Article Text

Download PDFPDF

Original article
Participation and yield of a population-based colorectal cancer screening programme in China
  1. Hongda Chen1,
  2. Ni Li1,
  3. Jiansong Ren1,
  4. Xiaoshuang Feng1,
  5. Zhangyan Lyu1,
  6. Luopei Wei1,
  7. Xin Li1,
  8. Lanwei Guo1,
  9. Zhaoxu Zheng2,
  10. Shuangmei Zou3,
  11. Yueming Zhang4,
  12. Jiang Li1,
  13. Kai Zhang5,
  14. Wanqing Chen1,
  15. Min Dai1,
  16. Jie He6
  17. the group of Cancer Screening Program in Urban China (CanSPUC)
  1. 1 Office of Cancer Screening, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
  2. 2 Department of Colorectal Surgery, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
  3. 3 Department of Pathology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
  4. 4 Department of Endoscopy, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
  5. 5 Department of Cancer Prevention, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
  6. 6 Department of Thoracic Surgery, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
  1. Correspondence to Professor Min Dai, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College , Beijing 100021, China; daimin2002{at}hotmail.com and Professor Jie He, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China; hejie{at}cicams.ac.cn

Abstract

Objective Colorectal cancer (CRC) screening has been widely implemented in many countries. However, evidence on participation and diagnostic yield of population-based CRC screening in China is sparse.

Design The analyses were conducted in the context of the Cancer Screening Program in Urban China, which recruited 1 381 561 eligible participants aged 40–69 years from 16 provinces in China from 2012 to 2015. 182 927 participants were evaluated to be high risk for CRC by an established risk score system and were subsequently recommended for colonoscopy. Participation rates and detection of colorectal neoplasms in this programme were reported and their associated factors were explored.

Results 25 593 participants undertook colonoscopy as recommended, with participation rate of 14.0%. High level of education, history of faecal occult blood test, family history of CRC and history of colonic polyp were found to be associated with the participation in colonoscopy screening. Overall, 65 CRC (0.25%), 785 advanced adenomas (3.07%), 2091 non-advanced adenomas (8.17%) and 1107 hyperplastic polyps (4.33%) were detected. Detection rates of colorectal neoplasms increased with age and were higher for men. More advanced neoplasms were diagnosed in the distal colon/rectum (65.2%). Several factors including age, sex, family history of CRC, dietary intake of processed meat and smoking were identified to be associated with the presence of colorectal neoplasms.

Conclusion The diagnostic yield was not optimal using colonoscopy screening in high-risk populations given the relatively low participation rate. Our findings will provide important references for designing effective population-based CRC screening strategies in the future.

  • colonoscopy
  • colorectal cancer
  • colorectal adenomas
  • cancer epidemiology
  • colorectal cancer screening
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known on this subject?

  • Screening for colorectal cancer has been proven effective in reducing the burden of this disease.

  • Colonoscopy is regarded as the gold standard for colorectal cancer screening, but the acceptance in general population is not optimal.

  • Evidence of adherence to screening colonoscopy and diagnostic yield in large-scale, population-based colorectal cancer screening programmes in China is sparse.

What are the new findings?

  • The adherence to screening colonoscopy in this high-risk population is relatively low, and factors associated with the willingness to accept colonoscopy screening were identified.

  • Colonoscopy is effective in terms of detecting colorectal cancer and precursors, and for both proximal and distal colonic lesions.

  • Risk factors associated with colorectal neoplasm in Chinese population were identified.

How might it impact on clinical practice in the foreseeable future?

  • Health promotion campaigns and adopting non-invasive screening tests such as faecal occult blood tests will be helpful for improvement of adherence and diagnostic yield in population-based screening programmes.

  • Participant rate and yield of colorectal cancer screening programmes will provide important references for evaluating the effectiveness and cost-effectiveness of cancer screening in China.

Introduction

Colorectal cancer (CRC) is the third most commonly diagnosed cancer in men and the fourth most commonly diagnosed cancer in women worldwide.1 With age-standardised incidence rate of 17.52 per 100 000 and age-standardised mortality rate of 7.91 per 100 000, respectively, in 2014, CRC is the fifth most commonly diagnosed cancer and the fifth most common cause of death by cancer in China.2 For the past decades, the incidence and mortality rates of CRC showed an increasing trend in China, and higher incidence and mortality rates were observed in urban areas than in rural areas.3 The trend could be explained by changes in diet, adoption of a Westernised lifestyle and deficiency in cancer early detection and early treatment.3 4

Randomised controlled trials and observation studies have shown that endoscopic or stool-based screening carries large potential for reducing the burden of this disease.5–7 To date, organised and opportunistic screening programmes have been widely implemented in many countries.8 However, such screening programmes were mainly implemented in high-income countries, which typically have high incidence rates of CRC. Colonoscopy is regarded as the gold standard of CRC screening, but it is an invasive procedure requiring high level of expertise. In countries having intermediate or low incidence rates of CRC and limited healthcare resources, risk-stratified scoring system was recommended to select high-risk patients for colonoscopy.9 However, evidence on the effectiveness of such strategies combining risk stratification and subsequent colonoscopy in population-based screening programmes is still sparse. In October 2012, the government of China initiated the population-based Cancer Screening Program in Urban China (CanSPUC), which targeted five types of cancer that are most prevalent in urban areas, including lung cancer, female breast cancer, liver cancer, upper digestive tract cancer (oesophageal cancer and gastric cancer) and CRC. Eligible participants are recruited in the communities of the study regions and invited to undertake cancer screening free of charge. Participants are first invited to take a cancer risk assessment by an established Clinical Cancer Risk Score System, and those who are evaluated to be high risk for specific types of cancer are recommended to take appropriate screening intervention per study protocol. For CRC screening, participants of high risk for CRC are recommended to take subsequent colonoscopy at tertiary-level hospitals designated by the programme.

For the present study, we report the results of CRC cancer screening conducted in the first 3 years of this cancer screening programme in China between October 2012 and October 2015. We aim to provide timely evidence on the participation and diagnostic yield of colonoscopy screening in high-risk populations in China and to provide important references for designing effective CRC screening strategies in the future.

Methods

Study design and study population

We performed a cross-sectional study under the framework of CanSPUC. CanSPUC is an ongoing national cancer screening programme which was initiated in October 2012. Briefly, residents aged 40–69 years old living in the selected communities of the participating cities were approached by trained staff by means of phone calls and personal encounter. Social media and community advertisement were used to raise public awareness of this cancer screening programme. After obtaining signed written informed consent, all the eligible participants were interviewed by trained staffs to collect information about their exposure to risk factors and to evaluate their cancer risk using an established risk score system. For the present screening programme, to optimise use of the limited healthcare resources and to enhance the detection rate of colorectal neoplasia, only participants who were assessed to be at high risk of CRC were recommended to undergo colonoscopy examination at a tertiary-level hospital designated by the programme free of charge.

For the present analyses, we used the data of the CRC screening conducted in the first 3 years between October 2012 and October 2015, which covered a total of 22 cities in 16 provinces. A map of the study sites is shown in online supplementary figure S1. Overall, there were 1 381 561 eligible participants recruited. After excluding participants with invalid risk assessment results (n=1071) and those not at high risk for CRC (n=1 197 563), 182 927 remaining participants were included in the present study. A flow diagram showing the recruitment of study population is shown in figure 1.

Figure 1

Flow diagram of participant recruitment in CanSPUC, 2012–2015. CanSPUC, Cancer Screening Program in Urban China.

Risk assessment

Participants were required to undertake risk assessment before colonoscopy. The rationale of the development of the cancer risk score system basically followed the Harvard Risk Index,10 but the included risk factors, relative risks and exposure rates of risk factors were adjusted according to the characteristics of the Chinese population. Briefly, the following factors were included in the risk score system: body mass index (BMI), dietary intake of whole grains, dietary intake of fresh vegetables, dietary intake of processed meat, habit of high-fat diet, history of gallstones, history of chronic colitis, family history of CRC in first-degree relatives, history of faecal occult blood test (FOBT) and history of colonic polyps. Each risk factor was allocated a score by the expert panel based on the magnitude of its association with CRC. The cumulative risk scores were calculated and were then divided by the average risk score in the general population to get the final individual relative risks. Individuals with relative risks over 1.50 were defined as high risk for CRC.

Clinical procedures

All colonoscopies were conducted in the tertiary-level hospital by experienced gastroenterologists (attending physician or above having experiences of endoscopy for at least 5 years). Abnormal findings during colonoscopy were carefully checked under standard clinical procedures and biopsies were collected for further pathology diagnosis. Any findings during colonoscopy were required to be photo-documented. Clinical information such as morphological feature, location (distance from the anus and segment), macroscopic diagnosis and size were collected and documented in the data system. Participants who had inadequate bowel preparation or incomplete colonoscopy were asked to retake the colonoscopy exam to meet the clinical standard for diagnosis.

Outcome ascertainment and quality control

All abnormal findings discovered during colonoscopy were confirmed by the pathological examination following up-to-date clinical guidelines. Pathologists were required to complete the highly standardised forms to collect pathology results. For difficult cases, consultation by the expert panel from the National Cancer Center of China was conducted and review reports were transferred to the respective physicians about the consultation results.

In this study, advanced adenomas are defined as at least one adenoma ≥10 mm or at least one adenoma with villous components or high-grade dysplasia. In addition, we categorised advanced neoplasms (CRC and advanced adenomas) into proximal colon and distal colon locations. If multiple similarly advanced lesions were present, the largest one was used to categorise the location. The proximal colon was considered to include the splenic flexure and all segments proximal to it, and the distal colon and/or rectum was considered to include all other segments.

Data acquisition

Paper-based standardised documentation forms (epidemiological questionnaire, colonoscopy report, pathology report and so on) were collected from trained staff and physicians. The validity of forms was checked and entered into the data management system by trained study staff. A consistency check was conducted, and mistakes were corrected by retrieving the original records if inconsistencies were identified. Each participant had a unique identification code which was used to track all of the individuals’ relevant documentation forms. All data were transmitted to the Central Data Management Team in the National Cancer Center of China, where the databases were constructed and analyses were performed.

Statistical analysis

In addition to the descriptive analyses regarding the characteristics of the study population, overall and group-specific participation rates (PRs) by common factors were calculated and respective 95% CIs were reported. Differences in PRs between different groups were compared using the χ2 test. Associations of factors—including age (categorised into 40–44, 45–49, 50–54, 55–59, 60–64 and 65–69 years), sex (male, female), educational background (low: primary school or below; intermediate: primary school to high school; high: high school or above), history of FOBT result (no, yes with positive results, yes with negative results, yes but results unknown), family history of CRC (yes, no) and previously detected colonic polyp (yes, no)—with PR in colonoscopy were quantified by ORs and their 95% CIs, which were derived by multiple logistic regression models after adjusting for factors including ethnicity, occupation, smoking and alcohol drinking. Diagnostic yield of screening colonoscopy—including colonic findings, age and sex detection rates, distribution of location of neoplasms—was calculated. Yield per 10 000 invitees and the number of colonoscopies to detect one colonic lesion were calculated. Associations of various characteristics with prevalence of colorectal neoplasms were likewise quantified by ORs and their 95% CIs using logistic regression models. All statistical analyses were performed with the R V.3.4.1 statistical software. All tests were two-sided and p values of 0.05 or less were considered statistically significant.

Results

Characteristics of the study population

As shown in figure 1, there were 1 381 561 eligible participants recruited in the CanSPUC in 2012–2015. After excluding participants with incomplete risk assessment questionnaire (n=1071) and those with risk assessment results showing not high risk for CRC (n=1 197 563), 182 927 remaining participants of high risk for CRC were included in the final analyses. The characteristics of the high-risk population of CRC are presented in table 1. Overall, more women (56.1%) were included in the study. The mean age was 54.7 years (SD=7.8 years), and the majority (68.1%) were between 40 and 69 years old. One-fifth of the high-risk population received FOBT previously and 37.2% of them (n=13 721) had positive FOBT results. About 23% of the participants (n=41 901) had family member being diagnosed with CRC and around a quarter (n=44 265) had previously detected colonic polyps.

Table 1

Characteristics of the study population and participation rates between different groups

PR in screening colonoscopy and its associated factors

For the 182 927 participants of high risk for CRC, 25 593 of them undertook colonoscopy as recommended by our study. The overall PR was 14.0% (95% CI 13.8% to 14.2%). The PRs stratified by potential associated factors are shown in table 1. Overall, the PRs were similar among men and women (14.1% and 13.9%, respectively) and were higher among participants aged 45–64 years than other ages. Univariate analyses showed that participants with a high educational background, history of positive FOBT, family history of CRC and history of colonic polyps had relatively higher PRs. We also conducted multivariable logistic regression models to explore the potential factors that were associated with PR, and the results are shown in table 2. With adjustment for factors including ethnicity, smoking and alcohol drinking, we found that age, educational background, history of FOBT, family history of CRC and previously detected colonic polyp were associated with PR. For instance, the odds of participants with a family history of CRC undertaking screening colonoscopy were 0.8-fold higher than participants with no family history of CRC (OR: 1.80, 95% CI 1.74 to 1.86). Participants with previously positive FOBT results were more willing to accept colonoscopy compared with participants who never received FOBT, with OR of 1.29 (95% CI 1.22 to 1.35). As PR varied between the study site and year of participant recruitment, these two factors were additionally adjusted in model II and the respective ORs did not change greatly (see table 2).

Table 2

ORs of factors associated with participation rate in colonoscopy in the screening programme

Colonic findings under screening colonoscopy

Table 3 presents the diagnostic yield of colonoscopy in our screening programme. Overall, there were 65 CRC, 785 advanced adenomas, 2091 non-advanced adenomas and 1107 hyperplastic polyps, yielding the detection rates for CRC, advanced adenoma, non-advanced adenoma and hyperplastic polyps at 0.25%, 3.07%, 8.17% and 4.33%, respectively. Furthermore, based on the sex-adjusted and age-adjusted detection rates by standard population of China (1982), we calculated that the numbers of colonoscopies to detect one CRC, advanced adenoma and non-advanced adenoma were 435, 39 and 54, respectively. In terms of diagnostic yield per invitee, 4 CRC, 43 advanced adenomas and 114 non-advanced adenomas could be detected per 10 000 invitees.

Table 3

Colonic lesions detected by colonoscopy in the screening programme

The detection rates for advanced neoplasms, non-advanced adenomas and any neoplasms increased with age and were higher for men than for women, as shown in figure 2. For instance, the detection rate of advanced neoplasms for men aged 65–69 years was 8.92% (95% CI 7.57% to 10.48%), significantly higher than the respective rate for women at the same age range (detection rate: 4.36%; 95% CI 3.45% to 5.50%).

Figure 2

Detection rates of advanced neoplasia, non-advanced adenoma and any neoplasms stratified by age and sex.

Regarding the anatomical location of CRC and advanced adenomas, more neoplasms were identified in the distal colon than in the proximal colon. In particular, 78.8% of CRC, 64.2% of advanced adenomas and 66.4% of advanced adenomas ≥10 mm were located in the distal colon. The distribution of anatomical location of advanced colorectal neoplasms is shown in figure 3.

Figure 3

Anatomical location of colorectal cancers and advanced adenomas detected by colonoscopy in the screening programme.

Factors associated with the neoplasm detection of screening colonoscopy

We further performed multivariate logistic regression analyses to identify potential risk factors that may be associated with colorectal neoplasms in the Chinese population, and the results are shown in figure 4 and online supplementary table S1. Older age, being male and having a family history of CRC were identified to be positively associated with both advanced colorectal neoplasms and any colorectal neoplasms. For instance, compared with individuals aged 40–44 years, the ORs (95% CI) of individuals aged 45–49, 50–54, 55–59, 60–64 and 65–69 years carrying advanced neoplasia were 1.46 (0.92 to 2.29), 2.19 (1.48 to 3.33), 2.60 (2.47 to 5.42), 4.24 (2.92 to 6.37) and 6.20 (4.24 to 9.37), respectively. High levels of dietary intake of processed meat and smoking were only positively associated with non-advanced colorectal neoplasms. Other factors including previously detected colonic polyps, history of FOBT, dietary intake of coarse grain food, dietary intake of fresh vegetables and alcohol consumption were found to have no association with the presence of colorectal neoplasms in our study population.

Figure 4

OR of risk factors associated with (A) advanced neoplasms and (B) any colorectal neoplasms. Analyses were adjusted for age, sex, body mass index (BMI), family history of colorectal cancer (CRC), previously detected colonic polyp, previous faecal occult blood test results, dietary intake of coarse grains, fresh vegetables and processed meat, smoking, alcohol drinking, ethnicity, educational background and year of recruitment.

Discussion

We reported the results of 182 927 participants undertaking CRC screening among a high-risk urban population in China. The analyses showed that the overall PR in screening colonoscopy among high-risk population was relatively low (14.0%, 95% CI 13.8% to 14.2%). For 25 593 participants undertaking colonoscopy, the detection rates for CRC, advanced adenoma and non-advanced adenoma were 0.25%, 3.07% and 8.17%, respectively. Higher detection rate was observed for men than women and increased with age. The majority of neoplasms were detected in the distal colon. Factors including age, sex, BMI, family history of CRC, dietary intake of processed meat and smoking were found to be associated with the presence of colorectal neoplasms. To our knowledge, our study is the first to present the participation and diagnostic yield of CRC screening using a strategy combining risk score stratification and colonoscopy based on the results from a large-scale cancer screening programme in China. Although great efforts have been made by previous studies to develop effective risk score systems, most studies mainly focused on optimisation of risk scores and few were truly validated in large-scale CRC screening programmes, especially in Asian populations.11 12

Colonoscopy is the gold standard for CRC screening which allows detection and removal of adenomas. However, the acceptance of colonoscopy in the general population is not optimal and is still a major challenge in population-based CRC screening,13 14 especially for screening programmes using colonoscopy as the main screening modality. For instance, in a randomised controlled trial conducted in four European countries (Poland, Norway, the Netherlands and Sweden), the PRs in colonoscopy ranged from 22.9% (the Netherlands) to 60.7% (Norway).15 Other CRC screening cohorts using FOBT as the main screening modality typically had higher adherence rates.16–18 For instance, in a randomised controlled trial conducted in Spain, the PRs of colonoscopy group and faecal immunochemical test group were 34.2% and 24.6%, respectively (p<0.001).17 A recent study from England reported that the PR using FOBT even increased to 66.4% in the National Health Service Bowel Cancer Screening Programme.16 Our analyses showed that previous FOBT was associated with increased uptake of subsequent colonoscopy. As detailed information such as the type of FOBT and the interval since the last one was not collected in our study, further studies are required to illustrate the role of FOBT in improving the uptake of CRC screening in China.

Previous studies suggested that factors such as individuals’ awareness and knowledge about CRC screening, physicians’ knowledge, attitude and recommendation, and logistic and financial considerations were key factors for a successful CRC screening uptake.19 20 In the CanSPUC, the overall PRs in other four types of cancer were much higher than CRC screening (data not shown in the results). The underlying reasons were probably due to the invasive nature of colonoscopy, requirement of bowel preparation, and poor awareness and knowledge about CRC screening. However, factors that were associated with the non-participants were not evaluated in our study and need to be further explored. Our results imply that public awareness campaigns would be necessary to improve PR in CRC screening in the future.

The overall adenoma detection rate (ADR) in our study population was 11.49%, which was lower than some studies conducted in Europe and North America. For instance, in a screening cohort with 9989 participants undertaking screening colonoscopy in the USA, the overall ADR was 37.2%.21 In another study using 4 407 971 opportunistic colonoscopy screening records from Germany in 2003–2012, the ADRs were 31.3% and 20.1% for men and women in 2012, respectively.22 ADR has been regarded as a primary benchmark of colonoscopy.23–26 A recent guideline from the USA (American College of Gastroenterology and the American Society for Gastrointestinal Endoscopy) recommended that the target ADR should be above 25% in a population aged ≥50 years undertaking screening colonoscopy.25 However, it should be noted that such target rate was defined using evidence mostly from the Western populations, which typically have higher prevalence of colorectal neoplasms.23 24 According to up-to-date statistics, the incidence rate of CRC in the USA (46.9 per 100 000 for men and 35.6 per 100 000 for women) is almost twofold higher than China (28.64 per 100 000 for men and 19.33 per 100 000 for women).2 27 Therefore, the low ADR in our study might be explained by the relatively low prevalence of colorectal neoplasm in China than other European and North American countries.

Our study shows that several sociodemographic factors including age, sex, BMI, family history, dietary intake of processed meat and smoking were associated with the presence of colorectal neoplasm. The associations of these factors with colorectal neoplasms have been extensively explored and our findings were in line with previous studies.28–31 As some factors (such as family history) were also included in the risk score system to select high-risk population, the magnitude of association might be underestimated. However, such analyses are indispensable to validate preincluded factors and explore new risk factors, with the purpose of further optimising the risk assessment model for future research.

It deserves to be noted that in our study we found that the overall detection rate for colorectal neoplasms was low in a high-risk population in urban China. Given the relatively low PR in screening colonoscopy, detection rates were even lower when calculated per invitee. To further improve the screening yield of CRC screening in China, the following issues need to be addressed in next-step research: (1) to optimise the risk assessment score based on the current study findings and other well-established risk prediction socres11 32 33; (2) to explore the role of non-invasive tests (such as FOBT) as supplement to screening colonoscopy; (3) to design novel risk-adapted screening strategies covering both high-risk and low-risk populations using appropriate screening modalities; and (4) to carry out multifactor interventions targeting multiple levels of care with the purpose of optimising CRC screening acceptance.

Specific strengths and limitations deserve careful attention when interpreting our results. A major strength of our study is the fact that our analyses were the first to illustrate the participation and diagnostic yield of screening colonoscopy in a large-scale, population-based cancer screening programme in China. Furthermore, detailed patient information including epidemiological questionnaire and clinical examination data (colonoscopy and pathology examination) were collected in a standardised manner by trained study staff to ensure the quality of data. Capacity training and central review of colonoscopy and pathology reports by expert panel were also conducted yearly to enhance the consistency and accuracy of clinical diagnoses. However, limitations include that although our study population was selected from 16 provinces, our study is not representative of the entire general population of China and therefore selection bias cannot be ruled out. Second, given that follow-up work for patients diagnosed with CRC is still under way, clinical disease information has not been fully obtained. Therefore, tumour stage information was not reported in our study.

In summary, in this large-scale CRC screening programme in China, we found that the diagnostic yield was not optimal using colonoscopy screening in high-risk populations given the relatively low PR. Further efforts to optimise the screening strategy by incorporating non-invasive tests (such as FOBT) and to enhance the screening acceptance by health promotion campaign are highly required. We further identified several factors associated with PR in screening colonoscopy and risk factors of colorectal neoplasms. Our findings will provide important references for designing effective population-based CRC screening strategies in the future.

Acknowledgments

We sincerely thank all the members of the Cancer Screening Program in Urban China from the National Cancer Center of China, provinces and external expert panels. We are also grateful to the participants for attending this study. We sincerely thank Dr. Hermann Brenner for his critical review and in-depth comments in preparation of this manuscript

References

View Abstract

Footnotes

  • HC and NL contributed equally.

  • Contributors Conception and design: HC, MD and JH. Statistical analyses: HC and NL. Data acquisition and data interpretation: HC, NL, JR, XF, ZL, LW, XL and LG. Drafting of the article: HC and NL. All authors revised the manuscript and approved its final version.

  • Funding The study was supported by the National Natural Science Foundation (81703309), CAMS Innovation Fund for Medical Science (2017-I2M-1-006), Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (2017PT32001), and National Key Research and Development Program of the Ministry of Science and Technology of China (2016YFC0905301).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval The study was approved by the Ethics Committee of National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.