Efficient combination of Human Papillomavirus Genotyping for the triage of women with Atypical Squamous Cells of Undetermined Significance in Chinese rural population: A population-based study

Objective: In this prospective, population-based study, we evaluated the utility of high-risk human papillomavirus (HR-HPV) genotyping for triaging women with atypical squamous cells of undetermined significance (ASC-US) in the Chinese rural area. Methods: A total of 40,000 women were recruited from rural areas of Shanxi Province, China, between June 2014 and December 2014. Women with Pap results of ASC-US underwent HPV genotyping, colposcopy and histopathological examination. For those with normal cervixes or cervical intraepithelial neoplasia (CIN) 1 on the initial evaluation, a 2-year follow-up study was performed. Results: The reporting rate of ASC-US was 5.76% (2,304/40,000) in the study population. The detection rates of CIN 2 or above (CIN2+) and CIN 3 or above (CIN3+) in women with ASC-US were 7.28% and 1.75%, respectively. HPV 16 (39.53%), HPV 58 (17.83%), and HPV 52 (15.50%) were the three most prevalent HR-HPV genotypes among all women with ASC-US cytology. The five most common HR-HPV genotypes in CIN3+ lesions were HPV16, HPV58, HPV33, HPV31 and HPV18. Compared with the 15 HR-HPV testing, genotyping for a combination of HPV16/18/31/33/58 increased specificity significantly with virtually no loss of sensitivity for detecting CIN2+ and CIN3+ lesions, as well as significantly reduced colposcopy referral rate (23.15% vs 33.70%, p<0.01). In addition, in the 2-year follow-up period, women with infection of HPV16, 18, 31, 33 or 58 genotypes were the most likely population (92%, 23/25) to develop CIN2 lesion. Conclusion: Our results demonstrate that genotyping for a combination of HPV16/18/31/33/58 provides a more efficient and cost-effective model to risk-stratify women with ASC-US in the Chinese rural population.


Introduction
Atypical squamous cells of undetermined significance (ASC-US) cytology is the most frequent abnormal interpretation on cervical Papanicolaou (Pap) test, accounting for ~5% of all cervical cytology in the western population [1][2][3] and 3-5% in Chinese population [4][5][6]. The management of women with ASC-US remains a clinical challenge, due to the equivocal cytology and variable underlying process from human papillomavirus (HPV)-unrelated, non-neoplastic conditions to the HPV-related cervical intraepithelial neoplasia (CIN) and carcinomas. Earlier studies have found that 3-36% of women with ASC-US cytology will have underlying CIN2 or more severe lesion (CIN2+) [7][8][9], and only a small percentage of ASC-US associated with cervical intraepithelial neoplasia 3 or worse (CIN3+) can be detected by follow-up colposcopic examination. Therefore, many efforts have been made trying to identify those who have cancer and significant precancerous lesions among women with ASC-US cytology, in order to improve the efficacy of screening and reduce the number of unnecessary colposcopic examination.
Based on the published evidence, the American Society for Colposcopy and Cervical Pathology (ASCCP) Guidelines have recommended women with an ASC-US cytology result should have a reflex HPV testing, and an immediate colposcopy referral or conservative follow-up based on patients' previous screening history is recommended for women with high-risk (HR) HPV-positive ASC-US [10,11]. Currently, the widely-used, US Food and Drug administration (FDA)-approved HPV testing platforms are largely pooled HR-HPV tests, which include 13 or 14 of the most common high-risk genotypes (16,18,31,33,35,39,45,51,52,56,58,59,68 and/or 66). However, the risk of progression to cervical cancer and CIN varies substantially among individual carcinogenic genotypes [12,13]. It has been reported that women with HPV16 or HPV18 positive ASC-US had approximately twice the risk of CIN3+ as women with ASC-US and high-risk HPV types other than 16 and 18 [2,14,15]. Thus, type-specific HPV identification appears to be a reasonable strategy to improve risk stratification of women with ASC-US and to reduce the colposcopy burden. This approach has been supported by the launches of additional HPV genotype 16, 18 and /or 45 in the recent FDAapproved Cobas 4800, Aptima and Onclarity tests.
It is well-known that the HR-HPV prevalence and distribution varies geographically worldwide. There are some distinctive characteristics in the Chinese population, including more frequent HPV52 and HPV58 genotypes, and less HPV45 and HPV18 genotypes in cervical cancer and precancerous lesions [16]. In order to evaluate whether selection of certain HPV genotypes may be more efficient and costeffective for the risk management of women with ASC-US in the Chinese rural population, in this study, we report the results of baseline and 2-year follow up of HPV genotyping and histologic correlation in 2,304 women with ASC-US cytology, from a large population-based cervical cancer screening program and a prospective cohort in the rural areas of Shanxi province, China.

Study population
The study design and methods have been described previously [17,18]. Briefly, we conducted free cervical cancer screening tests for eligible women who were permanently residing in two counties (Yangqu and Jiexiu) of Shanxi Province between June 2014 and December 2014. A total of 40,000 women aged 19-65 years completed an epidemiological questionnaire on potential risk factors for CIN and the Pap test. Women with Pap results of ASC-US underwent HPV genotyping, colposcopy and histopathological examination.
In Yangqu County, 1,034 women with ASC-US cytology were identified and 1,026 women underwent HPV baseline genotyping and immediate histopathologic examination. 77 women were diagnosed with CIN2 or worse including 4 squamous cell carcinomas on the histologic examination and all these patients received clinical treatment. 949 women were found to have normal cervixes or CIN1 and were followed-up for additional 2 years (Figure 1). Among 1,270 women with ASC-US cytology in Jiexiu County, 382 patients with normal cervix refused to have HPV genotyping test. 90 women with CIN2 or worse lesions received clinical treatment, while 798 women with normal or CIN1 lesion were followed-up for additional 2 years (Figure 2). All inspections and detections were implemented under double-blind conditions. The study was registered in the Chinese Clinical Trial Register (ChiCTR), registration number: ChiCTR-ROC-15006479. The study protocol and informed consent form were reviewed and approved by the Ethics Committee of Second Hospital of Shanxi Medical University.

Cytology testing
The liquid-based cytology (LBC) method was used in this study. The LBC preparation was produced by Lituo Biotechnology Corp. (Hunan, China; http://www.lituo.com. cn). The cytological evaluation was performed by two cytopathologists at The Second Hospital of Shanxi Medical University.
Cytology results were classified using the 2001 TBS criteria for reporting cervical cytology [19]. The cases with an interpretation of ASC-US or a worse condition were further reviewed by a senior cytopathologist who was blinded to the results. Ten percent of cases with a negative report were reviewed by another pathologist to provide quality control.

Colposcopy and histological examination
The colposcopy procedure and biopsy were performed by experienced gynecological specialists from The Second Hospital of Shanxi Medical University. The average time from Pap sampling to colposcopy was 15 days (range: 3-62 days). Colposcopy examinations were performed using the Preventive Oncology International microbiopsy protocol. During the colposcopy, the cervix was divided into quadrants and each quadrant was examined independently. All visually abnormal areas were biopsied. Quadrants with normal colposcopic impressions were biopsied at the squamocolumnar junction ("random biopsy"). Women with abnormal cytology results and negative or inadequate colposcopic findings underwent endocervical curettage (ECC). All histological slides were reviewed by two gynecological pathologists, who were blinded to the cytology results at The Second Hospital of Shanxi Medical University. Immunohistochemistry for p16INK4A (p16) (ZM-0205, ZSGB-BIO, Peking, China) was utilized in adjudicating the diagnosis. If the two pathologists disagreed on the diagnosis, a third senior pathologist reviewed the slides, and some difficult or equivocal cases were considered for the pathology panel consensus diagnosis.

Statistical analysis
The statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) version 22.0 software for windows (SPSS Inc., Chicago, IL, USA). Chi-squared test or Fisher's exact test were used for the categorical variables. Statistical tests were two-sided, and P <0.05 was considered statistically significant.
Of 2,296 women with histologic examination, 1,914 women had HPV genotyping test. Table 1 shows the relative frequencies of HPV infections in women with ASC-US Pap test according to histological diagnosis. The overall rates of HR-HPV and LR-HPV infections were 33.70% (645/1,914) and 1.83% (35/1,914). Among 645 women with HR-HPV infection, 445 patients were infected with a single HPV genotype and 200 patients were infected with multiple HPV genotypes including 20 patients co-infected with low-risk HPV genotypes. The prevalence of HR-HPV genotypes in women diagnosed with <CIN2, CIN2, CIN3 and SCC was 29.36%, 73.23%, 96.30% and 100%, respectively. LR-HPV infections alone, without HR-HPV infections, were not detected in any of the CIN3 and above cases.
The median age of women with ASC-US was 51 years (range: 20-65 years) in Yangqu County and 50 years (range: 20-65 years) in Jiexiu County. Table 2 presents age-stratified HR-HPV prevalence in women with ASC-US cytology. Among all age groups, women aged between 30 and 39 years had peak rate of HR-HPV infection in Yangqu (58.2%, 71/122) and Jiexiu (31.4%, 38/121) Counties. There was a dramatic decline in the HR-HPV infection rates after age 50 in both Counties.

Prevalence of different HR-HPV genotypes in women with ASC-US Pap test
In women with multiple infections, HPV genotypes were reported separately. The prevalence of HR-HPV genotypes in HR-HPV-positive cervical lesions in Yangqu County are presented in Table 3A. HPV16 (35.57%) was the most common HR-HPV genotype, followed, in order of decreasing frequency, by HPV52 (19.65%), HPV58 (17.91%), HPV53 (8.21%), HPV56 (6.97%), and HPV33 (6.97%). Additionally, HPV18 (4.48%) was in the 11th position. The distribution of HR-HPV genotypes by histologic outcome was different. HPV16 showed an increasing prevalence with the severity of the histological diagnosis (30.88% in <CIN2, 54.35% in CIN2, and 81.25% in CIN3+). HPV58 was the second-most prevalent genotype in CIN2+ lesions and the third-most prevalent genotype in <CIN2 lesions.

Two-year follow-up results of women with ASC-US cytology and initial baseline normal cervix or CIN1
Of 1,747 women with normal cervix or CIN1 at baseline colposcopic examination, 1,385 patients (79.28%) including 990 HR-HPV negative and 395 HR-HPV positive women underwent both HPV genotyping test and colposcopic examination at the 2-year follow-up. Among 990 women with initial negative HPV testing, 130 cases were found to have HR-HPV positive (13.13%) and 8 cases developed CIN2 at the 2-year follow-up. Among 395 women with baseline positive HPV testing, 50 cases had new HR-HPV infection (12.66%), and 52 cases had persistent infection (13.16%). 14 (26.9%, 14/52) women with persistent HR-HPV infection developed CIN2. Table 5 demonstrates the prevalence of HR-HPV genotypes of women who underwent 2-year follow-up testing including 180 women with new HR-HPV infection and 52 women with persistent HR-HPV infection. Among them, the most common genotypes were HPV16 (28.02%), HPV58 (20.26%) and HPV52 (17.67%). Of the 16 CIN2 cases with positive HR-HPV test, the 5 highest-risk HPV genotypes were HPV16 (68.75%), HPV58 (37.5%), HPV18 (12.5%), HPV31 (12.5%) and HPV33 (12.5%) ( Table 6). Table 5. Prevalence of HR-HPV genotypes at 2 years of follow-up in women with initial ASC-US cytology and normal cervix and CIN1 on immediate histologic examination cancer screening program in the Chinese rural population, we initiated a large population-based cervical cancer screening program and a prospective cohort in the rural areas of Shanxi province, China in 2014. We herein report the results of HPV genotyping, immediate histologic correlation and 2-year follow-up of 2,304 women with ASC-US cytology. To the best of our knowledge, this is the largest prospective study on ASC-US women with genotyping and histological correlation in the Chinese rural population. Table 6. Prevalence of HR-HPV genotypes in HR-HPV-positive cervical lesions at 2 years of follow-up in women with initial ASC-US cytology and normal cervix and on immediate histologic examination the most important risk factor for the development of cervical cancer and precancerous lesions [32][33][34][35]. In the present study, women infected with those 5 highest-risk HPV genotypes were the most likely population to develop CIN2. As shown in another study, the 3-year cumulative risks of developing CIN3+ lesions for women with HPV16 (16.0%), HPV18 (7.4%), HPV31 (7.0%), and HPV33/HPV58 (7.1%) were higher than the overall 5.2% risk associated with Hybrid Capture 2 (HC2) positivity [36]. These results further support the approach of genotyping for the combination of 5 highest-risk HPV genotypes could provide risk stratification for women with ASC-US. As mentioned earlier, due to the limited public health resource in the rural areas in China, the development of this specific HPV genotyping assay might significantly improve the cost-effectiveness of HPV testing by reducing unnecessary colposcopy burden as well as overtreatment, which would have significant impact in the Chinese rural population.
The current study has some limitations. First, this study population had not previously been vaccinated for HPV. As a result, the prevalence of HR-HPV genotypes in HPV-Vaccinated population may be different. Second, in this study, the selection of 5 HPV genotypes (HPV16/18/31/33/58) was based on a relatively small sample size, and larger sample sizes and study in different geographic regions are needed to validate this novel HPV genotyping assay.
In conclusion, in this large-scaled prospective study with cross-sectional analysis as well as 2-year follow-up analysis, the genotyping for a combination of HPV16/18/31/33/58 provides a more efficient and cost-effective risk-stratification model in women with ASC-US in our study population. Women who tested negative for the 5 highest-risk HPV genotypes (HPV16/18/31/33/58) may not require immediate colposcopy. Further prospective studies with larger sample sizes, as well as from different geographic regions are needed to validate this novel HPV genotyping assay in the triage of women with ASC-US in the Chinese rural population.