Abstract
The introduction of self-sampling in cervical cancer screening has raised the importance of HPV testvalidation on self-collected samples. This study aimed to evaluate the clinical performance of the OncoPredict HPV Screening (SCR) assay on self-collected vaginal and first-void urine (FVU) as part of the VALHUDES framework.
Vaginal (FLOQSwabs) and FVU (Colli-Pee) samples were self-collected by 500 women referred to colposcopy, followed by a clinician-collected cervical sample prior to colposcopy, which were all tested using OncoPredict HPV SCR.
OncoPredict HPV SCR demonstrated similar relative clinical sensitivity to detect cervical intraepithelial neoplasia grade 2 or worse (≥ CIN2) in urine (ratio: 0.95 [95%CI 0.88-1.02]) and vaginal self-samples (ratio: 0.96 [95%CI 0.90-1.02]) compared to cervical samples. The clinical specificity was lower in vaginal but not in urine samples compared to cervical which improved following cut-off optimization. A higher cellularity was found in vaginal as compared to cervical and FVU samples. Moderate to excellent agreement in HPV detection in self-collected samples and cervical scrapes was demonstrated (Kappa values: 0.53 to 1.00).
OncoPredict HPV SCR assay demonstrated similar accuracy on self-collected vaginal and FVU samples compared to cervical samples, although cut-off adjustment improved clinical specificity when applied to vaginal samples.
Key words (3 to 6): cervical cancer, self-sampling, diagnostic accuracy study, Human Papillomavirus (HPV), OncoPredict HPV SCR, European VALHUDES
Introduction
As proposed in the World Health Organization (WHO) call to action, screening of 70% of women a key target to support the elimination of cervical cancer by 2030 (1).
Human Papillomavirus (HPV) testing has been demonstrated to be more effective than cytology for secondary prevention of cervical cancer (2,3) and is therefore recommended as a primary screening tool in current screening algorithms (4). Additionally, meta-analyses have shown that the clinical accuracy of PCR-based HPV tests on self-samples is similar to that of cervical samples (5,6).
In 2021, 48 countries recommended primary HPV-based screening and 17 introduced self-sampling in their national programs or guidelines (7) as response to WHO Call to Action (1).
While several HPV tests are currently validated for use in cervical cancer screening (8), only a few are formally validated for use on self-collected specimens. The VALidation of HUman papillomavirus assays and collection Devices for Self-samples and urine samples (VALHUDES) Framework has defined a standardized protocol to assess the clinical performance of HPV tests in combination with self-collection devices (9). Results of a first installment of VALHUDES demonstrated similar accuracy of first void urine (FVU) collected with the Colli-Pee device and vaginal specimens compared to clinician-collected cervical samples using different HPV assays (10–15). We now report on a second iteration of VALHUDES undertaken in a different geographic setting and utilizing a different approach to vaginal sampling.
This present study aimed to evaluate the clinical performance of the OncoPredict HPV Screening (SCR) assay on vaginal self-samples collected with FLOQSwab resuspended in 5 ml eNat and FVU with Colli-Pee as compared to clinician-collected cervical scrapes to detect high-grade cervical lesions. Secondarily, we investigated the analytical performance of the assay and evaluated the adequacy of self-collected samples.
Material and Methods
2.1 Study design
Within the European VALHUDES Framework, 600 women, referred to colposcopy following a previous cervical abnormality or HPV positivity, were enrolled between July 2020 and February 2022 in four colposcopy centers (NHS Lothian, Edinburgh; ASST degli Spedali Civili di Brescia, Brescia, Italy; European Institute of Oncology IRCCS, Milan, Italy; U.O. Coordinamento Consultori Familiari, ASSL Sassari - ATS Sardegna, Sassari, Italy). Exclusion criteria have already been described (Latsuzbaia et al., submitted).
All women were asked to collect a urine sample followed by a vaginal specimen. FVU was collected using Colli-Pee FV5000 (Novosanis, Wijnegem, Belgium). The device captures approximately 13 ml of FVU that are mixed with 7 ml nucleic acid preservative included within the collection device. Vaginal self-collection was performed using a FLOQSwab (Copan Italia Spa, Brescia, Italy). During gynecological examination, a cervical specimen was collected by a clinician with Cervex-Brush (Rovers Medical Devices, Oss, The Netherlands) and immediately transferred in 20 ml PreservCyt (Hologic Inc., Bedford, Massachusetts, USA).
All women underwent colposcopy and biopsy was performed if clinically required. The histological result of the biopsy was used to determine the disease outcome.
Self-collected vaginal samples were transported dry to the laboratory together with the 20 ml PreservCyt vial containing cervical samples and the Colli-Pee tube containing FVU. All specimens were transported to the laboratories affiliated with the enrolling colposcopy centers. After arrival in the laboratories, cervical and FVU specimens were shaken for 30 seconds and divided into 1.5 ml aliquots. The dry vaginal swabs were resuspended in 5 ml of PreservCyt (Hologic Inc., Bedford, Massachusetts, USA) or 5 ml eNat (Copan Italia Spa, Brescia, Italy). Vaginal samples were further aliquoted into 0.4 ml volumes. All aliquots were stored at -20 °C until transferred to MIRRI-IT Biobank of the University of Milano-Bicocca where they were stored at -80 °C. Results reported in this manuscript are from the 500 women whose vaginal swabs were resuspended in eNat.
2.2 HPV testing
Testing of all specimens was performed at the Laboratory Clinical Microbiology and Virology, School of Medicine and Surgery, University of Milano-Bicocca (Monza, Italy). Nucleic acid extraction was performed using a Fluent 480 (Tecan, Männedorf, Switzerland) automated platform with Quick -DNA/RNA MagBead (Zymo, USA) starting from 400 µl of sample. Fluent 480 workstation was also used to set-up the real-time PCR plate of OncoPredict HPV SCR assay (Hiantis, Milan, Italy) according to manufacturer instructions with 10 µl of mastermix and 5 µl of sample’s DNA extract. The OncoPredict HPV SCR assay, previously validated for testing on cervical scrapes in a screening setting (16), is a partial genotyping assay targeting E6 and E7 DNA sequences of 13 high-risk Human Papillomavirus (hrHPV) types (HPV -16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, and -68). The test is composed of two separate real-time PCR reactions. A quality control tube (QC) well allows the assessment of nucleic acid extraction recovery with an exogenous control gene target added to the sample before preanalytical processing, additionally the QC tube determines sample adequacy in terms of human cellularity by the quantification ofC-C Motif Chemokine Receptor 5 (CCR5) gene. A second reaction well is used to assess the presence of HPV-16, HPV-18 individually and the 11 other hrHPV types as a pool. Both wells contain an amplification control to evaluate the potential PCR inhibition. The PCR was carried out using a CFX384 Touch Real-Time PCR Detection System (Bio-Rad, USA). All results were considered valid if HPV positive. In the case of HPV negative result(s) samples were defined as inadequate if i) the extraction efficiency was below 10%; ii) less than 400 cells/reaction in cervical samples (16) and 150 cells/reaction in urine and vaginal samples were detected and iii) there was PCR inhibition in any of the two reaction wells.
2.3 Statistical analysis
Clinical sensitivity was estimated for cervical intraepithelial neoplasia grade 2 or worse (≥CIN2) and for cervical intraepithelial neoplasia grade 3 or worse (≥CIN3). Specificity was estimated for <CIN2 outcome or accepting negative colposcopy as clinical endpoint when the gynecologist did not take a biopsy. We used McNemar tests to evaluate the accuracy differences between index and comparator tests with statistical significance accepted if p-values <0.05 or when the 95% confidence intervals excluded 1. Cohen’s kappa was employed to assess HPV test concordance between self- and clinician taken samples for the entire study population and according to disease status among specimens and categorized as: poor (0.00–0.19), fair (0.20–0.39), moderate (0.40–0.59), good (0.60–0.79), and excellent (0.80–1.00). Mann-Whitney test was used to evaluate differences in median Ct-values and median number of cells/reaction. All statistical analyses were conducted using Stata 16.1 (Statacorp, College Station, TX, USA).
2.4 Ethical Approval
The European VALHUDES study (ClincalTrail.gov: NCT04312737) was conducted in accordance with the Declaration of Helsinki and approved by the central Ethics Committee of the Coordinating Centre, ASST degli Spedali Civili di Brescia, Brescia, Italy (Ethics approval number: NP 3879- Studio WP6-HPVONC) on the 16th of July 2020, subsequently by the local Ethics Committees of the other participating centres. All women signed a written informed consent form before to enrolment.
Consorzio Italiano per la Ricerca in Medicina (C.I.R.M.), Milano, Italy, performed on site and remote monitoring of the study conduction, as previously described (Latsuzbaia et al., submitted).
Results
3.1 Study population
490 out of the 500 women were included in the study as reported in Figure 1. The median age of the women included in the study was 37 years (IQR: 31-47 years, range: 25-64 years), with median age of women with ≥CIN2 lesions being significantly lower of those with <CIN2 result, as previously described (Latsuzbaia et al., submitted). 489 women had colposcopy with the following outcomes: 134 (27.4%) negative, 245 (50.1%) minor colposcopy findings, 104 (21.3%) major colposcopy and 6 (1.2%) suspicion of cancer. 55% (271/490) of women underwent biopsy and diagnosis of ≥CIN2 was confirmed in 41.3% (112/271) of cases. Table 1 reports the characteristics of the study population by age group and colposcopy center. 28 cervical specimens, 13 vaginal swabs and 19 FVU samples were excluded from the analysis because they were inadequate and HPV-negative (Figure 1).