Comparison of cobas human papillomavirus test results from primary versus secondary vials of PreservCyt specimens and evaluation of potential cross-contamination

Authors


Abstract

BACKGROUND:

The preferred workflow for high-risk human papillomavirus (hrHPV) testing in the majority of laboratories involves cytology processing of samples collected in liquid-based cytology medium followed by hrHPV testing. The cobas HPV Test received approval from the US Food and Drug Administration in April 2011, and the supporting clinical trial design necessitated prealiquoting the sample used for hrHPV testing from the PreservCyt primary vial into a secondary vial that was placed on the cobas 4800 System.

METHODS:

To validate use of the postcytology residual sample in the primary vial, the authors presented the results of cross-contamination studies and a comparison of the cobas HPV Test results from the prealiquot in the secondary vial with results obtained from the postcytology primary vial on samples processed on either the Hologic ThinPrep 2000 System (T2000) or ThinPrep 3000 System (T3000).

RESULTS:

Cross-contamination checkerboard studies with 100 samples processed on the T2000 system and 120 samples processed on the T3000 system indicated no conversion of primary vial results from positive to negative or from negative to positive. For clinical samples, approximately 1100 archived specimens from the ATHENA (Addressing the Need for Advanced HPV Diagnostics) study and 1100 combined archived and fresh primary vial specimens that had been processed on the T2000 and T3000 processors, respectively, were compared. The overall percentage agreement between the primary and secondary vials was at least 93.5%.

CONCLUSIONS:

The postcytology residual of samples collected in PreservCyt primary vials and run on the T2000 and T3000 processors provided reliable cobas HPV Test results that were comparable to results from the precytology secondary vial, without any evidence of contamination. Cancer (Cancer Cytopathol) 2012. © 2012 American Cancer Society.

INTRODUCTION

The majority of cervical cancer cases and deaths can be prevented through early detection of precancerous changes in the cervix. Cytology has been central to cervical cancer screening programs for 50 years and has contributed to the 70% decline in the rates of cervical cancer reported in developed countries that have implemented screening.1 Human papillomavirus (HPV) is now recognized as a single, necessary cause of cancer of the cervix and is reported to be present in approximately 99.7% of cases.2 Approximately 13 HPV genotypes have been classified as carcinogenic or high risk (hrHPV) types (HPV-16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, and -68) and can cause cancer of the cervix, with genotype 66 classified as possibly carcinogenic.3 Awareness of this causal relationship between HPV and cervical cancer has led to the increased use of tests that detect infection with these hrHPV genotypes in cervical cancer screening programs.

The 2006 American Society for Colposcopy and Cervical Pathology consensus guidelines for the management of women with abnormal cervical cancer screening tests4 recognized the usefulness of combining cervical cytology, tests for hrHPV infection, and genotype-specific HPV testing (whenever such a test was to be approved) for women undergoing screening for cervical cancer. These guidelines have recently been revised and now recommend the combination of cytology and hrHPV testing (cotesting) as the preferred method of screening, with HPV type16/ 18 genotype-specific testing being an added option.5 The cobas HPV Test (Roche Molecular Systems, Pleasanton, Calif) is a new polymerase chain reaction (PCR)-based DNA amplification test that simultaneously identifies a pooled result for 12 hrHPV types (HPV-31, -33, -35, -39, -45, -51, -52, -56, -58, -59, -66, and -68) and individual results for HPV-16 and HPV-18. It was approved by the US Food and Drug Administration in April 2011 as a triage test in women aged ≥ 21 years with atypical squamous cells of undetermined significance (ASC-US) cytology results and as a cotest or adjunct test to cytology in women aged ≥ 30 years.6 The design of the supporting registrational trial, the ATHENA (Addressing the Need for Advanced HPV Diagnostics) study, necessitated manual prealiquoting of the sample used for HPV testing from the PreservCyt (Hologic, Inc, Bedford, Mass) liquid-based cytology vial (primary vial) into a secondary vial that was placed on the cobas 4800 instrument; the residual sample from the primary vial was then used for cytology processing. However, the workflow in many laboratories involves performing cytology testing first, followed by HPV testing on the residual sample. Because the cobas HPV Test provides detection of hrHPV DNA using real-time PCR technology, it is essential to ensure that DNA contamination is not introduced by cytology processing. To validate the use of the postcytology residual sample in the primary vial, we present herein the results of cross-contamination studies, as well as a comparison of the cobas HPV Test results from the prealiquot in the secondary vial with results obtained in the postcytology primary vial on samples that were processed on either the ThinPrep 2000 System (T2000; Hologic) or the ThinPrep 3000 System (T3000; Hologic).

MATERIALS AND METHODS

Study Populations

The majority of the samples for this subanalysis were selected from archived specimens from the ATHENA HPV study. The ATHENA study, described in detail previously,7 enrolled more than 47,000 women aged ≥ 21 years who presented for cervical cancer screening from May 2008 through August 2009. In the baseline phase, all eligible participants had both Papanicolaou (Pap) testing by liquid-based cytology (ThinPrep) and HPV testing (initially by first-generation Roche HPV tests [the AMPLICOR HPV Test and the LINEAR ARRAY HPV Genotyping Test] and then by the second-generation Roche HPV test [cobas HPV Test]). All women aged ≥ 21 years who had either abnormal cytology (defined as ASC-US or higher) or women aged ≥ 25 years who tested positive for hrHPV (using the AMPLICOR HPV Test or the LINEAR ARRAY HPV Genotyping Test) were to undergo colposcopy. In addition, a subset of women aged ≥ 25 years who tested negative for both cytology (negative for intraepithelial lesions or malignancies [NILM]) and hrHPV were randomly selected to undergo colposcopy. Women who achieved a diagnosis on colposcopic-directed biopsy of cervical intraepithelial neoplasia (CIN) of grade 2 or greater (≥ CIN2) were exited from the study; those women who were diagnosed with < CIN2 were invited to participate in the 3-year follow-up phase of the study, which is scheduled to be completed December 2012.

Cross-Contamination Studies

Test Sample Preparation

HPV-positive and HPV-negative samples were prepared at Roche Molecular Systems from specimens collected as part of the ATHENA study. The HPV-negative samples were prepared by pooling HPV-negative PreservCyt specimens; the HPV-positive samples were prepared by pooling and diluting HPV-positive PreservCyt specimens in an HPV-negative specimen background to produce an average cycle threshold (Ct) value below the median Ct value for the entire ATHENA test population, thereby achieving levels on the high end of typical target levels.

T2000 and T3000 Processing

Before cross-contamination testing, 25 HPV-negative samples were processed on the T2000 system and 40 HPV-negative samples were processed using the T3000 system to verify clean systems. Alternating between HPV-positive and HPV-negative samples in a checkerboard pattern, a total of 100 samples comprised of 50 HPV-positive and 50 HPV-negative samples were then processed on the T2000 system; in a similar manner, 120 samples comprised of 60 positive and 60 negative samples were processed on the T3000 system.

cobas 4800 System Testing

Before testing the T2000-processed and T3000-processed samples with the cobas HPV Test, a blank (negative) run comprised of 48 negative samples was performed on each cobas 4800 System to verify clean instruments. Upon verification, 1 system was used to test all of the HPV-negative samples and 1 system was used to test the HPV-positive samples. This approach eliminated the likelihood of cobas 4800 system-related events, and therefore allowed for the specific identification of any cross-contamination events resulting from the ThinPrep processors.

Cytology and HPV Testing

Cytology testing was performed on samples collected in PreservCyt solution at 4 accredited clinical laboratories in the United States, as described previously.8 Three of the 4 laboratories processed samples for cytology on the T2000 system, and the fourth laboratory used the T3000 system. Cleaning and decontamination procedures for the T2000 and T3000 processors were performed in accordance with the manufacturer's instructions.9, 10 It is interesting to note that, for the T2000 processor, the bleaching step recommended for the cobas Amplicor CT/NG was not used for the cobas HPV test. For the T3000 processor, samples were run in a routine batch mode with no manual interventions or special decontamination procedures and the routine cleaning procedures were performed daily and monthly as recommended in the operator's manual.

For the cobas HPV Test, the archived ThinPrep primary vial samples were prealiquoted for testing and processed for cytology on the T2000 or T3000 processor, as described.8 The prealiquoted samples (secondary vial) were previously tested with the cobas HPV Test at the respective laboratory or at Roche Molecular Systems on 3 different cobas 4800 instruments. No special preparation was performed on the samples before prealiquoting from the primary vial. The residual sample (postcytology) in the ThinPrep vial (primary vial), which had been stored at 2°C to 8°C, was brought to room temperature, vortexed for 5 to 10 seconds, and placed on the cobas 4800 instrument for testing. For the T3000 processor, testing of the fresh secondary vial samples was performed at Scott and White Memorial Hospital (Temple, Tex) and Roche Molecular Systems, and postcytology testing on the primary vial was performed at Roche Molecular Systems on 3 different cobas 4800 instruments.

T2000 Comparison of cobas HPV Test Results in the Primary and Secondary Vials

For the T2000 analysis, all samples were selected from a subset of women enrolled in the baseline phase of the ATHENA study whose cytology had been processed on the T2000 processor. These archived residual samples, which had been stored in the primary ThinPrep vial after cytology was performed, were randomly selected based on age, cytology, and secondary vial HPV results and were then tested with the cobas HPV Test.

T3000 Comparison of cobas HPV Test Results in the Primary and Secondary Vials

For analysis using the T3000 system, a subset of archived samples from women enrolled in the baseline phase (the ATHENA study) whose cytology was processed on the T3000 processor was tested in the primary vial. However, because an insufficient number of samples from the ATHENA study had been processed on the T3000 processor, an additional 1500 fresh samples were collected from women aged ≥ 21 years. A subgroup of these fresh samples was selected to supplement the archived samples so that the combined archived and supplemental subgroups had distributions of age, cytology, and HPV results that were similar to those in the baseline phase of the ATHENA study.

Statistical Analysis

Results from the primary vial and secondary vial were compared in the following populations for both T2000 and T3000 samples: ASC-US population aged ≥ 21 years, the NILM population aged ≥ 30 years, and the overall population (independent of cytology result) aged ≥ 25 years.

Using secondary vial results (positive, negative, or overall) as the reference percentage of the primary vial, results that were in agreement were calculated. The Wilson method for score confidence intervals (CI) was used for calculating the CI for the percentage agreement.11

RESULTS

T2000 Analysis Sample Selection, Randomization, and Testing

A total of 1100 samples from women enrolled in the baseline phase of the ATHENA study whose cytology was processed on the T2000 system were selected based on cytology results, secondary vial cobas HPV Test results, and participant age. Samples were randomly selected within each subgroup. In addition, 156 extra available ASC-US samples were added to achieve a sample size of 200 women in the ASC-US population aged ≥ 21 years. In all, 1256 samples were selected and randomly assigned to be tested, and 1252 valid results were obtained from the primary vials tested on 3 different cobas 4800 instruments at Roche Molecular Systems.

A total of 200 samples were evaluable for the ASC-US population aged ≥ 21 years, 865 samples were evaluable for the NILM population aged ≥ 30 years, and 1096 samples were evaluable for the overall population aged ≥ 25 years.

T3000 Study Sample Selection, Randomization, and Testing

For the archived subset, 352 archived samples from women enrolled in the baseline phase of the ATHENA study whose cytology was processed on the T3000 system were randomly selected to be tested in the primary vial (Fig. 1). The samples were selected to match the distribution of the screening population in the ATHENA study. In addition, all extra available ASC-US samples (21 samples) were added to increase the sample size for the ASC-US population. The 373 samples were randomly assigned to be tested on 3 different cobas 4800 instruments.

Figure 1.

A flow chart is shown for archived and fresh sample testing for the Hologic ThinPrep 3000 System (T3000). ASC-US indicates atypical squamous cells of undetermined significance; HPV, human papilloma virus; NILM, negative for intraepithelial lesions or malignancies; Pap, Papanicolaou; RMS, Roche Molecular Systems; S&W, Scott and White Memorial Hospital.

For the fresh subset, 1500 fresh supplemental samples were collected at the Scott and White Memorial Hospital (Fig. 1). Secondary vial prealiquots from these samples were removed and tested with the cobas HPV Test at Scott and White Memorial Hospital or at Roche Molecular Systems. Cytology was performed on the samples in the primary vial at Scott and White Memorial Hospital. Valid results for cytology and the secondary vial cobas HPV Test were obtained for 1459 of the 1500 samples. A subset of 748 residual samples in the primary vial from the set of 1459 samples with valid results were randomly selected so that the distribution of the combined archived and freshly collected samples was similar with regard to participant age, cytology, and secondary vial cobas HPV Test results to those in the baseline phase of the ATHENA study. Note that all remaining available ASC-US samples (31 samples) were also selected for primary vial testing to increase the sample size for the ASC-US population. These selected samples (779 samples) were randomly assigned to be tested on 3 different cobas 4800 instruments at Roche Molecular Systems.

Cross-Contamination Studies

For the T2000 analysis, the 25 HPV-negative samples processed before the cross-contamination evaluation all tested negative, indicating a clean T2000 processor. From the 100 alternating positive/negative samples processed on the T2000, none of the 50 hrHPV-negative samples tested positive for hrHPV and all of the 50 hrHPV-positive samples tested positive for hrHPV, resulting in a 100% agreement (Table 1).

Table 1. Checkerboard Run Comparison of Precytology Pooled HPV Samples With Postcytology cobas HPV Test Results
ProcessorProcess StageSample TypeHrHPV
PositiveNegative
  • Abbreviations: 95% CI, 95% confidence interval; HPV, human papillomavirus; hrHPV, high-risk human papillomavirus.

  • a

    Six samples (4 negative and 2 positive) failed to process on the T3000 and were not included in the current analysis.

T2000PreprocessNegative 50
Positive50 
PostprocessNegative 50
Positive50 
% Agreement100100
95% CI(92.9%-100.0%)(92.9%-100.0%)
T3000PreprocessNegative 56a
Positive58a 
PostprocessNegative 56
Positive58 
% Agreement100100
95% CI(93.8%-100.0%)(93.6%-100.0%)

For the T3000 analysis, the 40 hrHPV-negative samples processed before the cross-contamination evaluation all tested negative, indicating a clean T3000 processor. From the 120 positive/negative samples processed in a checkerboard pattern on the T3000 system, none of the 60 hrHPV-negative samples tested positive for hrHPV and all of the 60 hrHPV-positive samples tested positive for hrHPV, resulting in 100% agreement (Table 1). The additional 40 hrHPV-negative samples processed on the T3000 system subsequent to the checkerboard runs were all negative, resulting in a 0% run-to-run cross-contamination rate. Six samples (4 negative and 2 positive) failed to process on the T3000 system and were excluded from the analysis.

Comparison of cobas HPV Test Results in Primary and Secondary Vials

For the T2000 analysis, comparisons of results, along with estimates of agreement between the primary vial and secondary vial, are shown in Tables 2 through 4 for each of the following populations: the ASC-US population aged ≥ 21 years (200 patients), the NILM population aged ≥ 30 years (865 patients), and the overall population aged ≥ 25 years (1096 patients).

Table 2. Comparison of cobas HPV Test Results From the Primary and Secondary Vials in the ASC-US Population Aged ≥21 Years for the T2000 Study (n = 200)
 Secondary Vial, cobas HPV Test Result 
Primary Vial, cobas HPV Test ResultPositiveNegativeTotalPositive Percentage Agreement % (95% CI)Negative Percentage Agreement % (95% CI)Overall Percentage Agreement % (95% CI)
  1. Abbreviations: 95% CI, 95% confidence interval; ASC-US, atypical squamous cells of undetermined significance; HPV, human papillomavirus; T2000, Hologic ThinPrep 2000 System.

Positive6356895.5% (87.5%-98.4%)96.3% (91.6%- 98.4%)96.0% ((92.3%- 98.0%)
Negative3129132
Total66134200
Table 3. Comparison of cobas HPV Test Results From the Primary and Secondary Vials in the NILM Population Aged ≥30 Years for the T2000 Study (n = 865)
 Secondary Vial, cobas HPV Test Result 
Primary Vial, cobas HPV Test ResultPositiveNegativeTotalPositive Percentage Agreement % (95% CI)Negative Percentage Agreement % (95% CI)Overall Percentage Agreement % (95% CI)
  1. Abbreviations: 95% CI, 95% confidence interval; HPV, human papillomavirus; NILM, negative for intraepithelial lesions or malignancies; T2000, Hologic ThinPrep 2000 System.

Positive5025286.2% (75.1%-92.8%)99.8% (99.1%-99.9%)98.8% (97.9%-99.4%)
Negative8805813
Total58807865
Table 4. Comparison of cobas HPV Test Results From the Primary and Secondary Vials in the Overall Population Aged ≥25 Years for the T2000 Study (n = 1096)
 Secondary Vial, cobas HPV Test Result 
Primary Vial, cobas HPV Test ResultPositiveNegativeTotalPositive Percentage Agreement % (95% CI)Negative Percentage Agreement % (95% CI)Overall Percentage Agreement % (95% CI)
  1. Abbreviations: 95% CI, 95% confidence interval; HPV, human papillomavirus; T2000, Hologic ThinPrep 2000 System.

Positive106411092.2% (85.8%-95.8%)99.6% (99.0%-99.8%)98.8% (98.0%-99.3%)
Negative9977986
Total1159811096

The overall percentage agreement (OPA) and negative percentage agreement (NPA) were both ≥ 96.0% in all 3 populations. The positive percentage agreement (PPA) was 95.5% in the ASC-US population, 86.2% in the NILM population, and 92.2% in the overall population aged ≥ 25 years. The magnitude of PPA in the NILM population was affected by the higher percentage of low positive results in this subpopulation. Discrepant results between the primary and secondary vials for all 3 populations are shown in Table 5. The elbow Ct values indicated that the majority of samples were near the clinical cutoff (40.0, 40.5, and 40.0 for channels 1, 2, and 3, respectively) for the cobas HPV Test.

Table 5. Discrepant Results in the T2000 Study
Age, YearsPap ResultSecondary Vial HPV ResultPrimary Vial ResultSecondary Vial Ct Ch 1Primary Vial Ct Ch 1Secondary Vial Ct Ch 2Primary Vial Ct Ch 2Secondary Vial Ct Ch 3Primary Vial Ct Ch 3CPR Result
  1. Abbreviations: ASC-US, atypical squamous cells of undetermined significance; Ch, channel; CIN, cervical intraepithelial neoplasia; CPR, central pathology review; Ct, cycle threshold; HPV, human papillomavirus; NA, not applicable; NaN, not a number; NILM, negative for intraepithelial lesions or malignancies; Pap, Papanicolaou; T2000, Hologic ThinPrep 2000 System.

31NILMPositiveNegative37.6NaNNaNNaNNaNNaNNormal
45NILMPositiveNegative38.9NaNNaNNaNNaNNaNNormal
42NILMPositiveNegative38.840.4NaNNaNNaNNaNNormal
55NILMPositiveNegative39.341.7NaNNaNNaNNaNNormal
32NILMPositiveNegative29.8NaNNaNNaNNaNNaNNormal
30NILMPositiveNegative39.741.1NaNNaNNaNNaNNormal
32NILMPositiveNegative34.440.8NaNNaNNaNNaNNormal
39NILMPositiveNegative35.442.2NaNNaNNaNNaNNormal
27NILMPositiveNegative39.540.5NaNNaNNaNNaNNormal
47NILMNegativePositive41.437.3NaNNaNNaNNaNNormal
40NILMNegativePositive40.139.6NaNNaNNaNNaNCIN1
28NILMNegativePositiveNaN39.6NaNNaNNaNNaNNormal
21ASC-USPositiveNegative38.940.6NaNNaNNaNNaNNormal
64ASC-USPositiveNegative39.9NaNNaNNaNNaNNaNNormal
25ASC-USPositiveNegative38.842.3NaNNaNNaNNaNNA
24ASC-USNegativePositive40.340.0NaNNaNNaNNaNNormal
22ASC-USNegativePositiveNaN32.4NaNNaNNaNNaNNormal
21ASC-USNegativePositive43.535.9NaNNaNNaNNaNNormal
40ASC-USNegativePositive40.639.5NaNNaNNaNNaNNormal
32ASC-USNegativePositive40.636.9NaNNaNNaNNaNNormal

For the T3000 analysis, comparisons of results, along with estimates of agreement between the primary vial and secondary vial, are shown in Tables 6 through 8 for each of the following populations: the ASC-US population aged ≥ 21 years (92 patients), the NILM population aged ≥ 30 years (834 patients), and the overall population aged ≥ 25 years (1064 patients).

Table 6. Comparison of cobas HPV Test Results From the Primary and Secondary Vials in the ASC-US Population Aged ≥21 Years for the T3000 Study (n = 92)
 Secondary Vial, cobas HPV Test Result 
Primary Vial, cobas HPV Test ResultPositiveNegativeTotalPositive Percentage Agreement % (95% CI)Negative Percentage Agreement % (95% CI)Overall Percentage Agreement % (95% CI)
  1. Abbreviation: 95% CI, 95% confidence interval; ASC-US, atypical squamous cells of undetermined significance; HPV, human papillomavirus; T3000, Hologic ThinPrep 3000 System.

Positive4344795.6% (85.2%-98.8%)91.5% (80.1%-96.6%)93.5% (86.5%-97.0%)
Negative24345
Total454792
Table 7. Comparison of cobas HPV Test Results From the Primary and Secondary Vials in the NILM Population Aged ≥30 Years for the T3000 Study (n = 834)
 Secondary Vial, cobas HPV Test Result 
Primary Vial, cobas HPV Test ResultPositiveNegativeTotalPositive Percentage Agreement % (95% CI)Negative Percentage Agreement % (95% CI)Overall Percentage Agreement % (95% CI)
  1. Abbreviations: 95% CI, 95% confidence interval; HPV, human papillomavirus; NILM, negative for intraepithelial lesions or malignancies; T3000, Hologic ThinPrep 3000 System.

Positive4875584.2% (72.6%-91.5%)99.1% (98.2%-99.6%)98.1% (96.9%-98.8%)
Negative9770779
Total57777834
Table 8. Comparison of cobas HPV Test Results From the Primary and Secondary Vials in the Overall Population Aged ≥25 Years for the T3000 Study (n = 1064)
 Secondary Vial, cobas HPV Test Result 
Primary Vial, cobas HPV Test ResultPositiveNegativeTotalPositive Percentage Agreement % (95% CI)Negative Percentage Agreement % (95% CI)Overall Percentage Agreement % (95% CI)
  1. Abbreviations: 95% CI, 95% confidence interval; HPV, human papillomavirus; T3000, Hologic ThinPrep 3000 System.

Positive97910685.8% (78.2%-91.1%)99.1% (98.2%-99.5%)97.7% (96.6%-98.4%)
Negative16942958
Total1139511064

In the ASC-US population, the PPA and OPA were ≥ 93.5%. In this population, the magnitude of NPA (91.5%) was affected by the small sample size (47 patients) as well as by the presence of low positive samples. In the NILM and overall populations, the OPA and NPA were both ≥ 97.7 %. The PPA was ≥84.2% in these 2 populations and was also affected by the presence of low positive samples in these 2 populations. Discrepant results between the primary and secondary vials for all 3 populations are shown in Table 9. There were 27 samples for which the results between the primary vial and secondary vial disagreed. The elbow Ct values indicated that the majority of these samples were near the clinical cutoff (40.0, 40.5, and 40.0 for channels 1, 2, and 3, respectively) for the cobas HPV Test.

Table 9. Discrepant Results in the T3000 Study
Sample PapAge, YearsPap ResultSecondary Vial HPV ResultPrimary Vial ResultSecondary Vial Ct Ch 1Primary Vial Ct Ch 1Secondary Vial Ct Ch 2Primary Vial Ct Ch 2Secondary Vial Ct Ch 3Primary Vial Ct Ch 3CPR Result
  1. Abbreviations: ASC-US, atypical squamous cells of undetermined significance; Ch, channel; CPR, central pathology review; Ct, cycle threshold; HPV, human papillomavirus; NA, not applicable; NaN, not a number; NILM, negative for intraepithelial lesions or malignancies; Pap, Papanicolaou; T3000, Hologic ThinPrep 3000 System.

Archived47NILMPositiveNegative38.743.2NaNNaNNaNNaNNA
Archived27NILMPositiveNegative39.8NaNNaNNaNNaNNaNNA
Archived38NILMPositiveNegative38.141.9NaNNaNNaNNaNNormal
Archived38NILMPositiveNegative38.9NaNNaNNaNNaNNaNNA
Archived44NILMPositiveNegative40.0NaNNaNNaNNaNNaNNA
Archived35NILMNegativePositive40.639.5NaNNaNNaNNaNNA
Archived30ASC-USNegativePositive40.039.2NaNNaNNaNNaNNormal
Fresh54NILMPositiveNegativeNaNNaNNaNNaN35.1NaNNA
Fresh76NILMPositiveNegative38.7NaNNaNNaNNaNNaNNA
Fresh63NILMPositiveNegativeNaNNaN39.441.2NaNNaNNA
Fresh26NILMPositiveNegative36.140.1NaNNaNNaNNaNNA
Fresh25NILMPositiveNegative39.440.6NaNNaNNaNNaNNA
Fresh26NILMPositiveNegative40.842.236.9NaNNaNNaNNA
Fresh26NILMPositiveNegativeNaNNaN39.541.3NaNNaNNA
Fresh37NILMPositiveNegative39.1NaNNaNNaNNaNNaNNA
Fresh47NILMPositiveNegative35.1NaNNaNNaNNaNNaNNA
Fresh70NILMNegativePositiveNaN38.341.0NaNNaNNaNNA
Fresh46NILMNegativePositiveNaN41.541.339.3NaNNaNNA
Fresh30NILMNegativePositive44.938.5NaNNaNNaNNaNNA
Fresh58NILMNegativePositiveNaN38.9NaNNaNNaNNaNNA
Fresh33NILMNegativePositiveNaN39.9NaNNaNNaNNaNNA
Fresh60NILMNegativePositive41.338.3NaNNaNNaNNaNNA
Fresh53ASC-USPositiveNegative39.540.5NaNNaNNaNNaNNA
Fresh25ASC-USPositiveNegative36.8NaNNaNNaNNaNNaNNA
Fresh27ASC-USNegativePositiveNaN40.3NaN39.7NaNNaNNA
Fresh28ASC-USNegativePositive42.139.8NaNNaNNaNNaNNA
Fresh48ASC-USNegativePositiveNaN39.9NaNNaNNaNNaNNA

DISCUSSION

In the current study, the potential for contamination when a nucleic acid amplification test is performed on a liquid-based cytology sample after cytology processing was investigated. Results have been presented for cross-contamination studies and for the comparison of cobas HPV Test results in the secondary vial (precytology) and primary vial (postcytology) samples run on the T2000 and T3000 processors for patients with ASC-US cytology who are aged ≥ 21 years, patients with NILM cytology who are aged ≥ 30 years, and the overall patient population (independent of cytology) aged ≥ 25 years.

The cross-contamination studies that were run in a checkerboard pattern (alternating between positive and negative samples) indicated a 0% sample-to-sample contamination rate when testing cervical samples below the median Ct value for the entire population collected in PreservCyt solution and processed on either the T2000 or T3000 processor. Previous cross-contamination studies using contrived positive samples were performed at levels that represented at least 95% of a normal population of positive specimen target levels and demonstrated no cross-contamination; however, there was not a sufficient number of clinical specimens at this level. As a result, the HPV-positive samples were prepared by pooling and diluting HPV-positive PreservCyt specimens in HPV-negative specimen background to produce an average Ct value below the median Ct value for the entire ATHENA test population, thereby achieving levels on the high end of typical target levels.

The results of the secondary vial versus the primary vial studies were compared by estimating percentage agreements. The magnitude of percentage agreement estimates can be affected by 1) the percentage of low positive samples in the study, 2) the reproducibility of the assay, and 3) the sample size. In the current study, samples near the assay cutoff that gave positive results when tested in the secondary vial but negative results when tested in the primary vial did so because of assay variability affecting the PPA estimate. The variability of PPA to be expected can be estimated from the results observed in the ATHENA study and applied to the results observed herein.

In the ASC-US population aged ≥ 21 years in the ATHENA study, the percentage of low positive women among HPV-positive women was 18%. Approximately one-half of these women (9%) can be expected to remain positive at the time of retesting. The remaining 82% can be expected to remain HPV positive because their viral loads are far above the assay cutoff. Therefore, the expected estimated PPA is 91% (82% + 9%); the observed PPA for the ASC-US population was higher in both the T2000 and T3000 analyses.

By similar analyses using the Athena estimates of percentage of low positive samples for the NILM population aged ≥30 years, the observed PPA of 86.2% for the T2000 and 84.2% for the T3000 was higher than the expected percentage agreement between pre- and postcytology results of samples processed on either the T2000 or T3000 systems. Likewise, the estimation of the percentage of low positives samples in the overall population aged ≥25 years indicated that the observed PPA between pre- and postcytology samples of 92.2% and 85.8% in the T2000 and T3000 analyses, respectively also exceeded the expected agreement.

In comparing the results of the cobas HPV Test from the primary versus those of the secondary vial, NPA will be the best indicator of contamination because it will demonstrate the extent to which negative samples remain negative when tested after cytology. The data presented in the current study demonstrated the NPA to be consistently > 99% for all populations for both the T2000 and T3000 systems, with the exception of the ASC-US population aged ≥ 21 years. The marginally lower NPA estimates in the ASC-US population (96.3% and 91.5%, respectively, for the T2000 and T3000 systems) were likely due to the presence of low positive samples and to the smaller sample size in these subpopulations.

In those samples for which results between the secondary and primary vials were discordant and for which disease outcomes were available, we were also able to determine that there was no high-grade cervical disease detected. Again, this finding supported the explanation that discordance was due to the variation expected from samples with low viral loads, resulting in low positive results, which are of questionable clinical relevance.

The results of the current study demonstrate that the use of the primary postcytology residual samples collected in PreservCyt vials and run on the T2000 and T3000 processors provide reliable cobas HPV Test results that are comparable to results from the precytology secondary vial, without any evidence of contamination.

FUNDING SUPPORT

Supported by Roche Molecular Systems, Pleasanton, California.

CONFLICT OF INTEREST DISCLOSURES

Dr. Krevolin, Ms. Sun, Mr. Purner, Dr. Sharma, and Dr. Behrens are employees of Roche Molecular Systems. Drs. Rao and Young are Principal Investigators at two of the laboratory sites for the ATHENA (Addressing the Need for Advanced HPV Diagnostics) study (Scott & White & Tricore). Drs. Rao and Young are on the Molecular Advisory Board for Roche, and Dr. Young is on the advisory board for Quidel and has received honoraria for his participation.

Ancillary