ( ) Clinical center interpretation of the enrollment ThinPrep 1 Pap (n ¼ 1244) CIN1 (n ¼ 675) 24.8 ( ) 42.9 ( ) 22.6 ( ) 15.6 ( ) 10.7 ( ) * * 8.3 ( ) 3.9 ( ) HPV y (n ¼ 412) 17.8 ( ) 37.8 ( ) 15.8 ( ) 10.9 ( ) 12.5 ( ) * * * 9.1 ( ) Path QC review of the enrollment ThinPrep 1 (n ¼ 1196) CIN1 (n ¼ 931) 24.8 ( ) 42.8 ( ) 22.2 ( ) 19.8 ( ) 20.0 ( ) * * 12.5 ( ) 2.9 ( ) HPV y (n ¼ 254) 16.9 ( ) 50 ( ) 12.8 ( ) 4.9 ( ) 0 (0 24.7) * 0 (0 41.0) * * HC2 indicates Hybrid Capture 2; PCR, polymerase chain reaction; HPV, human papillomavirus; HPV cellular changes, Pap, Papanicolaou; CIN1, cervical intraepithelial neoplasia Grade 1; Path QC, pathology quality control. * Denominator 5. y HPV refers to HPV cellular changes. who had HC2-negative, PCR-negative results had the lowest absolute risk for CIN2 þ, and further differentiation by CIN1 and HPV changes could not be evaluated because of the small numbers. Type-specific PCR analyses demonstrated that, overall, HPV-16 was the most frequent HPV type found and was present in 21.7% to 24.1% of specimens with CIN1 and in 14.6% to 18.2% of specimens with HPV changes (Table 4). Nononcogenic PCR-positive specimens were more likely to be HC2-positive than HC2-negative samples, largely because of the known cross reactivity of the HC2 assay for HPV types when viral loads are high, which is typical for LSIL. Because the differences in HPV status between CIN1 and HPV changes were related primarily to a higher percentage of HPV types in women who had HPV changes compared with women who had CIN1, the distribution of individual HPV types was examined further. Of the HPV types we evaluated, there appeared to be a higher percentage of HPV-53, HPV-66, and HPV-6 among women who had HPV changes compared with women who had CIN1. Finally, we evaluated viral load by using HC2 analysis that was restricted to women who had a single type of HPV identified by PCR analysis. We excluded the women who had multiple types of HPV (53 67% of women with HPV changes, 56 60% of women with CIN1), because the contribution of each HPV type to the overall viral load could not be determined. Women with HPV changes who were HC2- positive/pcr oncogenic HPV-positive had higher viral load median values (RLU range, pg/ml) than women with CIN1 and the same HPV status (RLU range, pg/ml) for all cytologic interpretations except for the Pathology QC interpretation of the enrollment ThinPrep Pap test (RLU: HPV changes, 474 pg/ml; CIN1, 796 pg/ml). Similarly, when the analysis was restricted to HPV-16-only infections, the same 3 cytologic interpretations yielded higher viral load estimates (RLU range, pg/ml) for HPV changes than for CIN1 (RLU range, pg/ml). DISCUSSION The Bethesda System 17,18 for cytologic reporting combines HPV cellular changes and CIN 1 into a single group, that is, LSIL. The 2002 American Society for Colposcopy and Cervical Pathology consensus management guidelines 30 recommend that all women

8 HPV Cellular Changes and CIN1/Zuna et al. 295 TABLE 4 Distribution of Selected Human Papillomavirus (HPV) Types among Women with Positive Hybrid Capture 2-Positive Results Diagnosed as Low-Grade Squamous Intraepithelial Lesion Stratified by Cervical Intraepithelial Neoplasia Grade 1 and HPV Changes No. of women (%) * Diagnosis LSIL (Original referral pap, clinical center) LSIL (Original referral pap, path QC review) LSIL (ThinPrep, clinical center interpretation) LSIL (ThinPrep, path QC review) CIN Oncogenic HPV HPV (22.1) 232 (21.7) 163 (24.1) 215 (23.1) Non-HPV-16 y 789 (69.9) 756 (70.8) 524 (77.7) 746 (80.1) Nononcogenic HPV HPV (6.1) 77 (7.2) 48 (7.1) 94 (10.1) HPV (6.5) 78 (7.3) 55 (8.1) 92 (9.9) HPV-6 52 (4.1) 56 (5.2) 23 (3.4) 52 (5.6) HPV changes Oncogenic HPV HPV (18.2) 32 (14.8) 110 (19.3) 47 (17.7) Non-HPV-16 y 127 (59.4) 132 (61.1) 437 (76.8) 198 (74.7) Nononcogenic HPV HPV (8.4) 31 (14.4) 80 (14.1) 38 (14.3) HPV (8.9) 33 (15.3) 72 (12.7) 35 (13.2) HPV-6 15 (7.0) 11 (5.1) 49 (8.6) 28 (10.6) LSIL indicates low-grade squamous intraepithelial lesion; Pap, Papanicolaou; Path QC, pathology quality control; HPV, human papillomavirus; CIN1, cervical intraepithelial neoplasia Grade 1. * Women who had multiple types of HPV infection are represented in multiple rows. y HPV types 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68. with a cytologic interpretation of LSIL should be referred for colposcopy, primarily to exclude a higher grade lesion. However, the significance of lesions that harbor koilocytes with minimally enlarged nuclei remains in question. The current analysis was undertaken to compare the 2 cytomorphologic subcategories of LSIL that are used in the ALTS Study HPV cellular changes and CIN1 to determine whether the distribution of associated HPV types and/or the risk for histologically confirmed CIN3 warrants making this cytologic distinction. Referral smears (interpreted by the community laboratory and Pathology QC) and the enrollment ThinPrep tests (interpreted by the clinical center and Pathology QC) provided up to 4 interpretations for each woman. The majority of LSIL findings were subcategorized as CIN1 compared with HPV changes, but the ratio varied by interpretation group. Diagnostic consensus for the subcategorization of LSIL was low: Less than 1% of results that were interpreted as HPV changes were identified as such by all 4 interpretations; for results that were interpreted as CIN1, only 6% were interpreted as such by a consensus of all 4 groups. Most CIN1 and HPV changes were associated with oncogenic HPV types, however, CIN1 was associated somewhat more frequently with higher risk HPV categories compared with HPV changes. Specifically, CIN1 demonstrated a higher percentage of HC2-positive/ PCR HPV-16-positive findings (21 24%) compared with HPV cellular changes (14 18%). Conversely, the percentage of HPV-positive results was almost 2-fold higher for HPV changes compared with CIN1. HPV type-specific analyses supplemented these findings. We noted 3 HPV types (HPV-53, HPV-66, and HPV-6) that had higher prevalence in HPV changes compared with CIN1. Of particular interest was the higher proportion in HPV changes of HPV-66, which recently was reported as a probable high-risk type. 20,21 It is noteworthy that HPV-6, which has been associated with exophytic warts, 31 was present in only a minority of LSIL of either subcategory, although it was slightly more common in HPV changes. Nevertheless, both HPV changes and CIN1 still were associated predominantly with oncogenic HPV types, notably HPV- 16, consistent with previous reports. 32 Absolute risk calculations reveal that, regardless of cytologic interpretation HPV cellular changes or CIN1 women who had HC2-positive/HPV-16-positive PCR results had the highest absolute risks for CIN3 þ and CIN2 þ. However, given the same HPV status (e.g., both HC2 and PCR oncogenic HPV-positive), the absolute risks for CIN3 þ and CIN2 þ consistently, but very slightly, were higher for CIN1 than for HPV changes.

9 296 CANCER (CANCER CYTOPATHOLOGY) October 25, 2006 / Volume 108 / Number 5 Finally, Jovanovic et al. 33 previously reported that they observed the presence of cytoplasmic halos in perimenopausal or postmenopausal women that did not appear to be HPV-related. In our evaluation of whether HPV changes may be attributed to these characteristics, along with the use of oral contraceptives or hormones, we did not observe a correlation (data not shown). An unavoidable, real-life limitation of the current study is the poor reproducibility associated with cytologic distinctions between HPV changes and CIN1. 15,34,35 Therefore, we reported our results from 4 independent cytologic reviews that demonstrated the consistency of our findings. Because cross-reactivity of the HC2 probe B with HPV types has been reported previously, 24,36 we included results from both HC2 and the reverse-line blot PCR technique to define HPV risk status. We noted that the inclusion of HPV-66 as an oncogenic type did not alter our results. Our study strengths include the ALTS study design, which included the relatively large number of women with HPV changes and CIN1 for whom HPV typing data were available, and the ability to identify multiple HPV genotypes in a single sample. The longitudinal follow-up provided nearly complete ascertainment of disease outcome. Earlier studies were hampered by relatively small numbers of patients and/or limited numbers of HPV types included in the assays. These may have underestimated the number of multiple HPV types present per woman or the spectrum of types present, which significantly hampers the association of HPV type with morphologic pattern. The current results confirm that LSIL is overwhelmingly HPV-positive, but on an epidemiologic level, there are subtle distinctions between CIN1 and HPV changes with regard to the distribution of HPV types and the absolute risk for histologically confirmed CIN3 þ. However, there was considerable overlap in the HPV types associated with the 2 morphologic patterns, with HPV-16 the single most commonly identified type for both CIN1 and HPV changes. In addition, the interobserver reproducibility for rendering these interpretations was low. Therefore, on an individual patient level, clinically, such distinctions are not meaningful. Therefore, our data are consistent with the current Bethesda System practice of combining HPV changes and CIN1 into a single category. 16 REFERENCES 1. Koss LG, Durfee GR. Unusual patterns of squamous epithelium of the uterine cervix: cytologic and pathologic study of koilocytotic atypia. Ann NY Acad Sci. 1956;63: Ayre JE. Role of the halo cell in cervical carcinogenesis: a virus manifestation or premalignancy? Obstet Gynecol. 1960; 17: Meisels A, Fortin R. Condylomatous lesions of the cervix and vagina. I. Cytologic patterns. Acta Cytol. 1976;20: Meisels A, Fortin R, Roy M. Condylomatous lesions of the cervix: II. Cytologic, colposcopic and histopathologic study. Acta Cytol. 1977;21: Schiffman MH, Bauer HM, Hoover RN, et al. Epidemiologic evidence showing that human papillomavirus infection causes most cervical intraepithelial neoplasia. J Natl Cancer Inst. 1993;85: Bosch FX, Manos MM, Munoz N, et al. Prevalence of human papillomavirus DNA in cervical cancer: a worldwide perspective. International Biological Study on Cervical Cancer (IBBSCC) Study Group. J Natl Cancer Inst. 1995;87: Walboomers JM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol. 1999;189: Kuhler-Obbarius C, Milde-Langosch K, Loning T, Stegner HE. Polymerase chain reaction-assisted evaluation of low and high grade squamous intraepithelial lesion cytology and reappraisal of the Bethesda System. Acta Cytol. 1994;38: Casas-Cordero M, Morin C, Roy M, Fortier M, Meisels. Origin of the koilocyte in condylomata of the human cervix: ultrastructural study, Acta Cytol. 1981;25: Morin C, Meisels A. Human papilloma virus infection of the uterine cervix. Acta Cytol. 1980;24: Recher L, Strebnik. Histopathologic features of koilocytotic atypia: a detailed description. Acta Cytol. 1981;25: Jenson AB, Kurman RJ, Lancaster WD. Tissue effects of and host response to human papillomavirus infection. Dermatol Clin. 1991;9: Ziol M, Di Tomaso C, Biaggi A, et al. Virological and biological characteristics of cervical intraepithelial neoplasia Grade 1 with marked koilocytotic atypia. Hum Pathol. 1998;29: Sherman ME, Schiffman MH, Erozan YS, Wacholder S, Kurman RJ. The Bethesda System. A proposal for reporting abnormal cervical smears based on the reproducibility of cytopathologic diagnoses. Arch Pathol Lab Med. 1992;116: Lee KR, Minter LJ, Crum CP. Koilocytotic atypia in Papanicolaou smears: reproducibility and biopsy correlations. Cancer (Cancer Cytopathol). 1997;81: Kurman RJ, Malkasian GD, Sedlis A, Solomon D. From Papanicolaou to Bethesda: the rationale for a new cervical cytologic classification. Obstet Gynecol. 1991;77: Kurman RJ, Solomon D. The Bethesda System for Reporting Cervical/Vaginal Cytologic Diagnoses: Definitions, Criteria, and Explanatory Notes for Terminology and Specimen Adequacy. New York: Springer-Verlag; Solomon D, Davey D, Kurman R, et al. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA. 2002;287: Clifford GM, Smith JS, Aguado, Franceschi S. Comparison of HPV type distribution in high grade lesions and cervical cancer: a meta-analysis. Br J Cancer. 2003;89: Munoz N, Bosch FX, de Sanjose S, et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med. 2003;348: Cogliano V, Baan R, Straif K, et al., World Health Organization International Agency for Research on Cancer. Carcinogenicity of human papillomaviruses. Lancet Oncol. 2005;6:204. Abstract.

10 HPV Cellular Changes and CIN1/Zuna et al Schiffman M, Adrianza ME. ASCUS-LSIL Triage Study. Design, methods and characteristics of trial participants. Acta Cytol. 2000;44: Zuna RE, Wang SS, Rosenthal DL, Jeronimo J, Schiffman M, Solomon D. Determinants of human papillomavirus (HPV) negative low grade squamous intraepithelial lesions (LSIL) in the ASCUS LSIL Triage Study (ALTS). Cancer (Cancer Cytopathol). 2005;105: [No authors listed.] Human papillomavirus testing for triage of women with cytologic evidence of low-grade squamous intraepithelial lesions: baseline data from a randomized trial. The Atypical Squamous Cells of Undetermined Significance/ Low-Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group. J Natl Cancer Inst. 2000;92: Gravitt PE, Peyton CL, Apple RJ, Wheeler CM. Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by a single-hybridization, reverse line blot detection method. J Clin Microbiol. 1998;36: Gravitt PE, Peyton CL, Alessi TQ, et al. Improved amplification of genital human papillomaviruses. J Clin Microbiol. 2000;38: Bavin PJ, Giles JA, Deery A, et al. Use of semi-quantitative PCR for human papillomavirus DNA type 16 to identify women with high grade cervical disease in a population presenting with a mildly dyskaryotic smear report. Br J Cancer. 1993;67: Sherman ME, Schiffman M, Cox JT. Effects of age and human papilloma viral load on colposcopy triage: data from the randomized Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesion Triage Study (ALTS). J Natl Cancer Inst. 2002;94: Schiffman M, Wheeler CM, Dasgupta A, Solomon D, Castle PE, for the ALTS Group. A comparison of a prototype PCR assay and Hybrid Capture 2 for detection of carcinogenic human papillomavirus DNA in women with equivocal or mildly abnormal Papanicolaou smears. Am J Clin Pathol. 2005;124: Wright TC, Cox JT, Massad LS, Twiggs LB, Wilkinson EJ Consensus guidelines for the management of women with cervical cytological abnormalities. JAMA. 2002;287: Gissman L, Wolnik L, Ikenberg H, Koldovsky U, Schnurch HG, zur Hausen H. Human papillomavirus types 6 and 11 DNA sequences in genital and laryngeal papillomas and in some cervical cancers. Proc Natl Acad Sci USA. 1983;80: Kadish AS, Hagan RJ, Ritter DB, et al. Biologic characteristics of specific human papillomavirus types predicted from morphology of cervical lesions. Hum Pathol. 1992;23: Jovanovic AS, McLachlin CM, Shen L, Welch WR, Crum CP. Postmenopausal squamous atypia: a spectrum including pseudokoilocytosis. Mod Pathol. 1995;8: Woodhouse SL, Stastny JF, Styer PE, Kennedy M, Praestgaard AH, Davey DD. Interobserver variability in subclassification of squamous intraepithelial lesions: results of the College of American Pathologist Interlaboratory Comparison Program in Cervicovaginal Cytology. Arch Pathol Lab Med. 1999;123: Stoler MH, Schiffman M. Atypical Squamous Cells of Undetermined Significance Low-Grade Squamous Intraepithelial Lesion Triage Study (ALTS) Group. Interobserver reproducibility of cervical cytologic and histologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study. JAMA. 2001;285: Castle PE, Schiffman M, Burk RD, et al. Restricted crossreactivity of Hybrid Capture-2 with noncarcinogenic human papillomavirus types. Cancer Epidemiol Biomarkers Prev. 2002;11: