Established international validation guidelines (2009) contain performance criteria against which HPV tests, destined for application in cervical screening contexts, can be adjudicated and said criteria have made an important contribution to the practice and culture of HPV assay validation.¹ However, the guidelines do not include a piece for validation of the genotyping element of assays although this aspect is likely to have increasing importance.

Scotland, has a national, organised cervical screen- ing programme, associated with an uptake of around 70%. In line with the rest of the United Kingdom, women are screened every three years between the ages of 25-50 and every five years between the ages of 50-65.² HPV testing as a test of cure of treatment has been in place national- ly since 2012 and a key, planned development relates to the implementation of HPV-primary screening (to replace Pap) in 2019. The proposed algorithm is that all High Risk (HR)-HPV positive women will be triaged to cytology. However, incumbent on the programme is the assessment and consideration of emerging HPV assays, including those which may have a typing component, to determine how these could add value and efficiency to screening protocols. To this end, participation in international endeavours such as VALGENT designed to determine the performance of clinical HPV tests is apposite.3 Furthermore, the organised nature of screening in Scotland and the existence of a central IT system, the Scottish Cervical Call Recall System (SCCRs) which contains a woman’s entire screening record (including information on all management, recall, laboratory results, clinical results and vaccination status) facilitates population-based research and service evaluations designed to further refine and evolve cervical screening.

Additionally, the use of genotyping assays has been essential in determining early measures of the impact of the prophylactic HPV vaccines. Scotland has delivered a national HPV vac- cine programme since 2008 associated with high uptake (~90%) in the target group. Further- more, as women were initially screened aged 20 until 2016 in Scotland it has been possible to perform longitudinal surveillance assessment of HPV prevalence in successive birth cohorts, including those offered the vaccine. Notably, HPV16 and 18 prevalence reduced from 30.0% (95% Confidence Interval-CI: 26.9, 33.1%) in females born in 1988, who represented an unvaccinated baseline cohort, to 4.5% (95% CI: 3.5, 5.7%) in females born in 1995 cohort of whom around 90% were vaccinated. The 4.5% prevalence in the vaccinated 1995 cohort was observed using a genotyping assay with a high analytical sensitivity - the Optiplex HPV Genotyping Test, (Diamex, Heidelberg, Germany) and in fact, when a clinically validated assay (the RT HPV Test, Abbott Molecular, US) with genotyping capability was applied to samples from the 1995 cohort, HPV16/18 prevalence was 0.5%. These findings emphasise the influ- ence that assay choice can exert on observed prevalence and underline the differences between assays calibrated to detect cervical intraepithelial neoplasia grade 2 or worse (CIN2+) rather than a low amount of virus within a sample. In keeping with the theme of clinically relevant genotyping; one of the objectives of VALGENT is to gather data on type specific performance of assays to inform the creation of minimal re- quirements/characteristics of genotyping tests that can be applied to cervical screening.

As described elsewhere in this HPV World issue, the VALGENT endeavour is iterative; essential- ly, a “host/hub” site collates cervical samples, of which several aliquots are made and distributed to collaborating laboratories.³ From iteration 2 onwards the emphasis has been geared towards assays which would be suitable for service laboratories and screening/clinical applications, where- as VALGENT 1 also included assays with high analytical sensitivity designed for epidemiological use. In the context of VALGENT 2, which used samples from the Scottish Cervical Screening Population, one limited (Xpert HPV), one extended (Onclarity) and two full genotyping assays (Papil- loCheck and LMNX Genotyping Kit GP HR) were evaluated and pathology information obtained, with due process of governance, using the national IT system: SCCRs (described above). Analysis focused on the 14 HR-HPV types in common to all assays and clinical performance of these tests in terms of relative sensitivity and specificity for CIN2+ compared to the GP5+/6+ PCR-EIA is summarised in Table 1. VALGENT 2 also differed from VALGENT 1 in that all samples were collected in PreservCyt rather than SurePath. The HPV results obtained as a consequence of VALGENT 2 did not affect patient management nor was HPV primary screening in place at the time of sample collection so the preface/work-up which allowed case and control definition was driven entirely by cytology. Again, this differed from VALGENT 1 where HPV status at primary screen was known. Having the different biospecimen types reflected/represented in the VALGENT iterations, in addition to the different work-up strategies to identify disease is of value as it represents the heterogeneity of practice in the “field”.

With respect to the genotyping element of VALGENT 2, type specific concordance between assays was also determined. One of the comparisons drawn from VALGENT 2 focused on LMNX Genotyping Kit GP HR vs Onclarity and is presented in Table 2a and 2b.

Kappa values when computed for the types which are individually resolved by both assays (16,18,31,45,51,52) were excellent to good (over 0.7) although differences in agreement were observed according to whether an infection was present as a single or with other types. Ultimately VALGENT will generate a matrix of type-specific positivity according to assay, stratified by underlying pathology. Such a matrix will support the determination of what is clinically relevant HPV typing information. Furthermore, by depicting the level/extent of assay-driven differences in type specific concordance, such a matrix will inform on HPV prevalence comparisons where different HPV genotyping assays have been applied.³

As the VALGENT projects move into their fourth iteration (see Poljak & Ostrebenk and Bonde et al in this issue) and involve an even greater number of assays a sizeable data set will accrue which will provide insights to the community. Given the pace of change/implementation relating to HPV testing, internationally such a data set is timely.

References

1. Meijer CJ, Berkhof J, Castle PE, et al. Guidelines for human papillomavirus DNA test requirements for primary cervical cancer screening in women 30 years and older. Int J Cancer 2009;124:516-20.
2. Kavanagh K, Pollock KG, Cuschieri K, et al. Changes in the prevalence of human papillomavirus following a national bivalent human papillomavirus vaccination programme in Scotland: a 7-year cross-sectional study. Lancet Infect Dis 2017;17(12):1293-302.
3. Arbyn M, Snijders PJ, Meijer CJLM, et al. Which high-risk HPV assays fulfil criteria for use in primary cervical cancer screening? Clin Microbiol Infect 2015;21:817-26.
4. Geraets DT, Cuschieri K, de Koning MN, et al. Clinical evaluation of a GP5+/6+-based luminex assay having full high-risk human papillomavirus genotyping capability and an internal control. J Clin Microbiol 2014;52:39964002.
5. Cuschieri K, Geraets DT, Moore C, et al. Clinical and Analytical Performance of the Onclarity HPV Assay Using the VALGENT Framework. J Clin Microbiol 2015;53:3272-9.
6. Cuschieri K, Geraets DT, Cuzick J, et al. Performance of a Cartridge-Based Assay for Detection of Clinically Significant Human Papillomavirus (HPV) Infection: Lessons from VALGENT (Validation of HPV Genotyping Tests). J Clin Microbiol 2016;54:2337-42.
7. Heard I, Cuschieri K, Geraets DT, et al. Clinical and analytical performance of the PapilloCheck HPV-Screen- ing assay using the VALGENT framework. J Clin Virol 2016;81:6-11.