Abstract
-
Purpose
- Lung cancer is frequently observed as a second primary malignancy following gastric cancer, yet the genetic causality between them remains uncertain. This study aims to evaluate the causal relationship between gastric and lung cancers using Mendelian randomization (MR) analysis.
-
Materials and Methods
- Single nucleotide polymorphisms associated with gastric and lung cancers were selected from genome-wide association study in East Asian and European populations as instrumental variables. The causal effects between gastric and lung cancers were evaluated using univariable and multivariable MR analysis, with the inverse variance weighted (IVW) method serving as the primary criterion. Heterogeneity and sensitivity analyses were performed to ensure the robustness of the findings.
-
Results
- Univariable MR analysis demonstrated that genetic susceptibility to gastric cancer in the European population was significantly associated with an increased risk of lung cancer (IVW: odds ratio [OR], 1.285; 95% confidence interval [CI], 1.072 to 1.541; p=6.83E-03), which was consistently validated in the East Asian population (IVW: OR, 1.356; 95% CI, 1.114 to 1.651; p=2.40E-03). Multivariable MR analysis further indicated that the significant positive causal relationship between gastric cancer and lung cancer persisted in both populations after adjusting for confounding factors (all p < 0.05). Conversely, no significant causal relationship was observed for the risk of developing gastric cancer following the diagnosis of lung cancer in either population (p > 0.05).
-
Conclusion
- This study confirms that genetic susceptibility to gastric cancer increases the risk of lung cancer. This finding provides a theoretical basis for exploring the underlying biological mechanisms and suggests that enhancing lung cancer screening in patients with gastric cancer may be necessary to improve patient prognosis.
-
Key words: Stomach neoplasms, Lung neoplasms, Risk, Genetic, Mendelian randomization
Introduction
Gastric cancer (GC) remains one of the most common malignancies globally and is also one of the leading causes of cancer-related mortality. According to GLOBOCAN statistics, in 2020, there were 1,089,000 new cases of GC worldwide, resulting in 769,000 deaths, ranking fifth and fourth among all malignancies, respectively. The incidence of GC varies by region, with a higher concentration in East Asia, Eastern Europe, and South America [1-3]. The development of GC is closely related to risk factors such as Helicobacter pylori infection, smoking, and alcohol consumption, with specific genetic mutations also confirmed to play a role in its onset [4,5]. The widespread implementation of endoscopic screening has significantly improved the detection rate of early GC [6,7], while advancements in endoscopy, surgical resection, and perioperative adjuvant therapies have contributed to enhanced survival rates among GC patients [8-10].
With the prolonged survival of GC patients, the occurrence of second primary cancers has become more frequently observed. In a cohort study involving over 450,000 cancer survivors, Kjaer et al. [11] reported cumulative incidences of second primary cancers at 5, 10, and 15 years to be 6.3%, 10.5%, and 13.5%, respectively. Subsequent follow-up studies of diagnosed GC patients indicate that approximately 4.5% to 8.0% of them will develop a second primary cancer [12-14], with lung cancer (LC) being one of the most common types. The development of a second primary cancer significantly threatens the survival of patients initially diagnosed with the first primary cancer [15,16]. Although retrospective studies have reported an association between the occurrence of GC and second primary LC, the causal relationship between them remains unconfirmed due to the inherent limitations of observational studies, including confounding factors such as environmental influences.
Mendelian randomization (MR) analysis is a widely used method for causal inference [17], which assesses the causal effect of exposure factors on outcomes by using genetic variants, randomly allocated at conception, as instrumental variables (IVs). The genetic variants used as exposure factors are not influenced by aging or lifestyle changes. Compared to traditional retrospective analyses, MR can effectively reduce bias caused by confounding factors or reverse causality, making it a robust method for inferring causal relationships [18,19].
Existing literature has documented an association between the diagnosis of GC and the subsequent development of second primary LC. However, the causal relationship between them remains unclear, and whether this association is influenced by genetic susceptibility is still unknown. This study aims to investigate the causal relationship between GC and second primary LC at the genetic level using MR, with the objective of optimizing prevention and treatment strategies for GC patients and further improving their prognosis.
Materials and Methods
1. Data sources
Summary data on single nucleotide polymorphisms (SNPs) associated with GC and LC in European and Asian populations were obtained from the genome-wide association study (GWAS) database (https://gwas.mrcieu.ac.uk/) under the ID numbers ebi-a-GCST90018849, ebi-a-GCST90018629, ebi-a-GCST90018875, and ebi-a-GCST90018655. These data were provided by Sakaue S [20] (S1 Table). And Fig. 1 illustrated the study design and the Mendelian assumptions process.
2. Instrumental variables selection
IV selection is based on three key assumptions of MR studies: (1) The genetic instruments are strongly associated with GC; with a significance threshold of p-value < 5×10–8, ensuring that the selected instruments are significantly associated with the exposure. (2) The genetic instruments are independent of both known and unknown confounders; the linkage disequilibrium (LD) coefficient is adjusted to prevent redundancy and correlation between SNPs. Specifically, we applied criteria of R2 < 0.001 and a window distance > 10,000 kilobases. The parameter settings for R² and window size are well-established and widely used, setting a stringent standard that ensures minimal correlation between SNPs, thereby allowing them to be considered independent. While the window size effectively mitigates the “genetic drag effect” and long-range LD, thereby reducing the impact of negative LD and minimizing potential interference from variable interactions on causal inference. (3) The genetic instruments influence the outcome exclusively through the GC risk factor, rather than via other direct factors. The selected SNPs were assessed for direct associations with LC or other potential confounders using tools such as LDTrait (https://ldlink.nih.gov/?tab=home).
After fulfilling the screening criteria based on the three key assumptions, we further excluded weak IVs. When IVs are too weak, the two-stage least squares estimate in MR tend to be biased towards zero (i.e., underestimating causal effects), and the standard errors are underestimated, increasing the risk of false positives. The F-statistic is a crucial measure of IVs strength with an F-statistic ≤ 10 may lead to underestimating or inaccurate causal effect estimation. An F-statistic > 10 ensures that the instruments possess sufficient statistical strength to support causal inference without compromising the robustness of the estimates. These parameter settings adhere to widely recognized standards [21]. For SNPs missing in the outcome, we used highly linked SNPs (proxy SNPs) with an LD > 0.8 as substitutes. If no proxy SNP was identified, the missing SNP was excluded from the subsequent analysis. Additionally, SNPs with palindromic variants were excluded from the analysis because these variants could have two different genotype directions (e.g., A/T or T/A), potentially leading to confusion in genotype interpretation across two samples. This could lead to misinterpretation in the statistical model, thereby affecting the estimation of causal effects. Consequently, SNPs with palindromic variants (e.g., rs2920280) were excluded to ensure consistency in genotype orientation across all SNPs (S2 Table). We also identified genes closely associated with these SNPs using the National Library of Medicine (https://www.ncbi.nlm.nih.gov/). The stringent implementation of these standards ensures the scientific rigor, independence, and robustness of the selected IVs, while minimizing potential bias [19,22,23].
3. Univariable MR analysis
In this study, we primarily employed the inverse variance weighted (IVW) method to assess the causal relationship between GC and LC. To ensure the robustness of our findings, we also performed additional statistical analyses using the weighted median (WM) method and MR-Egger. The IVW analysis method is the primary approach in MR studies. When all SNPs meet the criteria for valid IVs, appropriate weights are assigned to each SNP to fit the analysis, thereby providing the strongest robustness in the results. The WM method requires only that more than 50% of SNPs to satisfy the three core MR assumptions to produce unbiased results, although its statistical power is lower than that of the IVW method.
The primary difference between MR-Egger and IVW is that MR-Egger regression analysis accounts for the intercept term. The slope of the MR-Egger regression represents the estimated causal effect of exposure on the outcome and does not require the regression line to pass through the origin, thereby allowing for pleiotropy in the included IVs. Thus, the intercept value from MR-Egger analysis can be used to evaluate the presence of horizontal pleiotropy. The Mendelian Randomization Pleiotropy RESidual Sum and Outlier (MR-PRESSO) method can perform outlier tests, enabling the removal of SNPs that exhibit horizontal pleiotropy before re-conducting the MR analysis, thereby obtaining an estimate closer to the true causal effect [24-26].
4. Sensitivity analysis
Cochran’s Q test and leave-one-out analysis are used for sensitivity analysis, with Cochran’s Q test assessing variability among the independent IVs. The magnitude of this variability reflects the level of heterogeneity among the IVs, with p < 0.05 indicating statistical heterogeneity [27]. Leave-one-out analysis involves removing a single SNP and re-estimating the causal effect on the outcome after excluding each individual SNP. If the remaining IVs significantly deviate from zero after removing the SNP, it suggests that the SNP has a substantial influence on the causal estimate, potentially compromising the robustness of the results [26]. Furthermore, the study employed Bayesian joint modeling to incorporate the effects of all SNPs, evaluating the combined effects and interactions among multiple SNPs.
5. Multivariable MR analysis
Considering the overlap in cancer risk factors for cancer, the study performed multivariable MR (MVMR) analysis on the shared and individual recognized risk factors for GC and LC, including lifestyle (smoking, alcohol consumption, physical activity, vitamin intake, and amino acid intake), environmental factors (industrial emissions, particulate air pollution), genetic factors (epidermal growth factor, tyrosine protein kinase), chronic lung diseases (interstitial lung disease, idiopathic pulmonary fibrosis, tuberculosis, chronic obstructive pulmonary disease), chronic gastric diseases (Helicobacter pylori infection, chronic gastritis, gastric ulcers), and metabolic-related diseases and indicators (diabetes, fasting blood glucose, 2-hour postprandial plasma glucose, glycated hemoglobin, total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and body mass index [BMI]) (S1 Table) [1,2,28]. Considering that SNPs may affect other traits, the study further used the LDTrait tool (https://ldlink.nih.gov/?tab=home) to examine whether the selected SNPs are strongly correlated with factors beyond the exposure, and incorporated these traits in the MVMR whenever possible (S3 Table). We applied MVMR-IVW to adjust for these confounding factors and subsequently evaluated whether the causal relationship between GC and LC remained robust, with p < 0.05 considered statistically significant.
In this MR analysis, odds ratios (OR) were served as the effect sizes, with 95% confidence intervals (CI) used to estimate the overall parameters. Statistical significance was determined at a threshold of p < 0.05. All MR analyses were performed using R software ver. 4.2.2 (http://www.r-project.org), with the primary analysis performed via the “Two Sample MR” package ver. 0.6.6. Additional R packages were employed for supplementary analyses.
Results
1. Univariable MR analysis
In the MR study, the number of SNPs identified as IVs for GC and LC in the European population was 9 and 12, respectively. In the East Asian population, nine SNPs were identified as IVs for both GC and LC, respectively. The F-statistic for all SNPs exceeded 10, indicating that the strength of the IVs was robust (S2 Table, S4-S6 Tables).
We used IVW and WM to assess the causal relationship between GC and LC in both European and East Asian populations. In the European population, genetic susceptibility to GC was significantly associated with an increased risk of LC (IVW: OR, 1.285; 95% CI, 1.072 to 1.541; p=6.83E-03 and WM: OR, 1.189; 95% CI, 1.057 to 1.337; p=3.88E-03) (Table 1). Similarly, in the East Asian population, GC was found to significantly promote the occurrence of LC (IVW: OR, 1.356; 95% CI, 1.114 to 1.651; p=2.40E-03 and WM: OR, 1.241; 95% CI, 1.062 to 1.449; p=6.46E-03) (Table 1).
Using the MR-Egger method, results from both populations suggested that GC is a protective factor against LC (MR-Egger [European]: OR, 0.904; 95% CI, 0.575 to 1.422; p=6.75E-01 and MR-Egger [East Asian]: OR, 0.942; 95% CI, 0.576 to 1.541; p=8.20E-01) (Table 1). However, after reanalysis using MR-PRESSO to remove outliers (S7 Table), the association between GC and an increased risk of LC remained significant in univariable MR analysis (IVW [European]: OR, 1.293; 95% CI, 1.076 to 1.554; p=6.18E-03 and IVW [East Asian]: OR, 1.345; 95% CI, 1.172 to 1.543; p=2.37E-05). Moreover, in the European population, all three MR analysis methods indicated that genetic susceptibility to GC is associated with an increased risk of LC (Table 1). The study also evaluated the genetic susceptibility of LC to the risk of GC, with results suggesting no significant causal relationship between LC and secondary primary GC (S8 Table).
Sensitivity analysis was performed after the removal of outliers, and no evidence of heterogeneity or horizontal pleiotropy was detected in the MR analysis across both populations (Table 1, Fig. 2). The leave-one-out test indicated that excluding any single SNP did not substantially alter the results (Fig. 3). Bayesian analysis was performed to evaluate the combined effects and interactions of multiple SNPs, with the results consistently supporting the positive causal effect of GC on LC (OR [European], 1.334; 95% CI, 1.219 to 1.472 and OR [East Asian], 1.385; 95% CI, 1.243 to 1.528) (S9 Table, S10 Fig.).
2. Multivariable MR analysis
The study further incorporated lifestyle (smoking, alcohol consumption, physical activity, vitamin intake, and amino acid intake), environmental factors (industrial emissions, particulate air pollution), genetic factors (epidermal growth factor, tyrosine protein kinase), chronic lung diseases (interstitial lung disease, idiopathic pulmonary fibrosis, tuberculosis, chronic obstructive pulmonary disease), chronic gastric diseases (Helicobacter pylori infection, chronic gastritis, gastric ulcers), and metabolic-related diseases and indicators (diabetes, fasting blood glucose, 2-hour postprandial plasma glucose, glycated hemoglobin, total cholesterol, LDL cholesterol, HDL cholesterol, BMI) and SNP-related traits for MVMR analysis (Fig. 4, S1 and S3 Tables, S11 Fig.). In the European population, after incorporating the aforementioned potential covariates into the MVMR-IVW analysis, a significant positive causal relationship between GC and LC remained evident (p < 0.05) (Fig. 4, S11 Fig.). In the East Asian population, due to the limitations of data for East Asians in the Open GWAS database (https://gwas.mrcieu.ac.uk/), we were unable to match the number of covariates included for the European population. However, after incorporating as many lifestyle factors, environmental factors, chronic lung diseases, chronic gastric diseases, and metabolic-related diseases and indicators as possible into the MVMR analysis, a significant positive causal relationship between GC and LC remained evident (p < 0.05) (Fig. 4, S11 Fig.).
Discussion
This study is the first to report a causal relationship between diagnosed GC and second primary LC from the genetic prediction perspective. By using SNPs of GC and LC as instrumental variables in MR analysis, it was confirmed that a diagnosis of GC increases the risk of developing LC in the European population. Additionally, reverse MR analysis did not find evidence of a causal relationship between LC diagnosis and second primary GC. Interestingly, these conclusions were consistently validated in the East Asian population. Furthermore, the study effectively ensured the robustness of the results through various MR analysis methods and sensitivity analyses.
Advancements in cancer diagnosis and treatment have improved patient survival rates, leading to an increase in reports of second primary cancers [12,29]. A second primary cancer is defined as a new cancer that occurs either simultaneously with or subsequent to the first diagnosed cancer, remaining pathologically independent from the original mal-ignancy. This secondary primary cancer can develop months or even years after the initial cancer diagnosis [30,31]. The question of whether a second primary cancer arises incidentally or as a consequence of the first primary cancer remains a topic of considerable debate. In addition to genetic predispositions, environmental factors are also closely linked to cancer development [32]. For instance, smoking and environmental pollution are recognized as potential risk factors for both LC and GC [2,33]. In the MVMR analysis, in addition to adjusting for common risk factors such as smoking and alcohol consumption, the study also comprehensively accounted for environmental factors such as occupational exposure, industrial pollution, lifestyle factors like diet and physical activity, and metabolic diseases. After addressing as many potential confounders as possible, the study found that the genetic evidence supporting an increased risk of LC following a GC diagnosis remained robust [34]. This suggests that genetic susceptibility may play a predominant role in the development of LC after a GC diagnosis.
However, the exact mechanism of LC occurrence following a GC diagnosis remains unclear. Notably, from an embryological perspective, both the respiratory and gastrointestinal tract originate from the endoderm, and numerous shared signaling pathways and molecules may participate in the development of visceral organs from the endoderm. This provides a theoretical foundation for the possibility that genetic variations may jointly influence the occurrence of both LC and GC [35]. Shi’s study identified key genes, including ADAM12, SPP1, COL1A1, and COL11A1, which may facilitate the development of GC and LC by regulating extracellular matrix-receptor interactions and cell adhesion signaling pathways [36]. Furthermore, GC, which is often associated with Lynch syndrome, frequently exhibits mutations in mismatch repair genes such as MLH1 and MSH2. The functional loss of these genes may increase genetic instability, thereby promoting the onset of LC [37-39]. In patients with primary GC, persistent systemic inflammation mediated by PTGER4 may further increase the risk of LC through the activation of the PGE2 signaling pathway [40]. Genetic variations in HLA-DQA1 may cause immune dysfunction, hindering the immune system of primary GC patients from effectively eliminating abnormal lung cells, thereby increasing the risk of second primary LC [41]. This suggests that the underlying mechanisms likely involve a combination of genetic variation, inflammatory response, immune regulation, and epigenetic factors. Future research should integrate multi-omics approaches to further investigate and validate these mechanisms.
It is noteworthy that previous studies have reported a potential association between the locations of second primary cancers and the sites of first primary cancers, suggesting that different first primary cancers may predispose individuals to specific second primary cancers. However, this remains a subject of controversy [42,43]. Li et al. [5] found that following a diagnosis of GC, the risk of second primary cancers in the esophagus, colon, and pancreas is increased, while the risk of prostate and breast cancers is decreased. In contrast, Lawniczak et al. reported that GC patients are more likely to subsequently develop colorectal cancer, LC, breast cancer, and prostate cancer [44]. Considering that these differences might be attributed to racial variations, our study included both European and East Asian populations, thereby mitigating the sampling biases that may occur when analyzing only a single population [45]. Additionally, opinions remain divided on whether LC increases the risk of GC. For instance, Ruan et al. [46] found, based on analyses of cancer associations in non-Hispanic white populations, that a diagnosis of LC is a risk factor for GC. Whereas Leung’s retrospective analysis of second primary cancers across multiple populations found no significant association between LC and second primary GC [47]. This lack of association may be influenced by specific environmental factors or genetic diversity among different populations. Furthermore, our study, through dual-population analysis, reinforces that there is no direct genetic link to the occurrence of GC following an LC diagnosis.
To our knowledge, this study is the first to analyze the association between genetic susceptibility to GC and the increased risk of LC based on GWAS data. Compared to traditional retrospective analyses, MR analysis offers robust support for causal inference. Furthermore, most current MR studies are based solely on European populations, limiting the generalizability of the findings due to population differences. To address this, we also conducted MR analysis on an East Asian population, thereby enhancing the robustness of our results. This dual-population approach represents a significant advantage of our study. However, there are also some limitations to this study. Firstly, the GWAS database is based on population-based prospective cohort studies, and the current number of observed cases during follow-up may affect the study’s statistical power. Additionally, the follow-up duration is closely related to the survival times of LC and GC, and insufficient follow-up time could impact the overall observations of the population. The study employed strict screening of instrumental variables, LDTrait checks, Bayesian models, and MVMR, among other methods, to cross-validate and enhance the robustness of the findings. However, some bias could not be fully eliminated. Despite adjusting for over 50 potential confounding factors in the multivariable analysis, missing covariates in the East Asian population could still potentially affect the degree of adjustment in the causal relationship between GC and LC. In conclusion, this study provides strong evidence of a significant positive causal relationship between GC and LC, suggesting that, in clinical practice, enhanced low-dose chest computed tomography screening for LC in patients diagnosed with GC could offer potential benefits. Building upon these findings, future prospective studies could assess the impact of genetic risk-based screening strategies on early LC diagnosis and survival rates, thereby providing valuable evidence to inform the refinement of clinical guidelines. Furthermore, exploring the underlying biological mechanisms and biomarkers through basic research, by which GC influences LC development, is crucial for the development of targeted therapies aimed at specific sites or markers, ultimately improving the prognosis of GC patients. Further research into the epigenetic modifications in gastric and LC patients is anticipated to provide new insights into cancer treatment strategies, reduce the incidence of second primary LC, and ultimately improve the overall prognosis of GC patients.
This study confirms that genetic susceptibility to GC increases the risk of LC, this finding not only provides a theoretical basis for exploring the underlying biological mechanisms but also suggests that enhanced LC screening in GC patients may be warranted, potentially improving their prognosis.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
NOTES
-
Author Contributions
Conceived and designed the analysis: Chen J, Zeng A, Yu Y.
Collected the data: Chen J, Yu Y, Sun S, Liao L, Huang S.
Contributed data or analysis tools: Yu Y, Sun S, Liao L, Huang S.
Performed the analysis: Zeng A, Sun S.
Wrote the paper: Chen J, Zeng A, Yang Z, Zhou J, Wu W.
Initiated and designed this work: Zeng A, Yu Y, Sun S, Liao L, Huang S, Yang Z, Zhou J, Wu W.
-
Conflicts of Interest
Conflict of interest relevant to this article was not reported.
-
Acknowledgments
We sincerely thank Libin Xu for her careful guidance on statistical analysis. Thanks to all the investigators of Within family GWAS consortium, UK Biobank study, FinnGen, GWAS Catalog, MRC-IEU for providing the data publicly. Thanks to Ben Elsworth et al. for their network technical support, https://www.biorxiv.org/content/10.1101/2020.08.10.244293v1.
Fig. 1.Study design overview. IV, instrumental variable; MR, Mendelian randomization; MR-PRESSO, Mendelian Randomization Pleiotropy RESidual Sum and Outlier; SNP, single nucleotide polymorphism.
Fig. 2.Scatter plot of causal effects of gastric cancer on lung cancers: European (A) and East Asian (B). MR, Mendelian randomization; SNP, single nucleotide polymorphism.
Fig. 3.Leave-one-out analysis for gastric cancer on lung cancers: European (A) and East Asian (B). MR, Mendelian randomization.
Fig. 4.Forest plot of the multivariable Mendelian randomization analyses exploring genetically determined gastric cancer (GC) with lung cancer (LC) and adjusted for confounding factors. CI, confidence interval; HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio; SNP, single nucleotide polymorphism.
Table 1.Univariate MR analysis of GC to LC
|
Ancestry |
Exposure-outcome |
nSNP |
Method |
OR (95% CI) |
p-value |
MR-PRESSO outlier corrected
|
Heterogeneity test
|
MR-Egger pleiotropy test
|
|
OR (95% CI) |
p-value |
Outliers |
Q value |
p-value |
Intercept |
p-value |
|
European |
GC-LC |
9 |
IVW |
1.285 (1.072-1.541) |
6.83E-03 |
1.293 (1.076-1.554) |
6.18E-03 |
6 |
0.421 |
0.810 |
0.012 |
0.889 |
|
GC-LC |
9 |
WM |
1.189 (1.057-1.337) |
3.88E-03 |
1.263 (1.013-1.573) |
3.78E-02 |
6 |
|
|
|
|
|
GC-LC |
9 |
MR-Egger |
0.904 (0.575-1.422) |
6.75E-01 |
1.197 (0.497 -2.882 |
7.57E-01 |
6 |
|
|
|
|
|
Asian |
GC-LC |
9 |
IVW |
1.356 (1.114-1.651) |
2.40E-03 |
1.345 (1.172-1.543) |
2.37E-05 |
4 |
3.431 |
0.488 |
0.080 |
0.190 |
|
GC-LC |
9 |
WM |
1.241 (1.062-1.449) |
6.46E-03 |
1.334 (1.125-1.581) |
8.99E-04 |
4 |
|
|
|
|
|
GC-LC |
9 |
MR-Egger |
0.942 (0.576-1.541) |
8.20E-01 |
0.972 (0.651-1.452) |
8.99E-01 |
4 |
|
|
|
|
REFERENCES
- 1. Smyth EC, Nilsson M, Grabsch HI, van Grieken NC, Lordick F. Gastric cancer. Lancet. 2020;396:635–48. ArticlePubMed
- 2. Thrift AP, Wenker TN, El-Serag HB. Global burden of gastric cancer: epidemiological trends, risk factors, screening and prevention. Nat Rev Clin Oncol. 2023;20:338–49. ArticlePubMedPDF
- 3. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209–49. ArticlePubMedPDF
- 4. Gullo I, van der Post RS, Carneiro F. Recent advances in the pathology of heritable gastric cancer syndromes. Histopathology. 2021;78:125–47. ArticlePubMedPDF
- 5. Li J, Woods SL, Healey S, Beesley J, Chen X, Lee JS, et al. Point mutations in exon 1B of APC reveal gastric adenocarcinoma and proximal polyposis of the stomach as a familial adenomatous polyposis variant. Am J Hum Genet. 2016;98:830–42. PubMedPMC
- 6. Choi KS, Suh M. Screening for gastric cancer: the usefulness of endoscopy. Clin Endosc. 2014;47:490–6. ArticlePubMedPMC
- 7. Asaka M, Kato M, Sakamoto N. Roadmap to eliminate gastric cancer with Helicobacter pylori eradication and consecutive surveillance in Japan. J Gastroenterol. 2014;49:1–8. ArticlePubMedPMCPDF
- 8. Macdonald JS, Smalley SR, Benedetti J, Hundahl SA, Estes NC, Stemmermann GN, et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med. 2001;345:725–30. Article
- 9. Japanese Gastric Cancer Association. Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric Cancer. 2011;14:113–23. ArticlePubMedPDF
- 10. Bang YJ, Van Cutsem E, Feyereislova A, Chung HC, Shen L, Sawaki A, et al. Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2-positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet. 2010;376:687–97. PubMed
- 11. Kjaer TK, Andersen EA, Ursin G, Larsen SB, Bidstrup PE, Winther JF, et al. Cumulative incidence of second primary cancers in a large nationwide cohort of Danish cancer survivors: a population-based retrospective cohort study. Lancet Oncol. 2024;25:126–36. ArticlePubMed
- 12. Zheng G, Sundquist K, Sundquist J, Chen T, Forsti A, Hemminki A, et al. Second primary cancers after gastric cancer, and gastric cancer as second primary cancer. Clin Epidemiol. 2021;13:515–25. ArticlePubMedPMCPDF
- 13. Lawniczak M, Gawin A, Jaroszewicz-Heigelmann H, Rogoza-Mateja W, Raszeja-Wyszomirska J, Bialek A, et al. Synchronous and metachronous neoplasms in gastric cancer patients: a 23-year study. World J Gastroenterol. 2014;20:7480–7. ArticlePubMedPMC
- 14. Buyukasik O, Hasdemir AO, Gulnerman Y, Col C, Ikiz O. Second primary cancers in patients with gastric cancer. Radiol Oncol. 2010;44:239–43. ArticlePubMedPMC
- 15. Narayana A, Vaughan AT, Fisher SG, Reddy SP. Second primary tumors in laryngeal cancer: results of long-term follow-up. Int J Radiat Oncol Biol Phys. 1998;42:557–62. ArticlePubMed
- 16. Gassenmaier M, Stec T, Keim U, Leiter U, Eigentler TK, Metzler G, et al. Incidence and characteristics of thick second primary melanomas: a study of the German Central Malignant Melanoma Registry. J Eur Acad Dermatol Venereol. 2019;33:63–70. ArticlePDF
- 17. Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23:R89–98. ArticlePubMedPMC
- 18. Burgess S, Thompson SG. Mendelian randomization: methods for using genetic variants in causal estimation. CRC Press; 2015.
- 19. Emdin CA, Khera AV, Kathiresan S. Mendelian Randomization. JAMA. 2017;318:1925–6. ArticlePubMed
- 20. Sakaue S, Kanai M, Tanigawa Y, Karjalainen J, Kurki M, Koshiba S, et al. A cross-population atlas of genetic associations for 220 human phenotypes. Nat Genet. 2021;53:1415–24. ArticlePubMedPMCPDF
- 21. Burgess S, Thompson SG; CRP CHD Genetics Collaboration. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011;40:755–64. ArticlePubMed
- 22. Hemani G, Tilling K, Davey Smith G. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet. 2017;13:e1007081ArticlePubMedPMC
- 23. Chen S, Luo X, Zhao J, Liang Z, Gu J. Exploring the causality between ankylosing spondylitis and atrial fibrillation: a two-sample Mendelian randomization study. Front Genet. 2022;13:951893.ArticlePubMedPMC
- 24. Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in Mendelian randomization with some invalid instruments using a weighted Median estimator. Genet Epidemiol. 2016;40:304–14. ArticlePubMedPMCPDF
- 25. Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44:512–25. ArticlePubMedPMC
- 26. Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018;50:693–8. ArticlePubMedPMCPDF
- 27. Hemani G, Zheng J, Elsworth B, Wade KH, Haberland V, Baird D, et al. The MR-Base platform supports systematic causal inference across the human phenome. Elife. 2018;7:e34408PubMedPMC
- 28. Leiter A, Veluswamy RR, Wisnivesky JP. The global burden of lung cancer: current status and future trends. Nat Rev Clin Oncol. 2023;20:624–39. ArticlePubMedPDF
- 29. Copur MS, Wedel W, Jonglertham P, Aprn CS, Horn A. Hypercalcemia in a patient with small lymphocytic lymphoma/chronic lymphocytic leukemia. Oncology (Williston Park). 2019;33:688884.PubMed
- 30. Warren S, Gates O. Multiple primary malignant tumors: a survey of the literature and statistical study. Am J Cancer. 1932;16:1358–414.
- 31. Keegan TH, Bleyer A, Rosenberg AS, Li Q, Goldfarb M. Second primary malignant neoplasms and survival in adolescent and young adult cancer survivors. JAMA Oncol. 2017;3:1554–7. ArticlePubMedPMC
- 32. Feller A, Matthes KL, Bordoni A, Bouchardy C, Bulliard JL, Herrmann C, et al. The relative risk of second primary cancers in Switzerland: a population-based retrospective cohort study. BMC Cancer. 2020;20:51.PubMedPMC
- 33. Tyczynski JE, Bray F, Parkin DM. Lung cancer in Europe in 2000: epidemiology, prevention, and early detection. Lancet Oncol. 2003;4:45–55. ArticlePubMed
- 34. Zeng M, Lin A, Jiang A, Qiu Z, Zhang H, Chen S, et al. Decoding the mechanisms behind second primary cancers. J Transl Med. 2025;23:115.ArticlePubMedPMCPDF
- 35. Nieto MA. Epithelial plasticity: a common theme in embryonic and cancer cells. Science. 2013;342:1234850.ArticlePubMed
- 36. Shi Y, Chang D, Li W, Zhao F, Ren X, Hou B. Identification of core genes and clinical outcomes in tumors originated from endoderm (gastric cancer and lung carcinoma) via bioinformatics analysis. Medicine (Baltimore). 2021;100:e25154ArticlePubMedPMC
- 37. Capelle LG, Van Grieken NC, Lingsma HF, Steyerberg EW, Klokman WJ, Bruno MJ, et al. Risk and epidemiological time trends of gastric cancer in Lynch syndrome carriers in the Netherlands. Gastroenterology. 2010;138:487–92. ArticlePubMed
- 38. Cani M, Novello S, Bironzo P. Mismatch repair deficiency in lung tumors: adding a new layer of complexity on pie slices. J Thorac Oncol. 2024;19:363–5. ArticlePubMed
- 39. Yang SR, Gedvilaite E, Ptashkin R, Chang J, Ziegler J, Mata DA, et al. Microsatellite instability and mismatch repair deficiency define a distinct subset of lung cancers characterized by smoking exposure, high tumor mutational burden, and recurrent somatic MLH1 inactivation. J Thorac Oncol. 2024;19:409–24. ArticlePubMedPMC
- 40. Echizen K, Hirose O, Maeda Y, Oshima M. Inflammation in gastric cancer: interplay of the COX-2/prostaglandin E2 and Toll-like receptor/MyD88 pathways. Cancer Sci. 2016;107:391–7. ArticlePubMedPMCPDF
- 41. Mahmud MT, Ahmed F, Rana MJ, Rahman MA, Atta A, Saif-Ur-Rahman KM. Association of HLA gene polymorphisms with Helicobacter pylori related gastric cancer: a systematic review. HLA. 2024;103:e15394PubMed
- 42. Muller J, Grosclaude P, Lapotre-Ledoux B, Woronoff AS, Guizard AV, Bara S, et al. Trends in the risk of second primary cancer among bladder cancer survivors: a population-based cohort of 10 047 patients. BJU Int. 2016;118:53–9. ArticlePubMed
- 43. Coyte A, Morrison DS, McLoone P. Second primary cancer risk - the impact of applying different definitions of multiple primaries: results from a retrospective population-based cancer registry study. BMC Cancer. 2014;14:272.ArticlePubMedPMCPDF
- 44. Wang X, Zeng M, Ju X, Lin A, Zhou C, Shen J, et al. Correlation between second and first primary cancer: systematic review and meta-analysis of 9 million cancer patients. Br J Surg. 2024;111:znad377.ArticlePubMedPDF
- 45. Moons KG, Kengne AP, Grobbee DE, Royston P, Vergouwe Y, Altman DG, et al. Risk prediction models: II. External validation, model updating, and impact assessment. Heart. 2012;98:691–8. ArticlePubMed
- 46. Ruan X, Huang D, Zhan Y, Huang J, Huang J, Ng AT, et al. Risk of second primary cancers after a diagnosis of first primary cancer: a pan-cancer analysis and Mendelian randomization study. Elife. 2023;12:e86379ArticlePubMedPMCPDF
- 47. Leung TH, El Helali A, Wang X, Ho JC, Pang H. Trends and age, sex, and race disparities in time to second primary cancer from 1990 to 2019. Cancer Med. 2023;12:22316–24. ArticlePubMedPMC
Citations
Citations to this article as recorded by
KCA
, Aiming Zeng3,4
PubReader
ePub Link
Cite
XML Download
