Feasibility and Accuracy of Menstrual Blood Testing for High-risk Human Papillomavirus Detection With Capture Sequencing

IMPORTANCE High-risk human papillomavirus (hrHPV) persistent infection is the major etiology of cervical precancer and cancer. Noninvasive self-sampling HPV testing is a promising alternative cervical cancer screening for avoiding stigma and improving patient willingness to participate. OBJECTIVE To investigate the feasibility and accuracy of menstrual blood (MB) hrHPV capture sequencing in hrHPV detection. 137 from 120 by performed the criterion standard for detecting hrHPV genotypes among enrolled women. Data analyzed from April 1 through June 2021. diagnosis for comparison of MB hrHPV–positive rates on different MCDs. When sample sizes were too small to support Pearson χ 2 test, Fisher exact test was used. We used κ test to analyze concordance of HPV detection methods. Sanger sequencing was used as the criterion standard for hrHPV genotypes.


Introduction
Persistent infection of high-risk human papillomavirus (hrHPV) is associated with various HPV-related precancers and cancers. 1 In recent years, HPV testing has gradually become the primary method for cervical cancer screening. 2 Although cervical cancer screening has been found to be helpful in decreasing the incidence and mortality rate of cervical cancer, various factors (including medical infrastructure, culture and mentality, and society) may influence women's acceptance of clinician sampling. 3 Among women with overdue screening, 29% were afraid of the stigma and 14% were had fear of pain in 2 studies. 4,5 Self-sampling HPV testing is a proposed alternative cervical cancer screening for avoiding stigma and improving participation. 6 However, to our knowledge, most existing self-sampling HPV studies were based on various sampling brushes inserted into the vagina, and patients may experience discomfort during sampling. [7][8][9][10][11] Compared with these methods, menstrual blood (MB) collection is associated with less stigma and pain.
MB offers a snapshot of cervical HPV infection status. Given that it is a biological fluid with periodic expulsion, MB is easy to collect. This suggests that HPV testing based on sanitary pads may be a convenient and noninvasive approach. More importantly, next-generation sequencing (NGS) for HPV detection, emerging as a highly sensitive method for HPV genotyping, has not yet been applied to MB HPV testing, to our knowledge. This study was designed to investigate whether MB hrHPV capture sequencing may be a feasible and accurate approach to detecting hrHPV infection among women who are premenopausal.

Methods
Ethics approval was obtained from the institutional review board of the Central Hospital of Wuhan.
All patients agreed to participate and provided signed informed consent before enrollment. This study is reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Study Population
From September 1, 2020, to April 1, 2021, we recruited women who were premenopausal and had tested positive for hrHPV in the Central Hospital of Wuhan, China. Cervical smears were collected from recruited patients by an experienced gynecologist (C.C.). The 21 HPV GenoArray Diagnostic Kit (Hybribio) was used to detect hrHPV infection in cervical smears. This kit has been used in many cervical screening studies, with approval and recommendation from the China National Medical Products Administration for HPV testing. [12][13][14] Residual cervical smears after HPV testing were collected and preserved at 4°C until further Sanger sequencing.
We excluded recruited patients who refused to participate in this study, had no menstruation within 3 months of hrHPV testing, or had a total menstrual bloodstain area on a sanitary pad of less than 3 cm × 3 cm (Figure 1). Enrolled patients were treated based on the 2019 American Society for Colposcopy and Cervical Pathology (ASCCP) consensus guidelines. 15 The diagnosis of cervical intraepithelial neoplasia (CIN) and cervical cancer were confirmed by the pathology of cervical biopsies or excisions.

Menstrual Sanitary Pads Collection
Self-provided commercial sanitary pads were used to collect MB from enrolled patients. The discarded sanitary pads were preserved in sterile sealed plastic bags that we distributed to enrolled patients in advance; pads were then placed in the refrigerator at −20°C within 1 hour after being removed and stored until further testing. The menstrual cycle day (MCD) was recorded according to patient self-report. Patients were encouraged to provide multiple sanitary pads for different MCDs of their current menstrual cycle.

MB DNA Extraction
A commercial Tiangen dried blood spot DNA extraction kit (Tiangen Biotech) was used to extract MB DNA from sanitary pads according to the manufacturer's protocol. The concentration of DNA was confirmed using the Qubit 4.0 fluorometer and Qubit dsDNA high-sensitivity (HS) assay kit (Invitrogen by Thermo Fisher Scientific), and the purity of DNA was determined by NanoDrop One spectrophotometer (Thermo Fisher Scientific). Extracted MB DNA was stored at −20°C.   products was confirmed using Qubit 4.0 fluorometer and Qubit dsDNA HS assay kit (Invitrogen); the size of PCR products was confirmed using Qsep100 bioanalyzer (BIOptic). Products in the final library ranged from 270 base pairs to 320 base pairs in length. Sequencing was performed using Illumina NovaSeq PE150.

HPV Genotyping and Validation
Our previous HPV detection reports 16,17 provided details on the HPV genotyping method. Briefly, after filtering of low-quality reads and duplicated reads, all clean reads were aligned to the merged 15 hrHPVs whole-genome sequence. HPV-positive genotypes were then identified with mapping coverage of 50% or more and effective mean depth of ×30 or more. Next, we compared genotyping results of MB hrHPV capture sequencing with those of conventional HPV testing. Here, we defined the degree of detection concordance between MB hrHPV capture sequencing and conventional HPV testing as follows: if there was at least 1 overlapping hrHPV genotype between testing methods among all genotypes found in 1 patient's sample, it was defined as concordance. If the 2 were identical, it was defined as complete concordance; otherwise, it was defined as incomplete concordance. If there was no overlapping hrHPV genotype between testing methods, it was defined as discordance. Sanger sequencing of cervical smears was performed as the criterion standard for hrHPV genotypes (Figure 1).

Statistical Analysis
Based on the results of previous studies, we estimated that MB HPV testing would yield 83% sensitivity and 98% specificity. 18 Using a 10% allowable error, a 5% α error, and an estimated 90% prevalence of hrHPV infection among the targeted population, the minimum sample size was calculated to be 90 patients.
SPSS statistical software version 25.0 (IBM) was used for statistical analysis. Normal distribution data were shown as mean (SD). P value < .05 was considered statistically significant, and P values were 2-sided. Pearson χ 2 test was used for comparison of the distributions of hrHPV genotypes detected by MB hrHPV capture sequencing and HPV testing. We also used Pearson χ 2 test for comparison of MB hrHPV-positive rates on different MCDs. When sample sizes were too small to support Pearson χ 2 test, Fisher exact test was used. We used κ test to analyze concordance of HPV detection methods. Sanger sequencing was used as the criterion standard for hrHPV genotypes.   Figure 2B). Distributions of hrHPV genotypes between detection methods were not statistically significantly different (P = .85) ( Figure 2C).   (Figure 3).

Results of MB hrHPV Capture Sequencing
There was no statistically significant difference in MB hrHPV-positive rate between patients providing sanitary pads from multiple MCDs and patients providing sanitary pads from 1 MCD (100 of 106 patients [94.3%]) (P = .71).
A total of 24 sanitary pads had incomplete concordance or discordance for the detection methods ( Table 1). Among individuals who provided these pads, 5 patients (20.8%) were correctly detected by conventional HPV testing and 18 patients (75.0%) were correctly detected by MB hrHPV capture sequencing. In 1 of these patients (4.2%), neither detection method correctly detected the genotype.
According to results of Sanger sequencing, MB hrHPV capture sequencing successfully  (Table 1).
Overall, MB hrHPV capture sequencing produced 126 true positive events, 3 false positive events, 3 false negative events, and 5 true negative events. The method had a sensitivity of 97.7% (95% CI, 95.0%-100%) ( Table 2). Sanitary pads of 14 patients were collected at multiple MCDs. There were 12 patients who tested positive for menstrual blood high-risk human papillomavirus (MB hrHPV) for each MCD on which they were tested, while patient 5 tested positive for MB hrHPV at MCD 1 but not MCD 2 and patient 14 tested positive for MB hrHPV at MCD 3 but not MCD 2.

Discussion
According to the ASCCP consensus guidelines, it is recommended that women aged 25 to 65 years undergo cervical cancer screening; this population largely overlaps with the population of women at childbearing age. 15 Menstruation is a physiological bleeding process during childbearing age.  infections, identification of true-negative samples, and identification of the real hrHPV genotypes among those misidentified by conventional HPV testing. These results were all validated by a third testing method, Sanger sequencing.
Our study also found that MB hrHPV-positive rates on different MCDs, from MCD 1 to MCD 5, were equivalent. Additionally, MB hrHPV-positive rates were similar among patients who provided 1 sanitary pad and those who provided more than 1 sanitary pad. These findings suggest that selfcollected MB at any day of the menstrual cycle is acceptable, which also suggests the convenience of MB hrHPV capture sequencing. However, we recommend using 1 sanitary pad from MCD 2 for MB hrHPV capture sequencing because of this day usually has a larger amount of bleeding.

Limitations
There are several limitations to this study. First, self-collection of MB could be performed only during the menstrual period; thus, self-sampling time was subject to certain restrictions. Second, there were slight differences in the materials used to make the sanitary pads. We found that certain materials may affect DNA extraction. Third, the proportion of women who did not have hrHPV and who had CIN or cervical cancer was relatively low in our study. Expanding the sample size and further exploring the value of MB hrHPV capture sequencing in the detection of women who do not have hrHPV and have CIN or cervical cancer is needed in future studies.

Conclusions
This study's findings suggest that sanitary pads with MB may serve as medical devices for HPV detection. With the obvious convenience of this method, MB hrHPV capture sequencing may be associated with decreased stigma of gynecological examination and increased participation in cervical cancer screening. MB hrHPV capture sequencing was associated with superior performance compared with cervical HPV testing not only in detection of additional hrHPV infections and multiple hrHPV infections, but also in identification of true negative events and hrHPV genotypes misidentified by conventional HPV testing. These results suggest that MB hrHPV capture sequencing may be a feasible and accurate self-collected approach for cervical cancer screening.