Clinical Performance of the BD Onclarity Extended Genotyping Assay for the Management of Women Positive for Human Papillomavirus in Cervical Cancer Screening

With adequate specificity and superior sensitivity to cytology (1–3), high-risk human papillomavirus (hrHPV) testing has become the primary cervical cancer screening test in several countries including Australia, England, the United States, and Turkey (4–7). Moreover, many more countries use human papillomavirus (HPV) testing as a cotest with cytology and/or to triage women with abnormal cytology (8, 9). Although HPV testing is part of an increasing number of cervical cancer screening programs, few assays differentiate between individual HPV genotypes in the management of results, and if they do, it is often confined to HPVs 16 and 18 (10). Yet, the risk of progression to cervical intraepithelial neoplasia grade 3 or cervical cancer (CIN3+) differs among hrHPV genotypes (10, 11). For example, the risk of HPV31 progressing to CIN3+ is lower than the risk conferred by HPV16 but higher than that of HPVs 51, 39, and 68 (11–14). Additionally, as HPV vaccination coverage increases, so does the importance of type-specific detection beyond HPVs 16 and 18.

The first HPV test approved by the FDA, Hybrid Capture 2 (HC2; Qiagen), detects 13 hrHPVs (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) as a pool without distinguishing individual genotypes. In February 2018, BD Onclarity HPV Assay [Becton Dickinson (BD) Diagnostics)–herein Onclarity–received approval from the FDA. Onclarity is one of five FDA-approved HPV tests, and one of two (the other is cobas) tests approved for use in primary HPV testing without cytology cotesting (15). The Onclarity assay is a real-time multiplex PCR-based assay targeting the E6/E7 region of the HPV viral genome. This assay detects HPVs 16, 18, 31, 45, 51, and 52 individually and eight genotypes as channels (33/58, 56/59/66, and 35/39/68; ref. 15). In contrast, the Linear Array assay (LA; Roche Molecular Diagnostics) offers a full genotyping test by detecting 36 HPV genotypes individually (16). Although there is no “gold standard” HPV genotyping test, LA has been widely used in epidemiologic studies, and has been validated for that purpose (16–18). In addition, within the VALidation of HPV GENotyping Tests (VALGENT) framework, the LA assay has also shown similar sensitivity and superior specificity (for the 13 hrHPVs detected by HC2) compared to HC2 to detect CIN grade 2 (CIN2+) lesions (19).

Given that most hrHPV infections resolve without intervention, and that the risk of progression to cervical cancer and its precursors differs among hrHPV genotypes, triaging women positive for specific hrHPV genotype(s) could circumvent unnecessary referrals to colposcopy and overtreatment of lesions (20). Evaluating the clinical performance of an extended genotyping test (i.e., Onclarity) would contribute to the evidence on how genotyping could permit finer risk stratification of women positive for a pool of hrHPV (18, 21). In this context, we compared–among women who tested positive for HPV by HC2 (with and without cervical abnormalities) in the Canadian Cervical Cancer Screening Trial (CCCaST)–the concordance between Onclarity and LA, overall and by viral load. We also evaluated Onclarity's clinical performance via 24 triage strategies for the detection of CIN2+ and CIN3+.

This study complies with the Standards for Reporting Diagnostic Accuracy (STARD) studies (22).

This study was conducted using data for a subsample of women enrolled in the CCCaST, which was originally designed to compare hrHPV DNA testing and Papanicolaou (Pap) cytology for the detection of CIN2+. Detailed descriptions of the study design and results have been published (23–26). Briefly, women (30–69 years old) attending routine cervical cancer screening in Montréal, Quebec and St. John's, Newfoundland, Canada, from 2002 to 2005 (Controlled Trials Number: ISRCTN57612064) were recruited. A total of 10,154 women eligible for the trial were randomly assigned to receive either a Pap test followed by an HPV test, or an HPV test followed by a Pap test. Of note, when the CCCaST was designed, it was considered unethical to withhold cytology and hence the order of tests was randomized. Women testing hrHPV⁺ or with abnormal cytology were referred for a colposcopic exam at participating clinics according to a standardized protocol, as previously described (24, 25). A histologic diagnosis was determined via biopsy, where biopsies were taken of all abnormal-appearing cervixes and at least one biopsy of normal-appearing ectocervical epithelium. Colposcopists and pathologists ascertaining disease endpoints were blinded to screening test results. If diagnosed with CIN2+, women were followed outside of the trial; if they were diagnosed with CIN1 or were lesion free, colposcopy was repeated 6 months later. Women were invited for a second round of screening 12 to 18 months after the initial visit, where the same process was repeated. The CCCaST was conducted in accordance with the Declaration of Helsinki as an ethical guideline. All participants provided written informed consent. Ethical approval for the CCCaST was obtained from McGill and Memorial Universities and all participating clinics and hospitals.

Cervical specimens were collected using Digene sampling kits (Qiagen, and tested for HPV with HC2, as previously described (24). For specimens that tested positive, HPV DNA was extracted using the MasterPure DNA Purification Kit (Epicentre Biotechnologies), and HPV genotyping was performed, as previously described, with the LA assay using the manufacturer proposed cut-offs for positive results (16). HPV52 infection was confirmed with a real-time PCR assay (27). Onclarity genotyping was performed on the residual specimen transport media of HC2-positive samples (stored at −80°C) with the BD Viper LT system according to the manufacturer's instructions, including recommended cut-offs for positive results. Samples with discordant results were retested with LA, where the repeat test results were included. No clinical interventions were applied to specimens between LA genotyping (performed in approximately year 2008) and Onclarity genotyping, which was performed around year 2016. When genotyping the samples with Onclarity, the test performers and readers were blinded to all previous HPV tests (HC2 and LA), cytology, and histology results. Since both LA and Onclarity genotyping tests were performed on precollected samples, there were no adverse events from performing these tests.

Cytology results were interpreted according to Bethesda 2001 terminology (28) and categorized for this analysis as: (i) negative for intraepithelial lesion or malignancy (NILM); (ii) atypical squamous cells–of undetermined significance (ASC-US); (iii) low-grade squamous intraepithelial lesion (LSIL); and (iv) high-grade squamous intraepithelial lesion (HSIL) or worse, which included HSIL, atypical squamous cells–cannot exclude HSIL (ASC-H), atypical glandular cells (AGC), adenocarcinoma in situ (AIS), and cancer. Cytotechnologists and cytopathologists examining Pap smears were blinded to HPV status of the samples, and they were unaware that the samples were collected from women participating in the CCCaST.

Several studies have evaluated Onclarity using hierarchical analyses to group genotypes by presumed risk (12, 18). For example, Demarco and colleagues utilized the classifications of HPV16, HPV18, HPV31/33/58/52/45, then HPV51/39/68/36/59/56/66 to delineate highest- to lowest-risk types, respectively (18). Although this approach is practical in a clinical setting that relies on simplified groupings to triage women, it could mask the importance of individual genotypes. We assessed 24 triage strategies: three abnormal cytology threshold strategies (ASC-US, LSIL, and HSIL) individually; the three aforementioned strategies combined with HPV16 and/or 18 (herein HPV16/18), and then combined with a non-HPV16/18 result; positivity for any Onclarity-detected HPV genotype/channel; and positivity to HPV16/18, 16/31, 16/45, 18/31, and 18/45.

To summarize the study population, women's cytology and Onclarity-detected genotyping results were tabulated for each histologic endpoint. We cross-tabulated Onclarity-detected genotypes by LA genotyping results. We then assessed concordance between the index (Onclarity) and reference (LA) tests by estimating expected agreement, agreement, and the κ statistic, with their respective 95% confidence intervals (CI). The concordance analysis was performed on hrHPV⁺ samples; women who tested hrHPV⁺ at both screening rounds contributed two samples. We also performed a sensitivity analysis considering women rather than samples as the unit of analysis; when a woman was hrHPV⁺ at both screening rounds, only her first result was included. The concordance analysis was performed overall and stratified by HPV viral load. The latter was expressed using the relative light unit (RLU) signal assessed by HC2; samples with greater than or equal to the median RLUs (was rounded to the nearest one) were categorized as “higher” viral load, and those with less than the median RLUs were categorized as “lower” viral load.

To assess the value of HPV genotyping at the point of care, we considered the screening round that contained the worst histologic outcome. For example, consider a woman testing hrHPV⁺ at the first and second screening rounds. With an hrHPV⁺ result at enrollment, she was referred to colposcopy; and if she was diagnosed with CIN1 at the first round and then CIN2 at the second round, she appeared in the analysis one time using the HPV, cytology, and histology results obtained at round two since they corresponded to her worst histologic endpoint. When both screening rounds produced the same histologic result, results from the first round were analyzed. Although HPV testing was performed on samples collected prior to colposcopic examination, this analysis is retrospective.

We compared the clinical performance of the various triage strategies evaluated by calculating their respective sensitivity, specificity, as well as positive and negative predictive values (PPV, NPV) for the detection of CIN2+ and CIN3+. Specificity, PPVs, and NPVs were calculated at two levels; the main analysis compared the CIN endpoint to all lower histologic grades (i.e., ≤CIN1 diagnoses vs. CIN2+ and ≤CIN2 diagnoses vs. CIN3+), whereas the sensitivity analysis compared the CIN endpoint with lesion-free women (i.e., normal vs. CIN2+ and normal vs. CIN3+). All analyses were performed using Stata 17 software (StataCorp).

Due to privacy/ethical restrictions, the individual study participant data the analyses are based on are not publicly available. The deidentified information can be derived from the various tables and supplementary tables available in the manuscript.

Among the 10,135 women with a valid HC2 test result at either screening round, 759 (7.5%) tested hrHPV⁺ via HC2 at one or both screening rounds. Of these, Onclarity and LA genotyping data were available for 696 (91.7%) women (Supplementary Fig. S1 shows the flow chart of participant/sample selection). Thirty-eight women tested HPV positive via HC2 at both screening rounds, resulting in 734 samples for the concordance analysis. Of the 696 women with Onclarity genotyping data, histology results were available for 617 of them (88.7%), among whom valid cytology was available for 606 (98.2%) included in the triage strategies analyses. As shown in Table 1, 64 women had CIN2+ and 46 had CIN3+. Most women's (n = 514) worst histologic diagnosis corresponded to that from the first screening round. The median age of women with histology result was 38 years (interquartile range: 33–45). The majority (75.9%) of women tested positive for one or more of the 14 genotypes/channels assessed by Onclarity (Table 1). The most frequently detected genotypes/channels were HPV56/59/66 (17.8%), followed by HPV16 (16.1%), HPV35/39/68 (14.9%), then HPV31 (11.0%).

Table 2 shows the high concordance observed between Onclarity and LA. Concordance was near perfect for the six individual HPV types detected by Onclarity, with κ values ranging from 0.98 for HPV31 to 0.92 for both HPV51 and HPV52; concordance was lowest for HPV35/39/68 (κ = 0.70; CI, 0.63–0.78). No notable differences were observed by HPV viral load, although concordance was slightly better among those with higher viral loads. Results were robust when women instead of samples were considered as the unit of analysis (Supplementary Table S1). Supplementary Table S2 contains a cross-tabulation of Onclarity-detected genotypes by the LA results. In terms of discordant samples (Supplementary Table S3), and considering the genotypes reported by Onclarity, 21 samples tested positive with LA but negative with Onclarity, 18 of which were single type/channel infections. Thirty-nine samples tested positive with Onclarity but negative with LA, of which 36 were single type infections. Twenty-seven of the 39 samples that tested negative via LA but positive for Onclarity, were positive for channel HPV35/39/68.

The performances of different strategies for the detection of CIN2+ or CIN3+ are presented in Tables 3 and 4, respectively. For both CIN endpoints, sensitivity decreased while specificity increased with higher cytological abnormalities. Cytology was more sensitive for the detection of CIN2+ or CIN3+ than cytology combined with HPV positivity to 16/18 and non-16/18 genotypes; however, this finding was only statistically significant for the ASC-US threshold with positivity to non-16/18 genotypes. Sensitivity to detect CIN2+ or CIN3+ among individual genotypes was highest for HPV16, followed by HPV31, then HPV18 (Tables 3 and 4). Examining the performances of 24 triage strategies for the detection of CIN2+ and CIN3+ revealed that the three strategies with the highest sensitivities were positivity to any Onclarity-detected genotype/channel, HPV16/31, and HPV16/18; however, positivity via any Onclarity-detected genotype/channel had a low specificity. In the sensitivity analysis comparing the CIN endpoint to lesion-free women, the specificity, PPV, and NPV values were similar (Supplementary Tables S4 and S5).

We found that among women positive for hrHPV, concordance between LA and Onclarity genotyping was near perfect for the genotypes/channels reported by Onclarity and did not differ by viral load. Among the 24 triage strategies for the detection of CIN2+ and CIN3+, HPV16/31 and HPV16/18 emerged as the better-performing strategies.

Demarco and colleagues assessed Onclarity and LA concordance among 1,919 women hrHPV⁺ via HC2 with prior ASC-US cytology and reported excellent agreement, with κ values ranging from 0.80 for HPV33/35/68 to 0.97 for HPV16 (18). Our study also found that HPV33/35/68 had the lowest κ value, and that concordance was excellent for HPV16. We additionally assessed concordance by viral load, as this had not yet been performed in previous studies; however, we found that ”higher” versus “lower” viral load (via HC2) did not significantly impact concordance in this study population.

Consistent with previous studies assessing Onclarity, our findings confirmed the importance of HPV16 (12, 18, 29). Since a recent systematic review highlighted that HPV31 carries CIN3+ risks comparable with HPV18 (10), we included the detection HPV16 and/or 31 as a triage strategy. Detection of HPV31 had comparable clinical performance to HPV18 to detect CIN2+ and CIN3+. Future studies should evaluate including HPV31 positivity (with and without other hrHPV types) as an indication for immediate colposcopy in a triage strategy.

We acknowledge some potential limitations of our study. First, study endpoints are surrogates for cancer. In fact, CIN2 may be linked to HPV types that, although are important for CIN2 development, are relatively less important for cancer development (12). However, the majority (71.9%) of CIN2+ cases were CIN3+ diagnoses. Second, since there is no accepted “gold standard” HPV genotyping test, our concordance analysis using LA as the reference test is susceptible to imperfect reference test bias. Third, having a sample size of four cancer cases excluded the possibility of examining Onclarity's performance in detecting cervical cancer, as done in an earlier study (12). In addition, we had intended to include genotype strategies of positivity to multiple genotypes, such as positivity to HPV16 and 18 or HPV16 and 31; however, the number of women positive for both genotypes was too few (2 and 6 women, respectively). A larger sample size would also have afforded greater statistical precision. Fourth, cervical specimens were tested via LA and Onclarity several years after HC2 testing; this may have implications for the quality of samples and could explain why less than 80% of samples tested by the genotyping assays were positive. Finally, only hrHPV⁺ specimens were genotyped and thus inferences are limited to a population screened as hrHPV⁺; this is, however, adequate to assess the value of triage strategies obtained with a generic HPV assay.

Our findings contribute information that can help identify women at greater risk of disease progression. These findings may help inform clinical guidelines regarding the value of HPV genotyping beyond HPV16/18 for women positive for hrHPV.

K.D. Volesky reports receiving an an honorarium from BD and Company for a conference presentation outside the submitted work. M. El-Zein has a patent for DNA methylation markers for early detection of cervical cancer pending. E.L. Franco reports grants and personal fees from Merck; and personal fees from GlaxoSmithKline outside the submitted work; in addition, E.L. Franco has a patent for Methylation markers pending to self. F. Coutlée reports grants and nonfinancial support from Becton Dickinson during the conduct of the study; in addition, F. Coutlée has received grants paid to the research center of the Centre hospitalier de l'Université de Montréal (CHUM; Montréal, Québec, Canada) for research projects from Roche Diagnostics and Merck Sharp and Dome, honorariums for presentations from Merck Sharp and Dome and Roche diagnostics; and has participated in an expert group by Merck Sharp and Dome. These activities took place more than 5 years ago. No disclosures were reported by the other authors.

K.D. Volesky: Formal analysis, writing–original draft. S. Magnan: Formal analysis, writing–review and editing. M.-H. Mayrand: Data curation, formal analysis, writing–review and editing. S.D. Isidean: Formal analysis, writing–review and editing. M. El-Zein: Supervision, writing–review and editing. E. Comète: Data curation, writing–review and editing. E.L. Franco: Conceptualization, supervision, writing–review and editing. F. Coutlée: Conceptualization, resources, data curation, writing–review and editing.

E.L. Franco received funding from the Canadian Institutes of Health Research (CIHR; grant nos. MCT-54063 and FDN-143347) for The CCCaST study. F Coutlée received funding from Becton Dickinson who provided the Onclarity kits for HPV detection and genotyping at no cost and contributed to the salaries of research assistants to perform testing.

Montréal Research Staff: L. Abruzzese, K. Bellegarde, M. Bernardin, E. Duarte, F. Ferdinand, E.L. Franco (Principal Investigator), S.D. Isidean, G. Kelsal, M.H. Mayrand, M. Paquin, S. Piché, J. Poirier, A.V. Ramanakumar, A. Rodrigues, N. Rousseau, C. Schwartz, N. Slavtcheva, E. Tunistky; St. John's Research Staff: A. Batstone, G. Condon, A. Fitzpatrick, P. Francis, B. Halfyard, C. Head, C. Leonard, D. Mason, J. McGrath, V. Moulton, E. Oates, W. Shea; Montreal Clinical Collaborators: M.Y. Arsenault, G. Asselin, L. Authier, S. Bagga, P. Bastien, L. Bazinet, F. Beaudoin, P. Beaulieu, M.J. Bédard, S. Belinski, S. Bélisle, J. Benoit, M. Bernard, S. Bianki, L. Biron, F. Bissonnette, R. Bou-Habib, J. Bourque, B. Bradbury, M. Champagne, Y. Charles, P. Choquette, J.N. Couture, H.Q. Dao, C. Desjardins, J. Desjardins, L. Desrosiers, S. DiTommaso, L. Dontigny, M. Doyle, J. Dubé, M.J. Dupuis, F. Durocher, F. Engel, B. Fafard, A. Ferenczy, G. Fortier, A. Fortin, C. Fortin, D. Francoeur, D. Frechette, P. Fugère, G. Gagné, S. Gascon, M.J. Gaudreau, D. Gaudron, K. Gemayel, L. Gilbert, S. Gilbert, J. Gill, I. Girard, A. Gobeil, L. Granger, E. Grou, F. Grou, G. Guertin, J. Guimond, R. Hemmings, N. Ifergan, C. Johnson, L. Johnson, L. Ketchian, Y. Korcaz, C. Lafortune, J. Lalande, J.F. Lanctot, D. Landry, M. Landry, D. Langevin, I. Langlois, L. Lanmy-Monnot, L. Lapensée, L. Larouche, D. Laurin, M.C. Lavigne, Y. Lavoie, M. Leduc, F. Leger, N. Leroux, G. Luskey, N. Mansour, J. Marceau, A. Masse, I. Mayrand, M.H. Mayrand, L.R. McLauchlin, S. Menard, C. Mercer, M. Messier, B. Michon, M. Nadeau, M. Nguyen, S. Ouellet, C. Paquin, R. Paré, S. Peloquin, Y. Piché, R. Pichet, C. Rivard, I. Rodrigues, S. Roman, L. Rusimovic, G. Sanche, D. Soulière, D. Sproule, M. Steben, S. Still, D. Theriault, G. Tondreau, D. Tremblay, T. Minh Dung Vo, V.M. Whitehead, M. Yaffe, A. DiZazzo, C. Ziegler. St. John's Clinical Collaborators: E. Bannister, E. Callahan, J. Collingwood, P. Crocker, L. Dawson, A. Drover, J. Dunne, F. Fifield, J. Fitzgerald, D. Fontaine, B. Grandy, M. Greene, K. Halley, L. Hatcher, J. Hickey, P. Horwood, J. Janes, F. Jardine, L. Kieley, S. King, C. Kirby, N. Kum, E. Mate, S. McGrath, C. McManamon, K. Misik, M. O’Dea, P. O'Shea, C. Peddle, M. Penton, C. Pike, P. Power, L. Rogers, K. Saunders, P. Skirving, T. Sullivan, J. Verge, P. Wadden, M. Watson, M. Young; HPV DNA Testing Laboratories: F. Coutlée (Montréal), S. Ratnam (St. John’s).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.