Screening History and Diagnostic Characteristics of Cervical Adenocarcinoma: a Retrospective Study


 Background: An increasing trend in incidence of cervical adenocarcinoma has been observed in recent years. This research aims to study the screening history and diagnostic characteristics of cervical adenocarcinoma.Methods: Cervical cancer patients hospitalized in the Gynecology Department of the Third Affiliated Hospital of Sun Yat-sen University between 1st Jan 2017 and 31st Mar 2020 were included. Cervical screening history investigation and analysis were carried out by medical record data inquiry and case investigation.Results: (1)The chief complaint of 72.0% of cervical squamous cell carcinoma was vaginal bleeding. 75.6% of adenocarcinoma came in for abnormal vaginal discharge, p<0.001. (2)The HPV infection rate in adenocarcinoma(74.2%) was lower than that in squamous cell carcinoma(92.9%), p<0.001. (3)The participation rate of cervical screening before diagnosis of adenocarcinoma(21.2%) was higher than that of squamous cell carcinoma(2.8%), p＜0.01. (4) The proportion of early-stage in adenocarcinoma (46.3%) was larger than in SCC (28.3%), p<0.01. Conclusion: Compared to cervical squamous cell carcinoma, HPV infection was found to be less closely associated with adenocarcinoma of uterine cervix. The participation rate and frequency of cervical screening of patients with cervical adenocarcinom were more than squamous cell carcinoma. The screening methods at present may not be sensitive enough for precancerous lesion of adenocarcinoma, but regular and multiple screening are still of great significance for early diagnosis of adenocarcinoma of uterine cervix.


Background
Cervical cancer is the fourth most frequent cancer and fourth in cancer deaths among women, accounting for 6.6% of all malignancies in women. In 2018, about 570,000 women were diagnosed with cervical cancer worldwide, of which 311,000 died [1] . China has a large amount of cervical cancer patients with approximately 62,000 new cases and 30,000 deaths in China every year [1] .
In the WHO (2014) classi cation of female reproductive system tumors [2] , cervical cancer is divided into squamous cell carcinoma (SCC), adenocarcinoma (AC), adenosquamous cell carcinoma (ASC), groundglass cell carcinoma, adenoid basal cell carcinoma, undifferentiated carcinoma, neuroendocrine tumor, etc. Among all pathological types of cervical cancer, SCC comes rst in incidence rate, AC comes second, and other pathological types of cervical cancer follow. Cervical intraepithelial lesion is a premalignant condition that can potentially progress to cervical cancer. Early detection of cervical intraepithelial lesions and timely intervention measures can effectively reduce the incidence of cervical malignancy [3,4] . Since the 1980s, HPV infection has been internationally recognized as an important pathogenic factor of cervical cancer [5,6] . With prevention methods of human papillomavirus (HPV) vaccination and cervical screening having signi cantly reduced the incidence of cervical cancer in the past few decades [8] , cervical cancer is largely preventable [7] . However, relevant studies in recent years have shown that the decrease of cervical cancer incidence is mainly attributed to the reduction of SCC. The number of AC incidences, on the other hand, has shown an obvious upward trend [9][10][11][12] . From the 1970s to the 2010s, the proportion of SCC incidence in cervical cancer diseases has decreased from 95% to 67-83% while that of AC has increased from 5% to 10-25% [12][13][14] . Thus, the increase of AC is not only re ected in its proportion in cervical cancer diseases but also in the absolute number of incidences [13][14] . So, what is the underlying reason for the increasing incidence of AC? And are there any differences between the screening history and diagnostic characteristics of AC and SCC? In this study, we aim to investigate the screening history (characteristics of cervical screening participation) and diagnostic characteristics of different histological types of cervical carcinoma.

Basic Information Collection
We collected data on all cervical cancer cases diagnosed in the Third A liated Hospital Of Sun Yat-sen University during 2017-2020. All the selected cases were con rmed to be cervical cancer by histopathological examination, and patients with different histological types of cervical cancer were grouped according to the the 2014 World Health Organization (WHO) classi cation of tumours of female reproductive organs [2] . According to the standards, patients with different histological types were divided into three groups: (1) History collection were carried out by electronic medical record data inquiry, telephone follow-up and case investigation. Data collection mainly included basic information (month/year of birth, month/year of diagnosis, age at menarche, menopausal age, reproductive history, etc), chief complaint, the history of participation in cervical cancer screening before diagnosis, the results of HPV test and cytology, histopathological results, and cancer staging. Patients who underwent surgery were staged according to the pathological stage of cancer at the time of surgery. Those who didn't receive surgery were staged by clinical stages. We de ned stage I as Early-Stage Cervical Cancer and stage II -IV as advanced cervical cancer.

Screening History
In 2018, consensus guidelines were issued for cervical screening in US populations [16][17] , recommending screening begin at age 21yrs; 3-yearly cytology for women aged 21-29yrs, and either 3-yearly cytology, 5yearly primary HPV testing or 5-yearly co-testing(co-occurring HPV and cytology testing) for women 30-65yrs. Most women aged >65yrs can cease cervical screening [16][17] . Based on the patient's participation in cervical cancer screening between 6 months and 5 years prior to diagnosis( since cancers histologically diagnosed within 6 months of an abnormal screening result were almost certainly present at the time of the diagnosis) [18] , we grouped the patients into four groups, including "Never screening, Irregular screening, Regular screening, and Frequent screening" (Figure 1). Females who had never attended the screening program before diagnosis were grouped into "Never screening"; females who had attended more than the frequency of recommendation were grouped into "Frequent screening"; females who had attended the screening as the frequency of recommendation were grouped into "Regular screening"; females who had attended screening before but hadn't participated in the most recent screening cycle were grouped into "Irregular screening".(A screening cycle is the recommended interval for cervical screening). Cervical cancer screening participation rate is de ned as the ratio of the number of patients participating in screening to the number of all patients.

Statistical Analysis
We presented the clinical characteristics, the features of HPV infection and the screening history in the past ve years before diagnosis of cancer among all cervical cancer cases. Quantitative data were described by Mean (Standard deviation), qualitative data was described by adoption rate or percentage. Comparison among different groups of quantitative data was conducted by T-test or ANOVA, and the chisquare test was used for the comparison of quantitative data. Analyses were strati ed by histopathological type. Data manipulation and statistical analyses were conducted in SPSS 25.0.
This study was approved by the medical ethics committee of the clinical research center of the Third A liated Hospital of Sun Yat-sen University.

Results
With the exclusion of those that had no histopathological results, a total of 291 women were diagnosed with cervical cancer in The Third A liated Hospital of Sun Yat-sen University between 1st January 2017 and 31st March 2020. Of these 291 cases, 81.1% were squamous, 15.5% adenocarcinoma and 3.4% other histopathological types. The latter group was excluded from further analysis.
The characteristics of different histopathological types of cervical carcinoma were compared in Table 1.
The mean age of SCC, AC and other histopathological types of cervical cancer were 50.31 ± 10.45 years, 46.44 ± 10.48 years and 56.50 ± 9.57 years. The menarche age of other histopathological types of cervical cancer (16.67 ± 6.51) was higher than that of SCC and AC(p 0.001), while there was no signi cant difference between SCC(13.64 ± 1.33) and AC(13.57 ± 0.85)(p 0.05). The majority of the patients sought medical attention for vaginal bleeding and abnormal vaginal discharge. The chief complaint of 72.0% of SCC was vaginal bleeding. 75.6% of AC came in for abnormal vaginal discharge.
The HPV infection rate in AC was 72.4%, and 92.9% in SCC which was signi cantly higher than AC (p < 0.001). The proportion of cancer stages differed for SCC and AC signi cantly (p < 0.05). The percentages of stage I to IV in SCC were 28.3%, 59.9%, 9.4%, 2.4% respectively, and in AC were 46.3%, 43.9%, 4.9%, 4.9% respectively. Compared to SCC, the proportion of early-stage was signi cantly larger in AC (19/41, 46.3%) than in SCC (60/212, 28.3%) (p < 0.01). From the results so far, we didn't nd any signi cant differences of histopathological types of cervical carcinoma in age, menopausal age, marital and reproductive history (including number of pregnancies, abortions, natural births and the age of rst birth or rst abortion) and educational levels (p 0.05).
In further analysis by cancer stage (Table 2), the reasons for hospitalizing behavior between SCC and AC, AC in different cancer stages, SCC in different cancer stages, SCC and AC in same cancer stage were compared. The results show that the percentage of patients with AC coming for abnormal cancer screening results was signi cantly higher than SCC (AC: 26.8%, SCC: 11.3%, p < 0.01). But there were no signi cant differences between AC in different cancer stages (early stage AC:31.6%, advanced AC: 9.1%, p = 0.070), SCC in different cancer stages (early stage SCC:18.3%, advanced SCC: 8.6%, p = 0.054), and SCC or AC in same cancer stage(p 0.05).
Among 214 cases with complete information about screening history and histopathological types (Fig. 2), this study found the participation rate of AC in cervical cancer screening(21.2%) was signi cantly higher than SCC(2.8%)(p 0.01). The percentage of Frequent screening and Regular Screening in AC(18.2%) was also higher than SCC(1.1%)(p 0.01). For early stage cancer, the percentage of Frequent screening and Regular Screening was signi cantly higher in AC(28.6%) than in SCC(3.9%)(p 0.05), as was for advanced stages where the percentage for AC was 10.5% and 0% for SCC (p 0.05). There was no signi cant difference between different cancer stages of AC in frequency of screening participation(p 0.05), nor was there in different cancer stages of SCC(p 0.05). Discussion HPV16, 18 are still the most common types in HPV positive cases. HPV infection rate is higher in SCC patients than in AC, indicating a closer relationship between SCC and HPV infection, this conclusion is also supported in previous studies [19][20][21] . Recently, an international group of gynaecological pathologists endeavoured to establish the International Endocervical Adenocarcinoma Criteria and Classi cation (IECC), a scheme supported by aetiology and based on the presence or absence of HPV infection related morphological features [22] . Some studies then found that worse disease-free survival (DFS) and diseasespeci c survival (DSS) correlated with non-HPV related adenocarcinoma (NHPVA) group [23] , which alert clinicians to take great care of AC with HPV negative results.
Regardless of histopathological types, most patients in this study came in with symptoms. Vaginal bleeding and abnormal vaginal discharge were the main clinical manifestations. Vaginal discharge was much more common in AC than in SCC, reminding clinicians to be alert of the possibility of AC when patients come in for vaginal discharge [24] . Interestingly, the percentage of patients that came in for abnormal cervical screening results in AC was signi cantly higher than in SCC. Figure 2 further compares the frequency of cervical screening before diagnosis in different histopathological types in different stages. Results show that whether irregular screening, regular screening or frequent screening, participation rates of cervical screening before diagnosis of AC were all signi cantly higher than that of SCC. The difference in screening frequency between early stage AC and advanced AC is not signi cant.
Combined with the characteristics of chief complaints of different histopathological types from Table 1 and Table 2, we suspect that the e cacy of the present cervical cancer screening methods is lacking in detection of AC, regardless of the cancer stage. Patients with squamous intraepithelial lesion could be detected promptly by cervical cancer screening and intervention could take place early on to prevent cancer development. As for patients with glandular intraepithelial lesion, a large proportion of AC patients were found to have had attended screening several times before but still could not be diagnosed timely. Recent study also found that being screened with normal results in two screening cycles was associated with 89% risk reduction for squamous cell cancer, compared to women unscreened, but only 60% reduction for adenocarcinoma. [22] Why does the screening method at present show different e cacy in different histopathological types? The reasons may be as follows: Firstly, the precancerous lesions of different histopathological types of cervical cancer have different primary sites. Glandular intraepithelial lesion (GIN), the precancerous lesion of AC, occurs in the depth of the cervical scale-column junction, where the cancer cells are arranged in an interlaced pattern. Thus, fewer abnormal adenocarcinoma cells may be in the cervical exfoliated cells, resulting in di culty in obtaining abnormal AC cells by cervical brush. Therefore, cytological screening has poorer detection on AC [26,27] . In contrast, cervical squamous cell carcinoma is easier to obtain abnormal exfoliated cells, which can be detected by cytological screening of precancerous lesions, so as to effectively prevent squamous cell carcinoma. Secondly, the heterogeneous changes in the nucleus in cervical adenocarcinoma, especially in early stage adenocarcinoma, is not as signi cant as that in squamous cell carcinoma. In particular, highly differentiated mucinous adenocarcinoma often leads to neglect of lm reading [28] . Thirdly, the HPV infection rate varies for different histopathological cervical cancers. In this study, we found the HPV infection rate higher in SCC than in AC (Table 1) [21] . In spite of the higher HPV infection rate, some research found that SCC also has a higher HPV viral load than AC [21] . Since the low viral infection rate in AC and the viral DNA integration in host cells, HPV virus is not easy to be detected in AC specimens [29] . Hence, to improve the detection accuracy of cervical cancer, co-testing (HPV and cytology testing) in screening program is an optimal choice. Moreover, further study of etiology and pathological development of cervical adenocarcinoma, as well as more e cient cervical screening methods, such as improved techniques of HPV DNA testing to con rm the presence or absence of HPV infection, reformed methods of brushing abnormal AC cells, and arti cial intelligent cervical screening program are needed.
In our study, the proportions of the stages of which SCC and AC were presented in were quite different. The percentage of stage I in SCC were lower than that in AC, but the percentage of stage II-IV in SCC were signi cantly higher than AC. We suspect that may be associated with the frequency of cervical screening. Figure 2 shows that the percentage of SCC that had never been screened was 97.2 %, which is signi cantly higher than that of AC (78.8%). Patients who had never attended screening before always sought medical attention because of symptoms, by which time the disease had always progressed to an advanced stage. And that may explain why the percentage of SCC in advanced stage is higher than early stage SCC. Table 2 shows that the percentage of AC patients who sought treatment for abnormal screening results in early stage was 31.6%, and 9.1% in advanced stage (p=0.07), which may suggest that while screening methods at present may not be sensitive enough for adenocarcinoma, regular or multiple screening is still of great signi cance for early detection. The reason why there was no signi cant difference may be related to the sample size. However, more evidence and lager studies are still needed to investigate whether different pathological types of cervical cancer have different cancer staging.

Conclusion
In conclusion, we found many consistent evidences for a relationship between the participation of cervical screening and different histopathological types of cervical cancer. Compared to SCC, HPV infection was found to be less closely associated with AC. The participation rate and frequency of cervical screening of patients with AC was more than SCC. The screening methods at present may not be sensitive enough for precancerous lesion of adenocarcinoma, and regular and multiple screening are still of great signi cance for early detection. In order to improve the detection accuracy of cervical cancer, cotesting (HPV and cytology testing) in screening program is an optimal choice. Further study of etiology and pathological development of cervical adenocarcinoma as well as more e cient cervical screening methods like reformed techniques of HPV DNA testing and cytological examination, arti cial intelligent cervical screening program are needed to be studied to decrease the incidence of cervical adenocarcinoma. This study was approved by the medical ethics committee of the clinical research center of Third A liated Hospital of Sun Yat-sen University. Due to anonymous analyses of the data, the committee concluded that informed consent was not required. The management and publication of patients' information in this study was strictly in accordance with the Declaration of Helsinki, including the con dentiality and anonymity.

Consent for publication
Not applicable.

Availability of data and materials
The data that support the ndings of this study are available from the corresponding author upon reasonable request.
Competing interests None declared.

Funding
No dedicated funding source was involved in this project.
Authors' contributions J L, MJ Y, HY and J H collected and analyzed patients data. J L and MJ Y were major contributors in writing this paper. All authors reviewed and edited the manuscript. All authors read and approved the manuscript.

Figure 1
Grouping based on the participation in cervical screening.