The role of human papillomavirus type 16/18 genotyping in predicting high-grade cervical/vaginal intraepithelial neoplasm in women with mildly abnormal Papanicolaou results†
Presented in part at the 59th Annual Scientific Meeting of the American Society of Cytopathology; November 4-8, 2011; Baltimore, MD.
The authors compared the predictive value of type 16 and/or 18 human papillomavirus (HPV) versus non-16/18 HPV types for high-grade (grade ≥2) cervical neoplasm/vaginal intraepithelial neoplasm and carcinoma (CIN/VAIN2+) in women with mildly abnormal Papanicolaou (Pap) results (ie, atypical squamous cells of undetermined significance [ASCUS] or low-grade squamous epithelial lesion [LSIL]).
The authors retrospectively selected Pap specimens with HPV testing results obtained from 243 women (155 with ASCUS and 88 with LSIL Pap results) in their Department of Pathology. HPV genotyping was performed using the EasyChip HPV blot assay. The Pap specimens with HPV16/18 and non-16/18 HPV types were compared with follow-up biopsy results. Follow-up duration ranged from 1 month to 58 months (mean, 26 months).
In total, 58 of 155 specimens (37%) that had ASCUS and 29 of 88 specimens (33%) that had LSIL were positive for HPV16/18. CIN/VAIN2+ biopsies were identified in 43 of 155 women (28%) with ASCUS and in 28 of 88 women (32%) with LSIL. Women with ASCUS and HPV16/18 had a significantly higher rate (43%) of CIN/VAIN2+ than women with ASCUS and non-16/18 HPV types (19%; P = .003; odds ratio, 3.10; 95% confidence interval, 1.48-6.53). There was no statistically significant difference in the rate of CIN/VAIN2+ between women who had LSIL and HPV16/18 (45%) and those who had LSIL and non-16/18 HPV types (29%; P = .16; odds ratio, 1.96; 95% confidence interval, 0.77-4.97).
HPV genotyping for HPV16/18 improved risk assessment for women with ASCUS Pap results and may be used to predict the risk of CIN/VAIN2+ to better guide follow-up management. Cancer (Cancer Cytopathol) 2013. © 2012 American Cancer Society.
High-risk human papillomavirus (HPV) is responsible for the development of >99% of cervical carcinomas.1 Currently, HPV DNA testing is used as a triage tool for women who have mildly abnormal Papanicolaou (Pap) results, ie, abnormal squamous cells of undetermined significance (ASCUS). According to American Society of Colposcopy and Cervical Pathology guidelines, reflex HPV DNA testing is a preferred option for the triage of women with ASCUS Pap results.2 Women who have positive HPV test results should be referred for colposcopy evaluation. To maximize the clinical sensitivity for high-grade (grade ≥2) cervical neoplasm/vaginal intraepithelial neoplasm and carcinoma (CIN/VAIN2+), the current HPV triage strategy is to collectively test 13 or14 high-risk HPV genotypes, including HPV16 and HPV18. The US Food and Drug Administration (FDA)-approved HPV testing assays, such as Hybrid Capture 2 (HC2) (Qiagen, Valencia, Calif) and Cervista HPV HR (Hologic, Marlborough, Mass), have been used for this purpose.
It has been well documented that women who have positive results for HPV types 16 and/or 18 (HPV16/18-positive results) have a significantly greater risk of developing cervical cancer than those who have other high-risk HPV types.3-7 Women with ASCUS Pap results and HPV16-positive results had a higher cumulative 2-year risk of developing grade ≥2 cervical intraepithelial neoplasm (CIN2+) than women who had other high-risk HPV genotypes.8 Therefore, genotyping for HPV16/18 may provide valuable risk assessment that may be used in patient follow-up and management. The current HPV triage strategy, which uses 13 or 14 pooled HPV genotypes, does not allow the detection of infection with specific HPV types and, thus, cannot be used to distinguish between women who have a significantly greater risk of developing CIN/VAIN2+ and those who have a relatively lower risk. Consequently, women who have a greater risk of CIN/VAIN2+ such as HPV16/18 infections, may be lost during follow-up because of low compliance in some patients. Conversely, women who have an average risk of CIN/VAIN2+ such as non-HPV16/18 infection, may be subjected to frequently unnecessary follow-up and biopsy.
Currently, the recommendation for clinical application of HPV16/18 genotyping is limited to women who have Pap-negative/HPV-positive test results for risk assessment.2 Recently, FDA-approved commercial HPV genotyping assays for detecting HPV16/18 in Pap specimens became available for this purpose. However, HPV16/18 genotyping has not been recommended for women who have ASCUS results for risk assessment, because there is limited clinical evidence indicating that HPV16/18 test information can improve risk assessment in these patients. To evaluate the risks of CIN/VAIN2+ between HPV16/18 and non-16/18 HPV types among women with mildly abnormal Pap results, we conducted a retrospective study to compare HPV genotypes and follow-up biopsies in women who had ASCUS or low-grade squamous epithelial lesion (LSIL) Pap results.
MATERIALS AND METHODS
Patients and Papanicolaou Specimen Collection
We retrospectively collected residual SurePath Pap specimens (BD Diagnostics, TriPath Imaging, Burlington, NC) obtained from 364 women after routine Pap testing between 2004 and 2009 in the Department of Pathology at The University of Texas MD Anderson Cancer Center. Of these, 252 women had ASCUS Pap results, and 112 had LSIL Pap results. All women underwent Hybrid Capture II (HC2) HPV DNA tests at the time of Pap cytology testing. The women ranged in age from 17 years to 90 years, and their mean age was 44 years. There were 155 Caucasian women, 39 Hispanic women, 37 African American women, 5 Asian women, and 1 woman of unknown ethnicity. The SurePath Pap specimens were screened by cytotechnologists and were verified by cytopathologists according to The Bethesda System.9 Reflex HPV DNA tests were performed using the HC2 assay for collective testing for 13 high-risk HPV DNA types. Upon completion of the cytology diagnostic evaluation, residual material from SurePath Pap specimens was collected and stored at −80°C, as described previously.10 The follow-up biopsies were processed and interpreted in the Department of Pathology at The University of Texas MD Anderson Cancer Center.
Human Papillomavirus Genotyping With the EasyChip Human Papillomavirus Blot
For HPV genotyping, DNA was extracted from the stored SurePath Pap specimens using the DNeasy kit (catalog no. 69506; Qiagen) according to the manufacturer's instructions. HPV genotyping was conducted using the EasyChip HPV Blot (King Car Yuanshan Research Institute, I-Lan, Taiwan) as described previously.11 The EasyChip HPV blot was designed to detect 38 HPV types (HPV types 6, 11, 16, 18, 26, 31, 32, 33, 35, 37, 39, 42, 43, 44, 45, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 66, 67, 68, 69, 70, 71,72, 74, 81, 82, 83, 84, and 85 as well as 3 intrinsic controls). First, HPV DNA was amplified with a polymerase chain reaction (PCR) assay using modified MY11/GP6-positive, biotinylated consensus primer sets to amplify a fragment of 192 base pairs in the L1 open reading frame of HPV; and forward and reverse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) biotinylated primer sets were used to amplify a 136-base pair segment for specimen validation, as described previously.11 Briefly, PCR for HPV DNA was performed in a final reaction volume of 26 μL with a 20-ng aliquot of genomic DNA in a PCR master mixture containing 15 mM Tris-HCl, pH 8.0; 2.0 mM MgCl2; 50 mM KCl; a 0.25-mM concentration of each deoxynucleotide triphosphate; 0.6 μM primer; and 0.5 U DNA polymerase (HP High Performance HotStart Taq DNA Polymerase; DNA Technologies Ltd., Cannock, United Kingdom). The PCR assay was performed as follows: 10 minutes at 95°C was followed by 40 cycles for 30 seconds at 95°C, 30 seconds at 45°C, and 30 seconds at 72°C; and a final extension for 5 minutes at 72°C. PCR for GAPDH was performed in a final reaction volume of 25 μL with a 10-ng aliquot of genomic DNA in a PCR master mixture containing 15 mM Tris-HCl, pH 8.0; 2.5 mM MgCl2; 50 mM KCl; 0.2 mM of each deoxynucleotide triphosphate; 0.2 μM primer; and 0.5 U HotStart Taq DNA polymerase (DNA Technologies Ltd.). The PCR assay was performed as follows: 10 minutes at 95°C was followed by 40 cycles for 15 seconds at 95°C, for 1 minute at 57°C, and for 30 seconds at 72°C; and a final extension for 5 minutes at 72°C. A 5-μL aliquot of PCR products was analyzed by electrophoresis on a 2% agarose gel and stained with ethidium bromide.
The hybridization was performed according to the manufacturer's instructions. The HPV types were determined by visual assessment protocol provided by King Car Yuanshan Research Institute.
The patients underwent follow-up biopsy and had a mean total follow-up of 26 months (range, 1-58 months). The majority of women (93%) underwent follow-up biopsies within 6 month of Pap/HPV testing.
Of the 364 women who initially were included in the study cohort, 121 women were excluded from data analysis. These included women in the following categories: 1) 34 women (9.3%) with Pap specimens that had negative HC2 results and negative EasyChip results; 2) 28 women (7.7%) with positive HC2 results but negative results for HPV DNA with the EasyChip HPV assay (no valid internal control signals were present in the specimens using the EasyChip assay, indicating insufficient DNA contents); and 3) 59 women (16.2%) who did not have follow-up biopsies.
Data Analysis and Statistics
Descriptive statistics were summarized for patient demographics. The chi-square test was used to test the association between 2 categorical variables. To evaluate the correlation between the incidence of CIN/VAIN2+ and HPV16/18 status, a multivariate logistic regression model was used. Two-sided P values < .05 were considered significant. All computations were carried out using the SAS statistical software package (version 8.0; SAS Institute, Cary, NC).
Distribution of Human Papillomavirus Type 16/18 and Non-16/18 Human Papillomavirus Types
In total, 243 cases were validated for data analysis. These included 155 specimens with ASCUS Pap results and 88 specimens with LSIL Pap results. In specimens that had ASCUS results, HPV16/18 types were positive in 37% (58 of 155 specimens); non-16/18, high-risk HPV types (HPV types 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) were positive in 48% (74 of 155 of specimens); and other types were positive in 15% (23 of 155 specimens). In specimens that had LSIL results, HPV16/18 types were positive in 33% (29 of 88 specimens); non-16/18, high-risk HPV types were positive in 50% (44 of 88 specimens); and other types were positive in 17% (15 of 88 specimens) (Table 1). Of all 87 specimens that had positive HPV16/18 results, HPV16 was detected in 59 specimens (68%), and HPV18 was detected in 24 specimens (27%); the remaining 4 specimens (5%) were positive for both HPV16 and HPV18. Of 155 women in the ASCUS group, a single non-16/18, high-risk HPV type was detected in 56 women (36%). Multiple non-16/18, high-risk HPV types were detected in 18 women (12%). Of 88 women in the LSIL group, a single non-16/18, high-risk HPV type was detected in 29 women (33%). Multiple non-16/18, high-risk HPV types were detected in 15 women (17%).
Table 1. Distribution of Human Papillomavirus Genotypes in Women With Mildly Abnormal Papanicolaou Results
|HPV16/18||58 (37)||29 (33)|
|Non-16/18 high-risk HPV||74 (48)||44 (50)|
|Non-high risk HPV||23 (15)||15 (17)|
Follow-Up Biopsies in Women With Atypical Squamous Cells of Undetermined Significance or Low-Grade Squamous Intraepithelial Lesion Papanicolaou Results
During follow-up, CIN/VAIN2+ was diagnosed in 43 of 155 women with ASCUS/HC2-positive results (28%) and in 28 of 88 women with LSIL/HC2-positive results (32%) (Table 2). Of 43 women who had ASCUS Pap results, the follow-up biopsies indicating CIN2/VAIN2+ were performed within 6 months in 34 (79%) patients and after 6 months in 9 (21%) patients. Seven patients had residual or recurrent CIN/VAIN2+ (patients with biopsies showing CIN/VAIN2+ during follow-up and rebiopsies showing CIN/VAIN2+ after treatment). Of the 28 women with LSIL Pap results, the follow-up biopsies showing CIN/VAIN2+ were obtained within 6 months in 20 women (71%) and after 6 months in 8 women (29%). Three women had residual or recurrent CIN/VAIN2+.
Table 2. Follow-Up Biopsies in Women With Atypical Squamous Cells of Undetermined Significance or Low-Grade Squamous Intraepithelial Lesion Papanicolaou Results
|CIN2 or CIN3+/Total||43/155 (28%)||28/88 (32%)|
Comparison of Follow-Up Biopsy Results With Human Papillomavirus Type 16/18 Versus Non-16/18 Human Papillomavirus Types in Women With Atypical Squamous Cells of Undetermined Significance or Low-Grade Squamous Intraepithelial Lesion
In the ASCUS group, 25 of 58 women (43%) with HPV16/18 underwent a follow-up biopsy that indicated CIN/VAIN2+, whereas 18 of 97 women (19%) with non-16/18 HPV types underwent a follow-up biopsy that indicated CIN/VAIN2+. The difference between the 2 groups was significant (P = .003; odds ratio [OR], 3.10; 95% confidence interval [CI], 1.48-6.53) (Table 3). In the LSIL group, 13 of 29 women (45%) with HPV16/18 underwent a follow-up biopsy that indicated CIN/VAIN2+, whereas 17 of 59 women (29%) who had non-16/18 HPV types underwent a follow-up biopsy that indicated CIN/VAIN2+. The difference between the 2 groups was not statistically significant (P = .16; OR, 1.96; 95% confidence interval, 0.77-4.97) (Table 4). In 10 women who had residual or recurrent CIN/VAIN2+, 6 women had HPV16/18, and 4 had non-16/18 HPV types.
Table 3. Human Papillomavirus Genotypes and Follow-Up Biopsies in Women With Atypical Squamous Cells of Undetermined Significance
|Non-16/18 HPV||18/97 (19)||—||1.00||—|
Table 4. Human Papillomavirus Genotypes and Follow-Up Biopsies in Women With Low-Grade Squamous Intraepithelial Lesion
|Non-16/18 HPV||17/59 (29)||—||1.00||—|
In the current study, we evaluated risk profiling among women who had mildly abnormal Pap results with or without positive HPV16/18 results by comparing the results from HPV genotyping tests with biopsy follow-up results. We observed that women with ASCUS and positive HPV16/18 test results had a significantly higher risk of having CIN/VAIN2+ lesions than those who had positive, non-16/18 HPV test results. Thus, we demonstrated that HPV16/18 genotyping may be used for risk assessment among women who have ASCUS Pap results and positive HPV DNA results. However, further studies with larger patients numbers and commercial HPV genotyping assays will be required to verify the clinical utility of HPV16/18 genotyping for women with ASCUS.
It is well documented that >60% of women who have CIN3 are positive for HPV16/18.12 However, how to use HPV16/18 genotyping information for women who have ASCUS or LSIL Pap results has not been established. Zuna et al reported a significantly increased rate of HPV16/18 detection in Pap specimens from LSIL to high-grade squamous intraepithelial lesion/carcinoma, suggesting the potential of using HPV genotyping as a risk-profiling component in the evaluation of women who have abnormal Pap results.13 In a screening population, Castle et al demonstrated that positive HPV16/18 results were equivalent to ASCUS or worse Pap results in terms of sensitivity for CIN3, indicating a remarkable risk associated with HPV16/18 for women aged ≥25 years.14 Recently, the potential clinical use of HPV16/18 genotyping in women with ASCUS has been reported. In a screening population, Stoler et al demonstrated that women who had ASCUS Pap results and positive HPV16/18 results had a significantly greater risk of CIN2+ (24.4%) compared with women who were positive for pooled high-risk HPV types (14%).7 Einstein et al reported a study using the Cervista HPV16/18 assay in women with ASCUS Pap results and demonstrated that women who had positive HPV16/18 results had 17.2 times the risk of CIN2+ compared with those who had negative high-risk HPV results; whereas women who had non-16/18 HPV types had only 4.0 times the risk of CIN2.15
Similarly, in the current study, we demonstrated a significant difference in the correspondence of positive HPV16/18 and non-16/18 HPV results with CIN/VAIN2+ biopsy results in women who had ASCUS Pap results. Our findings and the published data demonstrate that the risk of CIN/VAIN2+ in women who have ASCUS Pap results and positive HPV16/18 results is substantial enough to warrant close follow-up. The clinical implication for women with ASCUS results and positive HPV16/18 results is the higher risk level of CIN/VAIN2+ during follow-up. With HPV16/18 genotyping information, women who have HPV16/18 can be advised of the potential risk, which should lead to increased patient compliance with follow-up. Information on HPV16/18 status also may help clinicians avoid excessive and unnecessary follow-up of women who have non-16/18 HPV results. During clinical data review for the current study, we observed that many women who had mildly abnormal Pap results and subsequent positive HPV16/18 results were lost to follow-up partially because of the lack of relevant risk information for these patients. Conversely, in women who had non-16/18 HPV results or even those who had low-risk HPV types identified, more frequent colposcopy with negative biopsies were performed, and some patients underwent more than 1 colposcopy and biopsy within a year.
In our study, although there was a significantly high positive predictive value of 43% among women who had ASCUS and positive HPV16/18 results, a positive predictive value of 19% among women who had non-16/18 HPV results is high enough for triage to immediate colposcopy. Consequently, women with ASCUS/HPV-positive results, including those who have non-16/18 HPV genotypes, still need immediate colposcopy evaluation as recommended by the new guidelines for cervical cancer screening.16 However, HPV16/18 genotyping information can be used to refine the patient follow-up for women who have ASCUS Pap results, ie, to increase patient compliance for follow-up in women who have ASCUS/HPV16/18 results and to avoid excessive follow-up in women who have ASCUS/non-16/18 HPV results. Recently FDA-approved HPV testing assays, such as Cervista HPV16/18 and the Cobas HPV Test for HPV16/18 genotyping (F. Hoffmann-La Roche Ltd., Basel, Switzerland) have provided useful tools for clinicians in this regard.
In women who had positive HC2 results, we observed a small fraction of nonhigh-risk HPV types among women who had ASCUS (15%) or LSIL (17%) Pap results. These rates were comparable to those reported in a previously published study.17 Further separating this small fraction of non-high-risk HPV types from high-risk HPV types may not be clinically significant, because the group includes those who have borderline high-risk HPV types, such as HPV66, HPV67, HPV70, HPV73, and HPV82, which reportedly have certain risks of CIN/VAIN2+.18 Therefore, in our study, the non-high-risk HPV types were combined with non-16/18 high-risk HPV group as non-16/18 HPV types for data analysis.
For women who have LSIL Pap results, patient management can be challenging, partially because of the relatively low positive predictive value for CIN2+ in this group.2 Clinically, LSIL and ASCUS Pap results can occur alternatively in women who have multiple mildly abnormal Pap results during follow-ups. Therefore, a Pap classification of ASCUS or LSIL in the same patient during follow-up may have similar clinical implications. In our study, we observed that 37% of women who had ASCUS results and 33% of women who had LSIL results had HPV16/18 genotypes. The rate of HPV16/18 was comparable to the previously published rate.19, 20 Furthermore, the rates of CIN2+ in biopsies during follow-up also were comparable between women with ASCUS or LSIL results in the current study and in our previously reported study in a screening population.21 According to the current recommendation, women who have LSIL Pap results should be referred for colposcopy and biopsy. HPV DNA testing is recommended only for women who have LSIL after a negative colposcopy and biopsy.2 In our LSIL group, we observed a higher rate of CIN2+ among women who were positive for HPV16/18 compared with those who had non-16/18 HPV genotypes. However, no statistically significant difference was observed between the 2 groups, possibly because of the small sample size of our LSIL groups. Further studies with larger patient numbers will be required to verify whether HPV16/18 genotyping can be useful for risk assessment in women who have LSIL Pap results.
The limitations of our study are the small patient numbers, the origin of the study cohort from a cancer center composed of a relatively older population, a frequent history of cervical dysplasia, and the use of a PCR-based HPV testing assay that is not an FDA-approved method. More studies with large patient numbers from different populations should be performed with the available FDA-approved genotyping assay to confirm the clinical utility of HPV16/18 testing in women with ASCUS or LSIL for risk assessment and to determine the proper follow-up strategy.
We thank the American Society of Cytopathology for Cervical Cytology and the Human Papillomavirus (HPV) Partnership Investigator Award (American Society of Cytopathology, 2009-2011) and King Car Yuanshan Institute for providing reagents for the EasyChip HPV genotyping assay for the study. We also thank Ms. Jie Feng, Department of Pathology, for performing HPV genotyping tests and Mr. Walter Pagel, Scientific Publications, The University of Texas MD Anderson Cancer Center for editing the article.
This study was supported by a Cervical Cytology and HPV Partnership Investigator Award from the American Society of Cytopathology and an interim grant from The University of Texas MD Anderson Cancer Center to Dr. Guo.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.