The Utility of an Human Papillomavirus Genotype Assay for Cancer Screening in Self-Collected Urine and Vaginal Samples from Japanese Women

Most cervical carcinomas are associated with human papillomavirus (HPV) infection [1‒3], which is transmitted by sexual intercourse. Widespread cancer screening and HPV vaccination have been reported to contribute to a reduction in the incidence of cervical cancer (CCA) [4, 5], which is common among women of reproductive age. Moreover, the prevalence of CCA has decreased with the widespread use of the HPV vaccine. In 2005, the previous International Agency for Research on Cancer Handbook on CCA screening provided sufficient evidence that screening by conventional cytology reduced the incidence of and mortality associated with CCA based on cohort and case-control studies conducted in multiple countries [6]. In Japan, CCA screening using cytology was implemented as a national project in 1983, and similar scientific evidence has been obtained [7]. Currently, the significance of screening via the HPV test has been verified [8‒10]. In 2018, the US Preventive Services Task Force updated its screening guidelines to use high-risk HPV testing every 5 years for women aged 30–65 years. According to the 2019 ASCCP Risk-Based Management Consensus Guidelines, HPV tests play an important role in not only detecting CCA but also reducing the risk of lesions designated CIN3 or higher for surveillance following abnormal results [11].

However, the high incidence of invasive cancer among those who have not undergone CCA screening is considered a problem [12, 13]. In Japan, the morbidity and mortality rates among young women in their 20s and 30s are still high [14, 15]. One of the reasons is that the cancer screening rate is similarly low in young women. The Ministry of Health, Labor and Welfare of Japan recommends CCA screening for women aged 20 and over every 2 years. However, according to the 2019 Comprehensive Survey of Living Conditions by this ministry, the cancer screening rate for women aged 20–69 years was 43.7%, with particularly low rates in the 20–24 age group (15.1%) and the 25–29 age group (36.6%) compared with other age groups [16]. The mortality rates (per 100,000 population) of patients with CCA for those under 24 years of age and those aged 25–29 years were 0.1 and 0.5 in 2016, respectively, and 0.0 and 0.4 in 2019, respectively. For those aged 30–34, the rates were 1.4 in 2016 and 1.8 in 2019, indicating no significant decrease. Another reason is the suspension of the promotion of prophylactic HPV vaccination for young women through government policy. Specifically, the government effectively suspended the routine, publicly funded vaccination for the HPV vaccine for girls aged 12–16 years in June 2013 [17, 18].

An important challenge is to detect CCA and precursor lesions in women who do not receive medical care, such as those who do not undergo cancer screening or who do not receive the HPV vaccine, and this has been a worldwide challenge. In recent years, the usefulness of self-collected vaginal and urine HPV tests has been reported to solve this problem. The aim of this study was to investigate the usefulness of self-collected urine and a vaginal HPV test using the BD Onclarity^® assay (Onclarity assay) as a CCA screening tool in Japan.

In this study, we conducted a two-step verification analysis to investigate the usefulness of the urine HPV test. First, the appropriate container, temperature, and time for storage were verified until the urine HPV assay was performed in step 1. The Onclarity assay was used to detect 6 high-risk genotypes, namely, HPV16, 18, 45, 31, 51, and 52, and 3 groups (35 + 58, 39 + 68, and 56 + 59 + 66). Next, to verify the feasibility of the self-collected HPV test, the results of the urine analysis under those conditions, as well as the results of the cervical and vaginal HPV tests, were compared via the Onclarity assay for the patients who visited our hospital in step 2. The conditions used for each step are summarized in Table 1. Urine and vaginal samples were self-collected, and cervical samples were collected by gynecologists with a Cervex-Brush^® (Rovers Medical Devices). An Evalyn Brush^® (Rovers Medical Devices) was used for the self-collected vaginal samples. Urine and vaginal samples were first collected by the patients themselves on an outpatient basis before collection by a clinician. The participants were instructed to collect urine samples during their initial visit to the hospital after providing informed consent for the clinical trial. The collection was performed in the hospital’s restroom using a standard paper urine collection container. No specific instructions were given regarding clean catch urine (midstream) collection. After the urine was collected by the patient, the sample was promptly transferred by medical personnel into a sterile plastic container (Spitz) to minimize the risk of contamination.

The subjects were patients who underwent detailed inspection for an abnormal cervical smear, CIN, or CCA at Kagoshima University Hospital from May 2019 to December 2020. The primary aim of this study was to examine the feasibility of self-collected vaginal and urine HPV testing. Therefore, the concordance rate of cervical HPV detection with self-collected vaginal and urine samples was verified in a population with a high prevalence rate (over 50%), targeting a sample size of 100 patients. With respect to age, no strict restrictions were imposed, and patients who visited the hospital and provided written consent were consecutively enrolled. This study was approved by the Institutional Ethics Committee of Kagoshima University (Approval No. 25-161). A two-tailed p value <0.05 was considered to indicate statistical significance, and statistical analyses were performed with SAS for Windows (ver. 9.4, SAS Institute Inc., Cary, NC, USA) by a third-party organization.

The appropriate container, temperature, and time for storage until the urine HPV assay were determined to conduct a clinical trial in step 2, and validation of appropriate storage containers, duration, and temperature for HPV testing was performed via a mixed-effects logistic regression model.

A test container for urine samples suitable for the HPV test was selected. Each 2-mL sample of urine collected at each visit from 13 women was placed into four containers for storage and transportation with additive-free, EDTA, BD Sure Path^® dispersion liquid for liquid-based cytology (LBC), and the urine preservative transport reagent in the BD Probe Tec™ (QxUPT) kit within 30 min. The samples were stored at 4°C for 24 h, 72 h, or 1 week until the HPV assay. The HPV detection rates are shown in online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000541641). The urine samples stored in LBC containers had a significantly lower HPV detection rate (p = 0.011) than those stored without additives, whereas those stored in containers supplemented with EDTA and QxUPT had higher HPV detection rates. There was no difference in the sensitivity of HPV detection based on the storage time.

The storage temperature of the test container was determined. This temperature for the LBC containers was based on the results of step 1-1, whereas the storage temperatures of the other three test containers for 11 women were 4°C or at room temperature for 24 h, 72 h, or 1 week (data not shown). No difference in the HPV detection rate was observed according to either the use of 3 containers or the storage temperature. Regarding storage time, the detection rate was lower at 72 h (p = 0.0303) and 1 week (p = 0.0303) than at 24 h. Therefore, we decided to use a QxUPT kit specifically designed for urine tests.

The optimal time and temperature from urine collection to placement into a container and from storage to assay were determined (online suppl. STable 2). No significant differences in the HPV detection rate were observed according to either the time between urine collection and placement into the QxUPT (within 30 min vs. 24 h) or the storage temperature.

The optimal storage time and temperature were set after the urine was placed into the container until the HPV assay, assuming a midsummer environment. Three different temperatures (room, 37°C and 60°C) and three different times (24 h, 72 h, and 1 week) were compared. At 37°C for 24 h, 72 h, and 1 week, the HPV detection rates were 28.6, 42.9, and 14.3%, respectively, and at 60°C, these rates were 28.6, 42.9, and 14.3%, respectively. There were significant differences in the effects of temperature, storage time, and their interaction on the results of the analysis via generalized mixed-logistic regression analysis compared with those of the control (room, 24 h). Further detailed analysis revealed that there was no difference in temperature, but there was a significant difference in storage time at 72 h and 1 week compared with 24 h, and in terms of the interaction, there was a significant difference in storage conditions at 60°C for 1 week compared with the other conditions. Based on the above results, we decided to place the samples in a UPT storage container within 24 h of collection at room temperature and perform the HPV assay within 3 days thereafter in step 2.

The patient characteristics are shown in Table 2. This population included patients with CIN, CCA, and abnormal smears from the CCA screening. The median age was 43 (range: 20–80) years. Four (3.3%) women were pregnant. The percentage of patients with a history of CIN treatment, such as conization or laser vaporization, was 10.7%. The cervical smear by LBC was performed at the same time as the HPV test; NILM was found in 5 patients (4.1%), and the remaining findings were abnormal.

The HPV genotypes detected in the clinician-collected cervix, self-collected urine, and self-collected vaginal samples are shown in Figure 1. Patients with insufficient samples, unknown test results, or who were not tested were excluded from the analysis. The percentages of HPV types detected by the three collection methods were similar, with detection rates of 39.7% for 119 urine samples, 71.8% for 110 cervical samples, and 63.1% for 103 vaginal samples. HPV16 was detected most frequently, and HPV45 was not detected in any patient. Furthermore, when examining the concordance rates of the HPV genotypes detected in each case, the percentages of complete matches and partial matches between the cervix and urine were 80% and 20%, respectively. For the cervix and vagina, these rates were 87.1% and 12.9%, respectively. A complete match indicates that all detected genotypes were identical. A partial match means that not all detected genotypes were identical. For example, in 1 patient, HPV16 was detected in the cervix, whereas HPV 16 and HPV 33/58 were detected in the vagina. Among the three groups, there were eight discordant cases, and multiple HPV genotypes tended to be detected in the vaginal samples (online suppl. STable 3). There were no cases in which the genotypes of the three groups were completely discordant.

The concordance of self-collected urine and vaginal samples with clinician-collected cervix tests is shown in Table 3. The rate of concordance between the urine and cervical sample data was 63.9%, with a kappa coefficient of 0.34 (95% CI: 0.21–0.47), and that between the vaginal and cervical samples was 77.8%, with a coefficient of 0.68 (95% CI: 0.53–0.83), indicating good concordance. Logistic regression was performed with HPV positivity with the clinician-collected cervix as the objective variable and with only the HPV-positive results from urine and self-collected vaginal samples as factors (data not shown). The odds ratio of HPV in urine was significantly greater at 15.67 (95% CI: 3.49–70.28, p = 0.0003) and that in the vagina was 37.698 (95% CI: 10.818–131.36, p < 0.0001).

The final histological results are shown in Table 4. In total, 70.1% of the lesions were CIN2 or greater, 28.6% were invasive carcinomas, 1.7% were other malignancies, such as endometrial cancer, and 8% were nonmalignant (6.6%).

We divided the patients into two groups according to their histological diagnosis: presumed HPV-associated tumor/lesion or not. The former included CIN2/3, squamous cell carcinoma, in situ adenocarcinoma, and HPV-associated adenocarcinoma. The latter included HPV-independent adenocarcinomas, other cancers, and nondysplastic lesions. CIN1 is a quantitative assessment of dysplasia and is expected to include not only genital condyloma with low-risk HPV but also HSIL; however, the former excluded CIN1. The prevalence of HPV-associated tumors/lesions was 68.2% in the cervix, 70.6% in the urine, and 71.8% in the vagina.

The concordance between the HPV test results of clinician-collected cervical samples, self-collected urine samples, and self-collected vaginal samples and HPV-associated tumors/lesions is shown in Table 5. The absolute sensitivity, specificity, positive predictive value (PPV), and negative predictive value for detecting HPV-associated tumors/lesions in each HPV test are shown. The sensitivity values were 82.7, 46.4, and 75.7%, respectively. The PPVs for HPV-associated lesions were 78.5, 83.0, and 86.2%, and the negative predictive values were 58.1, 37.5, and 52.6%, respectively.

In step 1, we investigated the storage container, optimal time, and temperature for HPV testing, assuming that urine was collected at home and sent to a testing institution. Based on these results, the results of the HPV test using Onclarity assay in the urine, vagina, and cervical canal were obtained in step 2, suggesting that such testing would be useful.

To our knowledge, there have been some reports on HPV testing using urine, but there are few reports on tests conducted under changing conditions, such as temperature and time. We envision making urine HPV testing easier by using urine tests during health checkups or urine collection at home. Therefore, it was important to check the quality retention period and temperature of the samples from sample collection to transportation and the HPV assay. Based on these results, we believe that it is desirable to place the urine in a storage container within 24 h of collection and perform the HPV assay within 3 days at room temperature. In addition, the QxUPT kit, which is a storage container exclusively for normal urine, was the most suitable container for transporting urine.

In one study using the Cobas HPV test, all samples were kept at room temperature (18–24°C) [19], and in another study involving the Onclarity assay, all samples were kept cold in a cooler with ice packs within 10 min after collection, and all samples were stored at −20°C until the assay [20]. Alternatively, in some studies, samples were stored at −80°C after processing. Thus, the conditions were not consistent across studies [21].

In step 2 of our study, the rate of concordance between clinician-collected cervical HPV and self-collected urinary HPV was 63.9%, which was not bad, with a kappa coefficient of 0.34, and that between clinician-collected cervical HPV and self-collected vaginal HPV was 77.8%, which was good, with a kappa coefficient of 0.68. Several reports have shown that the concordance rate of vaginal and cervical HPV is high, with a kappa index of approximately 0.6 or higher [20, 22‒24]. It has also been reported that the concordance rates of urine HPV tests using hybrid capture II^® (HC-II) and Cervista™ are 38% and 37% (kappa: 0.22 and 0.26), respectively [25]. Moreover, the sensitivity (51%–92%) and specificity values (59%–98%) for high-risk HPV detection studies in urine compared with those in the cervix were reported in a systematic review and meta-analysis [26, 27].

There are several reports regarding the relative clinical sensitivity and specificity (for CIN2 or more) of HPV testing in self-collected urine/vaginas with respect to clinician-collected cervical samples in each study. In some reports, the sensitivity was approximately 80% (39–90.9%) in the urine, 95% (90.9–100%) in the cervix and 90% (67.8–96.4%) in the vagina, with the population in these studies consisting of patients undergoing examinations for cytological abnormalities or CIN. The specificity of each test was also inconsistent (30–67.8%). Our data demonstrated the sensitivity and specificity for detecting HPV-related lesions/tumors (Table 5). In any case, the sensitivity was lower than that of the other data. One reason for the low sensitivity in cervical samples may be collection errors by physicians. During sample collection, it is essential to thoroughly scrape the endocervical canal with a collection instrument. Another issue may stem from the criteria for CIN2 or higher lesions presumed to be HPV-related. The HPV positivity rate in cervical samples from patients with CIN2 or higher was 86.1%, and we closely examined the remaining 12 patients who were HPV-negative (4 adenocarcinomas, 6 CIN2, and 2 CIN3 patients). In adenocarcinoma cases, although clear cases of gastric-type adenocarcinoma and clear cell carcinoma were excluded, it is possible that HPV-independent adenocarcinoma was included. Additionally, the quantitative evaluation of dysplasia in CIN2 was based solely on traditional HE staining, which may include dysplasia lesions that exceed 30% of the basal layer caused by not high-risk HPV infection. All SCC lesions in the cervix were HPV-positive. Among these 12 cases, only two were positive for HPV in urine or vaginal samples, suggesting the possibility of low HPV detection or non-high HPV-related cases. In addition, many papers on urine reported first-void urine, but we used urine at any time, which may be a factor. Although meta-analyses have demonstrated that the accuracy of HPV detection is lower in both urine [27] and the vagina [28] than in the cervix, in the other review of 33 studies [29], most of these subjects reported a preference for self-sampling. Therefore, HPV testing of self-collected urine can be suggested as an additional screening strategy for women not undergoing screening.

With respect to testing methods, most studies have used the HC-II, cobas^® 4800 HPV test (Cobas^®), Cervista™ HPV HR, Abbott RealTime High-Risk HPV, Xpert HPV test, and Onclarity assay [19‒22, 24, 25, 30‒34], but some studies have used original methods [23, 35‒38]. HC-II is a commercial HPV detection test designed to detect 18 types of HPV, and the Cervista™ test is a diagnostic test for the qualitative detection of DNA from 14 high-risk types of HPV. The Cobas^® utilizes amplification of target DNA by polymerase chain reaction and nucleic acid hybridization for the detection of 14 high-risk HPV types in a single analysis. However, the meta-regression analysis of urine HPV tests revealed no differences in accuracy when covariates, including patient selection, purpose, sample timing, storage temperature and method of HPV testing, were used [27].

However, the incidence and type distribution of HPV-related cancers and precancer lesions differ among countries based on the data of the HPV Information Centre, which consists of the ICO and the IARC [39]. Among women with precancerous high-grade cervical lesions, 57.9% and 12.1% were in the USA, and 58.6% and 14% were in the UK. However, among women in Japan, the rates were 39% and 9%, respectively, and those of HPV52 and HPV58 were 21.7 and 16.7%, respectively. This finding is similar to our results and data from Asia; therefore, it is clear that there are regional differences in distribution. Therefore, we believe that identifying HPV types via Onclarity assay may be useful.

As mentioned in the introduction, government CCA screening programs in various countries have prevented the spread of HPV and CCA, which decreases premature deaths and costs per capita. Similarly, the widespread use of HPV vaccines has contributed to the reduction of HPV infections and CIN and CCA. However, a lack of vaccination is a problem not only in Japan but also in other regions of the world.

First, the rate of HPV detection was high in the vagina but not in the urine. Although this is probably due to the spot urine test, it was possible to identify a certain number of positive cases. It is necessary to improve the method and further verify it through screening for many populations with a low prevalence. Moreover, since the negative predictive value was low for this result, false-negative HPV results must be handled carefully, and it is necessary to ensure that patients undergo regular cancer screening, even if it is negative. Second, cytology alone is the mainstream screening method in Japan, and the percentage of LBC is not yet high. Similarly, HPV testing has a low penetration rate and is being implemented through our national policy as a preventive screening tool for CCA in municipalities. Specifically, it has just been announced that HPV testing alone will be promoted.

The conditions suitable for urine HPV testing in our study included the use of a container QxUPT kit within 24 h of collection and the performance of the HPV assay within 3 days at room temperature. The results of the self-collected urine and vaginal HPV tests had a high concordance rate with those of the clinician-collected cervical HPV test and were considered comparable. In spot urine, the detection rate was lower than that in first-void urine in other studies. Although the detection rate of HPV in spot urine is low and there is significant room for improvement, a simple self-collection test could be meaningful for women who, for various reasons, do not undergo cancer screening. This includes those who are not knowledgeable about CCA, have not received the HPV vaccine, are hesitant to undergo gynecological examinations, or face economic barriers. Further large-scale research is necessary for verification, and measures are needed to encourage these women to undergo testing.

We would like to thank the patients who participated in this study and the Japan Cancer Society for supporting the testing and analysis of this study. The results of this study were partially presented at the 73rd Annual Congress of the Japan Society of Obstetrics and Gynecology, Niigata, Japan, April 22–25, 2021.

This study was approved by the Institutional Ethics Committee of Kagoshima University (Approval No. 25-161) and was performed in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants.

The authors have no conflicts of interest to declare.

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Mika Mizuno: conceptualization, project development, research, sample collection, data interpretation, writing – original draft, and writing – review and editing; Masaki Kamio: conceptualization, project development, protocol design, and sample collection; Tadao Kakizoe: sample assay performance and data analysis; Hiroaki Kobayashi: principal investigator; Mika Sakihama, Shintaro Yanazume, and Shinichi Togami: informed consent and sample collection; all authors (except Tadao Kakizoe): data interpretation, manuscript review, and editing; and final manuscript approval: all authors.

All the data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.