Risk of cervical lesions in high-risk HPV positive women with normal cytology: a retrospective single-center study in China

Background To investigate high-risk HPV (hr-HPV) genotype distributions and the association between hr-HPV infection with severity of the cervical lesions in women with normal cytology. Methods In this cross-sectional study, the result of the hr-HPV test and biopsy of colposcopy of women with normal cytology from January 2012 to January 2019 were analyzed. The detection rate of high-grade squamous intraepithelial lesion (HSIL) and cervical cancer were calculated among different hr-HPV genotypes, viral load group, and age groups. Results Five thousand eight hundred eighty women were enrolled in this study. Overall, 59.97% had normal histological results, 19.32% had HSIL, and 1.07% had cervical cancer. The detection rate of HSIL or worse (HSIL+) in women with single HPV16(34.00%), HPV31(27.50%), HPV33(25.58%), and HPV52(20.88%) infection were higher significantly than single HPV18 (15.59%) infection, respectively. The HSIL+ detection rate between HPV16 single infection and multiple infections (excluding HPV18) was no significant difference (34% vs 35.47%, P = 0.638), contrary to HPV18(12.59% vs 21.67%, P = 0.022). In women without HPV16/18 infections, HSIL+ detection rates for single, double, and triple or more hr-HPV infections were 12.28, 20.31, and 37.50%, the risk of detection of HSIL+ significantly increasing. With the hr-HPV DNA load increases, the risk of detection of HSIL+ (χ2 = 91.01, P < 0.0001) and invasive cervical cancer (χ2 = 5.757, P = 0.016) increase. In age < 30, 31–40, 41–50, 51–60, > 60 group, HSIL+ detection rate were 24.80%、22.10%、19.59%、14.29, and 12.61%, respectively. Conclusion Women who have normal cytology with HPV 16/18/31/33/52/58 infections, multiple HPV infections and high viral load, have a higher detection rate of HSIL+.


Introduction
High-risk HPV (hr-HPV) infection is the main etiological factor for the development of cervical neoplasia [1], and routine cervical cancer screening includes hr-HPV and cervical cytology tests. Previous randomized controlled trials [2][3][4][5][6] have shown that the strengths of HPV testing are higher sensitivity and negative predictive value than cervical cytology. Given that cumulative incidence of cervical intraepithelial neoplasia (CIN) 3 or cancer 5 years after an HPV-negative test was lower than the risk 3 years after a negative Pap test [6,7], HPV testing was introduced as the supplementary test in conventional cytology-based screening, even gradually replacing cytology-based screening in primary screening for cervical cancer in some countries [8,9]. As HPV testing is widely used, the optimal clinical management of women with HPV infection, especially combined normal cytology, remains a challenge. The guidelines recommend managing HPV-positive/cytology-negative results by return testing at 1 year or HPV genotyping for HPV16 and HPV18 [10]. If women have negative cytology but HPV16 and HPV18 infection(s), immediate colposcopy is recommended instead of a 1-year return. HPV16 has higher relevance with cervical lesions than other hr-HPV types [9,11,12]. But the reason for choosing HPV18 is uncertain, partly because HPV18 infection has a correlation with cervical adenocarcinoma, which cannot be well detected by cytology [11]. The correlation with other types HPV is not clear yet. What's more, whether viral load can be used as a marker to predict the severity of cervical lesions is currently controversial [13].
Therefore, this study was designed to retrospectively analyze data for up to 7 years, to find the correlation between the severity of cervical lesions and hr-HPV genotypes, viral loads, and ages in HPV-positive/cytology-negative females in China.

Study population
This study was cross-sectional investigation based on data from women who accepted Thinprep cytologic test (TCT) test, HPV test, and colposcopy at Qilu Hospital of Shandong University from January 2012 to January 2019, after approval of the Qilu Hospital of Shandong University【2018(054)】. Clinical information of women who met the inclusion criteria was collected retrospectively.
Exclusion criteria: (1) women with history of treatments to cervical lesions, such as cervical surgery and medicine, et al.; (2) women with incomplete cervical cervix; (3) women with malignant tumors; (4) women with autoimmune diseases or receiving immunotherapy; (5) women with pregnancy; (6) women who had received HPV vaccine.

Cytology test
Specimen collection, specimen preparations, and results of the tests are performed, according to the instructions of the manufacturers, respectively.
One of those HPV tests could be chosen:

Colposcopy and guided cervical biopsies
All women, enrolled in this study, underwent colposcopy and guided cervical biopsies according to the American Society for Colposcopy and Cervical Pathology Colposcopy Standards recommendations [14]. The histologic diagnosis was regarded as gold standards for this study and was based on the consensus diagnosis of two experienced gynecologic pathologists who were blinded to results of TCT, HPV test, and colposcopic examination. The results of biopsy reports include: normal; low-grade squamous intraepithelial lesion (LSIL) (including CIN 1), high-grade squamous intraepithelial lesion (HSIL) (including CIN 2-3), and cervical cancer. HSIL or worse (HSIL+) includes HSIL and cervical cancer.

Statistical analysis
Statistical analyses were performed using the SAS 9.4 (SAS Institute, Cary, NC) for windows. The categorical variables were expressed by a percentage (%). The binary Logistic regression model was used to analyze the detection rate of HSIL+ and cervical cancer in different hr-HPV genotype groups. The relationships between the HPV viral load and the detection rate of HSIL+ were analyzed by the Cochran-Armitage Trend test. Results were considered statistically significant at p-values < 0.05.

Characteristics of the study population
A total of 5880 women with normal cytology and hr-HPV infection accepted colposcopic biopsy were enrolled in this study. 4332 women underwent HPV genotyping test and 1548 women were hr-HPV DNA positive by HC2 hr-HPV DNA test (Fig. 1). The median age was 39 years (26-81), and the average age was 39.29 ± 9.72 years. They were divided into 5 age groups. The prevalence of HPV infection among all age groups showed in Table 1 and Fig. 2. HPV16 and HPV18 were the most common, followed by HPV52 and 58, in all age groups. Among women with normal cytology and HPV infection, 3526(59.97%) women had normal pathology, 1155(19.64%) women had LSIL, 1136(19.32%) had HSIL. The remaining 63(1.07%) woman was diagnosed with cervical cancer, including 17 women with positive HC2-hr-HPV-DNA-test results, 31 women with single HPV16 infection, 4 women with HPV 18 infection, 8 women with multiple HPV 16 infections without HPV18, 1 woman with HPV 16 and HPV18 infections, and 2 women with HPV58 infection.

Multiple hr-HPV infections and histopathology in hr-HPVpositive/cytology-negative women
There was no significant difference in the detection rate of HSIL+ between single infection of HPV16 and multiple infections with HPV16 (excluding HPV18) (P = 0.599), not same as HPV18 (P = 0.022).

Discussion
Tests designed to stratify risks, such as HPV tests and cytological tests, play a vital role in managing patients. Most of the HPV-negative women is associated with a very low risk for developing cervical cancer [4][5][6][7]. One of the clinicians' work is to identify women at high risk of cervical lesions from HPV-positive and cytologynegative women.
This study chose HSIL+ as the primary endpoint, because the purpose of cervical cancer screening is to detect easily treatable precancerous lesions, most commonly defined as CIN2 and greater. And choosing CIN3 + as the endpoint would miss some invasive cancers which developed from a part of CIN2, although it will increase the rate of referral of colposcopy. However, the primary endpoint of HSIL + is not ideal enough. One of the reasons is that CIN2 is an equivocal diagnosis of precancer, part of CIN2 can regress spontaneously, and the rest may develop to CIN3 or even invasive cancer [15].
This study found that HPV 52 and HPV58 were the most common type of HPV infection apart from HPV16/18, which was same with previous studies showed HPV 52 and HPV58 were common in Asia, including Korea [23][24][25] and China [26][27][28]. Recently, high carcinogenicity of HPV58 and HPV45 has attracted much clinicians' attention. A meta-analysis showed that the pattern for HPV58 in Eastern Asia was different than elsewhere, and HPV58 continues to rise in the proportion of cervical diseases at all levels, accounting for 10.2% ± 3.9% of all invasive cervical cancer with HPV infection [29]. Another study showed that HPV-58 is more prominent in HSIL and invasive cervical cancer patients in Japan [30]. In this study, in women with single hr-HPV infection, excluding HPV16/18, cervical cancer only occurred in 2 women infected with HPV58, consistent with previous studies in southwestern China that HPV58 is common in ICC [31]. HPV45 is a close relative of HPV18 [19,32], and some studies reported that HPV45 was the third most common genotype involved in invasive cervical carcinoma, found in 6% of cervical cancers [33]. However, this study did not observe the high carcinogenicity of HPV45 in women with hr-HPV infection and normal cytology, which was related to the abnormality of cytology in women infected with HPV45. And a study reported that only 0.5% of women with HPV45 infection had normal cytology [34].
Multivariate analysis showed that the risk of cervical cancer was 19.9 times higher in women with single HPV infection, and 31.8 times higher in those with multiple HPV infection [35]. This suggests that multiple infections have a cumulative effect on developing to HSIL+. This study analyzed the association between the detection rate of HSIL+ and multiple hr-HPV infections, HPV infection and ≥ triple hr-HPV infection (P = 0.019), which means that more hr-HPV genotypes infected in women, the higher the HSIL detection rate. It is controversial whether the viral hr-DNA load can be used as a reliable marker to predict cervical lesions. Previous studies have reported that HPV viral load has   [39,40]. This study found that the higher viral load, the higher the detection rate of HSIL+ in women with normal cytology(χ2 = 91.01, P<0.0001),  same as the detection rate of cervical cancer (χ2 = 5.757, P = 0.016), which may be correlated to the fact that patients with low-load HPV infection are more likely to be eliminated, while those with high-load HPV infection are more difficult to eliminate [41]. This study found that the absolute value of HSIL+ detection rate decreased with age, while the absolute value of cancer detection rate increased with age, in women with normal cytology and hr-HPV infection. The time of hr-HPV infection is short, and cervical cell morphology has not changed in young women, what's more, the proportion of CIN2 that can regress is higher in young women [14]. However, hr-HPV infection time is longer, and the proportion of CIN2 that progresses to CIN3 or invasive cancer is higher, while the percentage of CIN2 that can regress is smaller, in older women [14,40]. Therefore, older women had lower HSIL+ detection rate and higher cervical cancer detection rate in women with normal cytology and HPV infection.

Conclusion
In conclusion, about one in five hr-HPV-positive/cytologynegative women are diagnosed with HSIL+. Women with HPV 16/18/31/33/52/58 infections, multiple HPV infections whether combined with HPV16/18 infections, as well as high viral load, have a higher detection rate of HSIL+, although they have normal cytology.

Limitation
This study was a cross-sectional study without follow-up and it could not estimate the cumulative risk of cervical lesions within a period. The population enrolled in this study is the patients who underwent colposcopy in Qilu Hospital, Shandong University. So, the results of this study can not be directly applied to the general screening population, but it has certain reference significance for the diagnosis and treatment of patients coming to the hospital. Besides, we can not ignore the heterogeneity of 3 HPV assays and specifically analyze women infected with 12 h-HPV.