Clinical performance of the novel full‐genotyping OncoPredict HPV Quantitative Typing assay using the VALGENT framework

Clinical validation of human papillomavirus (HPV) assays according to international criteria is prerequisite for their implementation in cervical cancer screening. OncoPredict HPV Quantitative Typing (QT) assay (Hiantis Srl, Milan, Italy) is a novel full‐genotyping multiplex real‐time PCR quantitative assay targeting E6/E7 genes, allowing individual viral load determination of 12 high‐risk (HR) HPV types. Quality controls for sample adequacy, efficiency of nucleic acid extraction and PCR inhibition are included in the assay. Clinical performance of OncoPredict HPV QT test was assessed as part of the “Validation of HPV Genotyping Tests” (VALGENT‐2) framework, consisting of 1300 cervical liquid‐based cytology (LBC) samples of women aged between 20 and 60 years who had originally attended for routine cervical screening in Scotland. The clinical accuracy of the OncoPredict HPV QT (index test) for the detection of CIN2+ was assessed relative to the GP5+/6+ Enzyme ImmunoAssay (GP5+/6+ EIA) (comparator test), using noninferiority criteria. Intra‐ and interlaboratory reproducibility of the assay was assessed on a subpopulation, comprising 526 samples. The relative sensitivity and specificity for OncoPredict HPV QT vs GP5+/6+‐PCR‐EIA were 1.01 (95% CI: 0.99‐1.03) and 1.03 (95% CI: 1.0‐1.06) respectively. The P‐values for noninferiority were ≤0.001. The intra‐ and inter‐laboratory reproducibility demonstrated a high concordance (>98.7%) with kappas for individual types ranging from 0.66 to 1.00. OncoPredict HPV QT fulfills the international validation criteria for the use of HPV tests in cervical cancer screening.

Testing for high-risk HPV infection is an effective screening tool for the prevention of invasive cervical cancer.Many HPV assays are commercially available, but few have been fully validated according to international guidelines.Here, the authors evaluated the clinical accuracy and reproducibility of OncoPredict HPV Quantitative Typing assay, a novel full-genotyping assay targeting the 12 oncogenic HPV types.They showed that OncoPredict HPV QT performed with comparable accuracy to the well-established GP5+/GP6+ Enzyme ImmunoAssay and had good reproducibility within and between labs.Its ability to determine the genotype-specific viral load makes the OncoPredict HPV QT also useful for the risk stratification and follow-up monitoring of HPV-positive women.

| INTRODUCTION
Cervical cancer is the second most common cause of death in women of reproductive age worldwide. 1Persistent infection with high-risk HPV (HR-HPV) is now known to be associated with progression to precancer and ultimately invasive cancer. 27][8] Consequently, many countries worldwide have already, or are now shifting to HR-HPV primary screening, using HPV tests that have undergone clinical validation according to international criteria. 9In recent years many HPV assays have become commercially available but only a relatively small number have been fully validated according to the international guidelines. 10,11The European VALidation of HPV GENotyping Tests (VALGENT) collaborative framework was designed to support robust evaluation and validation of HPV tests including those with genotyping capacity.It involves the application of HPV tests to samples derived from a screening population and a disease-enriched population, where clinical outcomes are known. 12Our study aimed to evaluate the clinical accuracy and intra-and interlaboratory reproducibility of the novel quantitative full-genotyping assay, OncoPredict HPV QT, which allows normalized genotype-specific viral load determination.Evaluation of clinical performance and reproducibility was performed through the VALGENT-2 iteration.

| Study population and VALGENT-2 panel
The VALGENT-2 framework/collection is based on samples collected from the organized screening program in Scotland in 2012.Dimensions of the sample are described in detail elsewhere. 13  types (viral genomic units/10 4 cells).PCR was carried out using a CFX384 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA).Samples were considered adequate for the HR-HPV analysis if at least 400 cells/reaction were detected, if nucleic acid extraction efficiency was ≥10% and if the amplification control cycle threshold (Ct) was ≤30.Samples were then considered to be HR-HPV positive if a viral load ≥50 copies/10 4 cells was detected.In the case of multiple HR-HPV infections, the HPV type with the highest viral load or predominant HR-HPV type was used for assessing positivity.

| Testing of samples with standard comparator test
In VALGENT-2, the GP5+/6+ PCR enzyme immunoassay (GPEIA) 14 was used as the standard comparator test to assess noninferior clinical accuracy of HR-HPV testing with OncoPredict HPV QT.LMNX Genotyping Kit GP HR (LMNX; Diassay BV, Rijswijk, the Netherlands), subsequently referred to as "LMNX Diassay," 15 was used to evaluate the genotyping capability of OncoPredict HPV QT.The overall HR-HPV prevalence in both the screening and disease-enriched populations according to baseline cytology and age was determined using Onco-Predict HPV QT and LMNX Diassay.Type-specific agreement was assessed using the kappa statistics for the 12 HR-HPV types resolved by the OncoPredict HPV QT assay.
Testing of VALGENT-2 samples with the GPEIA and LMNX Diassay was performed at DDL Diagnostic Laboratory, Rijswijk, The Netherlands from April to September 2013 16 (Figure 1).

| Reproducibility assessment
The intra-and interlaboratory reproducibility of the OncoPredict HPV QT was assessed on a subset of 526 samples, randomly selected from the original VALGENT-2 panel imposing 30% HR-HPV GPEIA positives samples, as indicated by the validation guidelines. 9The reproducibility panel included 157 HR-HPV positive and 369 HR-HPV negative samples, as indicated in Figure 1.
Set-up of OncoPredict HPV QT assay composed by quality control (QC) for the evaluation of sample adequacy and 4 Quantitative modules (QT1 to QT4) for the assessment of type-specific viral loads.
The remaining extracted DNA following initial testing was frozen and subsequently thawed to perform reproducibility analysis.Intralaboratory reproducibility of the OncoPredict HPV QT testing was assessed by repeated testing at UniMiB, whereas the interlaboratory reproducibility assessment was performed by repeating the testing at Parco Tecnologico Padano (PTP) Laboratory, Lodi, Italy.

| Clinical outcomes
The British Society for Clinical Cytopathology (BSCC) terminology and the cervical intraepithelial neoplasia (CIN) nomenclature were used for reporting cytology and histology results, respectively. 16,17Women with abnormal cervical cytology were managed according to the Scottish guideline criteria.In particular, women with abnormal cytological results were referred to colposcopy and colposcopy-directed biopsies were taken when clinically indicated according to local protocols.HPV test results did not influence patients' management.
The clinical sensitivity of OncoPredict HPV QT assay was based on test performance for the detection of histologically confirmed diagnosis of cervical intraepithelial neoplasia (CIN) grade 2 or worse (CIN2+) within 18 months of sample collection.The clinical specificity was determined using samples from women who had two consecutive negative cytology samples across two screening rounds (control group).Clinical performance of the index test was assessed for women irrespective of age as well as separately for women ≥30 years of age.

| Statistical analyses
The absolute clinical sensitivity for CIN2+ and specificity for ≤CIN1 of the OncoPredict HPV QT were calculated with 95% CI for women irrespective of age and in women aged 30 years and older.Noninferior clinical accuracy of the index test compared to the standard comparator test was assessed as proposed by Tang 18 applying the benchmarks of 0.90 for relative sensitivity and 0.98 for relative specificity.

Differences in accuracy between index and comparator tests
were also assessed by 95% confidence intervals (95% CIs) around the relative sensitivity and specificity taking the paired design into The reproducibility was expressed as the overall percentage agreement, which is the proportion of the number of concordant results (positive on both assays + negative on both assays) overall test results.The reproducibility validation criterion was considered as fulfilled when the left 95% CI bound for HR-HPV concordance exceeded 87% and the kappa >0.5. 9e analytical concordance between OncoPredict HPV QT and LMNX Diassay was assessed using Kappa statistics 19

| Cytopathological findings, adequacy of specimens for HPV testing
The cytopathological findings in the screening population were as follows: 89.8% were cytology negative; 5.4% had borderline nuclear changes; 3.8% low-grade dyskaryosis; 1% moderate to severe (highgrade) dyskaryosis.The disease-enriched population was composed by design 13  Overall, there were a total of 95 women with CIN2+ (denominator for clinical sensitivity); of these 50 women had CIN3+ or worse.A total of 720 women had two consecutive cytology negative results (denominator for clinical specificity).

| Clinical accuracy of OncoPredict HPV QT
Table 1 provides a summary of the data on clinical sensitivity and specificity for both assays as well as relative sensitivity and specificity for women of all ages and for women >30 years of age.
In the total study population, OncoPredict HPV QT showed a relative sensitivity for CIN2+ of 1.01 (95% CI: 0.99-1.03)and a relative specificity for ≤CIN1 of 1.03 (95% CI: 1.00-1.06)compared to GPEIA.The sensitivity for CIN2+ and specificity for ≤CIN1 of OncoPredict HPV QT was noninferior to GPEIA (P = <0.0001for both).Similar results were found when restricting the analysis to women aged 30 years and older.

| Intra-and interlaboratory reproducibility of OncoPredict HPV QT assay
The intra-and interlaboratory reproducibility results are reported in Table 2.Among the subset of 526 samples, included in the reproducibility panel, 3 samples were excluded because they were invalid on at least of one retesting.Out of the 523 valid samples, 519 were concordant over the two intralaboratory runs, while 516 concordant results were obtained on repeating testing at PTP.The intra-and interlaboratory reproducibility was 99.2% (95% CI: 98.1-99.8;k = 0.98) and 98.7% (95% CI: 97.3-99.5;k = 0.96), respectively.
Reproducibility was also high at type-specific level with general concordance >99% and kappa values ranging from 0.66 to 1.00, as reported in Table 3.
Values of genotype-specific viral load (copies/10 4 cells) are reported in the Table S3.

| DISCUSSION
Worldwide cervical cancer screening programs are transitioning from cytology to primary HPV testing resulting in molecular HPV assays becoming increasingly available commercially; however only a relatively small number of them have been assessed according to international criteria designed to validate HPV DNA tests for use in cervical cancer screening. 10,21is clinical performance study demonstrated the noninferior accuracy of OncoPredict HPV QT compared to GPEIA for the detection of cervical precancer.The assay also showed excellent intra-and interlaboratory reproducibility of qualitative HPV detection.
Furthermore, OncoPredict HPV QT was evaluated as part of the 2019 HPV Labnet international proficiency study, which reported it to be 100% proficient in the detection of the targeted HR-HPV types. 22coPredict HPV QT assay is a novel quantitative full-genotyping assay targeting E6/E7 viral genes, allowing normalized genotypespecific viral load determination of 12 HR-HPVs defined as 1A carcinogens by the International Agency for Research on Cancer. 23A key and unique feature of this assay is the extent of quality controls included to assess sample adequacy, efficiency of nucleic acid extraction and amplification.Sample adequacy is evaluated through the quantitative assessment of a single-copy human gene, CCR5, located on chromosome 3, allowing the number of human cells present in the clinical sample to be determined as well as allowing the accurate normalization of viral loads, as previously described. 24,25In particular, CCR5 is not a pseudogene, as in the case of several housekeeping genes used as internal controls in nucleic acid amplification tests (NAATs) which have variable copy numbers in the human genome.[29][30] Studies are ongoing to evaluate the potential value of OncoPredict HPV QT in the molecular triage of HPV-positive women following primary HPV-based screening.Moreover, HR-HPV genotype-specific viral load is expected to be a useful marker in the post-treatment management of women with high-grade dysplasia (test-of-cure).In the present study the genotyping performance of OncoPredict HPV QT assay has been compared to that of GP5+/6+ PCR EIA, according to the current validation guidelines. 9The latter has been replaced operationally by newer assays in HPV-based screening programmes 21 and as a result guidelines are in the process of being updated to include alternative potential comparators in addition to HC2 and GP5+/6+ PCR EIA.In case of multiple infections, the HPV genotype with the highest viral load was defined as the predominant HPV-type and was used to assess OncoPredict HPV QT positivity, as previously reported for RIATOL qPCR assay. 26,31Moreover, validation of RIATOL qPCR assay showed similar results in terms of sensitivity and specificity to detect CIN2+ when considering either the cumulative hrHPV viral loads or the genotype with the highest viral load. 31ijer Guidelines 9 suggest that clinical validation of new candidate HPV tests for their use in cervical cancer screening should be performed on a selected panel of CIN2+ cases and <CIN2 controls from women of at least 30 years of age.However, in the current validation study, conducted as part of the VALGENT-2 framework, 13 cervical samples were collected in 2012 in Scotland, when at the time, routine screening was offered to women from 20 to 60 years of age; representing a "real-life" screening population.We also note that samples collected in VALGENT-1 and VALGENT-4 framework came from an HPV-primary screening setting, while those collected during VALGENT-2 and VALGENT-3 from a cytology-based program; however, a multivariate analysis confirmed that primary screening modality did not affect validation outcomes; this aids in extrapolating the findings to contemporary settings. 21ile the present study focused on clinician-taken samples, there is an increasing global practice and appetite to apply HPV tests to self-taken samples.In the future, there will be a greater onus on demonstrating tests are validated for both clinician-and self-collected specimens. 32To this end, accuracy of type-specific viral load determination with OncoPredict HPV QT on self-collected as compared to clinician-collected specimens is currently being addressed in the VAL-HUDES framework. 33 conclusion, OncoPredict HPV QT is a novel full-genotyping HR-HPV test with the unique features of allowing genotype-specific normalized viral load determination as well as independent quality controls to assess sample adequacy and preanalytical/analytical processing.The present study demonstrated noninferior accuracy to detect cervical precancer compared to GPEIA in addition to robust intra-and interlaboratory reproducibility.OncoPredict HPV QT, therefore, fulfills all three international validation criteria for HPV tests suitable for primary cervical cancer screening.
However, in brief, the population included 1300 samples comprising of 1000 consecutive samples collected from the routinely screened population (screening population) enriched with 300 cytologically abnormal samples (disease-enriched population).All liquid-based cytology (LBC) samples resuspended in PreservCyt liquid medium (Hologic, Bedford, MA) were aliquoted and stored at À80 C in the Scottish HPV Archive.A separate aliquot of the original LBC specimen was defrosted only once before its use for the validation of an HPV DNA assay as part of the VALGENT-2 Framework.When considering the total population (screening and disease-enriched) women's median age was 38 years (ranging from 19 to 68 years).

Figure 1 2 . 2 |
Figure 1 describes the complete flowchart of sample collection and testing procedure used in this validation study.
account and by McNemar (McN) tests.
of 100 samples with borderline nuclear changes, 100 with low-grade dyskaryosis and 100 with high-grade dyskaryosis.Out of the total 1300 samples comprising the VALGENT-2 panel, 1286 were available for the purpose of this clinical performance study.Out of the 1286 available samples, a further 46 samples (25 from the screening and 21 from the disease-enriched population) had to be excluded from the analysis due to insufficient starting volume (17 samples' aliquots contained <400 μL volume required for nucleic acid extraction and testing according to OncoPredict HPV QT protocol) or to "invalid result" due to low sample cellularity or reduced nucleic acid extraction efficiency (6 samples with <400 cells/reaction; 23 with nucleic acid recovery <10%).Testing of the VALGENT-2 panel with GPEIA at DDL Diagnostic Laboratory yielded two samples that were previously excluded due to operational issues.The final number of matched samples where valid results were available for both GPEIA and OncoPredict HPV QT were n = 1239.
which could compromise the accuracy of viral load determination.The efficiency of nucleic acid recovery, through an external control introduced in the sample before nucleic acid extraction, provides information on the quality of this important preanalytical step.It also supports evaluation, validation and comparison of the assay with different extraction methods/protocols.Finally, an external amplification control is included in each of the five separate real-time PCR reaction wells, part of the analytical step, allowing potential PCR inhibition and/or reagent failures to be discerned.The OncoPredict HPV QT assay has been CE marked for its use in full automation, combining high-throughput automated DNA extraction and amplification/detection by real-time PCR, as well as for a manual preanalytical and analytical set-up, thus allowing the use of the test both in large scale screening and in smaller molecular diagnostics laboratories.
Absolute and (b) relative sensitivity for CIN2+ and CIN3+ and specificity for ≤CIN1 of OncoPredict HPV QT compared to GPEIA. a a Positivity for the detection of 12 HR-HPV types.b Women with two consecutive negative cytology results.c P for the McNemar test for a difference between matched proportions; P values of >.05 indicate that the sensitivity or specificity of the OncoPredict HPV QT assay is not significantly different from that of GPEIA.d P for the test of noninferiority; P values of <.05 indicate that the sensitivity or specificity of the OncoPredict HPV QT assay is not significantly lower than that of GPEIA.T A B L E 2 Intra-(a) and inter-(b) laboratory reproducibility of HR-HPV testing with the OncoPredict HPV QT assay evaluated by the laboratories of UniMiB and PTP.
T A B L E 3 Intra-and interlaboratory reproducibility of HR-HPV and type-specific HPV testing with OncoPredict HPV QT evaluated by the laboratories of UniMiB and PTP.Concordance of OncoPredict HPV QT and LMNX Diassay results in the total population for types individually resolved by both assays.