Dietary and Circulating Vitamins, Polymorphisms of Vitamin Metabolism Genes, and the Risk of Gastrointestinal Cancers: A Systematic Review and Meta-Analysis.

INTRODUCTION
Gastrointestinal cancers, including cancer of stomach, esophagus, colorectum, liver, and pancreas, rank among the most common cancer types worldwide (1), accounting for 24.0% of the global incident cancer cases and 33.2% of all cancer-related deaths (2). Previous evidence from mechanistic studies and epidemiologic studies has indicated an association of vitamins, both from dietary intake (3–6) and supplementation (7–9), with the risk of gastrointestinal cancers. Moreover, circulating vitamin levels have also been inversely associated with gastrointestinal cancer risk (10,11), suggesting a potential protective effect of vitamins against gastrointestinal cancer.
Vitamins are essential nutrients that play a critical role in the maintenance of genomic stability, particularly in the repair processes of DNA strand break and damaged DNA methylation patterns (12,13). Studies have also shown that vitamin metabolism may be regulated by host genes, such as genes expressing methylenetetrahydrofolate reductase (MTHFR), methionine synthase (MTR), and methionine synthase reductase (MTRR) controlling B-vitamin metabolism (14,15) and vitamin D receptor (VDR) for D-vitamin metabolism (16,17). Considering the complex interrelationship between vitamin metabolism and genetic predisposition, the inverse associations of dietary and circulating vitamins with gastrointestinal cancer risk may be modified by the variations of essential genes involved in vitamin metabolism (18–22). In line with the evidence from basic research, molecular epidemiologic studies have demonstrated that the associations of vitamins with gastrointestinal cancer risk seemed particularly evident for individuals carrying specific genotypes (23–27), but previous studies have yielded inconsistent findings (28–30). Two meta-analyses examined the association of dietary folate intake with the risk of esophageal cancer (EC) and colorectal cancer (CRC), respectively, for individuals with different MTHFR C677T polymorphisms, but both were published more than 1 decade ago and no other meta-analyses were available for dietary and circulating vitamins and the risk of gastrointestinal cancers by genetic variations (31,32).
Given the knowledge gap, it remains to be clarified whether heterogeneity in the associations between vitamin levels and the risk of gastrointestinal cancers exists for individuals with different genetic background. In our study, we conducted a systematic review and meta-analysis of published data to determine whether genetic polymorphisms involved in vitamin metabolism would modify the associations of dietary and circulating vitamin levels with the risk of gastrointestinal cancers.

METHODS
This study was conducted according to the requirements of the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement (33). The PRISMA checklist is provided in Supplementary Table S1 (see Supplementary Digital Content 1, http://links.lww.com/CTG/B353). The protocol was registered in PROSPERO (https://www.crd.york.ac.uk/prospero/) with the identifier number of CRD42024545415.

Search strategy
We systematically reviewed the PubMed, Embase, and Web of Science databases from inception until August 1, 2024. Search terms included “gene*,” “polymorphism,” “SNP,” and major gastrointestinal cancers including gastric, esophageal, colorectal, liver, and pancreatic cancers. Detailed search strategies are provided in the Supplementary Table S2 (see Supplementary Digital Content 1, http://links.lww.com/CTG/B353). The literature search systematically filtered non-English publications and studies involving nonhuman species.

Inclusion and exclusion criteria
Eligible studies examined the associations between vitamins, genetic polymorphisms, and gastrointestinal cancer risk. Eligibility assessment was performed independently by 2 independent researchers, with discrepancies resolved by a third investigator. The literature search was strictly limited to case-control, nested case-control, and cohort studies. Case studies, case reports, Mendelian randomization analyses, editorials, conference abstracts, and reviews (including meta-analysis) were excluded. Given that our primary aim was to ascertain whether genetic polymorphisms implicated in vitamin metabolism could modify the associations between dietary or circulating vitamin levels and the risk of gastrointestinal cancer, we confined our inclusion criteria to studies that provided information on dietary vitamin intake or circulating levels (in serum/plasma) and vitamin metabolism genetic polymorphisms. These studies were required to report effect estimates (e.g., odds ratios [ORs] and 95% confidence intervals [CIs]) for major gastrointestinal cancer (CRC, EC, gastric cancer [GC], liver cancer, and pancreatic cancer) associated with vitamins across different genotypes or had required data to yield such effect estimates (Figure 1).
As a secondary aim, we also examined the associations between dietary or circulating vitamin levels and the risk of gastrointestinal cancers. Studies without extractable effect estimates suitable for meta-analysis were excluded from this analysis (see Supplementary Figure S1, Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

Data extraction
Information on authors, publication year, type of study, location, cancer type, vitamin type, vitamin exposure, genetic variants, risk allele, and effect values were independently extracted by 2 investigators. Discrepancies were resolved through cross-checks and consultation with a third researcher. Only studies that presented effect estimates with reference categories compatible with those of other studies, or studies for which the reference categories could be harmonized by transforming the reference using available sample distribution data, were included in the meta-analysis.

Study quality assessment
The quality of included studies was assessed using the Newcastle-Ottawa Scale (34), including 8 items across 3 key domains of Selection, Comparability, and Exposure (case-control studies) or Outcome (cohort studies) (see Supplementary Table S3, Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

Statistical analysis
The meta-analysis was conducted by pooling effect estimates for the association of gastrointestinal cancer risk with high dietary or circulating vitamin level category, with the low-level category as the reference. The meta-analyses stratified by genetic polymorphisms were only conducted if there were at least 3 studies examining the interaction between a specific vitamin (either circulating or dietary) and genetic polymorphism on gastrointestinal cancer risk.
To investigate the heterogeneity among studies, we calculated I2 and tested the uncertainty of the heterogeneity using a χ2 test. Depending on the degree of heterogeneity between studies, a fixed-effect model was used when I2 ≤ 50% and a random-effects model otherwise. Heterogeneity tests were also performed for the associations between individuals with different genotypes of a polymorphism. Subgroup analyses by continent of study origin (America, Asia, or Europe) and study quality were conducted to identify potential sources of heterogeneity.

Statistical analysis on publication bias
To evaluate the publication bias, funnel plots were generated and Egger's tests were conducted for each meta-analysis, with a P-value of <0.05 indicating significant publication bias. In case of asymmetric funnel shape, Duval and Tweedie “trim-and-fill” analysis (35) was performed, which detects putative missing studies to rebalance the distribution. This analysis provides an adjusted pooled estimate taking the additional studies into account, thus correcting the analysis for publication bias. We further conducted 2 sets of sensitivity analyses to assess the robustness of the results. One was conducted by using the leave-one-out method to assess the influence of individual studies, and the other was conducted to exclude studies with a sample size of fewer than 300 participants, based on the distribution of sample sizes.
All meta-analyses were performed using the “meta” package in R 4.1.3. Detailed methods are presented in the Supplementary Materials and Methods.

Certainty of evidence
Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) criteria (36) were used to construct evidence profiles for the certainty of evidence of all analyses. Detailed methods are presented in the Supplementary Materials and Methods (see Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

RESULTS

Study characteristics
Our literature search resulted in a total of 4,183 records, with 3,715 records left after the initial filtering, which was further reduced to 3,148 unique studies after removal of duplicates. Among these, 2,871 studies were excluded based on title/abstract screening. After application of the inclusion criteria, 64 studies were finally included (Figure 1 and Table 1) (20,23–25,27,29,30,37–50) (51–80) (81–93), including 17 from the America, 32 from Asia, 14 from Europe, and 1 from Oceania. Of them, 59 were case-control studies, 4 were nested case-control studies, and 1 was cohort study. These studies primarily centered on 3 gastrointestinal tumors, specifically CRC (52 studies), GC (7 studies), and EC (6 studies), with 1 study concurrently examining both GC and EC. No enrolled studies were conducted on pancreatic cancer or liver cancer.
Most of the enrolled studies was on vitamin B complexes (47/64 studies, 73.4%), which investigated specific B vitamins including folate, vitamin B2, B6, and B12. Of them, a substantial portion (36 studies) examined vitamin B intake, whereas a significant minority (13 studies) explored circulating levels of vitamin B. Two studies concurrently assessed both dietary and circulating levels of vitamin B. Furthermore, the research encompassed 14 studies on vitamin D, with 8 studies on vitamin D intake and 6 studies on circulating vitamin D levels. In addition, 2 studies focused solely on vitamin E, 1 study on vitamin A, but none on vitamin C. Based on these studies, associations of dietary or circulating levels of vitamin B and vitamin D with the risk of gastrointestinal cancers were examined.
Among the 64 enrolled studies, a total of 40 single nucleotide polymorphisms in 18 vitamin metabolism genes were involved. As the primary aim of our study, meta-analyses were conducted by genetic polymorphisms that had 3 or more studies on a particular type of gastrointestinal cancer and vitamin. A total of 32 articles were eligible for stratified meta-analysis, involving 4 polymorphisms of vitamin B metabolism genes (MTHFR C677T, MTHFR A1298C, MTRR A66G, and MTR A2756G), and 4 polymorphisms of vitamin D metabolism genes (VDR ApaI, BsmI, FokI, and TaqI) (see Supplementary Table S4, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). The excluded 32 studies from this meta-analysis were not systematically different in study design, continent of study origin, or quality from those included (see Supplementary Table S5, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). In addition, the effect estimates of associations between the genetic polymorphisms not included in the meta-analysis and their interactions with vitamins in original studies are provided in Supplementary Table S6 (see Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

Association between dietary and circulating vitamins and the risk of gastrointestinal cancers
A total of 49 studies were eligible for the meta-analysis on the association between vitamins and gastrointestinal cancer risk. Individuals receiving high level of vitamin B intake had a notably decreased risk of CRC (OR 0.87, 95% CI 0.81–0.94), GC (OR 0.61, 95% CI 0.53–0.71), and gastrointestinal cancer overall (OR 0.84, 95% CI 0.79–0.90), and a nonstatistically significant reduction in EC risk (OR 0.73, 95% CI 0.53–1.01). Among vitamin B subclasses, a significant inverse association was found for dietary intake of folate, vitamin B2, and vitamin B6, and the inverse association generally persisted across individual gastrointestinal cancers (see Supplementary Table S7, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). Meta-analysis on circulating vitamin B complex yielded inverse associations with the overall and individual gastrointestinal cancer risk (see Supplementary Table S7, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). Among its subclasses, circulating folate, vitamin B2, and vitamin B6 were generally inversely associated with the risk of gastrointestinal cancers overall or the specific cancer type, although some associations were not statistically significant. Neither dietary nor circulating levels of vitamin B12 was associated with an altered risk of gastrointestinal cancers overall or CRC risk, but an inverse association for EC was observed based on only 1 available study (OR 0.18, 95% CI 0.07–0.42).
We also examined dietary and circulating vitamin D levels (see Supplementary Table S7, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). Those with a higher vitamin D intake (OR 0.69, 95% CI 0.53–0.90) or circulating vitamin D level (OR 0.74, 95% CI 0.59–0.94) had a significantly decreased risk of gastrointestinal cancers. Aside from CRC, the meta-analysis for GC or EC associated with vitamin D levels was restricted by the limited number of studies.

Association between dietary and circulating vitamin B and CRC risk in populations with different genotypes
We conducted meta-analyses stratified by genetic polymorphisms of the MTHFR C677T, MTHFR A1298C, MTRR A66G, and MTR A2756G. High-level intake of vitamin B was associated with a reduced risk of CRC in individuals carrying different MTHFR C677T genotypes, without statistically significant heterogeneity (P-het = 0.33 in the dominant model and P-het = 0.30 in the recessive model) (Figure 2a and b). In the stratified analysis by the MTHFR A1298C polymorphism, no significant heterogeneity in CRC risk associated with high vitamin B intake was found in the dominant genetic model (P-het = 0.97) (Figure 2c). Nevertheless, individuals harboring the MTHFR 1298 AA/AC genotype exhibited heterogeneity in their association between vitamin B intake and CRC risk based on the recessive genetic model which only enrolled studies on folate intake (P-het = 0.04) (Figure 2d and see Supplementary Table S8, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). No heterogeneity was observed according to the MTRR A66G and MTR A2756G polymorphisms, respectively (Figure 2e–f). Other than folate intake, meta-analyses did not reveal heterogeneity in the associations of other vitamin B subclasses with CRC risk for individuals with different genotypes of these polymorphisms (see Supplementary Table S8, Supplementary Digital Content 1, http://links.lww.com/CTG/B353).
As for circulating vitamin B levels, an inverse association with CRC risk was found for individuals with the MTHFR 677 TT genotype (OR 0.57, 95% CI 0.33–0.97) but not for those possessing the CC/CT genotype (OR 0.98, 95% CI 0.80–1.21), demonstrating a heterogeneity (P-het = 0.06, Figure 3). In addition, 1 original study also conducted stratified analysis based on the recessive genetic model of the MTHFR C677T polymorphism and did not find significant heterogeneity between these 2 genotypes either (71). Owing to the absence of available studies, we were unable to explore the associations of circulating vitamin B levels with CRC risk by the MTHFR A1298C, MTRR A66G, or MTR A2756G polymorphisms.
In subgroup analyses by continent of study origin (see Supplementary Table S9, Supplementary Digital Content 1, http://links.lww.com/CTG/B353) and study quality (see Supplementary Table S10, Supplementary Digital Content 1, http://links.lww.com/CTG/B353), the association between vitamin levels and CRC risk across different genotypes remained generally consistent with the overall findings. An exception was noted for the MTHFR C677T polymorphism under the recessive model. Specifically, significant heterogeneity was observed in the European population only regarding the association between vitamin B intake and CRC risk when comparing the MTHFR 677 CC/CT and TT genotypes (P-het = 0.005) (see Supplementary Table S9, Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

Association between dietary and circulating vitamin D levels and CRC risk in populations with different genotypes
Categorizing the population by the VDR genotypes, high vitamin D intake was associated with a decreased risk of CRC across populations carrying different genotypes of the VDR ApaΙ, BsmΙ, FokΙ, and TaqΙ polymorphisms, without statistically significant heterogeneities (all P-het >0.10) (Figure 4). Based on available literature, we examined the association of circulating vitamin D with the risk of CRC according to the VDR ApaΙ, BsmΙ, FokΙ, and TaqΙ polymorphisms. An inverse association was exclusively yielded in individuals with the VDR TaqI Tt/tt genotype (OR 0.52, 95% CI 0.28–0.95), other than the TT genotype (OR 0.91, 95% CI 0.70–1.19, P-het = 0.10) (Figure 5).

Association between vitamin B and D intake and the risk of GC and EC in population with different genotypes
Eligible studies examined the association between folate intake and risk of GC or EC according to the MTHFR C677T, MTHFR A1298C, MTRR A66G, and MTR A2756G polymorphisms. Such studies were not available for other vitamin B subclasses. Among them, only 1 study was available on the association of folate intake with GC risk according to the MTRR A66G polymorphism, which found an inverse association only among MTRR 66 AA carriers (OR 0.44, 95% CI 0.26–0.74), but not among AG/GG carriers (OR 0.92, 95% CI 0.55–1.53), indicating statistically significant heterogeneity (P-het = 0.05). No such heterogeneity was found across genotypes of the MTHFR C677T, MTHFR A1298C, and MTR A2756G polymorphisms (see Supplementary Table S8, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). In addition, 2 original studies reported significant interactions between the MTHFR C677T polymorphism and folate intake on GC risk (both P-for-interaction <0.05) (75,79), but no data are available for the stratified meta-analysis in our study. For EC, the enrolled studies only examined its association with folate intake by the MTHFR C677T polymorphism, while no significant heterogeneity was observed between different genotypes.
Among included studies on dietary vitamin D, no study was available for EC and only 1 study was available for GC, which reported no significant interaction between vitamin D intake and VDR BsmΙ, rs2239179, and rs4516035 polymorphisms on GC risk (all P-for-interaction >0.10) (30). The available literature precluded from the possibility of examining circulating vitamin B and D levels associated with EC or GC by the abovementioned genetic polymorphisms.

Publication bias
Egger's tests indicated the presence of the publication bias for most comparisons assessing the association between dietary vitamin B, vitamin D, and CRC risk in populations with different genotypes (see Supplementary Figures S2 and S3, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). We further applied the trim-and-fill method, which imputed a small number of missing studies from the distribution. After rebalancing by including additional studies, the adjusted pooled estimates seemed consistent with the original findings (see Supplementary Table S11, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). For circulating vitamin B associated with gastrointestinal cancers by MTHFR C677T genotypes, the estimated number of missing studies was zero, indicating the absence of publication bias among included studies (see Supplementary Figure S4 and Supplementary Table S11, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). Sensitivity analyses using the leave-one-out method and by excluding studies with a sample size showed no material changes in the results, supporting the robustness of our findings (see Supplementary Table S12, Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

Quality assessment of studies
Based on the Newcastle-Ottawa Scale, the average score of included studies was 6.27, ranging from 5 to 8. Notably, 84% of the studies were deemed to be of high quality (see Supplementary Table S3, Supplementary Digital Content 1, http://links.lww.com/CTG/B353). However, past studies suffer from several limitations that could affect the interpretation and generalizability of the findings. Approximately half of the studies used hospital-based selection of control subjects, and furthermore, the majority failed to report the response rate, specifically the proportion of participants who successfully completed the required data collection processes (such as questionnaires, blood sample collection, and genotype sequencing), or presented differing response rates among comparison groups without offering a clear explanation (see Supplementary Figure S5, Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

Certainty of evidence
Based on the GRADE framework, which initially assign observational studies as low-quality evidence, 8 analyses retained their original ratings. We did not downgrade their certainty of evidence as no critical limitations were identified. The quality of other analyses was upgraded as having “low-to-moderate” quality owing to the presence of a dose-response gradient and adjustments for potential biases and confounding factors (see Supplementary Table S13, Supplementary Digital Content 1, http://links.lww.com/CTG/B353).

DISCUSSION
Based on a systemic review and meta-analysis of 64 studies, dietary and circulating vitamin B and vitamin D levels were inversely associated with the risk of major gastrointestinal cancers. The association of CRC risk with dietary and circulating vitamin B levels varied by the MTHFR A1298C and MTHFR C677T polymorphism, respectively, whereas the association between circulating vitamin D level and CRC risk differed based on the VDR TaqI polymorphism, suggesting genetic variations of vitamin metabolism pathways may modify the effect of vitamin levels on gastrointestinal cancer.
Our findings on vitamin B complex and its subclasses associated with gastrointestinal cancers are consistent with previous mechanistic studies. High levels of vitamin B have been shown to enhance DNA methylation, improve DNA repair and replication, and may also decrease cell proliferation and inflammation, thereby reducing the risk of gastrointestinal cancers (6,9,94). Among major vitamin B subclasses, an inverse association was found for dietary or circulating levels of folate, vitamin B2, and vitamin B6 associated with the risk of major gastrointestinal cancer, although the associations were not all statistically significant. However, there was no indication that dietary or circulating vitamin B12 levels may be associated with altered risk of CRC or total gastrointestinal cancers in our meta-analysis. These results are consistent with previous studies (95,96), while the biological mechanism underlying the association of vitamin B12 with gastrointestinal cancer was poorly understood.
In line with previous findings on biological mechanisms, our study offers evidence that polymorphisms in vitamin metabolism genes can affect the relationship between vitamin B, vitamin D, and gastrointestinal cancer risk. Accumulating evidence suggested that vitamin B levels may interfere with genetic variants involved in 1-carbon metabolism, such as the MTHFR, MTRR, and MTR polymorphisms, on gastrointestinal cancer risk (19,24,25). These genetic polymorphisms alter 1-carbon metabolism enzyme activity, influencing folate distribution away from nucleotide synthesis to homocysteine remethylation (97). In our study, the association of CRC risk with dietary and circulating vitamin B levels varied by the MTHFR A1298C and MTHFR C677T polymorphism, respectively. Both of these common genetic polymorphisms affect MTHFR enzyme activity, leading to altered levels of 5-methylenetetrahydrofolate (5-MTHF) and thymidine. With high levels of vitamin B, the recycling pathway from 5-MTHF to 5,10-MTHF would be fully functional, leading to the accumulation of 5,10-MTHF and thymidine and facilitating the subsequent metabolism of the common methyl donor S-adenosylmethionine essential for DNA synthesis and methylation (98,99). These metabolic variations could modify the efficiency of vitamin B coenzymes and affect the utilization of vitamin B, thereby affecting the protective effect of vitamin B intake against colorectal carcinogenesis (100).
Pervious research demonstrated that polymorphisms of VDR gene may interact with vitamin D on CRC risk, though the evidence was not conclusive (23,26–28). We observed a significant heterogeneity in CRC risk associated with circulating vitamin D levels between individuals with the VDR TaqΙ TT and Tt/tt genotypes, suggesting that VDR TaqΙ polymorphism may influence the protective effect of circulating vitamin D against CRC. Genetic factors could modulate the influence of vitamin D levels on VDR enzyme regulation. In this scenario, the response to vitamin D intake differs widely among individuals with different VDR variants (101). The VDR TaqI polymorphism seems to be in linkage disequilibrium with a series of polymorphisms at the 3′ end of the VDR gene, where alterations in this region (such as 3′ untranslated region) often affect transcriptional regulation, mRNA stability, or protein translational efficiency. Consequently, individuals with different genotypes of the VDR TaqI polymorphism may exhibit varying VDR expression and activity, leading to interindividual differences in the biological response to vitamin D intake, and ultimately affecting the risk of CRC (101,102).
Our study has several limitations. First, due to differences in thresholds of vitamin levels across included studies, we simply classified the vitamin levels into groups of “low,” “medium,” and “high” without setting specific cutoffs and were unable to analyze the dose-response association. Second, the studies we included only examined CRC, GC, and EC, with the majority (81.3%, 52/64) focusing only on CRC. The meta-analyses for some vitamin B subclasses associated with EC and GC by proposed genetic polymorphisms were not feasible. We were also not able to examine pancreatic cancer, liver cancer, or other gastrointestinal cancer. Third, our study only focused on vitamin B and vitamin D. Further studies are still warranted to clarify the association of vitamin A, vitamin C, and vitamin E levels associated with gastrointestinal cancer risk in individuals with distinct genotypes. Fourth, in the primary meta-analysis, studies were excluded either due to an insufficient number of studies examining the interaction between a specific vitamin and genetic polymorphism on gastrointestinal cancer risk or because some effect estimates could not be transformed for consistent comparison because of the lack of genotype-specific data on sample distribution. This may have resulted in missing data in the meta-analysis. Fifth, we sought to ascertain could the factors accounting for the observed heterogeneity across studies, which revealed a reduction in heterogeneity for studies of specific subgroups, suggesting that geographical difference and study quality may be potential sources of heterogeneity. However, the heterogeneity persisted within some subgroups. All studies included in the meta-analysis were in case-control or nested case-control designs, precluding subgroup analyses by study design. In addition, the lack of demographic data or other potential confounders rendered subgroup analyses by alternative factors unfeasible. Further studies are warranted to better understand the associations between dietary and circulating vitamin levels and the risk of gastrointestinal cancers across different genotypes, with adjustment for potential confounding factors.
In conclusion, our systematic review with meta-analysis identified the inverse association of high levels of dietary and circulating vitamin B and vitamin D with gastrointestinal cancer risk. Individuals harboring different variants of vitamin metabolism genes, such as the MTHFR C677T, MTHFR A1298C, and VDR TaqI polymorphisms, may exhibit heterogeneities in their associations with CRC risk. Our findings suggest a potential beneficial impact of vitamin B and vitamin D in reducing gastrointestinal cancer risk, emphasizing the significance of interpreting these effects in the context of genetic background and the importance of tailored, genetically informed gastrointestinal cancer prevention strategies.

CONFLICTS OF INTEREST
Guarantor of the article: Wen-Qing Li, PhD.
Specific author contributions: W.-Q.L. contributed to the study conception and design. X.-L.W. and H.-M.X. contributed to data collection, interpretation and statistical analysis. W.-Q.L., X.-L.W., and H.-M.X. wrote the draft of the manuscript. Z.-Q.H. and K.-F.P. contributed to administrative, technical, or material support and data collection. W.-Q.L. and K.-F.P. revised the manuscript. All authors read and approved the final draft submitted.
Financial support: This study was funded by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (No. 2023ZD0501400-2023ZD0501402), National Natural Science Foundation of China (No. 82273704), Beijing Hospitals Authority's Ascent Plan (DFL20241102), China Postdoctoral Science Foundation (2024M760152), Beijing Hospitals Authority Clinical Medicine Development of Special Funding Support (No. ZLRK202325), Peking University Medicine Fund for World's Leading Discipline or Discipline Cluster Development (No. BMU2022XKQ004), Science Foundation of Peking University Cancer Hospital (No. BJCH2024BJ02, XKFZ2410, and BJCH2025CZ04).
Potential competing interests: None to report.
Ethics statement: The study was approved by the Institutional Review Board of Peking University Cancer Hospital.

Supplementary Material