![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Department of Medical Microbiology, Nijmegen University Centre for Infectious Diseases, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands,1 Department of Obstetrics and Gynaecology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands2
Received 10 March 2006/ Returned for modification 9 May 2006/ Accepted 19 June 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
value, 0.79). The genotyping assays were therefore highly comparable and reproducible. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For these screening purposes, several tests have been developed in order to distinguish high-risk HPV infections from no HPV infection. Among these are the signal amplification method Hybrid Capture II (hc2) (Digene Corp., Gaithersburg, Maryland) and the recently developed target amplification method Roche AMPLICOR HPV test (Roche Molecular Systems, Inc., Branchburg, NJ) (35). Although both tests are commercially available and Conformité Européenne (CE) marked, hc2 is currently the only FDA-registered HPV screening assay (7). Both tests differentiate between an infection with one or more of 13 hr-HPV genotypes (genotypes 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) and no hr-HPV infection—an "hr-HPV plus/minus" screening. Although these tests are not designed to detect the recently described probable hr-HPV or any lr-HPV infection, some cross-reactivity outside of the spectrum of 13 hr-HPV genotypes has been reported for the hc2 assay (5). Neither the hc2 nor the AMPLICOR HPV assay allows the identification of specific genotypes (26), nor do they have the ability to identify infections involving multiple genotypes.
However, recent studies have provided evidence for a difference in oncogenic potential between the different hr-HPVs (6), arguing for the importance of HPV genotyping in addition to the "hr-HPV plus/minus" screening. Outside of the clinical setting, HPV genotyping is a key characteristic of studies evaluating the epidemiology of HPV infections worldwide. Although a number of HPV genotyping assays have been used in such studies, a reliable comparison between the diagnostic and epidemiological data generated is difficult, since data on the intertest comparisons between the different genotyping assays are limited.
The SPF10-INNO LiPA assay is capable of amplifying up to 43 different genotypes and providing type-specific genotype information for 25 different HPV genotypes simultaneously, has been extensively tested, and has proven to be highly sensitive and specific (15, 25). The Roche Linear Array (LA) HPV genotyping test (Roche Molecular Systems, Inc., Branchburg, NJ) is a recently launched new HPV genotyping assay able to genotype 37 HPV types, concurrently assessing human ß-globin. The full spectrum of HPV genotypes amplified by the PGMY primer system (13) used in the Roche Linear Array HPV genotyping test has not been assessed beyond the 37 genotypes probed. In essence, both assays could be used for genotyping analysis.
This study was designed to compare these two well-known and commonly used commercially available genotyping assays with HPV DNA-positive samples.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Specimen preparation. For isolation of DNA from cervical scrapes in liquid-based cytology medium, the MagNAPure LC isolation station (Roche Diagnostics GmbH, Roche Applied Science, Mannheim, Germany) was used; 200 µl of material was isolated using the MagNA Pure LC Total Nucleic Acid Isolation kit (Roche Diagnostics GmbH, Roche Molecular Biochemicals, Mannheim Germany), as described by the manufacturer. With each set of 28 cervical-scrape samples, four negative controls (distilled water) were used to monitor the DNA isolation procedure and to assess contamination. Nucleic acid was resuspended in a final volume of 50 µl; 10 µl was used for each of the various PCR analyses.
SPF10-INNO LiPA HPV detection and genotyping (DNA enzyme immunoassay [DEIA] and LiPA). (i) PCR amplification of HPV DNA. Broad-spectrum HPV DNA amplification was performed using a short-PCR-fragment assay (SPF10 HPV PCR; Labo Bio-Medical Products B.V., Rijswijk, The Netherlands). This assay amplifies a 65-bp fragment of the L1 open reading frame and allows detection of at least 43 different HPV types (16, 25). The SPF10 PCR system was used in a final reaction volume of 50 µl containing 10 µl of the isolated DNA sample and 40 µl of the PCR mixture, which contained 10 mmol/liter Tris-HCl (pH 9.0), 50 mmol/liter KCl, 2.0 mmol/liter MgCl2, 0.1% Triton X-100, 0.01% gelatin, 200 µmol/liter of each deoxynucleoside triphosphate (dATP, dCTP, dGTP, and dTTP), 15 pmol each of the forward and reverse primers tagged with biotin at the 5' end, and 1.5 units of AmpliTaq Gold (Applied Biosystems, Foster City, CA). Activation of AmpliTaq Gold for 9 min at 94°C, was followed by 40 cycles of 30 s at 94°C, 45 s at 52°C, and 45 s at 72°C, with a final extension of 5 min at 72°C. Appropriate negative and positive controls were used to monitor the performance of the PCR method in each experiment.
(ii) HPV detection by DEIA. The presence of HPV DNA was determined by hybridization of SPF10 amplimers to a mixture of general HPV probes recognizing a broad range of high-risk, low-risk, and possible high-risk HPV genotypes in a microtiter plate format, as described previously (15, 25). All HPV DNA-positive samples (by SPF10 DEIA) were genotyped using the INNO-LiPA HPV genotyping assays and the Roche Linear Array HPV genotyping test as described below. Twenty randomly selected DEIA-negative samples that had previously tested negative by the Roche AMPLICOR HPV test (35) were also assessed using both genotyping assays.
(iii) HPV genotyping by reverse hybridization using the INNO-LiPA HPV genotyping system. The 28 oligonucleotide probes that recognize 25 different types (Table 1) were tailed with poly(dT) and immobilized as parallel lines to membrane strips (Labo Bio-Medical Products B.V., Rijswijk, The Netherlands). The HPV genotyping assay was performed as described previously (15). The LiPA strips were manually interpreted using the reference guide provided.
|
Linear Array HPV genotyping test. The LA HPV genotyping test (Roche Molecular Systems, Inc., Branchburg, NJ) is a new qualitative in vitro test for the determination of 37 anogenital HPV DNA genotypes (Table 1). The LA test was applied to all samples that tested positive for HPV by DEIA and to 20 randomly selected DEIA-negative samples.
(i) PCR amplification of HPV DNA. The LA test uses biotinylated PGMY primers to amplify a 450-bp fragment within the polymorphic L1 region of the HPV genome. The PGMY amplification system has been described previously (13). The PGMY primers are present in the "master mixture" (containing buffer, nucleotides [dATP, dCTP, dGTP, and dUTP], MgCl2, and <0.02% AmpliTaq Gold DNA polymerase) and amplify HPV DNA from 37 HPV genotypes, including 13 high-risk types (Table 1). Amplicons incorporate dUTP, allowing the use of AmpErase enzyme (uracil N-glycosylase), which is included in the master mixture to prevent PCR carryover contamination. Capture probe sequences are located in polymorphic regions of L1 bound by these primers. An additional primer pair targets the human ß-globin gene (268-bp amplicon) to provide a control for cell adequacy, extraction, and amplification.
PCR was performed in a final reaction volume of 100 µl, containing 50 µl HPV master mixture, 40 µl PCR water, and 10 µl isolated DNA. The mixture was incubated for 2 min at 50°C and for 9 min at 95°C, followed by 40 cycles of 30 seconds at 95°C, 1 min at 55°C, and 1 min at 72°C, with a final extension at 72°C lasting from 10 min to a maximum of 1 h. The provided HPV-positive and -negative controls were used with each set of 10 samples to assess the performance of the reaction.
(ii) Hybridization and detection. Following amplification, the HPV and human ß-globin amplicons were denatured by immediately adding 100 µl denaturation solution to each PCR tube. Hybridization and HPV genotyping were performed as described by the manufacture (Roche Molecular Systems, Inc., Branchburg, NJ). The strips were manually interpreted using the Linear Array HPV reference guide, by reading the individual types down the length of the strip. Samples that were both SPF10 DEIA and LA ß-globin positive yet were not reactive to any of the genotype probes on the LA strip were considered "LA negative."
Design of the study. Previously, the samples had been assessed in an analysis comparing only high-risk HPV types detected by the Roche AMPLICOR HPV test and the INNO-LiPA HPV detection and genotyping assay (35). Since the present study compares two genotyping assays, only the DEIA HPV-positive samples and 20 randomly selected DEIA (and Roche AMPLICOR) HPV-negative samples were assessed. In order to have the most accurate comparison between the two genotyping tests, only the HPV genotypes identified by both assays (i.e., lr-HPV 6, 11, 40, 42, 54, and 70; possible hr-HPV 53 and 66; and hr-HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59) were considered for direct comparison of the individual HPV genotypes (Table 1). These will be referred to as assay-common genotypes. High-risk HPV genotypes 68 and 73 were not taken into account for individual comparison, since these types are identified by a single probe in the LiPA assay and thus cannot be distinguished. Moreover, the classification of HPV 53 as possibly high risk is currently disputed. When comparing the two genotyping assays, results were termed concordant, compatible, or discordant based on the following definitions. If the analyses yielded identical assay-common genotypes in both tests, the results were termed concordant. Results were termed compatible if one or more additional assay-common genotypes were not detected by either of the assays. Genotyping results were termed discordant if there were no similarities in the assay-common genotypes between the two tests. Assay results for HPV genotypes uniquely identified by each of the two assays (i.e., assay-unique HPV genotypes 34, 43, 44, and 74 detected only by LiPA and the assay-unique HPV genotypes 26, 55, 61, 62, 64, 67, 69, 71, 72, 81, 82, 83, 84, IS39, and CP6108 detected solely by the LA test) were not considered in determining concordant, compatible, or discordant status.
From all compatible and discordant samples, a reextracted DNA sample was randomly retested in a blind approach in a discrepancy analysis using both genotyping assays. Eleven concordant samples (six single infections, four double infections, and one triple infection) and six double-negative (i.e., DEIA-positive, LiPA X-type, and LA-negative) samples were used as positive and negative controls for both inter- and intra-assay performance control.
All HPV tests were performed by investigators unaware of the results of the comparative HPV detection or genotyping tests.
Statistics. All data were analyzed using SPSS version 12.0.1. for Windows. Agreement was measured by absolute agreement and Cohen's kappa statistics, a measure of the agreement between two methods that is in excess of that due to chance.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Of the 210 DEIA-positive samples tested using both genotyping assays, 40 samples were excluded, since one of the tests was negative whereas the comparative test detected an assay-unique genotype or LA was negative and LiPA showed an X type (Table 2).
|
|
Compatible cases. All 36 compatible cases were multiple infections. The LiPA assay did not detect a total of 41 genotypes in 30 separate clinical samples. In 23 cases, 1 type was missed; in 5 cases, 2 types were missed; and in 2 cases, 4 types were missed (13 low-risk, 1 possible high-risk, and 27 high-risk genotypes were not detected by the LiPA test). The Linear Array assay, on the other hand, did not detect 12 genotypes in eight separate samples. In six cases, 1 type was missed; in one case, 2 types were missed; and in one case, 4 types were missed (2 low-risk, 1 possible high-risk, and 9 high-risk HPV types). Table 4 gives an overview of the individual types that were not detected. Fifteen of the 16 cases in which LiPA missed an hr-HPV type were samples infected with multiple hr-HPV types that tested positive for another high-risk type, which was also detected in the LA.
|
The genotypes that were detectable by both assays among all 170 samples (112 concordant, 36 compatible, and 22 discordant) were individually compared, as summarized in Table 5. The overall strength of agreement between the two assays for the individual genotypes was considered good ( = 0.792). Although HPV 16 was detected in 45 samples using the LA test and in 39 samples using the LiPA, agreement between the tests was considered very good, with a value of 0.874. The agreement between the two assays for the other high-risk and probable high-risk genotypes varied between "good" and "very good." The agreement between the two tests for the low-risk genotypes was "moderate" to "perfect." The agreement for HPV 54 was moderate, since LiPA and LA shared 8 samples harboring the low-risk genotype whereas LA detected it in 10 additional samples. Also, the agreement for lr-HPVs 11 and 42 was moderate, while HPV 70 was detected in equal amounts by both assays. Low-risk HPV 40 was not detected in either of the tests; thus, no agreement could be calculated. The difference in detection of lr-HPV 54 was statistically significant (P < 0.05; McNemar's test). Although the differences for hr-HPV 16, 51, and 59 and lr-HPV 42 between the assays were large, they were considered not quite statistically significant (P > 0.07; McNemar's test). In the individual comparison of the other genotypes, no statistically significant differences were detected.
|
|
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Generally, a separate screening is needed preceding genotyping in order to assess a sample's HPV DNA positivity, i.e., an HPV plus/minus screening. An advantage of the LiPA is the use of the same amplicon for both detection of 43 different lr-, probable hr-, and hr-HPV genotypes and genotyping of 25 different HPVs. For the LA, a prescreening test with the PGMY primers is available using a generic HPV probe labeled with digoxigenin in a microtiter plate-based assay, as recently described (18). Without the need for further amplification, this amplicon can be directly used for the Linear Array genotyping assay. However, the efficiency of such a combination has not been studied. The recently launched HPV Roche AMPLICOR test for HPV plus/minus screening is not meant for an LA screen. It could also be used as a pretest, but the assay detects only high-risk HPV types (35).
In the initial comparison, i.e., prior to the discrepancy analysis, LiPA did not detect 27 high-risk genotypes in 30 compatible cases. Evidently, all the cases involved were multiple infections, i.e., containing two or more HPV types. Apparently, if an infection encompasses multiple genotypes, the SPF10-INNO LiPA assay is less sensitive than the LA. After finding analogous results using the LiPA assay, Van Doorn et al. propounded the idea of PCR competition between genotypes in mixed infections and suggested a combined testing algorithm using broad-spectrum and type-specific PCRs for HPV 16 and HPV 18 (L. J. van Doorn, A. C. Molijn, B. Kleter, W. G. V. Quint, and B. Colau, Abstr. 22nd IPV Conf., abstr. N-01, 2005). The complexity of assessing multiple genotypes was addressed previously (34). Amplification and identification of two genotypes present in equimolar amounts are likely possible. However, "primer competition" between genotypes might occur if one genotype is present in molar excess, outcompeting the other (34). In the present study, this was demonstrated by the samples harboring multiple infections that were not identically genotyped by both assays. Also, LA detected hr-HPV 16 in seven samples that were LiPA HPV 16 negative; after the second LA, however, five samples no longer showed HPV 16. Moreover, in a previous study Van Doorn and colleagues detected HPV 16 and HPV 18 using type-specific PCR in samples negative for these genotypes (but not for other genotypes) using general primer sets (34). In the present study we observed similar results (data not shown). Although the viral load was not determined in the present study, low-copy-number samples have previously shown more discrepancies in intralaboratory and interlaboratory comparisons (17).
The LA assay is unable to distinguish hr-HPV 52 from other high-risk genotypes (33, 35, and 58). This could be inconvenient in future studies using the Linear Array, since hr-HPV 52 is prevalent in approximately 5% of the HPV-positive women with normal cytology (8) and causes 2.2% of all cervical cancers (27). In 19 samples from the present study, hr-HPV 52 positivity could not be excluded based on LA genotyping. However, in these cases, the comparative LiPA tests did not detect this specific genotype. Two samples were considered Linear Array HPV 52 positive based on the LiPA results.
Among the 22 discordant cases, the number of hr-HPV genotypes detected by the Linear Array was not higher than the number detected by LiPA. All but three of these samples were single infections, predominately HPV 33, HPV 51, and HPV 52. A higher inclusivity level has been observed for some high- and low-risk HPV genotypes, particularly hr-HPV 33 and hr-HPV 56, when the PGMY amplification system is used (see the product insert for the CE-marked Linear Array HPV genotyping test, European market). The inclusivity level equates to the lowest concentration (copies/ml) that shows a 100% positive hit rate in a replicate of six tests or the concentration that is the probit-predicted 95% positive hit rate. This could explain some of the differences between the two assays observed in our study. Thus, the LA seems to be less sensitive than the LiPA if a sample has a single infection with some specific HPV genotypes that are poorly amplified by PGMY. Even though the majority of samples were cytologically classified as normal, proper HPV assessment, including genotyping, remains essential, particularly for healthy women with normal cytology (35), especially since Wallin and colleagues observed a strong concordance between the HPV type found in baseline smears with normal cytology and the eventual type found in histological samples of invasive cancers (37). In the present study, hr-HPV 51 was missed by LA in four of the discordant cases; this genotype accounted for approximately 0.9% of all squamous cell cervical cancers in previous studies (27). Curiously, the inclusivity level for HPV 51 is lower than the level for HPV 16 using PGMY primers, suggesting highly sensitive detection (see the product insert for the CE-marked Linear Array HPV genotyping test, European market). The observed difference in HPV 51 detection between the two assays thus cannot be explained by a lower efficiency of the Linear Array PGMY primer.
After discrepancy analysis of the compatible and discordant cases, both LiPA and LA detected more concordance (Table 6). Some previously undetected genotypes, for example, appeared in the second test run, and vice versa. This could be due to low copy numbers or to sampling, as DNA reextracts were used for the analysis. Also, it could possibly indicate the suggested competition between genotypes present in more or less molar excess. However, results from a discrepancy analysis should generally be handled with care and interpreted carefully. Discrepancy analyses are not perfect, since an analysis is easily biased in favor of the new test, and hard and fast rules do not exist (23). Moreover, the interpretation of results that cannot be dichotomized (i.e., concordant, compatible, and discordant) is less straightforward.
Failing to detect genotypes will lead to underestimation of the prevalence of certain genotypes and will cause false-negative results. Studies concerning (i) the epidemiology of HPV, (ii) HPV vaccination/surgical treatment trials, and (iii) cervical cancer screening and triage, especially, will be negatively affected by this. In epidemiologic studies, genotyping is compulsory in order to evaluate type-specific HPV DNA prevalence among infected women (3), to assess geographic heterogeneity in HPV type distributions (8), and to study type-specific HPV concordance between sexual partners (4). The importance of suitable algorithms for HPV detection and genotyping, in addition to the introduction of type-specific antiviral therapies or monovalent vaccines, was already addressed by Koutsky and colleagues (19). Moreover, current extensive trials testing multivalent vaccines, comprising multiple commonly occurring HPV types, demand accurate, unequivocal, and sensitive methods and algorithms detecting and specifically genotyping HPV (14, 19, 22). These algorithms are also compulsory for clinical trials monitoring surgical treatment of HPV-induced CIN lesions (24, 26) or monitoring persistent infections in consecutive smears, because persistence has been identified as an important risk factor (10, 28). Finally, according to Snijders and colleagues, adding general hr-HPV testing could be beneficial for the efficacy of the population-based screening programs for cervical cancer (32). Castle and colleagues, however, observed that ASCUS women infected with hr-HPV 16 had a 2-year cumulative absolute risk for developing CIN of at least grade 3 of 32.5% compared to the 8.4% risk of developing CIN of at least grade 3 for other high-risk HPV types (6). This underlines the potential importance of assessing the specific genotype causing the HPV infection. Triaging patients using cytology and genotyping assays might have a cost benefit over cytology combined with hr-HPV testing alone. The existence of triage management of ASCUS women in the United States depends solely on an accurate genotyping test (1). Both tests assessed in the present study could be suitable as triage tests.
In addition to accurate genotyping, the appropriate detection of multiple infections seems to be an important application of tests when they are implemented into any format of population-based screening for the prevention of cervical cancer, especially since the presence of multiple human papillomavirus genotypes in a single sample—suggesting repetitive exposure—is suspected to be associated with an increased risk for progressive disease (2). Moreover, mixed infections appear to be more frequent than previously suspected; 35% of the HPV-positive samples and more than 50% of human immunodeficiency virus-positive women are infected with multiple HPV types (12, 21). Multiple infections were less prevalent in cervical carcinomas (15).
In conclusion, the two genotyping assays are handled equally well and have been shown to be highly comparable. All of the HPV genotypes detected in either one or both of the assays, regardless of the analytical or clinical sensitivity and specificity of the tests, should not be trivialized, since their natural behaviors and cancerous potentials in both single and mixed infections remain ambiguous.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Antimicrob. Agents Chemother. | Clin. Microbiol. Rev. |
|---|---|
| Clin. Vaccine Immunol. | ALL ASM JOURNALS |
|---|
